From 8ac992696564547de95655ec250d16dc4bd41782 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Sat, 30 Oct 2021 00:00:43 -0400
Subject: [PATCH 01/65] Replace the `fewshot_description` API with a
 `description_dict` based interface

---
 lm_eval/base.py                               | 18 +++--
 lm_eval/evaluator.py                          | 26 +++---
 lm_eval/tasks/anli.py                         |  4 -
 lm_eval/tasks/arc.py                          |  4 -
 lm_eval/tasks/cbt.py                          |  4 -
 lm_eval/tasks/coqa.py                         |  5 +-
 lm_eval/tasks/drop.py                         |  4 -
 lm_eval/tasks/glue.py                         | 17 ----
 lm_eval/tasks/headqa.py                       |  4 -
 lm_eval/tasks/hellaswag.py                    |  5 --
 lm_eval/tasks/hendrycks_math.py               |  3 -
 lm_eval/tasks/lambada.py                      |  4 -
 lm_eval/tasks/lambada_cloze.py                |  3 -
 lm_eval/tasks/logiqa.py                       |  4 -
 lm_eval/tasks/mathqa.py                       |  4 -
 lm_eval/tasks/mc_taco.py                      |  3 -
 lm_eval/tasks/mutual.py                       |  4 -
 lm_eval/tasks/naturalqs.py                    |  4 -
 lm_eval/tasks/openbookqa.py                   |  4 -
 lm_eval/tasks/piqa.py                         |  4 -
 lm_eval/tasks/prost.py                        |  8 +-
 lm_eval/tasks/pubmedqa.py                     |  5 --
 lm_eval/tasks/qa4mre.py                       |  3 -
 lm_eval/tasks/quac.py                         |  5 --
 lm_eval/tasks/race.py                         |  4 -
 lm_eval/tasks/sat.py                          |  5 --
 lm_eval/tasks/sciq.py                         |  5 +-
 lm_eval/tasks/squad.py                        |  4 -
 lm_eval/tasks/storycloze.py                   |  6 --
 lm_eval/tasks/superglue.py                    | 32 --------
 lm_eval/tasks/translation.py                  |  6 --
 lm_eval/tasks/triviaqa.py                     |  5 --
 lm_eval/tasks/truthfulqa.py                   |  8 +-
 lm_eval/tasks/unscramble.py                   |  3 -
 lm_eval/tasks/webqs.py                        |  7 +-
 lm_eval/tasks/wikitext.py                     |  6 +-
 lm_eval/tasks/winogrande.py                   |  4 -
 lm_eval/tasks/wsc273.py                       |  4 -
 main.py                                       | 18 +++--
 scripts/cost_estimate.py                      |  2 +-
 scripts/fewshot_description_experiment.py     | 79 -------------------
 scripts/get_prompts.py                        |  1 -
 scripts/write_out.py                          | 15 +++-
 task-guide.md                                 | 54 +++++++++----
 .../test_description_option/descriptions.json |  3 +
 .../test_description_option.py                | 39 +++++++++
 tests/test_evaluator.py                       |  4 +-
 tests/test_version_stable.py                  |  2 +-
 48 files changed, 147 insertions(+), 318 deletions(-)
 delete mode 100644 scripts/fewshot_description_experiment.py
 create mode 100644 tests/test_description_option/descriptions.json
 create mode 100644 tests/test_description_option/test_description_option.py

diff --git a/lm_eval/base.py b/lm_eval/base.py
index 317c70a67e..000f2f2765 100644
--- a/lm_eval/base.py
+++ b/lm_eval/base.py
@@ -2,6 +2,7 @@
 import random
 import numpy as np
 import re
+from lm_eval import tasks
 
 from lm_eval.metrics import mean, perplexity, weighted_perplexity, weighted_mean
 
@@ -224,11 +225,15 @@ def higher_is_better(self):
         pass
 
     def fewshot_description(self):
+        import warnings
+        warnings.warn(
+            "`fewshot_description` will be removed in coming versions. Pass " \
+            "any custom descriptions to the `evaluate` function instead.",
+            DeprecationWarning)
         return ""
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
-        raw_description = self.fewshot_description()
-        description = (raw_description + "\n===\n\n") if provide_description and raw_description else ""
+    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
+        description = description + "\n\n" if description else ""
 
         if num_fewshot == 0:
             labeled_examples = ""
@@ -295,16 +300,13 @@ class PerplexityTask(Task, abc.ABC):
     def has_training_docs(self):
         return False
 
-    def fewshot_description(self):
-        return ""
-
     def fewshot_examples(self, k, rnd):
         assert k == 0
         return []
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
+    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
         assert num_fewshot == 0
-        assert not provide_description
+        assert description is None 
         return ""
 
     def higher_is_better(self):
diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index f4c7dee84c..63679d854b 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -1,5 +1,6 @@
 import collections
 import itertools
+import json
 import random
 import lm_eval.metrics
 import lm_eval.models
@@ -7,7 +8,7 @@
 import lm_eval.base
 import numpy as np
 
-def simple_evaluate(model, model_args, task_names, num_fewshot=0, batch_size=None, device=None, no_cache=False, limit=None, bootstrap_iters=100000):
+def simple_evaluate(model, model_args, task_names, description_path=None, num_fewshot=0, batch_size=None, device=None, no_cache=False, limit=None, bootstrap_iters=100000):
     random.seed(1234)
     np.random.seed(1234)
 
@@ -19,7 +20,12 @@ def simple_evaluate(model, model_args, task_names, num_fewshot=0, batch_size=Non
         lm = lm_eval.base.CachingLM(lm, 'lm_cache/' + model + '_' + model_args.replace('=', '-').replace(',', '_').replace('/', '-') + '.db')
     
     task_dict = lm_eval.tasks.get_task_dict(task_names)
-    results = evaluate(lm, task_dict, False, num_fewshot, limit)
+    description_dict = {}
+    if description_path:
+        with open(description_path, 'r') as f:
+            description_dict = json.load(f)
+
+    results = evaluate(lm, task_dict, num_fewshot, limit, description_dict)
 
     # add info about the model and few shot config
     results["config"] = {
@@ -28,6 +34,8 @@ def simple_evaluate(model, model_args, task_names, num_fewshot=0, batch_size=Non
         "num_fewshot": num_fewshot,
         "batch_size": batch_size,
         "device": device,
+        # TODO (jon-tow): Should we add the description info to `results["config"]`?
+        # "description_dict": description_dict,
         "no_cache": no_cache,
         "limit": limit,
         "bootstrap_iters": bootstrap_iters
@@ -36,9 +44,7 @@ def simple_evaluate(model, model_args, task_names, num_fewshot=0, batch_size=Non
     return results
 
 
-def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_iters=100000):
-    assert not provide_description # not implemented. todo: implement proper description-providing system
-
+def evaluate(lm, task_dict, num_fewshot, limit, description_dict=None, bootstrap_iters=100000):
     # TODO: completely refactor this entire function to not be a huge mess, ideally breaking it down into smaller pieces
 
     task_dict_items = [(name, task) for name, task in task_dict.items() if(task.has_validation_docs() or task.has_test_docs())]
@@ -73,16 +79,16 @@ def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_i
         rnd.seed(42)
         rnd.shuffle(task_docs)
 
+        description = description_dict[task_name] if description_dict and task_name in description_dict else ""
+
         for doc_id, doc in enumerate(itertools.islice(task_docs, 0, limit)):
             docs[(task_name, doc_id)] = doc
-
             ctx = task.fewshot_context(
                 doc=doc,
-                provide_description=provide_description,
                 num_fewshot=num_fewshot,
-                rnd=rnd
+                rnd=rnd,
+                description=description
             )
-
             reqs = task.construct_requests(doc, ctx)
             if not isinstance(reqs, (list, tuple)): reqs = [reqs]
             for i, req in enumerate(reqs):
@@ -168,4 +174,4 @@ def make_table(result_dict):
     # todo: make latex table look good
     # print(latex_writer.dumps())
 
-    return md_writer.dumps()
\ No newline at end of file
+    return md_writer.dumps()
diff --git a/lm_eval/tasks/anli.py b/lm_eval/tasks/anli.py
index 1304c5da2b..13c4044560 100644
--- a/lm_eval/tasks/anli.py
+++ b/lm_eval/tasks/anli.py
@@ -33,10 +33,6 @@ def test_docs(self):
         if self.has_test_docs():
             return self.data["test_r" + str(self.SPLIT)]
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     def doc_to_text(self, doc):
         # OA does this a bit weirdly: they prepend "anli 1:  anli 1:  " to the beginning
         # of the prompt (yes, repeating it!). also, " True, False, or Neither?" is directly 
diff --git a/lm_eval/tasks/arc.py b/lm_eval/tasks/arc.py
index a0d13abc59..2a8a999842 100644
--- a/lm_eval/tasks/arc.py
+++ b/lm_eval/tasks/arc.py
@@ -29,10 +29,6 @@ def _convert_standard(self, doc):
         }
         return out_doc
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     def doc_to_text(self, doc):
         return doc["query"]
 
diff --git a/lm_eval/tasks/cbt.py b/lm_eval/tasks/cbt.py
index 8837caff6d..e239a630b4 100644
--- a/lm_eval/tasks/cbt.py
+++ b/lm_eval/tasks/cbt.py
@@ -17,10 +17,6 @@ class CBTBase(HFTask):
 
     VERSION = 0
 
-    def fewshot_description(self):
-        # TODO: Figure out description.
-        return ""
-
     def detokenize(self, text):
         text = text.replace(" '", "'")
         text = text.replace(" \n", "\n")
diff --git a/lm_eval/tasks/coqa.py b/lm_eval/tasks/coqa.py
index beba53a663..095220bfe6 100644
--- a/lm_eval/tasks/coqa.py
+++ b/lm_eval/tasks/coqa.py
@@ -36,10 +36,7 @@ def validation_docs(self):
 
     def test_docs(self):
         pass
-    
-    def fewshot_description(self):
-        return "Given a passage and a conversation so far, answer the next question in the conversation."
-    
+
     def doc_to_text(self, doc):
         # Given a passage p, the conversation history {q1, a1, . . . qi−1, ai−1} 
         # and a question qi, the task is to predict the answer ai
diff --git a/lm_eval/tasks/drop.py b/lm_eval/tasks/drop.py
index aaf4201abd..5e1e85529e 100644
--- a/lm_eval/tasks/drop.py
+++ b/lm_eval/tasks/drop.py
@@ -40,10 +40,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     def _load_docs(self, docs):
         for doc in docs:
             for qa in doc["qa_pairs"]:
diff --git a/lm_eval/tasks/glue.py b/lm_eval/tasks/glue.py
index a9243dafbb..df8194eea6 100644
--- a/lm_eval/tasks/glue.py
+++ b/lm_eval/tasks/glue.py
@@ -21,10 +21,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO
-        return ""
-
     def doc_to_text(self, doc):
         return "{}\nQuestion: Does this sentence make sense?\nAnswer:".format(doc["sentence"])
 
@@ -69,9 +65,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        return "Indicate if the sentiment of each sentence is positive or negative."
-
     def doc_to_text(self, doc):
         return "{}\nQuestion: Is this sentence positive or negative?\nAnswer:".format(
             general_detokenize(doc["sentence"]),
@@ -342,9 +335,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        return "Indicate if both sentences mean the same thing."
-
     def doc_to_text(self, doc):
         return "Sentence 1: {}\nSentence 2: {}\nQuestion: Do both sentences mean the same thing?\nAnswer:".format(
             general_detokenize(doc["sentence1"]),
@@ -395,9 +385,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        return "Indicate if both questions ask the same thing."
-
     def doc_to_text(self, doc):
         return "Question 1: {}\nQuestion 2: {}\nQuestion: Do both questions ask the same thing?\nAnswer:".format(
             doc["question1"],
@@ -448,10 +435,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return True
 
-    def fewshot_description(self):
-        return "Indicate if both sentences mean the same thing from a scale of 0-5, " \
-           "where 5 means identical and 0 means unrelated."
-
     def doc_to_text(self, doc):
         return "sentence 1: {}\nsentence 2: {}\nAnswer:".format(
             doc["sentence1"],
diff --git a/lm_eval/tasks/headqa.py b/lm_eval/tasks/headqa.py
index 3c66dc064b..c258ff1d07 100644
--- a/lm_eval/tasks/headqa.py
+++ b/lm_eval/tasks/headqa.py
@@ -25,9 +25,5 @@ def _convert_standard(self, doc):
         }
         return out_doc
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     def doc_to_text(self, doc):
         return doc["query"]
diff --git a/lm_eval/tasks/hellaswag.py b/lm_eval/tasks/hellaswag.py
index 762ce47337..56450cf3e6 100644
--- a/lm_eval/tasks/hellaswag.py
+++ b/lm_eval/tasks/hellaswag.py
@@ -35,10 +35,5 @@ def _convert_standard(self, doc):
         }
         return out_doc
 
-    def fewshot_description(self):
-        return "Label for the relevant action: Sentences describing the " \
-            "context, with an incomplete sentence trailing\nanswer that " \
-            "plausibly completes the situation."
-
     def doc_to_text(self, doc):
         return doc["query"]
diff --git a/lm_eval/tasks/hendrycks_math.py b/lm_eval/tasks/hendrycks_math.py
index 379e727d61..bb864e86a4 100644
--- a/lm_eval/tasks/hendrycks_math.py
+++ b/lm_eval/tasks/hendrycks_math.py
@@ -55,9 +55,6 @@ def validation_docs(self):
     def test_docs(self):
         return self._load_docs(self.DATASET_PATH / "test" / self.get_file_info())
 
-    def fewshot_description(self):
-        return "Given a mathematics problem, determine the answer. Simplify your answer as much as possible."
-
     def doc_to_text(self, doc):
         return "Problem: " + doc["problem"] + "\nAnswer:"
 
diff --git a/lm_eval/tasks/lambada.py b/lm_eval/tasks/lambada.py
index bcb4ae019c..f9514ff920 100644
--- a/lm_eval/tasks/lambada.py
+++ b/lm_eval/tasks/lambada.py
@@ -47,10 +47,6 @@ def doc_to_text(self, doc):
 
     def doc_to_target(self, doc):
         return " " + doc['text'].rsplit(' ', 1)[1]
-    
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
 
     def construct_requests(self, doc, ctx):
         ll, is_greedy = rf.loglikelihood(ctx, self.doc_to_target(doc))
diff --git a/lm_eval/tasks/lambada_cloze.py b/lm_eval/tasks/lambada_cloze.py
index 90bd4f10ca..dc1d4b168b 100644
--- a/lm_eval/tasks/lambada_cloze.py
+++ b/lm_eval/tasks/lambada_cloze.py
@@ -13,6 +13,3 @@ def doc_to_text(self, doc):
 
     def doc_to_target(self, doc):
         return " " + doc['text'].rsplit(' ', 1)[1]
-    
-    def fewshot_description(self):
-        return "Fill in blank:\n"
diff --git a/lm_eval/tasks/logiqa.py b/lm_eval/tasks/logiqa.py
index ac544ca7c5..36b8b99e4f 100644
--- a/lm_eval/tasks/logiqa.py
+++ b/lm_eval/tasks/logiqa.py
@@ -80,9 +80,5 @@ def validation_docs(self):
     def test_docs(self):
         return self._load_docs(self.DATASET_PATH / "Test.txt")
 
-    def fewshot_description(self):
-        # TODO: figure out actual description
-        return ""
-
     def doc_to_text(self, doc):
         return doc["query"]
diff --git a/lm_eval/tasks/mathqa.py b/lm_eval/tasks/mathqa.py
index 84e5ab9eca..a02a5b59bb 100644
--- a/lm_eval/tasks/mathqa.py
+++ b/lm_eval/tasks/mathqa.py
@@ -29,9 +29,5 @@ def _convert_standard(self, doc):
         }
         return out_doc
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     def doc_to_text(self, doc):
         return doc["query"]
diff --git a/lm_eval/tasks/mc_taco.py b/lm_eval/tasks/mc_taco.py
index c9b2dd91fc..64a36a01f7 100644
--- a/lm_eval/tasks/mc_taco.py
+++ b/lm_eval/tasks/mc_taco.py
@@ -39,9 +39,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return True
 
-    def fewshot_description(self):
-        return "Determine whether the candidate answer is plausible (\"yes\") or not (\"no\")"
-
     def doc_to_text(self, doc):
         return f"{doc['sentence']}\nQuestion: {doc['question']}\n"\
             f"Answer: {doc['answer']}\nPlausible:"
diff --git a/lm_eval/tasks/mutual.py b/lm_eval/tasks/mutual.py
index 17274a46fd..5fe6ffcdb9 100644
--- a/lm_eval/tasks/mutual.py
+++ b/lm_eval/tasks/mutual.py
@@ -70,10 +70,6 @@ def validation_docs(self):
     def test_docs(self):
         return NotImplemented
 
-    def fewshot_description(self):
-        # TODO: figure out fewshot description
-        return ""
-
     def doc_to_text(self, doc):
         return self.detokenize(doc["article"])
 
diff --git a/lm_eval/tasks/naturalqs.py b/lm_eval/tasks/naturalqs.py
index f31875240f..e7a381dcd4 100644
--- a/lm_eval/tasks/naturalqs.py
+++ b/lm_eval/tasks/naturalqs.py
@@ -21,10 +21,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     def training_docs(self):
         # Cache training for faster few-shot.
         # Data is too large to fit in memory.
diff --git a/lm_eval/tasks/openbookqa.py b/lm_eval/tasks/openbookqa.py
index 40fc7a026b..5f87d8a8ec 100644
--- a/lm_eval/tasks/openbookqa.py
+++ b/lm_eval/tasks/openbookqa.py
@@ -25,9 +25,5 @@ def _convert_standard(self, doc):
         }
         return out_doc
 
-    def fewshot_description(self):
-        # TODO: figure out fewshot description
-        return ""
-
     def doc_to_text(self, doc):
         return doc["query"]
diff --git a/lm_eval/tasks/piqa.py b/lm_eval/tasks/piqa.py
index 8b43d1af03..bdf3ec35dc 100644
--- a/lm_eval/tasks/piqa.py
+++ b/lm_eval/tasks/piqa.py
@@ -18,10 +18,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out fewshot description
-        return ""
-
     def _convert_standard(self, doc):
         out_doc = {
             "goal": doc["goal"],
diff --git a/lm_eval/tasks/prost.py b/lm_eval/tasks/prost.py
index 1a634d17c8..f905de7266 100644
--- a/lm_eval/tasks/prost.py
+++ b/lm_eval/tasks/prost.py
@@ -36,13 +36,9 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return True
 
-    def fewshot_description(self):
-        # TODO: figure out fewshot description
-        return ""
-
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
+    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
         assert num_fewshot == 0, 'PROST is designed to probe models in a zero-shot fashion only.'
-        return super().fewshot_context(doc, num_fewshot, provide_description, rnd)
+        return super().fewshot_context(doc, num_fewshot, rnd, description)
 
     def _convert_standard(self, doc):
         out_doc = {
diff --git a/lm_eval/tasks/pubmedqa.py b/lm_eval/tasks/pubmedqa.py
index 14335a5de0..c597064f0c 100644
--- a/lm_eval/tasks/pubmedqa.py
+++ b/lm_eval/tasks/pubmedqa.py
@@ -23,11 +23,6 @@ def test_docs(self):
             # HF is labelled as train but its really just for testing
             return self.data["train"]
 
-    def fewshot_description(self):
-        # Average ctx length in labelled dataset is 238.9
-        # 2 few-shot exmamples pushes it beyond context window
-        return ""
-
     def doc_to_text(self, doc):
         ctxs = "\n".join(doc["context"]["contexts"])
         return "Abstract: {}\nQuestion: {}\nAnswer:".format(
diff --git a/lm_eval/tasks/qa4mre.py b/lm_eval/tasks/qa4mre.py
index 42e1b98c51..448710317e 100644
--- a/lm_eval/tasks/qa4mre.py
+++ b/lm_eval/tasks/qa4mre.py
@@ -67,9 +67,6 @@ def load_docs(self, textfilename, tfds=False):
                 out_doc['source'] = src
                 yield out_doc
 
-    def fewshot_description(self):
-        return ""
-
     def test_docs(self):
         return self.load_docs(f"data/qa4mre/QA4MRE-{self.YEAR}-EN_GS.xml")
 
diff --git a/lm_eval/tasks/quac.py b/lm_eval/tasks/quac.py
index bb02b1c4e3..c7ce752233 100644
--- a/lm_eval/tasks/quac.py
+++ b/lm_eval/tasks/quac.py
@@ -51,11 +51,6 @@ def validation_docs(self):
     def test_docs(self):
         raise NotImplementedError("QuAC has no test docs.")
     
-    def fewshot_description(self):
-        # TODO: figure out fewshot description
-        desc = "TITLE: Title of the context passage - subtitle of the passage\nPARAGRAPH: Passage describing the relevant information for answering questions.\n\nQ: Text of a question.\n\nA: Answer to the question, based on the passage. If it cannot be answered based on the passage, write CANNOTANSWER"
-        return desc
-
     def load_doc(self, myjson):
         docs = []
         for item in myjson:
diff --git a/lm_eval/tasks/race.py b/lm_eval/tasks/race.py
index d379884669..7145fa6f14 100644
--- a/lm_eval/tasks/race.py
+++ b/lm_eval/tasks/race.py
@@ -63,10 +63,6 @@ def validation_docs(self):
     def test_docs(self):
         return self._collate_data("test")
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     @classmethod
     def get_answer_option(cls, problem):
         answer = cls.letter_to_num[problem['answer']]
diff --git a/lm_eval/tasks/sat.py b/lm_eval/tasks/sat.py
index e4411edfd8..d75d7923b5 100644
--- a/lm_eval/tasks/sat.py
+++ b/lm_eval/tasks/sat.py
@@ -61,10 +61,5 @@ def validation_docs(self):
             }
             yield doc
 
-    
-    def fewshot_description(self):
-        # TODO: figure out actual description
-        return ""
-
     def doc_to_text(self, doc):
         return "{} is to {} as".format(*doc['query'])
diff --git a/lm_eval/tasks/sciq.py b/lm_eval/tasks/sciq.py
index 8993251d78..7e24f03ff6 100644
--- a/lm_eval/tasks/sciq.py
+++ b/lm_eval/tasks/sciq.py
@@ -50,9 +50,6 @@ def load_docs(self, textfilename):
         for record in docs:
             yield self._convert_standard(record)
 
-    def fewshot_description(self):
-        return ""
-
     def training_docs(self):
         return self.load_docs("data/sciq/SciQ dataset-2 3/train.json")
 
@@ -63,4 +60,4 @@ def test_docs(self):
         return self.load_docs("data/sciq/SciQ dataset-2 3/test.json")
 
     def doc_to_text(self, doc):
-        return "{}\nQuestion: {}\nAnswer:".format(doc["source"], doc["query"]).strip()
\ No newline at end of file
+        return "{}\nQuestion: {}\nAnswer:".format(doc["source"], doc["query"]).strip()
diff --git a/lm_eval/tasks/squad.py b/lm_eval/tasks/squad.py
index 72e1a19b0e..2a69a67c7b 100644
--- a/lm_eval/tasks/squad.py
+++ b/lm_eval/tasks/squad.py
@@ -41,10 +41,6 @@ def training_docs(self):
     def validation_docs(self):
         return self.data["validation"]
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     def doc_to_text(self, doc):
         return 'Title: ' + doc['title'] + '\n\n' + 'Background: ' + doc['context'] + '\n\n' + 'Question: ' + doc['question'] + '\n\n' + 'Answer:'
 
diff --git a/lm_eval/tasks/storycloze.py b/lm_eval/tasks/storycloze.py
index 3e178facb4..2cc16cf66d 100644
--- a/lm_eval/tasks/storycloze.py
+++ b/lm_eval/tasks/storycloze.py
@@ -27,18 +27,12 @@ def load_doc(self, filename):
             filereader = csv.reader(file)
             return list(filereader)
                 
-
     def validation_docs(self):
         return self.load_doc("data/storycloze/cloze_test_val__winter2018-cloze_test_ALL_val - 1 - 1.csv")
 
     def test_docs(self):
         return self.load_doc("data/storycloze/cloze_test_test__winter2018-cloze_test_ALL_test - 1.csv")
 
-    
-    def fewshot_description(self):
-        # TODO: figure out fewshot description
-        return ""
-    
     def doc_to_text(self, doc):
         return ' '.join([*doc[1:5]])
 
diff --git a/lm_eval/tasks/superglue.py b/lm_eval/tasks/superglue.py
index 33598f2301..c6d0bcfa2c 100644
--- a/lm_eval/tasks/superglue.py
+++ b/lm_eval/tasks/superglue.py
@@ -26,10 +26,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out actual description
-        return "Read the following passages and answer each question with a yes or a no."
-
     def doc_to_text(self, doc):
         return f"{doc['passage']}\nQuestion: {doc['question']}\nAnswer:"
     
@@ -78,11 +74,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out actual description
-        return "Given a premise and a hypothesis, classify whether the author of the premise is committed" \
-            "to the truth of the hypothesis. The three possible labels are true, false or neither."
-
     def doc_to_text(self, doc):
         return "{}\nQuestion: {}. True, False or Neither?\nAnswer:".format(
             doc["premise"],
@@ -150,11 +141,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out actual description
-        return "Given a premise and one alternative with a causal relation to the premise and another without," \
-            "choose the more plausible alternative"
-
     def doc_to_text(self, doc):
         # Drop the period
         connector = {
@@ -215,10 +201,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out actual description
-        return "READING COMPREHENSION ANSWER KEY"
-
     def doc_to_text(self, doc):
         return f"{doc['paragraph']}\nQuestion: {doc['question']}\nAnswer:"
 
@@ -270,10 +252,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out actual description
-        return ""
-
     def training_docs(self):
         # In ReCoRD, each doc manifests multiple "examples" in the context of few shot example packing.
         # Each doc consists of multiple answer candidates, each of which is scored yes/no.
@@ -363,10 +341,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return False
 
-    def fewshot_description(self):
-        # TODO: figure out actual description
-        return ""
-
     def doc_to_text(self, doc):
         return "Sentence 1: {}\nSentence 2: {}\nQuestion: Is the word '{}' used in the same way in the" \
                " two sentences above?\nAnswer:".format(
@@ -432,12 +406,6 @@ def training_docs(self):
                 ]
             return self._training_docs
 
-    def fewshot_description(self):
-        return "Final Exam with Answer Key\n" \
-           "Instructions: Please carefully read the following passages. " \
-           "For each passage, you must identify which noun the pronoun marked in *bold*" \
-           " refers to.\n====="
-
     def doc_to_text(self, doc):
         raw_passage = doc["text"]
         # NOTE: HuggingFace span indices are word-based not character-based.
diff --git a/lm_eval/tasks/translation.py b/lm_eval/tasks/translation.py
index de02946334..2e70b03a84 100644
--- a/lm_eval/tasks/translation.py
+++ b/lm_eval/tasks/translation.py
@@ -166,12 +166,6 @@ def higher_is_better(self):
             "ter": False,
         }
 
-    def fewshot_description(self):
-        language_codes = self.sacrebleu_language_pair.split("-")
-        src_lang = code_to_language(language_codes[0])
-        tar_lang = code_to_language(language_codes[1])
-        return f"Translate these {src_lang} phrases to {tar_lang}."
-
     def __str__(self):
         language_codes = self.sacrebleu_language_pair.split("-")
         src_lang = code_to_language(language_codes[0])
diff --git a/lm_eval/tasks/triviaqa.py b/lm_eval/tasks/triviaqa.py
index e61a40bdde..0579cd4f5f 100644
--- a/lm_eval/tasks/triviaqa.py
+++ b/lm_eval/tasks/triviaqa.py
@@ -36,10 +36,6 @@ def validation_docs(self):
     def test_docs(self):
         raise NotImplementedError()
     
-    def fewshot_description(self):
-        # TODO: figure out fewshot description
-        return ""
-    
     def doc_to_text(self, doc):
         return f"Question: {doc['Question']}\nAnswer:"
 
@@ -56,7 +52,6 @@ def _remove_prefixes(self, aliases):
                 ret.append(alias)
 
         return ret
-        
 
     def construct_requests(self, doc, ctx):
         ret = []
diff --git a/lm_eval/tasks/truthfulqa.py b/lm_eval/tasks/truthfulqa.py
index e3c621c4e2..7380c65555 100644
--- a/lm_eval/tasks/truthfulqa.py
+++ b/lm_eval/tasks/truthfulqa.py
@@ -85,9 +85,9 @@ def doc_to_text(self, doc):
     def doc_to_target(self, doc):
         return " "
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
+    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
         assert num_fewshot == 0, "TruthfulQA is intended only for the zero-shot setting."
-        return super().fewshot_context(doc, num_fewshot, provide_description, rnd)
+        return super().fewshot_context(doc, num_fewshot, rnd, description)
 
     def construct_requests(self, doc, ctx):
         """ Uses RequestFactory to construct Requests and returns an iterable of
@@ -213,9 +213,9 @@ def doc_to_text(self, doc):
     def doc_to_target(self, doc):
         return " "
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
+    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
         assert num_fewshot == 0, "TruthfulQA is intended only for the zero-shot setting."
-        return super().fewshot_context(doc, num_fewshot, provide_description, rnd)
+        return super().fewshot_context(doc, num_fewshot, rnd, description)
 
     def construct_requests(self, doc, ctx):
         """ Uses RequestFactory to construct Requests and returns an iterable of
diff --git a/lm_eval/tasks/unscramble.py b/lm_eval/tasks/unscramble.py
index dc742a2cee..5e90dd358c 100644
--- a/lm_eval/tasks/unscramble.py
+++ b/lm_eval/tasks/unscramble.py
@@ -45,9 +45,6 @@ def validation_docs(self):
         file = self.BASE_PATH / self.FILENAME
         return (json.loads(line) for line in open(file).read().splitlines())
 
-    def fewshot_description(self):
-        return "Please unscramble the letters into a word, and write that word:"
-
     def doc_to_text(self, doc):
         return doc["context"]
 
diff --git a/lm_eval/tasks/webqs.py b/lm_eval/tasks/webqs.py
index 51ed016758..ebab7c8968 100644
--- a/lm_eval/tasks/webqs.py
+++ b/lm_eval/tasks/webqs.py
@@ -17,10 +17,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return True
 
-    def fewshot_description(self):
-        # TODO: figure out description
-        return ""
-
     def doc_to_text(self, doc):
         return "Question: " + doc['question'] + '\nAnswer:'
 
@@ -40,7 +36,6 @@ def _remove_prefixes(self, aliases):
                 ret.append(alias)
 
         return ret
-        
 
     def construct_requests(self, doc, ctx):
         ret = []
@@ -62,4 +57,4 @@ def aggregation(self):
     def higher_is_better(self):
         return {
             "acc": True
-        }
\ No newline at end of file
+        }
diff --git a/lm_eval/tasks/wikitext.py b/lm_eval/tasks/wikitext.py
index 24f9ec3507..3ef0b4b6ed 100644
--- a/lm_eval/tasks/wikitext.py
+++ b/lm_eval/tasks/wikitext.py
@@ -49,10 +49,6 @@ def download(self):
             download_file("https://s3.amazonaws.com/research.metamind.io/wikitext/wikitext-2-raw-v1.zip", "data/wikitext/wikitext-2-raw-v1.zip", "ef7edb566e3e2b2d31b29c1fdb0c89a4cc683597484c3dc2517919c615435a11")
             sh("cd data/wikitext/ && unzip wikitext-2-raw-v1.zip")
 
-    def fewshot_description(self):
-        # TODO: figure out fewshot description
-        return ""
-
     def has_validation_docs(self):
         return True
 
@@ -87,4 +83,4 @@ def doc_to_target(self, doc):
     
     def count_words(self, doc):
         # count number of words in *original doc before detokenization*
-        return len(re.split(r"\s+", doc))
\ No newline at end of file
+        return len(re.split(r"\s+", doc))
diff --git a/lm_eval/tasks/winogrande.py b/lm_eval/tasks/winogrande.py
index 106c826a69..2e188d7b30 100644
--- a/lm_eval/tasks/winogrande.py
+++ b/lm_eval/tasks/winogrande.py
@@ -29,10 +29,6 @@ def has_test_docs(self):
     def doc_to_text(self, doc):
         return self.partial_context(doc, doc["option" + doc["answer"]])
 
-    def fewshot_description(self):
-        # TODO: redo description
-        return "Winograd schema sentence including a either a ___ blank with a missing word, making the pronoun ambiguous, or the same with the word filled in."
-
     @classmethod
     def partial_context(cls, doc, option):
         # Substitute the pronoun in the sentence with the specified option
diff --git a/lm_eval/tasks/wsc273.py b/lm_eval/tasks/wsc273.py
index 20dd5175b6..505557b15c 100644
--- a/lm_eval/tasks/wsc273.py
+++ b/lm_eval/tasks/wsc273.py
@@ -53,10 +53,6 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return True
 
-    def fewshot_description(self):
-        # TODO: redo description
-        return "Winograd schema sentence with correct continuation. True. Winograd schema sentence with incorrect continuation. False."
-
     def fewshot_examples(self, k, rnd):
         # NOTE: `super().fewshot_examples` samples from training docs which are
         # not available for this test-set-only dataset.
diff --git a/main.py b/main.py
index 7a973e197b..efdf29ba2c 100644
--- a/main.py
+++ b/main.py
@@ -13,7 +13,7 @@ def parse_args():
     parser.add_argument('--model', required=True)
     parser.add_argument('--model_args', default="")
     parser.add_argument('--tasks', default="all_tasks")
-    parser.add_argument('--provide_description', action="store_true")
+    parser.add_argument('--description_path', default=None)
     parser.add_argument('--num_fewshot', type=int, default=0)
     parser.add_argument('--batch_size', type=int, default=None)
     parser.add_argument('--device', type=str, default=None)
@@ -26,8 +26,6 @@ def main():
 
     args = parse_args()
 
-    assert not args.provide_description # not implemented
-    
     if args.limit:
         print("WARNING: --limit SHOULD ONLY BE USED FOR TESTING. REAL METRICS SHOULD NOT BE COMPUTED USING LIMIT.")
 
@@ -36,7 +34,17 @@ def main():
     else:
         task_names = args.tasks.split(",")
 
-    results = evaluator.simple_evaluate(args.model, args.model_args, task_names, args.num_fewshot, args.batch_size, args.device, args.no_cache, args.limit)
+    results = evaluator.simple_evaluate(
+        args.model,
+        args.model_args,
+        task_names,
+        args.description_path,
+        args.num_fewshot,
+        args.batch_size,
+        args.device,
+        args.no_cache,
+        args.limit
+    )
 
     dumped = json.dumps(results, indent=2)
     
@@ -46,7 +54,7 @@ def main():
         with open(args.output_path, "w") as f:
             f.write(dumped)
 
-    print(f"{args.model} ({args.model_args}), limit: {args.limit}, provide_description: {args.provide_description}, num_fewshot: {args.num_fewshot}, batch_size: {args.batch_size}")
+    print(f"{args.model} ({args.model_args}), limit: {args.limit}, num_fewshot: {args.num_fewshot}, batch_size: {args.batch_size}")
     print(evaluator.make_table(results))
 
 if __name__ == "__main__":
diff --git a/scripts/cost_estimate.py b/scripts/cost_estimate.py
index 4339b8dbd2..5f15079315 100644
--- a/scripts/cost_estimate.py
+++ b/scripts/cost_estimate.py
@@ -51,7 +51,7 @@ def main():
     values = []
     for taskname in task_list.split(","):
         lm.tokencost = 0
-        evaluator.evaluate(lm, {taskname: tasks.get_task(taskname)()}, False, 0, None, bootstrap_iters=10)
+        evaluator.evaluate(lm, {taskname: tasks.get_task(taskname)()}, 0, None, bootstrap_iters=10)
 
         print(taskname, lm.tokencost)
         values.append([taskname, lm.tokencost, lm.tokencost / 1000 * 0.0008, lm.tokencost / 1000 * 0.0012, lm.tokencost / 1000 * 0.006, lm.tokencost / 1000 * 0.06])
diff --git a/scripts/fewshot_description_experiment.py b/scripts/fewshot_description_experiment.py
deleted file mode 100644
index e6ad97b340..0000000000
--- a/scripts/fewshot_description_experiment.py
+++ /dev/null
@@ -1,79 +0,0 @@
-import json
-import numpy as np
-import random
-import logging
-from lm_eval import models, tasks, evaluator, base
-
-logging.getLogger("openai").setLevel(logging.WARNING)
-
-
-fewshot_descriptions = [
-    "foo",
-    "bar"
-]
-
-task = "lambada"
-num_fewshot = 0
-model = "gpt2"
-model_args = ""
-limit = None
-no_cache = False
-
-
-class CustomDescTask:
-    def __init__(self, task, desc):
-        self.task = task
-        self.desc = desc
-
-        def fewshot_description():
-            return self.desc
-        
-        self.task.fewshot_description = fewshot_description
-
-    def __getattr__(self, attr):
-        return getattr(self.task, attr)
-
-
-def main():
-    random.seed(42)
-    np.random.seed(42)
-
-    lm = models.get_model(model).create_from_arg_string(model_args)
-    
-    if limit:
-        print("WARNING: --limit SHOULD ONLY BE USED FOR TESTING. REAL METRICS SHOULD NOT BE COMPUTED USING LIMIT.")
-
-    if not no_cache:
-        lm = base.CachingLM(lm, 'lm_cache/' + model + '_' + model_args.replace('=', '-').replace(',', '_') + '.db')
-
-    task_dict = tasks.get_task_dict([task])
-
-    for desc in fewshot_descriptions:
-        custom_task_dict = {k: CustomDescTask(v, desc) for k, v in task_dict.items()}
-
-        results = evaluator.evaluate(lm, custom_task_dict, True, num_fewshot, limit)
-
-        dumped = json.dumps(results, indent=2)
-
-        print('Description:', desc)
-        print(dumped)
-
-        # MAKE TABLE
-        from pytablewriter import MarkdownTableWriter
-
-        writer = MarkdownTableWriter()
-        writer.headers = ["Task", "Metric", "Value"]
-
-        values = []
-
-        for k, dic in results.items():
-            for m, v in dic.items():
-                values.append([k, m, '%.4f' % v])
-                k = ""
-        writer.value_matrix = values
-
-        print(writer.dumps())
-
-
-if __name__ == "__main__":
-    main()
diff --git a/scripts/get_prompts.py b/scripts/get_prompts.py
index cdda6dc43e..56a9ff79f4 100644
--- a/scripts/get_prompts.py
+++ b/scripts/get_prompts.py
@@ -9,7 +9,6 @@
     print('#', tname)
     docs = islice(task.validation_docs() if task.has_validation_docs() else task.test_docs(), ct)
     print()
-    print('**Zero-Shot Prompt**:', "\n```\n" + task.fewshot_description() + "\n```\n")
     for i in range(ct):
         print()
         doc = next(docs)
diff --git a/scripts/write_out.py b/scripts/write_out.py
index b7eb30c15a..bd63168acf 100644
--- a/scripts/write_out.py
+++ b/scripts/write_out.py
@@ -1,5 +1,6 @@
 import argparse
 import numpy as np
+import json
 import os
 import random
 from lm_eval import tasks
@@ -12,7 +13,7 @@ def parse_args():
     parser = argparse.ArgumentParser()
     parser.add_argument('--output_base_path', required=True)
     parser.add_argument('--tasks', default="all_tasks")
-    parser.add_argument('--provide_description', action="store_true")
+    parser.add_argument('--description_path', default=None)
     parser.add_argument('--sets', type=str, default="val") # example: val,test
     parser.add_argument('--num_fewshot', type=int, default=1)
     parser.add_argument('--seed', type=int, default=42)
@@ -29,6 +30,12 @@ def main():
     else:
         task_names = args.tasks.split(",")
     task_dict = tasks.get_task_dict(task_names)
+
+    description_dict = {}
+    if args.description_path:
+        with open(args.description_path, 'r') as f:
+            description_dict = json.load(f)
+
     os.makedirs(args.output_base_path, exist_ok=True)
     for task_name, task in task_dict.items():
         rnd = random.Random()
@@ -47,14 +54,16 @@ def main():
 
         docs = join_iters(iters)
 
+        description = description_dict[task_name] if description_dict and task_name in description_dict else ""
+
         with open(os.path.join(args.output_base_path, task_name), "w") as f:
             for i, doc in zip(range(args.num_examples), docs) if args.num_examples > 0 else enumerate(docs):
                 f.write(EXAMPLE_DIVIDER.format(i=i))
                 ctx = task.fewshot_context(
                     doc=doc,
-                    provide_description=args.provide_description,
                     num_fewshot=args.num_fewshot,
-                    rnd=rnd
+                    rnd=rnd,
+                    description=description
                 )
                 f.write(ctx + "\n")
 
diff --git a/task-guide.md b/task-guide.md
index 5ea43fc2f4..7bee0ab2c2 100644
--- a/task-guide.md
+++ b/task-guide.md
@@ -87,8 +87,7 @@ There are 2 standard approaches we follow for downloading data:
     ```
    These methods return `True`/`False` whether or not your task dataset provides documents for each split type. __Note__: if the test set doesn't have publicly available labels, please do not put it down as having a test set.
 
-	Lastly, we need to load the documents. In our terminology, a document (`doc`) is a single natural language data example stored in a Python `dict`. E.g.:
-	`{“question”: “What is the capital of France?”, “answer”: “Paris”}`. Override the following methods to load your data splits from their storage location in `DATASET_PATH`:
+	Lastly, we need to load the documents. In our terminology, a document (`doc`) is a single natural language data example stored in a Python `dict`. E.g.: `{“question”: “What is the capital of France?”, “answer”: “Paris”}`. Override the following methods to load your data splits from their storage location in `DATASET_PATH`:
     ```python
     def training_docs(self):
         return #...
@@ -117,7 +116,7 @@ class TaskName(..., MultipleChoiceTask):
 
 This will require you to format your documents such that they contain `gold` and `choices` fields. They can also have other fields, but those will be ignored by `MultipleChoiceTask`. `choices` should be a list of possible continuations, and `gold` should be an integer specifying the index of the correct completion.
 
-See [this task](https://github.com/EleutherAI/lm-evaluation-harness/blob/105fa9741ff660f6a62c2eef0d2facfde36dda41/lm_eval/tasks/sat.py#L56) for an example. When used in combination with `HFTask`, it may be useful to override [`_convert_standard`](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/common.py#L28), which will be applied to every document in the HF dataset. See [this task](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/headqa.py) for an example of this.
+See [this task](https://github.com/EleutherAI/lm-evaluation-harness/blob/105fa9741ff660f6a62c2eef0d2facfde36dda41/lm_eval/tasks/sat.py#L56) for an example. When used in combination with `HFTask`, it may be useful to override [`_convert_standard`](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/common.py#L28), which will be applied to every document in the HF dataset. See task](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/headqa.py) for an example of this.
 
 You can now skip ahead to <a href="#Registering-Your-Task">registering your task</a>.
 
@@ -125,17 +124,9 @@ You can now skip ahead to <a href="#Registering-Your-Task">registering your task
 
 <br>
 
+In the case your task is _not_ multiple-choice, override the following methods for your task class:
 
-In the case your task is not multiple-choice, override the following methods for your task class:
-
-Put the natural language task description as a single line (no `\n`s) string here. E.g. `"Translate English to French:"`
-
-```python
-def fewshot_description(self):
-    return ""
-```
-
-Format your document into a single query prompt __without the answer__ here. This method takes a single `doc` example (in dictionary form) . You should concatenate its members into a nicely formatted prompt.
+Format your document into a single query prompt __without the answer__ here. This method takes a single `doc` example of type `dict` with `str` key-value members. You should concatenate these `doc` item values together into a neatly formatted prompt.
 
 ```python
 def doc_to_text(self, doc):
@@ -151,6 +142,41 @@ def doc_to_target(self, doc):
 
 Understand that the strings from `doc_to_text` and `doc_to_target` will be concatenated together to build up labeled examples in the k-shot setting where k > 0. Design with that in mind 👍.
 
+### Formatting Prompts
+
+If you'd like to prepend your few-shot examples with a natural language description or provide a lone custom prompt under a zero-shot setting, you can do this on a per-task basis via the `description_dict` arg of `evaluator.evaluate` which is accessible through the `evaluator` module. This `description_dict` must adhere to the following key-value structure:
+
+- **key**: the task name as specified in the lm-eval-harness task registry (see the following section on task registry).
+- **value**: the corresponding description/prompt for the task identified by **key**.
+
+E.g.
+
+```python
+description_dict = {
+    "task_name_1": "task_name_1 custom prompt or few-shot task description",
+    "task_name_2": "task_name_2 custom prompt or few-shot task description",
+    ...
+}
+```
+
+At a higher level, one can interface with `evaluator.evaluate` by simply passing a JSON file path to the `description_path` arg of the command-line interface program, `main.py`. The JSON file pointed to should be structured the same way as the aforementioned `description_dict`. E.g. for some file at `/your/path/descriptions.json` you might have:
+
+```json
+{
+    "cycle_letters": "Please unscramble the letters into a word, and write that word:",
+    "copa": "Given a premise and one alternative with a causal relation to the premise and another without, choose the more plausible alternative"
+}
+```
+
+which can then be hooked up to the evaluator through the `main.py` CLI as:
+
+```python
+python main.py  \
+--tasks cycle_letters,copa \
+--description_path /your/path/descriptions.json \
+...
+```
+
 ### Registering Your Task
 
 Now's a good time to register your task to expose it for usage. All you'll need to do is import your task module in `lm_eval/tasks/__init__.py` and provide an entry in the `TASK_REGISTRY`  dictionary with the key as the name of your benchmark task (in the form it'll be referred to in the command line) and the value as the task class. See how it's done for other tasks in the [file](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/__init__.py).
@@ -161,7 +187,7 @@ After registering your task, you can now check on your data downloading and veri
 
 ```bash
 python -m scripts.write_out \
-    --task <your-task> \
+    --tasks <your-task> \
     --output_base_path <path> \
     --sets <train | val | test> \
     --num_fewshot K \
diff --git a/tests/test_description_option/descriptions.json b/tests/test_description_option/descriptions.json
new file mode 100644
index 0000000000..6bace5dac5
--- /dev/null
+++ b/tests/test_description_option/descriptions.json
@@ -0,0 +1,3 @@
+{
+    "hellaswag": "Label for the relevant action:\nSentences describing context, with an incomplete sentence trailing answer that plausibly completes the situation."
+}
diff --git a/tests/test_description_option/test_description_option.py b/tests/test_description_option/test_description_option.py
new file mode 100644
index 0000000000..fd6b7b16f9
--- /dev/null
+++ b/tests/test_description_option/test_description_option.py
@@ -0,0 +1,39 @@
+import json
+import argparse
+import lm_eval.tasks
+import lm_eval.models
+from lm_eval.evaluator import evaluate
+
+
+def parse_args():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--description_path', default=None)
+    parser.add_argument('--num_fewshot', type=int, default=0)
+    parser.add_argument('--limit', type=int, default=None)
+    return parser.parse_args()
+
+
+def main():
+    args = parse_args()
+
+    task_names = ['hellaswag', 'copa']
+    task_dict = lm_eval.tasks.get_task_dict(task_names)
+    lm = lm_eval.models.get_model('dummy')()
+
+    description_dict = {}
+    if args.description_path:
+        with open(args.description_path, 'r') as f:
+            description_dict = json.load(f)
+
+    num_fewshot = args.num_fewshot
+    results = evaluate(
+        lm,
+        task_dict,
+        num_fewshot,
+        args.limit,
+        description_dict
+    )
+
+
+if __name__ == '__main__':
+    main()
diff --git a/tests/test_evaluator.py b/tests/test_evaluator.py
index 8b6ba9c2e3..0936d77a0a 100644
--- a/tests/test_evaluator.py
+++ b/tests/test_evaluator.py
@@ -47,8 +47,8 @@ def ll_perp_fn(reqs):
     lm.loglikelihood_rolling = ll_perp_fn
 
     limit = 10
-    e1 = evaluator.evaluate(lm, task_dict, False, 0, limit, bootstrap_iters=10)
-    e2 = evaluator.evaluate(lm, task_dict, False, 0, limit, bootstrap_iters=10)
+    e1 = evaluator.evaluate(lm, task_dict, 0, limit, description_dict=None, bootstrap_iters=10)
+    e2 = evaluator.evaluate(lm, task_dict, 0, limit, description_dict=None, bootstrap_iters=10)
 
     # check taht caching is working
     assert e1 == e2
diff --git a/tests/test_version_stable.py b/tests/test_version_stable.py
index 12933d61c8..d1e7a46f1a 100644
--- a/tests/test_version_stable.py
+++ b/tests/test_version_stable.py
@@ -81,5 +81,5 @@ def greedy_until(reqs):
     lm.greedy_until = greedy_until
 
     limit = None
-    res = evaluator.evaluate(lm, task_dict, False, 0, limit, bootstrap_iters=10)
+    res = evaluator.evaluate(lm, task_dict, 0, limit, description_dict=None, bootstrap_iters=10)
     assert_target(f"{taskname}-v{Task.VERSION}-res", res)

From 1d04c42d2c9b92094c0fd6b7497120ce76984590 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Wed, 15 Dec 2021 18:18:55 -0500
Subject: [PATCH 02/65] Merge

---
 lm_eval/base.py      | 17 +++++++++++++----
 lm_eval/evaluator.py | 10 +++++-----
 main.py              |  3 ++-
 scripts/write_out.py |  8 +++++---
 4 files changed, 25 insertions(+), 13 deletions(-)

diff --git a/lm_eval/base.py b/lm_eval/base.py
index 0ad023335b..e9661164d0 100644
--- a/lm_eval/base.py
+++ b/lm_eval/base.py
@@ -452,12 +452,17 @@ def higher_is_better(self):
     def fewshot_description(self):
         import warnings
         warnings.warn(
-            "`fewshot_description` will be removed in coming versions. Pass " \
+            "`fewshot_description` will be removed in futures versions. Pass " \
             "any custom descriptions to the `evaluate` function instead.",
             DeprecationWarning)
         return ""
 
-    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
+        assert not provide_description, (
+            "The `provide_description` arg will be removed in future versions. To provide "
+            "custom descriptions on a per-task basis, supply the `description_dict` "
+            "arg with your task-to-description dictionary."
+        )
         description = description + "\n\n" if description else ""
 
         if num_fewshot == 0:
@@ -531,9 +536,13 @@ def fewshot_examples(self, k, rnd):
         assert k == 0
         return []
 
-    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert num_fewshot == 0
-        assert description is None 
+        assert not provide_description, (
+            "The `provide_description` arg will be removed in future versions. To provide "
+            "custom descriptions on a per-task basis, supply the `description_dict` "
+            "arg with your task-to-description dictionary."
+        )
         return ""
 
     def higher_is_better(self):
diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index c84e4528f7..dc2dae4e05 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -48,12 +48,13 @@ def simple_evaluate(model, model_args, task_names,
         )
     
     task_dict = lm_eval.tasks.get_task_dict(task_names)
+
     description_dict = {}
-    if description_path:
-        with open(description_path, 'r') as f:
+    if description_dict_path:
+        with open(description_dict_path, 'r') as f:
             description_dict = json.load(f)
 
-    results = evaluate(lm, task_dict, num_fewshot, limit, description_dict)
+    results = evaluate(lm, task_dict, False, num_fewshot, limit, description_dict=description_dict)
 
     # add info about the model and few shot config
     results["config"] = {
@@ -62,8 +63,6 @@ def simple_evaluate(model, model_args, task_names,
         "num_fewshot": num_fewshot,
         "batch_size": batch_size,
         "device": device,
-        # TODO (jon-tow): Should we add the description info to `results["config"]`?
-        # "description_dict": description_dict,
         "no_cache": no_cache,
         "limit": limit,
         "bootstrap_iters": bootstrap_iters
@@ -140,6 +139,7 @@ def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_i
             ctx = task.fewshot_context(
                 doc=doc,
                 num_fewshot=num_fewshot,
+                provide_description=provide_description,
                 rnd=rnd,
                 description=description
             )
diff --git a/main.py b/main.py
index 1466e6a26a..e55015b4a2 100644
--- a/main.py
+++ b/main.py
@@ -12,13 +12,14 @@ def parse_args():
     parser.add_argument('--model', required=True)
     parser.add_argument('--model_args', default="")
     parser.add_argument('--tasks', default="all_tasks")
-    parser.add_argument('--description_path', default=None)
+    parser.add_argument('--provide_description', action="store_true")
     parser.add_argument('--num_fewshot', type=int, default=0)
     parser.add_argument('--batch_size', type=int, default=None)
     parser.add_argument('--device', type=str, default=None)
     parser.add_argument('--output_path', default=None)
     parser.add_argument('--limit', type=int, default=None)
     parser.add_argument('--no_cache', action="store_true")
+    parser.add_argument('--description_dict_path', default=None)
     return parser.parse_args()
 
 
diff --git a/scripts/write_out.py b/scripts/write_out.py
index bd63168acf..b39fd64541 100644
--- a/scripts/write_out.py
+++ b/scripts/write_out.py
@@ -13,11 +13,12 @@ def parse_args():
     parser = argparse.ArgumentParser()
     parser.add_argument('--output_base_path', required=True)
     parser.add_argument('--tasks', default="all_tasks")
-    parser.add_argument('--description_path', default=None)
+    parser.add_argument('--provide_description', action="store_true")
     parser.add_argument('--sets', type=str, default="val") # example: val,test
     parser.add_argument('--num_fewshot', type=int, default=1)
     parser.add_argument('--seed', type=int, default=42)
     parser.add_argument('--num_examples', type=int, default=1)
+    parser.add_argument('--description_dict_path', default=None)
     return parser.parse_args()
 
 
@@ -32,8 +33,8 @@ def main():
     task_dict = tasks.get_task_dict(task_names)
 
     description_dict = {}
-    if args.description_path:
-        with open(args.description_path, 'r') as f:
+    if args.description_dict_path:
+        with open(args.description_dict_path, 'r') as f:
             description_dict = json.load(f)
 
     os.makedirs(args.output_base_path, exist_ok=True)
@@ -62,6 +63,7 @@ def main():
                 ctx = task.fewshot_context(
                     doc=doc,
                     num_fewshot=args.num_fewshot,
+                    provide_description=args.provide_description,
                     rnd=rnd,
                     description=description
                 )

From ee53be219bf3f1ef485a725db6b7047b09d0eb63 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Wed, 15 Dec 2021 18:26:19 -0500
Subject: [PATCH 03/65] Add `provide_description` arg for backward compat

---
 lm_eval/evaluator.py                          |  9 ++++++--
 lm_eval/tasks/hendrycks_ethics.py             |  6 -----
 lm_eval/tasks/hendrycks_test.py               |  4 ----
 lm_eval/tasks/prost.py                        |  4 ++--
 lm_eval/tasks/truthfulqa.py                   |  8 +++----
 scripts/cost_estimate.py                      |  2 +-
 task-guide.md                                 | 15 ++++++------
 .../descriptions.json                         |  0
 .../test_description_dict.py}                 | 23 +++++++++----------
 tests/test_evaluator.py                       |  4 ++--
 10 files changed, 35 insertions(+), 40 deletions(-)
 rename tests/{test_description_option => test_description_dict}/descriptions.json (100%)
 rename tests/{test_description_option/test_description_option.py => test_description_dict/test_description_dict.py} (52%)

diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index dc2dae4e05..4c224edef3 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -11,7 +11,8 @@
 
 def simple_evaluate(model, model_args, task_names,
                     num_fewshot=0, batch_size=None, device=None,
-                    no_cache=False, limit=None, bootstrap_iters=100000):
+                    no_cache=False, limit=None, bootstrap_iters=100000,
+                    description_dict_path=None):
     """Instantiate and evaluate a model on a list of tasks.
 
     :param model: str
@@ -32,6 +33,8 @@ def simple_evaluate(model, model_args, task_names,
         Limit the number of examples per task (only use this for testing)
     :param bootstrap_iters:
         Number of iterations for bootstrap statistics
+    :param description_dict_path:
+        Path to a JSON file containing `task_name: description` key-values for custom prompts
     :return
         Dictionary of results
     """
@@ -71,7 +74,7 @@ def simple_evaluate(model, model_args, task_names,
     return results
 
 
-def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_iters=100000):
+def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_iters=100000, description_dict=None):
     """Instantiate and evaluate a model on a list of tasks.
 
     :param lm: obj
@@ -86,6 +89,8 @@ def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_i
         Limit the number of examples per task (only use this for testing)
     :param bootstrap_iters:
         Number of iterations for bootstrap statistics
+    :param description_dict:
+        Dictionary of task descriptions of the form: `task_name: description` 
     :return
         Dictionary of results
     """
diff --git a/lm_eval/tasks/hendrycks_ethics.py b/lm_eval/tasks/hendrycks_ethics.py
index 50e94a508c..75369ed237 100644
--- a/lm_eval/tasks/hendrycks_ethics.py
+++ b/lm_eval/tasks/hendrycks_ethics.py
@@ -237,9 +237,6 @@ def process_doc(self, docs):
         for doc in docs:
             yield {"activity": doc[0], "baseline": doc[1], "rating": ""}
 
-    def fewshot_description(self):
-        return "Rate how pleasant each of the following activities is on a scale from 1 (very unpleasant) to 10 (very pleasant).\n\n"
-
     def fewshot_examples(self, k, rnd):
         # Overwriting fewshot examples as k can be max 5
         assert k <= 5, "There are only 5 possible shots for this task. Refer to the V2 for more."
@@ -350,9 +347,6 @@ class EthicsVirtue(Ethics):
     def get_prefix(self):
         return "virtue/virtue"
 
-    def fewshot_description(self):
-        return "The following is a list of sentences and traits, along with whether the trait is exhibited in that sentence.\n\n"
-
     def process_doc(self, doc):
         # Append identifiers before shuffling to calculate exact matches lateron & skip the first element of headers
         return [x + [i] for i, x in enumerate(doc[1:])]
diff --git a/lm_eval/tasks/hendrycks_test.py b/lm_eval/tasks/hendrycks_test.py
index 46c0306fcd..a505f3a134 100644
--- a/lm_eval/tasks/hendrycks_test.py
+++ b/lm_eval/tasks/hendrycks_test.py
@@ -114,9 +114,5 @@ def fewshot_examples(self, k, rnd):
 
         return rnd.sample(list(self._fewshot_docs), k)
 
-    def fewshot_description(self):
-        subject = self.subject.replace("_", " ")
-        return f"The following are multiple choice questions (with answers) about {subject}."
-
     def doc_to_text(self, doc):
         return doc["query"]
diff --git a/lm_eval/tasks/prost.py b/lm_eval/tasks/prost.py
index f905de7266..d53ece825b 100644
--- a/lm_eval/tasks/prost.py
+++ b/lm_eval/tasks/prost.py
@@ -36,9 +36,9 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return True
 
-    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert num_fewshot == 0, 'PROST is designed to probe models in a zero-shot fashion only.'
-        return super().fewshot_context(doc, num_fewshot, rnd, description)
+        return super().fewshot_context(doc, num_fewshot, provide_description, rnd, description)
 
     def _convert_standard(self, doc):
         out_doc = {
diff --git a/lm_eval/tasks/truthfulqa.py b/lm_eval/tasks/truthfulqa.py
index b02595d223..9fe941f614 100644
--- a/lm_eval/tasks/truthfulqa.py
+++ b/lm_eval/tasks/truthfulqa.py
@@ -85,9 +85,9 @@ def doc_to_text(self, doc):
     def doc_to_target(self, doc):
         return " "
 
-    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert num_fewshot == 0, "TruthfulQA is intended only for the zero-shot setting."
-        return super().fewshot_context(doc, num_fewshot, rnd, description)
+        return super().fewshot_context(doc, num_fewshot, provide_description, rnd, description)
 
     def construct_requests(self, doc, ctx):
         """ Uses RequestFactory to construct Requests and returns an iterable of
@@ -217,9 +217,9 @@ def doc_to_text(self, doc):
     def doc_to_target(self, doc):
         return " "
 
-    def fewshot_context(self, doc, num_fewshot, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert num_fewshot == 0, "TruthfulQA is intended only for the zero-shot setting."
-        return super().fewshot_context(doc, num_fewshot, rnd, description)
+        return super().fewshot_context(doc, num_fewshot, provide_description, rnd, description)
 
     def construct_requests(self, doc, ctx):
         """ Uses RequestFactory to construct Requests and returns an iterable of
diff --git a/scripts/cost_estimate.py b/scripts/cost_estimate.py
index 5f15079315..4339b8dbd2 100644
--- a/scripts/cost_estimate.py
+++ b/scripts/cost_estimate.py
@@ -51,7 +51,7 @@ def main():
     values = []
     for taskname in task_list.split(","):
         lm.tokencost = 0
-        evaluator.evaluate(lm, {taskname: tasks.get_task(taskname)()}, 0, None, bootstrap_iters=10)
+        evaluator.evaluate(lm, {taskname: tasks.get_task(taskname)()}, False, 0, None, bootstrap_iters=10)
 
         print(taskname, lm.tokencost)
         values.append([taskname, lm.tokencost, lm.tokencost / 1000 * 0.0008, lm.tokencost / 1000 * 0.0012, lm.tokencost / 1000 * 0.006, lm.tokencost / 1000 * 0.06])
diff --git a/task-guide.md b/task-guide.md
index 7bee0ab2c2..bf63dac7d4 100644
--- a/task-guide.md
+++ b/task-guide.md
@@ -144,7 +144,7 @@ Understand that the strings from `doc_to_text` and `doc_to_target` will be conca
 
 ### Formatting Prompts
 
-If you'd like to prepend your few-shot examples with a natural language description or provide a lone custom prompt under a zero-shot setting, you can do this on a per-task basis via the `description_dict` arg of `evaluator.evaluate` which is accessible through the `evaluator` module. This `description_dict` must adhere to the following key-value structure:
+If you'd like to prepend your few-shot examples with a natural language description or provide a lone custom prompt for a zero-shot task, you can do so on a per-task basis via the `description_dict` arg of `evaluator.evaluate` which is accessible from the `evaluator` module. This `description_dict` must adhere to the following key-value structure:
 
 - **key**: the task name as specified in the lm-eval-harness task registry (see the following section on task registry).
 - **value**: the corresponding description/prompt for the task identified by **key**.
@@ -153,13 +153,13 @@ E.g.
 
 ```python
 description_dict = {
-    "task_name_1": "task_name_1 custom prompt or few-shot task description",
-    "task_name_2": "task_name_2 custom prompt or few-shot task description",
+    "task_name_1": "fewshot description",
+    "task_name_2": "fewshot description",
     ...
 }
 ```
 
-At a higher level, one can interface with `evaluator.evaluate` by simply passing a JSON file path to the `description_path` arg of the command-line interface program, `main.py`. The JSON file pointed to should be structured the same way as the aforementioned `description_dict`. E.g. for some file at `/your/path/descriptions.json` you might have:
+One can also interface with `evaluator.evaluate` from a higher level by simply passing a JSON file path to the `description_dict_path` arg of the command-line interface program, `main.py`. The JSON file pointed to should be structured the same way as the aforementioned `description_dict`. E.g. for some file at `/your/path/descriptions.json` you might have:
 
 ```json
 {
@@ -173,7 +173,7 @@ which can then be hooked up to the evaluator through the `main.py` CLI as:
 ```python
 python main.py  \
 --tasks cycle_letters,copa \
---description_path /your/path/descriptions.json \
+--description_dict_path /your/path/descriptions.json \
 ...
 ```
 
@@ -187,11 +187,12 @@ After registering your task, you can now check on your data downloading and veri
 
 ```bash
 python -m scripts.write_out \
-    --tasks <your-task> \
     --output_base_path <path> \
+    --tasks <your-task> \
     --sets <train | val | test> \
     --num_fewshot K \
-    --num_examples N
+    --num_examples N \ 
+    --description_dict_path <path>
 ```
 
 Open the file specified at the `--output_base_path <path>` and ensure it passes
diff --git a/tests/test_description_option/descriptions.json b/tests/test_description_dict/descriptions.json
similarity index 100%
rename from tests/test_description_option/descriptions.json
rename to tests/test_description_dict/descriptions.json
diff --git a/tests/test_description_option/test_description_option.py b/tests/test_description_dict/test_description_dict.py
similarity index 52%
rename from tests/test_description_option/test_description_option.py
rename to tests/test_description_dict/test_description_dict.py
index fd6b7b16f9..ad6dc13b6b 100644
--- a/tests/test_description_option/test_description_option.py
+++ b/tests/test_description_dict/test_description_dict.py
@@ -5,15 +5,13 @@
 from lm_eval.evaluator import evaluate
 
 
-def parse_args():
-    parser = argparse.ArgumentParser()
-    parser.add_argument('--description_path', default=None)
-    parser.add_argument('--num_fewshot', type=int, default=0)
-    parser.add_argument('--limit', type=int, default=None)
-    return parser.parse_args()
-
-
-def main():
+def test_cli_description_dict_path():
+    def parse_args():
+        parser = argparse.ArgumentParser()
+        parser.add_argument('--description_dict_path', default=None)
+        parser.add_argument('--num_fewshot', type=int, default=0)
+        parser.add_argument('--limit', type=int, default=None)
+        return parser.parse_args()
     args = parse_args()
 
     task_names = ['hellaswag', 'copa']
@@ -21,14 +19,15 @@ def main():
     lm = lm_eval.models.get_model('dummy')()
 
     description_dict = {}
-    if args.description_path:
-        with open(args.description_path, 'r') as f:
+    if args.description_dict_path:
+        with open(args.description_dict_path, 'r') as f:
             description_dict = json.load(f)
 
     num_fewshot = args.num_fewshot
     results = evaluate(
         lm,
         task_dict,
+        False,
         num_fewshot,
         args.limit,
         description_dict
@@ -36,4 +35,4 @@ def main():
 
 
 if __name__ == '__main__':
-    main()
+    test_cli_description_dict_path()
diff --git a/tests/test_evaluator.py b/tests/test_evaluator.py
index 070590d3f8..85e1449f08 100644
--- a/tests/test_evaluator.py
+++ b/tests/test_evaluator.py
@@ -48,8 +48,8 @@ def ll_perp_fn(reqs):
     lm.loglikelihood_rolling = ll_perp_fn
 
     limit = 10
-    e1 = evaluator.evaluate(lm, task_dict, 0, limit, description_dict=None, bootstrap_iters=10)
-    e2 = evaluator.evaluate(lm, task_dict, 0, limit, description_dict=None, bootstrap_iters=10)
+    e1 = evaluator.evaluate(lm, task_dict, False, 0, limit, bootstrap_iters=10, description_dict=None)
+    e2 = evaluator.evaluate(lm, task_dict, False, 0, limit, bootstrap_iters=10, description_dict=None)
 
     # check that caching is working
     assert e1 == e2

From d7a8ab24ab8778351ae8651093d5883ef72f1ec7 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Wed, 15 Dec 2021 22:37:27 -0500
Subject: [PATCH 04/65] Add basic `description_dict` test

---
 tests/test_description_dict.py                | 47 +++++++++++++++++++
 tests/test_description_dict/descriptions.json |  3 --
 .../test_description_dict.py                  | 38 ---------------
 3 files changed, 47 insertions(+), 41 deletions(-)
 create mode 100644 tests/test_description_dict.py
 delete mode 100644 tests/test_description_dict/descriptions.json
 delete mode 100644 tests/test_description_dict/test_description_dict.py

diff --git a/tests/test_description_dict.py b/tests/test_description_dict.py
new file mode 100644
index 0000000000..bd4148477c
--- /dev/null
+++ b/tests/test_description_dict.py
@@ -0,0 +1,47 @@
+import random
+import lm_eval.tasks
+import lm_eval.models
+
+
+def test_description_dict():
+    seed = 42
+    num_examples = 1
+    task_names = ["hellaswag", "winogrande"]
+    description_dict = {
+        "hellaswag": "Label for the relevant action:\nSentences describing context, with an incomplete sentence trailing answer that plausibly completes the situation.",
+        "winogrande": "Winograd schema sentence including a either a ___ blank with a missing word, making the pronoun ambiguous, or the same with the word filled in."
+    }
+
+    task_dict = lm_eval.tasks.get_task_dict(task_names)
+    for task_name, task in task_dict.items():
+        rnd = random.Random()
+        rnd.seed(seed)
+
+        if task.has_training_docs():
+            docs = task.training_docs()
+        elif set == "val" and task.has_validation_docs():
+            docs = task.validation_docs()
+        elif set == "test" and task.has_test_docs():
+            docs = task.test_docs()
+
+        description = (
+            description_dict[task_name]
+            if description_dict and task_name in description_dict
+            else ""
+        )
+
+        for _, doc in (
+            zip(range(num_examples), docs)
+            if num_examples > 0
+            else enumerate(docs)
+        ):
+            ctx = task.fewshot_context(
+                doc=doc,
+                num_fewshot=1,
+                provide_description=False,
+                rnd=rnd,
+                description=description,
+            )
+            print(ctx + "\n\n")
+            assert description in ctx
+test_description_dict()
\ No newline at end of file
diff --git a/tests/test_description_dict/descriptions.json b/tests/test_description_dict/descriptions.json
deleted file mode 100644
index 6bace5dac5..0000000000
--- a/tests/test_description_dict/descriptions.json
+++ /dev/null
@@ -1,3 +0,0 @@
-{
-    "hellaswag": "Label for the relevant action:\nSentences describing context, with an incomplete sentence trailing answer that plausibly completes the situation."
-}
diff --git a/tests/test_description_dict/test_description_dict.py b/tests/test_description_dict/test_description_dict.py
deleted file mode 100644
index ad6dc13b6b..0000000000
--- a/tests/test_description_dict/test_description_dict.py
+++ /dev/null
@@ -1,38 +0,0 @@
-import json
-import argparse
-import lm_eval.tasks
-import lm_eval.models
-from lm_eval.evaluator import evaluate
-
-
-def test_cli_description_dict_path():
-    def parse_args():
-        parser = argparse.ArgumentParser()
-        parser.add_argument('--description_dict_path', default=None)
-        parser.add_argument('--num_fewshot', type=int, default=0)
-        parser.add_argument('--limit', type=int, default=None)
-        return parser.parse_args()
-    args = parse_args()
-
-    task_names = ['hellaswag', 'copa']
-    task_dict = lm_eval.tasks.get_task_dict(task_names)
-    lm = lm_eval.models.get_model('dummy')()
-
-    description_dict = {}
-    if args.description_dict_path:
-        with open(args.description_dict_path, 'r') as f:
-            description_dict = json.load(f)
-
-    num_fewshot = args.num_fewshot
-    results = evaluate(
-        lm,
-        task_dict,
-        False,
-        num_fewshot,
-        args.limit,
-        description_dict
-    )
-
-
-if __name__ == '__main__':
-    test_cli_description_dict_path()

From 3fdff22f5986a895766bcb905493886b3156b781 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Wed, 15 Dec 2021 22:38:43 -0500
Subject: [PATCH 05/65] Remove `print` from test

---
 tests/test_description_dict.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tests/test_description_dict.py b/tests/test_description_dict.py
index bd4148477c..7d6c34ecfc 100644
--- a/tests/test_description_dict.py
+++ b/tests/test_description_dict.py
@@ -42,6 +42,5 @@ def test_description_dict():
                 rnd=rnd,
                 description=description,
             )
-            print(ctx + "\n\n")
             assert description in ctx
 test_description_dict()
\ No newline at end of file

From 09cd76c76be8f22d1b80e656069028306ee73e1b Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Wed, 15 Dec 2021 23:07:06 -0500
Subject: [PATCH 06/65] Fix assertion error string

---
 lm_eval/base.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/lm_eval/base.py b/lm_eval/base.py
index e9661164d0..db45068929 100644
--- a/lm_eval/base.py
+++ b/lm_eval/base.py
@@ -459,9 +459,9 @@ def fewshot_description(self):
 
     def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert not provide_description, (
-            "The `provide_description` arg will be removed in future versions. To provide "
-            "custom descriptions on a per-task basis, supply the `description_dict` "
-            "arg with your task-to-description dictionary."
+            "The `provide_description` arg will be removed in future versions. To prepend "
+            "a custom description to the context, supply the corresponding string via the  "
+            "`description` arg."
         )
         description = description + "\n\n" if description else ""
 

From d1319950d7836102b9a2eaa359319143b2f1e00c Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Wed, 15 Dec 2021 23:25:59 -0500
Subject: [PATCH 07/65] Add newline to end of file

---
 tests/test_description_dict.py | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/tests/test_description_dict.py b/tests/test_description_dict.py
index 7d6c34ecfc..e94028e868 100644
--- a/tests/test_description_dict.py
+++ b/tests/test_description_dict.py
@@ -9,7 +9,7 @@ def test_description_dict():
     task_names = ["hellaswag", "winogrande"]
     description_dict = {
         "hellaswag": "Label for the relevant action:\nSentences describing context, with an incomplete sentence trailing answer that plausibly completes the situation.",
-        "winogrande": "Winograd schema sentence including a either a ___ blank with a missing word, making the pronoun ambiguous, or the same with the word filled in."
+        "winogrande": "Winograd schema sentence including a either a ___ blank with a missing word, making the pronoun ambiguous, or the same with the word filled in.",
     }
 
     task_dict = lm_eval.tasks.get_task_dict(task_names)
@@ -31,9 +31,7 @@ def test_description_dict():
         )
 
         for _, doc in (
-            zip(range(num_examples), docs)
-            if num_examples > 0
-            else enumerate(docs)
+            zip(range(num_examples), docs) if num_examples > 0 else enumerate(docs)
         ):
             ctx = task.fewshot_context(
                 doc=doc,
@@ -43,4 +41,6 @@ def test_description_dict():
                 description=description,
             )
             assert description in ctx
-test_description_dict()
\ No newline at end of file
+
+
+test_description_dict()

From 10dd7d38535a3695546fc3d762fc37596950ad93 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Wed, 15 Dec 2021 23:36:47 -0500
Subject: [PATCH 08/65] Make `evaluate` and `simple_evaluate` description args
 consistent

---
 lm_eval/evaluator.py | 13 ++++---------
 main.py              |  6 ++++++
 task-guide.md        |  4 ++--
 3 files changed, 12 insertions(+), 11 deletions(-)

diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index 4c224edef3..9963bb5eb0 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -12,7 +12,7 @@
 def simple_evaluate(model, model_args, task_names,
                     num_fewshot=0, batch_size=None, device=None,
                     no_cache=False, limit=None, bootstrap_iters=100000,
-                    description_dict_path=None):
+                    description_dict=None):
     """Instantiate and evaluate a model on a list of tasks.
 
     :param model: str
@@ -33,8 +33,8 @@ def simple_evaluate(model, model_args, task_names,
         Limit the number of examples per task (only use this for testing)
     :param bootstrap_iters:
         Number of iterations for bootstrap statistics
-    :param description_dict_path:
-        Path to a JSON file containing `task_name: description` key-values for custom prompts
+    :param description_dict:
+        Dictionary of custom task descriptions of the form: `task_name: description` 
     :return
         Dictionary of results
     """
@@ -52,11 +52,6 @@ def simple_evaluate(model, model_args, task_names,
     
     task_dict = lm_eval.tasks.get_task_dict(task_names)
 
-    description_dict = {}
-    if description_dict_path:
-        with open(description_dict_path, 'r') as f:
-            description_dict = json.load(f)
-
     results = evaluate(lm, task_dict, False, num_fewshot, limit, description_dict=description_dict)
 
     # add info about the model and few shot config
@@ -90,7 +85,7 @@ def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_i
     :param bootstrap_iters:
         Number of iterations for bootstrap statistics
     :param description_dict:
-        Dictionary of task descriptions of the form: `task_name: description` 
+        Dictionary of custom task descriptions of the form: `task_name: description` 
     :return
         Dictionary of results
     """
diff --git a/main.py b/main.py
index e55015b4a2..c7f9ce6d8b 100644
--- a/main.py
+++ b/main.py
@@ -35,6 +35,11 @@ def main():
     else:
         task_names = args.tasks.split(",")
 
+    description_dict = {}
+    if args.description_dict_path:
+        with open(args.description_dict_path, 'r') as f:
+            description_dict = json.load(f)
+
     results = evaluator.simple_evaluate(
         model=args.model,
         model_args=args.model_args,
@@ -44,6 +49,7 @@ def main():
         device=args.device,
         no_cache=args.no_cache,
         limit=args.limit,
+        description_dict=description_dict
     )
 
     dumped = json.dumps(results, indent=2)
diff --git a/task-guide.md b/task-guide.md
index bf63dac7d4..c649aba331 100644
--- a/task-guide.md
+++ b/task-guide.md
@@ -159,7 +159,7 @@ description_dict = {
 }
 ```
 
-One can also interface with `evaluator.evaluate` from a higher level by simply passing a JSON file path to the `description_dict_path` arg of the command-line interface program, `main.py`. The JSON file pointed to should be structured the same way as the aforementioned `description_dict`. E.g. for some file at `/your/path/descriptions.json` you might have:
+One can also interface with `evaluator.evaluate`/`evaluator.simple_evaluate` from a higher level by simply passing a JSON file path to the `description_dict_path` arg of the command-line interface (CLI) programs, `main.py` and `write_out.py` . The JSON file pointed to should be structured the same way as the aforementioned `description_dict`. E.g. for some file at `/your/path/descriptions.json` you might have:
 
 ```json
 {
@@ -168,7 +168,7 @@ One can also interface with `evaluator.evaluate` from a higher level by simply p
 }
 ```
 
-which can then be hooked up to the evaluator through the `main.py` CLI as:
+which can then be used, for example, in the `main.py` CLI as:
 
 ```python
 python main.py  \

From e3ddcfc277a449b4974159c4b275c0fd0cfc72c5 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Thu, 16 Dec 2021 00:04:55 -0500
Subject: [PATCH 09/65] Remove needless call

---
 tests/test_description_dict.py | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/tests/test_description_dict.py b/tests/test_description_dict.py
index e94028e868..aaec13fbce 100644
--- a/tests/test_description_dict.py
+++ b/tests/test_description_dict.py
@@ -41,6 +41,3 @@ def test_description_dict():
                 description=description,
             )
             assert description in ctx
-
-
-test_description_dict()

From 09e1de936d7deb7938a313679b45057dfe31ef82 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Thu, 16 Dec 2021 00:19:00 -0500
Subject: [PATCH 10/65] Fix assertion error message

---
 lm_eval/base.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/lm_eval/base.py b/lm_eval/base.py
index db45068929..442d8d5d8e 100644
--- a/lm_eval/base.py
+++ b/lm_eval/base.py
@@ -452,7 +452,7 @@ def higher_is_better(self):
     def fewshot_description(self):
         import warnings
         warnings.warn(
-            "`fewshot_description` will be removed in futures versions. Pass " \
+            "`fewshot_description` will be removed in futures versions. Pass "
             "any custom descriptions to the `evaluate` function instead.",
             DeprecationWarning)
         return ""
@@ -539,9 +539,9 @@ def fewshot_examples(self, k, rnd):
     def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert num_fewshot == 0
         assert not provide_description, (
-            "The `provide_description` arg will be removed in future versions. To provide "
-            "custom descriptions on a per-task basis, supply the `description_dict` "
-            "arg with your task-to-description dictionary."
+            "The `provide_description` arg will be removed in future versions. To prepend "
+            "a custom description to the context, supply the corresponding string via the  "
+            "`description` arg."
         )
         return ""
 

From 3f06b6032194e75ec3b6b9fdbd08597d37af6a55 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Thu, 16 Dec 2021 00:25:05 -0500
Subject: [PATCH 11/65] Remove unused import

---
 lm_eval/evaluator.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index 9963bb5eb0..e8b7c030c4 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -1,6 +1,5 @@
 import collections
 import itertools
-import json
 import random
 import lm_eval.metrics
 import lm_eval.models

From e54380d882436c5136a2b474e341dfb26cd672a1 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Thu, 16 Dec 2021 00:53:30 -0500
Subject: [PATCH 12/65] Fix task example link

---
 task-guide.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/task-guide.md b/task-guide.md
index c649aba331..7f43de59d3 100644
--- a/task-guide.md
+++ b/task-guide.md
@@ -116,7 +116,7 @@ class TaskName(..., MultipleChoiceTask):
 
 This will require you to format your documents such that they contain `gold` and `choices` fields. They can also have other fields, but those will be ignored by `MultipleChoiceTask`. `choices` should be a list of possible continuations, and `gold` should be an integer specifying the index of the correct completion.
 
-See [this task](https://github.com/EleutherAI/lm-evaluation-harness/blob/105fa9741ff660f6a62c2eef0d2facfde36dda41/lm_eval/tasks/sat.py#L56) for an example. When used in combination with `HFTask`, it may be useful to override [`_convert_standard`](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/common.py#L28), which will be applied to every document in the HF dataset. See task](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/headqa.py) for an example of this.
+See [this task](https://github.com/EleutherAI/lm-evaluation-harness/blob/105fa9741ff660f6a62c2eef0d2facfde36dda41/lm_eval/tasks/sat.py#L56) for an example. When used in combination with `HFTask`, it may be useful to override [`_convert_standard`](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/common.py#L28), which will be applied to every document in the HF dataset. See this [task](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/headqa.py) for an example of this.
 
 You can now skip ahead to <a href="#Registering-Your-Task">registering your task</a>.
 

From 744482bb183deab3abf5c2206b217422221aaee1 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Thu, 16 Dec 2021 00:54:25 -0500
Subject: [PATCH 13/65] Fix task example link

---
 task-guide.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/task-guide.md b/task-guide.md
index 7f43de59d3..c750ab5a1c 100644
--- a/task-guide.md
+++ b/task-guide.md
@@ -116,7 +116,7 @@ class TaskName(..., MultipleChoiceTask):
 
 This will require you to format your documents such that they contain `gold` and `choices` fields. They can also have other fields, but those will be ignored by `MultipleChoiceTask`. `choices` should be a list of possible continuations, and `gold` should be an integer specifying the index of the correct completion.
 
-See [this task](https://github.com/EleutherAI/lm-evaluation-harness/blob/105fa9741ff660f6a62c2eef0d2facfde36dda41/lm_eval/tasks/sat.py#L56) for an example. When used in combination with `HFTask`, it may be useful to override [`_convert_standard`](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/common.py#L28), which will be applied to every document in the HF dataset. See this [task](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/headqa.py) for an example of this.
+See [this task](https://github.com/EleutherAI/lm-evaluation-harness/blob/105fa9741ff660f6a62c2eef0d2facfde36dda41/lm_eval/tasks/sat.py#L56) for an example. When used in combination with `HFTask`, it may be useful to override [`_convert_standard`](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/common.py#L28), which will be applied to every document in the HF dataset. See [this task](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/headqa.py) for an example of this.
 
 You can now skip ahead to <a href="#Registering-Your-Task">registering your task</a>.
 

From 1bc6cdb160efa0bbc3198a15d2cf428a06158aec Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Thu, 16 Dec 2021 03:44:05 -0500
Subject: [PATCH 14/65] Add type info to doc-string

---
 lm_eval/evaluator.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index e8b7c030c4..6feb356788 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -32,7 +32,7 @@ def simple_evaluate(model, model_args, task_names,
         Limit the number of examples per task (only use this for testing)
     :param bootstrap_iters:
         Number of iterations for bootstrap statistics
-    :param description_dict:
+    :param description_dict: dict[str, str]
         Dictionary of custom task descriptions of the form: `task_name: description` 
     :return
         Dictionary of results
@@ -83,7 +83,7 @@ def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_i
         Limit the number of examples per task (only use this for testing)
     :param bootstrap_iters:
         Number of iterations for bootstrap statistics
-    :param description_dict:
+    :param description_dict: dict[str, str]
         Dictionary of custom task descriptions of the form: `task_name: description` 
     :return
         Dictionary of results

From acf76b50b6f69917a09379415111c0af3f00c5b3 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Thu, 16 Dec 2021 19:53:12 -0500
Subject: [PATCH 15/65] Add `description_dict` docs and update `task-guide`

---
 README.md                           |   3 +-
 docs/description_guide.md           |  49 ++++++++++++++++++++++++++++
 docs/img/fewshot_example_gpt3.png   | Bin 0 -> 315681 bytes
 task-guide.md => docs/task_guide.md |  35 --------------------
 4 files changed, 50 insertions(+), 37 deletions(-)
 create mode 100644 docs/description_guide.md
 create mode 100644 docs/img/fewshot_example_gpt3.png
 rename task-guide.md => docs/task_guide.md (88%)

diff --git a/README.md b/README.md
index fc6c4e4ba7..112595d8da 100644
--- a/README.md
+++ b/README.md
@@ -55,7 +55,7 @@ To evaluate mesh-transformer-jax models that are not available on HF, please inv
 
 ## Implementing new tasks
 
-To implement a new task in eval harness, see [this guide](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/task-guide.md).
+To implement a new task in eval harness, see [this guide](./docs/task_guide.md).
 
 ## Cite as
 
@@ -298,7 +298,6 @@ To inspect what the LM inputs look like, you can run the following command:
 ```bash
 python write_out.py \
 	--tasks all_tasks \
-	--provide_description \
 	--num_fewshot 5 \
 	--num_examples 10 \
 	--output_base_path /path/to/output/folder
diff --git a/docs/description_guide.md b/docs/description_guide.md
new file mode 100644
index 0000000000..c97a82ae0c
--- /dev/null
+++ b/docs/description_guide.md
@@ -0,0 +1,49 @@
+# Description Guide
+
+![fewshot-example](./img/fewshot_example_gpt3.png)
+(Figure from [Brown et al., 2020](https://arxiv.org/pdf/2005.14165.pdf))
+
+Task descriptions provide in-context task instruction for your language model. If you'd like to prepend a natural language description to your few-shot examples and prompt, you can do so on a per-task basis via the `description_dict` arg of [`evaluator.evaluate`](../lm_eval/evaluator.py). This `description_dict` must adhere to the following key-value structure:
+
+- **key**: the task name (`str`) as specified in the lm-eval-harness task registry (see the following section on task registry).
+- **value**: the corresponding (`str`) description/prompt for the task identified by **key**.
+
+```python
+description_dict = {
+    "task_name_1": "description",
+    "task_name_2": "description",
+    ...
+}
+```
+
+Note that a task's description will be separated from its following few-shot examples and prompt by a new line as such:
+
+```python
+"""
+<description>
+
+<examples>
+
+<prompt>
+"""
+```
+
+## Descriptions in File
+
+One can also interface with the aforementioned [`evaluator.evaluate`](../lm_eval/evaluator.py) (or `evaluator.simple_evaluate`) method from a higher level by simply passing a JSON file path to the `description_dict_path` arg of the command-line interface (CLI) program, `main.py`. The JSON file pointed to should be structured the same as the `description_dict`. E.g. for some file at `/your/path/descriptions.json` you may have:
+
+```json
+{
+    "cycle_letters": "Please unscramble the letters into a word, and write that word:",
+    "copa": "Given a premise and one alternative with a causal relation to the premise and another without, choose the more plausible alternative"
+}
+```
+
+which can then be supplied to the CLI as:
+
+```python
+python main.py  \
+--tasks cycle_letters,copa \
+--description_dict_path /your/path/descriptions.json \
+...
+```
diff --git a/docs/img/fewshot_example_gpt3.png b/docs/img/fewshot_example_gpt3.png
new file mode 100644
index 0000000000000000000000000000000000000000..b199736867aff6d1aa8ad407ba31810f4016fb75
GIT binary patch
literal 315681
zcmeFZbySqw+cyqFg9=D@Nq2WCA)V5VbT>$j2uMgHAgM?R!T^FaLrWtF3@zQ=>3icj
za=y<w%V(YS`{VgNYrXHR8D_KN-uvFyb$zbd_k^n|%icvNMu&rgyDRrtN*xXk1p)_$
zAcBev)Xd<EFTlawaj<#xNLB99BT7|QCrcZ93plvP;Yr#kIyiD9nXGb;A6tV_AEPy(
z4HYF~>A}GlJ}-FjAXe2(5tZrKwcxuhwp#IO;mfT<#6?wX=I*z!3sXsHLChmn+@!#V
zpl372nLm6p$9+yRoEodR>xAK~k6mcvMKm#j{9M31Vj?MmV!W^9))8?{;0U3lCI*>`
zEfJBC$f}{|O+9raaB7FzWO!BAg4f2<VeGMca1ujQ{fj-`XIT93K~9Tu?~&lJ@ao$?
zfx<xp!8+RL2rsG?!4*l~pFvhkE(_VpW+Ei>J#YnTH3l|(aKtB+11yUckA#V{C_V8H
zq|M-XWF<+VIId?{%w`lxS>n=QX3DCN^ZPGNMWDh5Bx&P;6<LDzSv`H^9o_ZF(yUSs
z5*A&oF0*paX5+KyDfUh8l00E}^-_Bg)6SdOOFj{#LA8x<)zQXed)FoFWofZ69D_7N
z%wzTS92pTJl2J@Cw;-W|8Y(p`5=m0J{>X5xy;R+I9QWFsE3=2Ck<z7J9#I)l$A-`q
zxL3*Nw-x!lt<+WLQ)tgxkw>{=@)M0+p)1ooo<&6>P4ZXFd7P30;zTm*d8p=kr;Qj(
zxXTsiou-g^#EZ=oDzZWmBkU&Hwmp!GH)KdmD~0xMr1o0!Vo&IPdlkBLIw78zCZ%GO
z>L`kE-@?5POt)q>jJD$URKlR~)77^Ir_W<mEw*89U;#!+i+mMzEYi~SNR0<RE=>MY
z2C)Ps-}|UMlkw;Zmgl>WoL!i<7IfpES4&x>d#l~!=TauZY`iapz=sO2NEyb8qX;i3
z@om&_?Cwq?jcE|ZdrA`2pqvPzH*s77)S+XDs)*M?kKvxS9y~{pMast7&kECB8o_>g
zHyZz5{}&_`s(utTO3axa6sz#C2HIfySnR7-VFEt#D1J=7_S2@7Fj(mw46d(CGGdEy
zg@W(p?$Cx}(Y!R35}v}sgkJ4hd*6NDhj-H9AG&|~LB(6%nqGw3fZiR;Q3anhHh^GB
zXApiXtk5>@%z@R}&UzRDv*YTmhc%Ja_fROdS6G0KH|60bylxefqx6%M+3{<A<XMrs
zN#x_HQSq{|!C$Io2fjhZb_uKA6&!s&QtNmlXyA0b`{m?%d{#3H^P*sfw~Q8Mx=HGJ
z->#XsObFVhY;vT-?5PM2(;iD`9^qYN4dBQ$u$Dc<3^Yhmk*2li=j}J_&q_2$dcKAf
z@v1YEI+Y5~0s|w3<<tX<L(nX9m?5vIq6h;n!s*ri(a~$p)4P&{1dXoq&rT;;Lw)T>
zw4*0M5@eS$*8Zp6x~*3qq!GS*<EgN*YPwd>KMZn0Q>q@m6NCnVtKy*0{*HDjfrW|`
zM~VO7xh4mimM-d84Xz?W1=LUVj(oE`1m7M`AL<Dq*+dAFAhSlKZ`O6jHb5S0HmrUo
z)T}UwFar)ilzfbO_bm<bOGdL00m_n>{P%-iP=9?a5+CuDa{MJ>98E=R$YaKG;_)E+
zM<3#{;&Q$K@@gaNNP`*2ws?0RyIqy(LnLv|=&JQM5{tKL*3?4SeKPs)q*tlk1@*vM
zZ+*Y&9X@!5-4N}W&HmNsj>$S90(UDZr&H9cb~76te7g2!6IvZXt&puw;_q&a5gsPf
zFqHI{!!C5&sM^kmnJ*4@3r|V?h#KK%UyaS@?-iXssz>?pFdI7x^&X-zD2yXBRw+e>
zfi4lp^L{i|KeXYSoVV<Q`$SSC?+B(M3)@KN73aC;>E_|haLO`HQ(y|zad#fydl0D3
z60X2RiI*^%a2&53TM{cI=Wn}in`c{MYies_YgUP5yJYM8ExO21%{OmUWt3V!&cCbd
zshAJ450Vc`4cISyN^U#fTEm{<I#Ejgo`qQM^YSd$P}i7!Ekna;gK1g+^xDV%na@Jb
zC65l)F-OEzJqnIB`FJy!^4apY-|?wAY0M5;^VsQOvA)uh&{r<X%XlZ6ePlDpo6aLZ
z?DwjMIElEDSewVlCZ6{rkM5fUo=Bb^UQ9cqq0LchJ6*f7%4$8gPfV53w$tC%NAm3I
ztUGO2?AV8Aa;D#18HBwz|A_OUbxikth2ctKzqydi=j@6#k@JV=_s)r)<+G`c$t35b
z<>-H?Q~9j)Syn}Mz>r<nsn#fWhO^-watFlHnb0Q+FX~YgNmL9?lq^IBRfdJEEU+yJ
zd*?`Y9c2yUP~i}B)p+cOHIxUH_Y1F%g{qbFEEfMKeo`xeKIe)#qSo!WTAC}FeVNx$
zB0T%i*tSodaGiLDt?Sf^)Jn_3U8>94yc_ly7aVay<znTc<x;y`64*3qZGsBFeCj=X
zv4)XHp3vTFuEn9k?b{pGOWnI@IoUhWd*0{Jo6t*;Lc!7GBw4mw^`)Gl!n$g;jHi;+
zPOzuEV?P>iyn9mKy0U*geZ<#}WCM4jebi%gz~RtAfKQr_m+!>hrV3o8;ZS5>I*z?n
zzva5cH*Q!ZyCu3Aa-4LWOsa~af)dwC7j~@nLQO2EE$4&GwEKIZa#9R3J5mZE&h&Q~
z_yQw>4a5$DLJo%?zOKjwx*R+AyUon`jY)UX4q1Q|f_q4`eYG>~ayNMS-{_}I8hn2C
zd2IXMHt3{p`?JU@xw$C5FORR^@!@{rv+a}3{qn=^-HBC?Vf``Te&5s+AByKHkO6}%
znJmo4%toZfFuxy%ZihWblM5$)QGJBXDP-p47ikx6SD(+k5IGRIKvl@@h;NWw5sE+*
zNV<r7!Jf#X=Hq3qI=;H4&ohE^n`NKRR``!sj4M&;JobJp_gKF`*;;)ybP&aK*>v2L
z*Hm<le2%*{jT#kO4ZDci(KSj?M430Aeu&xCb8(HqR`cPv&;r@7$6tq5L(5iSU&`z{
zW+KZXusgrCKi5&-d;`01;d9Yg2q2RsCEzb{of`H|KGoe@xxb2*fOGtyUg3_iL7X>#
z7oS|_xAfxF^h|?{OEKC979R4Qr(YMWVwzHHQ<&sGs?g-p=5Hx(WMB!IZoTMtk2#z_
zB*Af@Q_g~Mbvdigyf;tarFWyJ=%;-8{^@4dY1^wMEf-&I!Bi4%UvK*vp&5k$tfr`0
z+FgQTe37300lkbPkx6l1CC_9IRUNiX{4!}0sZuv;*Q&+(fmeOgNy4$o@m`e1tl~63
z3Up-CnZ-ogpLQQIZrw+bD|>JLVdq<=hVOf|p$sl5J~b`lY>dEl{iUjXibYHe1axMM
zi#%EtVG%>$TJL8YcP@w&tVEd6na^$L#1i-`d|r<ipH-0eTlUTHp`<1V=KGYiPf5E8
zuY?Rtie_<%dtEzWq!Sby?T)o1hTM&eBIwgUJPq5Wu#Mo2$Zh8*xL}<rM>M9^8>;%|
zuy&qnEPqN%AfKVoqae<@z*@JP(-VUyevf59Z>Flnt$jJEEonZ9Ukj;ZqCwBs<Li2-
zT}MeFtNH7s#H&ir{fXDEC;O}$3>%9ZLQL<N*T?irjBBokk>^4;lA4oH_=p`0YWJ?_
zuJf4Fg86o4&FXXAdpbY9`}nyk#Hr|Gc?r|3@|02a)mro-*-F6vb>Xx8x&EPyEU$1c
zkIm>=|JsS7+GjCeGx2z>U)AT?Ii1h+Ts%A;+3}ic+Jak<Sd(;C4cKiTAQGPy!1X9N
zgPhD=HR)WH@1b5jmSa^H&Z){*))`aF(whz7*|&BJHg1hG&Ip(uu^!d&QQUz=EF>;O
zMq{Qk2z{M=V6!fk>c8)Du(({8-;q`8`D|*fid)U9KkKfr?>gzmRQSR53W>g-ezehU
zqpH`}12cyr*^RhakpSVL<#8?Ft_}4oi>#izTjmon#%T4f`?g<3D#irH*o<OoW&&u|
zIzFbY@e4cdIf*u9cwHQ2v<P2$txVHQG1ruOG+lzf27j2{3J@Qh8+X|>{n9!rix#sj
zwk#TOCUv~1vGOkeQ~uI`V%8IJFaH@B-l)e}`wM~$c3Sp}y2u}c*X9e%liE#}M2A*(
z1E-n&V?$#IPOlqyz3VpEPnzENE<B^C?+G0H(fEEK+vx;l2XjR%&(CjEd<$nyqur`*
z;I8qz#wn676oj$Pxl_%?@OSulm@$I*>)^6;K@wpreBFpky3VRtOJ^Qf;2$rNR@x=^
z5KVk>;KC2!UV^XiVZ~IVcR&(*7+Hjfs{Ehl^6u-tPAeM1$IE<y>W_fn^a`Gm7j9Qq
zh-0wH#^n3KJ;F@~xHoxB1KH?QE!?eI?Yd}Fuo`#rQ{}{!VJFwiMwAHDspgZbCb%s1
zXhq+={Q$uKh_KL+vs6-oV*<*kaES24a7aK29{7mCliZeN;2Gc$e%6EF;KFR+5dZ9>
z41C|bVu8=im_NP|;zQw3fdB3TAMb3?uihw-Y=mEB1QDPOPEzBMoE-42VeV>S;pk@V
z<W6`s^9`s#bAGJr1_y^vd-H*pQ-81vTz||)Q^#FLNm0n$$$`!Esgs!no414W&2`{J
zyoG?GgN3^(rMH8<qnnVoDD}@CLO}VZnw^^RXBT%nQED9}Rmw+Bt`?MhY#eMH)MDtA
zl$0W_Pc4PirDXmb4*VxdZSC&vEX2<4<>kfZ#m(mAYQ@eeC@9Fz!NtzS#R~Lbb@Op_
zH}z(9bffv>BEPO9W#MM-YUAu~<K#$rb6rz2Cl7Z~YU-PbZeM?_)56>4ubCX({=6;V
z2H9^~*g4ra*l({53>CSl6;ieFwy@WgvT*?R446YqfJ;E+Xa9e+{59i$jMVvSB&Q(X
ze;N88O@ANy#LdF>k&^>3r@Po+^!jt&|7`qopa}cTt^dbX{ISnJYk`dxLl<GcC7Kv|
z@e_f1fR3a#QYxCjH^5~#FZfL055phdKpEtGzf^y~6b?=TPEJZv(;I%X0WFAB*5}yR
z*IK)-c4t`VW9w^{$^CFA5@QpRsAkf8VVqQSZSdC*uw1dkq7hF)6mqEWW(h%Pd?~H4
z4!%a#+0QKFub)nzu<EYNOk9bp9E~bE**vSV@_J_JdCXY&xvsXTXy*Khhlf(rk=7G;
zD3KsG7W}_`GjM^?86lAdAbBH1DT#mghTe|_>ElWMzB6ShqQIKDw6rvn^6U=V>56}#
z`kX<%BSlh4ozcq%_TU{auK&fAQwm_Rytn56b_cP*@33P*b=Ip%5Y(4J|L!f35`+sg
zB^JW}_jDqL2gnp^gckWPrwBrQf%f?S-Ww^fc5>P&C@6T0Hu&|cX-j}tv-r+M0{Y6l
z$3=nr^wGmje<Hu8g#TChfA@wg8R7pwG5mLs^*@Ct5|Vf@GBUEfckigv8>6NFTa|?V
zm~z`|LLF0E;%VMD-L@o^-XX*`7B74=I!hbST2A#5_3Yfx{Re@GPUDey#{2Jw7!Ubk
z@V2(L$V_uxHD6r&%<f3-4_sUqICr>zS4R>8b$mL&G^m^QfH5}{r7!X_G#4VT2<G4O
z>MV4wmPuw!uY1j!{t7YBp#GTLJDq$c6O(b-4J|pofm_Ur{b05G#fs~_F&`Cy#!QYP
z*Hm|`XO);MFWOW3&Bv@G@%j^})f-FZ;*anmrKjBMmc?on>p~Sq3AB&RLN?y`a=Lxg
zui}#g?qGg&@{mkKT~tF9!&YkQ9EU%BUH45;%2NNs-ZqwLKKjmwM=XL>YbieO8)a(6
zIxVpI4YY^Oj@$iZBJfJ=0@@XiSPHPuN@A8Y1_p`K<O4c4E-lw>qke3j2FOAI6f_hF
zq~K2vG#0yRywf<FyOecuB^TnhSL?4?yZwyn_@=V!`~P&MtFbD^rgO_RXI=In(VOKe
zn!MbP)1N#{gh8}?LMF1=UCSOdE%ysb3*{LHFw)A^ael8kSEJ#MMw3pTYd_I7{ZQs&
z`~=MX&C1KGqU0~tDG`33;R7fQe^gWXGwB+~D^`a*Y+^e-V+CD9;z|qEyTWzr>#t>O
z9;5%(2zW^cE%m-E>GpqOky}{Uq4mTDZ)(u4C;V1BOJq^SJ^_<i+}~bda}@BRNcz?X
zDHlV`J(|;k=?otdD_=3l7VsRan<UE!{x|b{Nea@i)Cli$fJ>bCG|2WT@?xv{TPt3?
zeqqH+=B$)lw3oWuENzMa+YHINY)Ch@=F&dCW|%D}U~Zg=@BRSYc38Z=k`=-m3wvAb
z_#8B;r{6O*(W!Yqz|&&GjyJ?x4mb40Z>>eJ5!Uz5VtN*sFP<F336FSNu$v8kq^1P@
z`Hp1)KWBubafo`GZ(t9@MQ)7JY;`Tmm#u^Zu|N)<tHu4+9Fma3Drj>B^QxL)LWU!I
zymKF*OFz8rO)8yDiN0xB?BbtYhJ<dfV`vw$k%RtLQ{^$(@B=MP&2R10>=i_TB=4&r
zTj5IYIeNUEnyFfiUsh^%Al738us@`xZ`00}+hJ=<Z=n`*gi$!eTPe2?TA=vf7|q~)
z8dVfpPeYF+#+D)25{ic4{a@Vog#_G&3zMA?zLPYD0-52B$wltdPsL-Q)-@rxQ+fZk
z-`}?w{0^(&;|DGegb#|9pJ38+rBb&@7fl!Y5>$sbI_gWJr|701X0x*@AoL_-4F~Z(
zbQ`Dr0o85KzwbkKC!zqnIo|a*ZF6~-;#XvzWT#amR-qkcAs9Vx4qeLjQ+C`FmT{xS
zf3k;jX1o}i7WqMj<4}6b-irRQSvx_$#!zQ_w|V2R(i;tdYrh_&;?ou%`Sa?}KE*fo
z6sp(5chrXi7Xis?nSn)ReWdm_yt*D^-|SAFRBToQJeSTKWwf0O-KLe8X8Xk@x9sw(
z;Qw4Aga{5}vzd3qQ)qAZWS(I!rx8;&L}kAI#x<}x|6k!ROs!0VE}VO?D^Nzq%Ioi$
zZ>Mzz+ZLT~5(YQQ8L0mSCAR|qM?qo_^;-+Nuk))$H2hM;#PVOt6%I{`36k?^O^lBq
zx}!tMVasx04u3e5v2E#G41E4ZKB{`N$a3F5tI74(HvJ`MziLA0R56)1&(}XZr`u@c
z^h&la4BJq?_A_Glz;{{dp-<SJk6_jS>e#Z*>W$V}=DU+s=%(B8wDZ(4#|*^|d>U?a
zBzu-(Z6tGvVlFwGz8DQ2te!DBs*=>^vql(y6<OSfBj3FL>YF^Qokt<GYQN<)N=!tK
zglZynee(}4v6#cJ4H`LD+}8K~^3za0mF0Z-yxAVX?iqT31<~>8^{Vxez!r}9$xLQe
z0+{fTNu4(xasSz|K{^hAOnz5AgB_=q+Oa+d7bmsq`}EL$n}$3r_C57@?|*K}FDwcK
z@jgk)yQc+1fF%E=9Ka88zC?m9uy6tC{EDt$;)nVLROi>NE{Oe0g8w0z)ri(;h8Rmn
zl#2A3z;;)!F{S?^Vi0PqL@6H3VT;U?f3(4|<9v+llwMm55KOkLa!uvG%j2!>LP-Gf
zWZHWJU*6J>^rj`^7t1V$++6K~Sl`rt+BUoMH8VE|PzSnTCgI=SCRPF*YpGz;Y3u|%
z=Z-4JU+U=#CfJA(GWoW0#;1Cmr}<p3<PsHt%K(cHq<`t8Ulad{v=RoO6uGH8NqC+H
z2I%y^<TaKa)Szyn;nJ|-obYvM%!ldK{YV+~&HKz);WWSc{Nl*lb%s2ELpp}8GjV%M
zc#LWa3klsB>)Y;M5;pK!TRM2wXu+=KclLD1(BLPyY~3(u=1-eUEY+Pr`zJ8f^;n6O
z{zRdlT>GS0WLbE6dp~txqamr`9hv2~xqj&zlkHHB!^1mmd~)!4zZ`=?y1d9Hrs29P
zA-_R#ntU71jWN}OL&pd?Kve&Ft5E)4kf5Md4QYpjVHOLnCzSqZs3sggu-Ls)ru$L9
zqr#LPIkuHE<S~X(c5+jR{6Afpd_k^n^4aWsw;@bl$y{GwK#VLI8}R;n3OJJR{-XHb
z7<UyOEU&|Ux9}q$0u!^Z^JmAuI2=6nC-|_1=T*^Ib~fX;I?LW|asz+b8Ae>UW4bF`
z;Cf5Re-^Hip_GbLM*P`KoN&|rk)CMD&jjeLsE%{{sZ0DqOqPaT-k^@#8Dm@H4ZxIQ
z6&nz{STG$qvx$eml7%9x_wk-5DbU<NPX7}9u$Vfy52{8a49f6$q>wh=x7Nzt{=v;c
zQUV`lp%&Hs;xA$Q8`%z1;I!q6TZLdps%GYVd+CPX_$|fXsbjA6q`hzY32kyR0aHI`
z4rq)&{5dmi*OwUKpED0=$6u>kVQV)&AUbpDOe*}=dUn=5k_F%mp01mIosG4q*)eaG
zr{NEwb>xg4H$)i0$I_2Oa3NyauJ?6*84<S<{j0FAke&D8KKdq;#z&jB0}QRQTVZ?p
zV2}psc<wa~1S)2Bud+(6IThQ^Cp{6#(<d><mFl;W)Ig>B1e*b3S5$e=(@9|6=oi7k
z-iXiML-Uq+Ss^yzA8F$kz}t~;kw>)5f*xWnBdtfKCiVW4bhkA`Xqxjh9AnUB&H4Wz
zbu15<oDp(jvSp`dLb3K8oxycU9w5V#UU_%&Z^_U`uL_TJEWuEUA{owliyD+5Ih@rv
zknh->r5%{{%44WqW9l>)%ElI$`+G|u_hAtFjTw+j9ldzV&J5-tQFW6Yfp7N2oFUZ)
zJ~%*dt-eRQ<}XnH8+3yO5Ct&l^D$%8O}wB(;aU^ZHGY5+Z16K0_^F6)6kgF!MKry{
zoi%pA>RUGdQ(H3qw0^wKdRLBqYN!3K_jT0~H`;0Pr@H)He)1UHfP-#d0%p~oxk34z
z{U_VMNBMj3Ob{Ees1!cT#hEec*9J<k197-%g`g$BI$YS2L2Y;1ua>{TaS&=74JDGN
zh%!3uFYbt?0L7ti+UHG|L;Eo=)njrS^M*7#QqJ?g9Z;S9b|Dgo&y*=`kUXoLzM2*)
zh#KgOzK>DJkTy7rzcK!Ps{Zu+e^o{QLjaK0AT^={&}64hW`M3qvQ-U}_AOiey%385
zo?(FGG;&_70U%*K^Wf?yNXT*RFGSyf#KA$*zY^7w>tgiHf-mo((bO7ZE6lM;=Mzk=
z4gaBs{{{^&C4SIKAbFCFLdlDHAs0@xs{Mc3F4;jhVl(4XFwRp&&X}Dd|A7$kDcJDN
zy8NwY@Jp6&_ftL+h|cCdA&-~&Q|`MnBw}!3<xsX0TkJ(ARR>xYh(++UxV(N1?%#Ov
z_Fu3ys1gDtr^eW{mZ+t^r6)h2;aBcjoJl{x*xYB<SbF?hWO1dX#Mm?lXRQClEfP+E
zLZ2piW)V<|3$rkSc4z)-`6nMyN&ph7Ytl)XOL<G8GfIHD#28-NkOD@AO*51JZOh*a
zu?X-*(2^G1`N|Q&sod<png?$Tl#?j^rfLQB%>cFD?)Sf4!qc1O7kj&_{klQQ0YHHI
zxd&Y%l+ia`nUt2Kf3^H$|DVZ7pl@cn3>)6gEJ}k_2;E*?<-o(kBQl@%0jT+jyvtjO
zz9&q=AMoFz3J{hO6qclZdr2uu67YLG*m$FK*b<<bOm!2)*lw(5g&RMonk(s9^-p)`
zUsPQ#JVxR}kg-rix2G`oIkx~CW-fg&xP|?HFeV$~vm9vjdB<kH@%o&fPr|>^OY%d;
zKh5_CG~(MgixUtxt~sv&{0Hg;kZ^?JS$h7KC@UnVc(;c>zdQW5G<Ye&2Wn6=v5M3l
zyX9>uN}L<a_cB4>lTo6|47bwFNI>$`#mndVy>TZ>4agSY{Sp~yOy^v9B>Xku;=^oN
zbbDQsmFLwvp=6pE81|P5zr|^I?A2LB+qZyFAbI{MLF{h$^E$XM0dYeDO%o)<A$c0J
z@OH8M5|@9NnP)d#b-Ax;f6E!;0Jvsks5CVS6;#O|k6qOwGaTGD{=Hxlg$w(DbXyzp
z0FVeOo0GCnPr=ij=!)~Vmia%Z2tq7?lIw&pw%Ju8dxE_2a&PxQRs!7~^JTpaAQ}0(
ztG}1$b7nxOMu}lpm8VHge=2nb0l=)oh_2sxz<}gA%g8Lp^xHd^19D8ljO7v9zgQfg
z<6|UGS=q3{f`TvZG2{P#%S06g$L1_Nl4lN(*m_{$!-BC{&Z1=@|7*Gb5OP=Q*iRr~
zdQ;q6<0^X-?eH;}6&P4b5L$Ke2^wAA!X60fetWyu|0uu#)c^SGmQ7S2-<S?Ppt@LT
z&3b^Sh9bxEy1!k4d<YyK%wyz<rL&Sp;mZetw=&Y5@+aIc5My^wWb5ge=<9C|%8~t+
zJ%UgVg+Y}-Fe?sJH=9()^^1YqZy=Ym;3SU60z55?E^qeRD@Z&?<kB_qa(Bn|Ka(^@
zp52N4-;p5wBNw{Bh6q@iUDvf3*gm(r^!Gyc4{ShTZwKNg-lk5$4{tJ~s_!pMgvz1Y
z!69^UzmF&k{HKIvem$nH^_<$^dmTIZ3_I5m@26f}4Mt?5<er6}ax2=L`T*If12jpa
zzGt`s8<OYz%cOs(!CNU#6!}B;q}lT6-UniloP<HhR;LcCiS(JPXfjAVwZfa@bKJUL
zLC21qPW4W!?Fvd<4tWet8w<?yNIY2m8e`|;O|YZ6(@S?<s1ATATI*TQ>zD^EAqV$8
z(v|XdMMLf+BSt!>&-sPKV59=}BT2t;mEQwww)EtJ1$(ZLoizu64rvQsl++QkqE@~K
zpT3HtPO94bcglgMy}CbqkCz7<Ixotr{+?B%0f=m8y3bU|Z$26BK*7gUL{e61k`qFf
zgg>=9CAPa(`cpO}0v|y6J-rv|sEVQFNIzcmzuJi#@&NR`<2bG%7mz?ULcISENxUUw
zgfgIaAL7C1Bj(c{$=EZOZ8eXM)|gsl;EqJUIQ|1RUC;z_`I1HvAS32!p<6okt&kc?
zH31gvF~Hf0e$<y&mNXpqC5ijSU9E9X;{WBn+}cII3NHcN-1JU=Mjj!*{u%PIy^541
zAB|nq0`wyO4+X+N0wU#+_PSHGQ1+_1qp(Av_2{QDzKP5TQOH5WiGRLu3iz%{(jhX3
zv@dkN{$8}QZlo~2UCM4_1U72EaQOXG)G)uUMC_mQpu!WgXrzi_ldzUsnJs);X!rO7
zL#jix3cdmAonc5`_V<KI{tH9a(8pIEOx_L8>a0Q_$-kRFq!(d#I}18|<JK8D@$)T2
zd$%$?C%#xf30r_)ac!VDs<N^`Mr_tmXg{+2{w6&C#1~=V_d42`G~Y;Iug!AGC6xbM
znfxobSrzlLjBp3isuoPhho8|}6`bj>U*lXRR(H>e@nOo6F`G|4Tv{z&y?;!w&}~%g
z?9&36ry^iYo?DIh_rhQp+Aep#6PWDU$e_6RjO#RjJ~?pagdVTIedCynY-4ZBpR-=S
zY@~V^X?o#6{OM!*wTo%I|EYjU=y#EP{FQQi3`N;ak$U0v>7{Qofm!DP1cNtOuIk-C
zlDpdvO;jD8V0UP`V@s0%mLIXe)Hncqv6VmeXF0?Fa=}%_U&d>`kXZc7T~=L3gdu!C
zhv7g3?KWG!upoJw3Y>Il0+s`Pq|(FR$AgW~b`OrAOh5poh!^69;tTg#1pG>CjB0pg
zie`xH&ka|V0;LNs+S_wOXvMJr+nO5RqmZ5z$+IvycJa6N6nGI$2dQclx(&bTP`$!w
zuW6=z^gUqum$Lyl*9~Akdn*%Rc>Qt()5qYuZLYF7-*uqRl&AuD2s8eQ5B-wKKV5Ke
zfJ{(U@fF}l5j`h=>czV4L#$83r)zWqm;Gl@i9rywi9OiY+`-sxw<gyn2hKnp-G8YG
zL=M!VyF>n=rGEkQVJMuqHYN}OQG8**oY=h~4U0BO*eAwX`~!iyBj9E|(&aBn?LER5
zi)GO_$->?!nKK(Vx-KaKVByMUxTN%3L7cy_kI3^WzcFo+x6-wHuzkbT`q??V6x%uk
z+G!??ISH-Dm<3=;nE^rqxFWu$iRv#ljYh^a-$bB$e3CCTco6y9rQ^N<0N5pbc>mnR
zp6fb7L2>SnHJ@|AH||)P=g@ivjJ0jYf%NfV{xw|f$%go_Jzd7%yGM!4K&${TCB1U%
zh~UYaN28QvB&WX)dHm#0uqzeRsKo_VO?MnrZy}GqS^t)$=_SCNO<Yf+Z_)bi1xjtu
zG$RBEy_)qcrtCP^Ea&jgRrCbFsm%s&LDp6a;3Bs*K;`dVY#5bahA{jLeO$kEr_gA$
zh#@gPiASQ^Co`UMuLy~}anyTWclsGK|I91!C)@9}-^9x-ddzo9T$JqfOetbmz0oWK
zo$@py(Y3$M%*<eC(g`t{ocd{D*dJS()n_YBIj@N%_Ny5Hd3QFI_rPxBJeK~wgW;b7
zC+>ER51ZR*Z|{C{lLT@-1Cl^HEtL{BSColgKS4`e)hwnupV&XG7gV9YIl80J@n6pK
zTy9k`@}45TH%wH#$?BW!0$F{7)|y>M3cRekfI~5+9m50vL;tT<e!$qe;r}xB|I=BZ
zn0o{aQo=%PtryM_b?FlqJBG7a=!8P6aW@iq51VD7RkHF)SqF6@C&=RWTzf7GRs=k0
zOj@MxKBTNlXYPXpXqs=vjAx%3`B(&xH#HRwb2YZyA=Wf6skH?Yi9TcC8ga95WW*74
zeNW~yYn$os`O3ylp>$g8*&)`!wpW(*`nqio-%hRpTf7qDIURgef<1d_8(;O9E?|g_
zq4w~DfDn$;IFr@ik^a_(#!7}0m%}Xn!1c(sXklhZrNu!?tY+IoU%bny#<HSY<Z^do
z`TD|Gff2&Y%-oAcugFDl`F#zreI+}3G{Dmn&)SwRW_*pgQ;G>H&`gjf;ggo%C*-A&
zIMrjFDKl60R32XFH%?4Us11-PYG|YpaXqC(X1pVjYLzAWZi=uhlb^3D#s7j>(ZeIR
zu(Y%iL93<WNwR(>Mp~Bc<82Z9qy~AuB9L(R3qaBu%x0_{UxVrfaFY=6z(zRVd&ZUU
z7e{T-RWQuzR!<Z5-ZWS%3OoH&^k5c5SpO~!T*hoiHQ#E1U4|0IBt;dstu0Gq+coS2
z9;4OnV5EaS)ByVi$|VRu_F=-tNrhgtv;MYWbtdtE36PD72xFztxr^kfv<*b3oO%QJ
z<VDljOF5MUJjAQWVEaXvXg`dEgr*Nu)iIM_^PJ_(lojwaU2Q2WzYTj$r1xO>$&)8<
zxcgHaOB&Cga+|eLrmtPP*p}w~IAoi47ikxGx5>)-aXGok?<<Smr=jK4fw{rA9(+3#
zr_fZajmIBUhBBhgEV20;w->utj)1s;ZT0dp-$4t;MsF}?BRrC05s_H^OuFX;kK0^`
zo1zlt#EF2t!Yd>cTl?7nD3?yHdj%PU)2iRwt&6rZ%{-cjh+ZK)jFHzeAatS2i_<Y}
zd-hF&IIy`)K>j4>OX7e}z~|yI-1#MqY#_++lnT0iV(M-*^Duu2bylo@FZPurfAohP
zDr<I`kufRvrMhs5WWc;XM}lX{KH>c|QPok&YH#-NW3;Hl753Huq2eqb!&?f?t9Mvt
zPs_1$cDyJYIc0Z)`;#>A(4D4Y^)!Mp1^!gFUD(!-CVi`yUIJ;J`*%pCHG09u9uRaQ
z-U7Z{n6fgc{ilQV6%i(VG6s#lH*H)}|2#@8%t0cs%Q<zq0u)hic@SwVg4sUvjKNNK
z^3{`NL!Tv(+l)_zNva^0)$@a6v(!sdEt_!Pi!FA&Tsl^@!YLB*fC9mjDc$A$9u+BN
z!0w}Kci1?Nz?)Q?VNYSzKOVPt*<Gx5dDMF}mgQjCN374L+ZP{$Ug5Pvb?{B2sITFK
zjW-<}bjLa*0NDMlcz8L+{R-44p~*1EJDtal_8y}BS-3D!8SQKSZcyE2kB%TN{C(?|
zb3QM&aocTg+YB-vf++60bf+d`>1o5}b>`=W)wYV9ZtZe7y-EZFN*?<zx<9SGTC9Q(
z#XS7tm7|xJWh0gnPo9mUjFoV3#H<z2ExDXdkb5_JE*Ts|pCzbQNOIUyk$s$H5KP)z
zT1{z^X7)<m5!=nP3E6)UO~Hfw4U^-?J~*LQ#IMaH^~)bIti>z|S|hfN6LP*lKl8ep
zqcUDgti0_{r6h-yR!SrmjXv!zmb=iYkK_mJg=RqbbF~x^hPhb?q&2uEL+qy`KURuV
zjFq}RF`ap~5&Oi1-LRJS$e{>-cxl$TRnU`o&o#p+wR=5@*(=3&Ti5J0`>Z99&;X0s
zjS9GpTXd#{VC3h}RkH_OeNa!du+*K25vy&v-1#Av{Dv#N=}dR5w0a)#px-M~_4V2I
zn?m^t+Y!El4dW(zt<sqgp{#2tw|NKQ0YRV(l_ll(AiSCxob3B)!6q}K3k83Mfd_>i
zmDJyyj5TZ|B?^QTj__1g<%^kBlzCBnn(j0lyY#3T^@%gB2r*yZ79+VAxG4&YAvU04
zE?Axanna4m^H#KJE2(c;oZQfXuUUvoo4+`(d&+zL!9hFoa$2;fhppRE{%!+3l+S0Z
zd0CTWZNjXh_jtIfT>#RVaks+G*pP5Lr)7`Wz0?v0g|{>j*T;G3Gmu<_+!p9-pL8Z3
z=r>FlwcbBYw#LD4EIeDk@xBWpvBB~sou=z5lu4&?tMd+?a2eZBz#+J3jv{*P8(nlU
z|43}`(FX%kJXlOI%v8*0jLYp53m4WFb>|)h5A^cbJpT<2{oM-x1lO+8eYH%)Q)QT2
zHIe^)P2+EK1y$3$G}}b|w4}3Cyw;sEW?HDe#y;;fOVAaAcG5=bFw^QdA%`S>+V5hN
zw1XQvYdWL(J{xdT08zhFylS09ASN&I5^gqC-6Sk1MZBegC~ZUOMJe9SDaW~FofpOW
zkkg7LEP?K>3hRod$eKLw5_*N>G~2=Kk1L)$yqT=~@pcG5MvM(Xp6Tk<pSlHUpR!|C
z<L~Fhu<@&ily%!rR?Uva;rAlDov#h|@{Xn3(<u_FRr&1D){6NE<H0aRc57>nMpt^j
zH&^murn>~;ckcKMbfrA>HR?(}F}iaq;vCe&j~Lyw0qDhESlJ^N>V7vp8>2w~gqp3Y
zmA+^)6pbzFzy+l&u?fFrZh~=-X2Z*glh<&8!-MF)nlaStfJvwE^5msQR&6+i*}5Ia
z{G|W*n#K@4GfS%dv!dop+b;YH?}dQjD9rWC&Nbe%x#YeFXLA<_dFJ<?&Y?Jx*E#!f
zs`e${>&&@#K{XZPmUTQlM<^R+0}T(Xvn>(y?8Wluz}Q`PVh1>@;@M<RTJOnQ=*$!b
z(($YWcp3<7uq!<JeDVV7$bdI{71?eH$}7kG+-@&>Mr>Y3{OPl#e(y0KDcdxDe(PJ6
zEZYjfq<sk;5&!GW!sVyMA*}*SuJo#})IR=*@LVneLcvWx5D!Xl=<yW6eecAgnr(kv
zTs}KEQE}q1RH0$QmB~?|Fl{_PCR%<zzF?xXJp-%ZW~>y3#rP1SZ<1s~9s9GPVrUc%
zgfhyA@~O~!e!Ztt6tB~&URP-bide7s9^d=&t!c&FdB5%Law(e-epBj)T0crQ3%>Y|
zpU?!5jzv{X0<QuXAYV)0F4)Vi^pQFQ>^I`{n&@=?#2$$<R0shBQdLe98VaT<vnWET
zw$ZrB^;*_8_lQnudXMTZ$>QOfd9QY(D{a7-EoCBebIFCo+oJ^vanm;@5Gh$l1B{cr
zG&Y_bnY(jDzKv7Y&gY9+*M~$5hgB)9I`R>-!UfRA)ps?^k0^v#%Xs5@UYQ$rVQw5L
zuscA4d5KAMXD)M}_NEIi^mdAQ;v=&-cdrbUY<NxW#n)VHYjYlX)%h#ihz6|XcBaMn
zEgF28>DY8%VCozF=vy{(<yw4kop;45d&i!lw02#JVK1g)%Zfj4)zuzUY(8ApAbLzT
z7X(O&&1y9UJP_(@j@utxw_EBDUd|#ncDRPvTj?oepvQfOZ+a~mx4)a;&x1jSsnYQD
zzM`7i2?FR7w6S3r<ZF(J2g`bnbWBi^f?}6Hw7qy3JiqCs`;_1MqEUwX$3_fcZ)K|D
z_4u*bs}tkK6>qn%>3hd0hFoY#Ud5xJi#=tG%}fUWuZ>feVnE8*JV`wyyPd+#$zIq*
z?4&N&a_{)jWd!!+>(;Pr%w8gkM%N<{H-5#tU3WnNzBx?xSw}Jd6E2-JuS)1O84H9!
z0^CPY?@F!OjaaoQXs>*}cc4Q<d!lP$wt|7#liL~1BYys!(b75t7gkooXrcwW%&0*=
z8v6`g-eg~af-ag&fMjVzSn(YXAWKB43))1I2P<`G#)^bH&{ho*_fAYrS!Vh=zIsdT
zIhd_?AsiyV38)7Xrp=1nD5-D@PrT)Jf-&&qyKkShlye%K=QJ7G`*~VJM;5B~FUh1V
zj!qg+uviAYNS0H@Pjsx<>g+Qszbus*_+Y~-R$_h5*G5vf0_u3B%&x?az7@jqa*N{>
zcstWMgX~5$_Igukb&PF9ucN(sgZF)?+M>A&3P8a;dVS|dzE}hD6?S8i8Zk}*>ACX`
zPH8rDD$B(Pu4lAR#&Py`KDHHW-}lf_+<8H_?F4=xKT!iq2VIb)ARl#E)b+yGPhl*>
zw4aA;v}Qd^)8zn8IF18{1^or-zeNaL&Ug4PrzQz2aVL&T=8L$YB)u*ksIM8WuWu3z
zc>y4iHWFV}famMo2F@k<jTJlj;&L&VB+~I%0aW=5Q@DfEG~7YYNq+*Uj(|6OXFzta
zUj#^oqaBXDMd^(66310hDVcTBGVe%8tRlut&v<w#rA)Ow9UGgBzG<bk#j{guUf6st
zXsBLee`#_sg2~Fm&6(s~BVY%9wI8l-lEr|r`Q@D@x2ea4(e(SP2c=zj(Lqm4eyrBC
zFhEL$NA8dTd5k&v&-EJdf8^{pnh^E$^mgBpYB@@9SPy2Gh?8Fs;$KvtYBdDEW%;rw
zF!)vXz@ql%q)>X~XRHf}<A%rUC$Q72wv9^<^tnL#RT2MhY_oI6+YTHmVl_jT%lz#6
zb`LuSjtp&{oC37@jGmSA;_<0e<3Wm)ADNEt@>ki}Pa5bNj54x4Pnle1k4kX1OoCo;
zdW4IGi-b1mhu};$H8uNf=m)lGcB9x|ea9UJ7{*j|5ed?gnXs1{&1@`?@jhmuTXTIb
zetB_z#U#1X&MUPFI5xBRMbh%})HH7E5hJItdjZ;Tmbl%=F9=AUVNb*YPg||bCN5IH
z82O3bMX^`)>8l+O5=X9>cKxoxr!CLI)kUP;sCYu>!4@07)1v{&h+tiBNUgzC1Fsng
zw`4>XC`$p(GZLRKcd7+6BsP+S&0V33v?HlOXKo)62mtg_o|(=m@!0>^-DNshEhi)N
zUL)?fgC{;^yUQi?%8<uU@%c$b$$JCXLVMTggi!CJEYV380c5)|sUtCUmo*8*T6&Wr
zI-%9{Dh>mSY0lC~a|5TAo-2|j@8-)%k~$6SIw0CJi92uiwabSn;v-GfUDV7;;>(mK
zLYlO&hvVZ%HOg~Ck|4xl&_xj@2{)_)>bX_)5vmBQEMzIY(`ndJQ$;xEO}?=<b$orc
zJikCkIub=6n_K=Wena*^Hq=2|$gL}Y;Ucwv9P`w>vC(x0{oo>ZGKy$WkOh&(+ObOG
zC4m<Bs-;$RIipF>>`vYXuW0!$1$s>0ABW8$%PFR$a(d+bj1Z?Fq$=@{)18t|Vi*o^
zcsVb=L?$DE2p^+||GPOIfzh^munD6NmE|46=Zt+1pG+B!icx{|A#@XRKhTgo-Nq#8
z)33!(GvgM7q1{BaVIko^2>Qq^9Itn6EOp3wlE;>$vb_+?m8tGwY-Sk*Gq#nz55KW*
zW+_$G)NF`JNP45UCKX%fT2C8Z@xF{nea<Tz_MLZ#f0iV%wqL^}tcif2=;=*>A)yBd
zFf4@7AoU)a8FgGV<@Go{@<f^U%ZnuG@4^X;p~d9i;*1d~2X({nLWkG6Vv-7cvrgN_
zRkszx{F--6ZW2I4z&YFN_boNv(_&19UJajpm2>?%$pG$GWV&1vNIC`F5Q~1n6&=W8
zvA$->dyR<ruA8Xm9U%OXWhX!|vz0UNHWut3qFZSjRw)vQ5!`9xpp;9_s#@B3{=jX&
zOGDiwZ^M7j@8-CF`~2^Y`%j0N133F;i}>@W$Ov{aYJmun{-#p;fV@hJ9{pu2abM)_
zlj*9){TfoP_x*m?2L$HbfhA>8{Pt4%>1Xu|-}eI#uTpm&$r}kZiZjKyv`SBB#T@*6
z>|hR<a<4Vci7dr1m=7uVOvvz?FU)YIQ5GMjhaemU@ryw*?PDE_N^)Xz8z2())j#2X
z<IB<4k*0rdUO&a|dwHh^z?|e$yRU6f^_}NNi()A3kfo!Hj?ACRf3#+7bpEQy?eNVc
zDF*ZLwG<j8X9f{z=ySkX{8`MBueuEqJ}hAicRrnpat0@^wn2rQBPQMQq}I5p1Iabm
z2&2C*2z2C?B|4zUL=&P_NPsn~;fd}t3nC$y;q_HC%A0tK@)3WhEhb%Pyf=%ZxF=>*
zePx(nDF@`cwf$&$Lx=fdd+f9M-heU^pS#Lu1L0YNC#o3s=Er6>Ax_g5p=Uc|LMR`{
z`_)3u=Ff(E^Vw4@SXE(bY$#TEFdc*o4d7z}91XSrG)JNxN5hT^xxS59;14GEZvy3n
z_l{0CU5=K+vi#CgdEB!{?_n0;E|vl)JXish3A~y&mrdPci1ZUF@zb(=MPh$_SZtF=
zPO;yUgT(N7k0Fi)$g>$$TtI3t>`Vd?b|ivzqgQRecBB_EwpGLq(_%A7f>u#J@$jbh
zrSZK9M!}q)#$mvgjV88;PURXoS5RGdl%Bun*EaF;)fh*DatSZ<*-uSUAk!wcnvkvd
z?)dlBpDiCJ^3vvaT=iPbI9m3TYCJ4|{hHz8m^Ve=(Wkpur;IfhBI;K??swLsZRYTO
zPfJZp@BWIy8~KCNg`U0)FE-#@Z6Dvxr!N`&PVqiFz65V<hGjM`WE)kPGl3w_CY(aD
zV--FO<sQ$(dAA#l%(pbBfCEuWC>(L+YO-#rBt1U-Y+Aa^rH=jHY}U>3sR<`W2o7_5
zd_;9)o;rRbHtkSoN8ASTXJ~B<jnz0$ZD&c3i&a|qESn{UOu7I3iBrmva80PCRhufZ
z*fSh4`+KEfGwh=A>@R~k$;tS4NRcBm;XS3Y6)L&KwxlQ}LB$TQ2*b|;1CJukuJ=D7
z9jp3jUK|WVkdKASkUhm3(dkpd{G|J=4$JT+`;?lDEj`JWF$A3Ho6$F{0$rk00h9Mp
z{AuRaM%8EHJ~RV_Ix+hsO;hsIVU6BzUm?yCU?iHpV7{b>HKW|tP6(W>bKV_>3z_my
zl#{d<5(}GI&(q-GU#~s$gXV)S*i^t|#bD9G>?@qnHy%0(hh<51ecgs(iU^~q!d@K=
zZdAyRqsXO-FtpR|2dRB$$N*hbK%g`JSPmo=p(rK-qsjS5$DljkzaLPJB|+gnVTk*h
zRi14=Dtg3yK<G9Odv?G+mK80j<WZG1cEolTPYofmtIfDvAQ-ffYsylx0|Ev#7=5Y3
zvTtatH>Wi=CVaZ@AKe3f{fOQJ({bL-_*LW0(Pirx!iUO;&s2i6R5VR4>`cZU#rgO5
zeqEN;L#R86#S2Eo4T76x)iiRNt|vkOAuRDGeEjf%t5Y$}HuNr;5gcQ@+;blV*Snkc
zv+r63nT!OE;k6znc9tcXp=1x$@3!bM&=_+Hdy#*V*X|GCb(ml{_!`sLr<nf5!%#Qk
zE88ljw6T}=>D>gEn(KrMxy2uR5hPiWpX)>y4w{5H9mJ>Y_nyYvtn?!+3(xut4x*<<
z+r&--P6uAKMW7{jdotsdUsz9KsIcr=Fl!&(f!=7+^6_L=>!*Vv$hT%}XQ_*jE1#7<
zt}!Fu2qK|R{W3(_p2>}q!bIuf6c$VK5fdkTeFfeN>6n?3+1C)|jxHya@$-Gs4A0m`
z3N{<@gM|s??Qt#35?@(`Au^2Ry%2)Y)T|$uy2DeXcSu*x$29Vzc>)jklPSndGY8!1
zDH0UMno73J*&H+*+w$rY>7APor>48G>0_cwt>zed5bd>1BAei>-h)WWV0SFLH3PLw
zP?9(Z1?6OxHhDt4kB62NzK9jinUwrkpFl=l-yC%4a1R}uYHTZ>450K!5zAw}9zwE2
zAARF|H$Cz>m{7?Iw^>xu{n*d>VW4!jGosGbn?c<#Pr?2Rg$as*Zyu%(ESnR4z0CJ&
zG`%+jsvFMG6Ng?dviIM^g_V#w6I80afSdfXrD`(pgE$W>ES_~~g1HwXOSDIkLioXz
z`6%y1iM>H^+P+t&Ni`%=fb8D8g-h&gE^Ah??}1y_{C%RDCT{31_-GSZX{DmgKzyV{
ziK^eAjD0G!S-1`|Iz)B(cwUiV0eLJ18}byY(sF50k7(wKz9}Ap<VlmSxPpLutZDM&
z*gWMSshvrFAY4)ql$ILA3JI*W9#KQzROtMcwlm`#S_zF7HdhIMTBs5k=YqcX8n7O+
zG93HUc<Dra<9+=M6P$#nzdUObrW}Faxsz}8Rz*7@9=EXvcb*Dk6W-dqKN#tl_07Zj
z2PrIz&|kj=(4+J)JV+(GrtItHbREHl?RHx?b;?83ox8{MvN1$<9nmCKABtLo9k1#X
zCI_@FRq$YbPL{yQu@&HnFCsSzshRo6jF(x4_s2x{a47S&{50MxaBpzLh=xi+y;#Bb
z%ll?}yuQfk3@vvv6Qq14oZ2E?$W;%Q*PE5j?J3N)$(S4*e9mi**<$v&i%GapV!pQQ
z(;+z!_I$;yQ6C`Wa+Mgi#--xB=Wz7BnbL4(z-9==%;%L_N^blb_q6v>G4mu422Bk-
z4HMq;L$1h*U4xh&Eh`w;PL)JXW)$C1S(?BmWI_@*@{z5r=yJ_AP_wU#6z})QQ&A4v
z6^>pS9*kwnqI!h<?h@q@mFS?krcJ<Zux#;@B8?!|fuo@sN1-zz;9<z@;%J@7?dcTV
zk8i;WE>BX3?|dN<@y=cn#0u3-|Ik2hv9yS`6I?s)c+iLvtXtE_{kpHxqJ>BH)d$(D
zf{+l$q`tt@nM~3!U5+LP4(A*sVU%cRI0^38TxR~o?IvS$ts9V$N%tJwnz>UyA3};T
zYx+l{+#Pf4IIUbe>5@S<2jkfyG_v=q2B^n8j)^q1v_~ZkErerhSe?tOYo!GZua+zB
z2lO+d<j;%Wp~<>Rsd-%v$sppfrp!AfM`d0U`UYNSJE=Zx_3doQB{Ep3Xgi;F%=9*+
zMK-KAS)j(O5_MaN=I5VsR<<%iXy+7sVxC=-71%-0b~i_AYNrHc%d}pI=u6^tkN4?j
z#jE*e6Mo1(igd++#Tb6qFfA7jea*VeXifoFmX;OW#;D}za_&1>tjU;}R)k^*q_M5M
zt!Zy2?2Q;PmIy_v`Qq>H3?#A*TmTOs{tz>t3DY1|#T&4AVj?U3^#?Aj_|#kOb5-i=
z(bxSQ?1z#f=$aH_1D0=%G+wS$BJ%4Z9T!<%-P1Ff#Gln$W4WG3T24_$@-)KHd`cZ_
z1@dclMc;EWL*JwD)_6JlvM{O18;J#ShR{GEAl6KT2UFd@TpU(4S&#)An4<3b<0(>Y
zGehp;!mRXx#Q6g?rZZ}2^oh{r;1#jh%<O?)J}50N=c*>x?e1nZ*|lF+o40ky&d`!w
zdiRX0+o_f*dB^5`FsQBw$uk=oisX4_Q8(5O%K)YAyf~D7ZbYAtT1yqH3-)hgf(&Hv
z@_mGE(~%qE0}uZ=1B+{LFFF%KcBx@(JBu4)g3LzpLLO1XW3Xlo0jc%XObHG1gz>xW
zM~$D4&EOt<Br3!$;A@qM3gR;LTRF^|?$r1|1-rXBtMPIN0qk3bq*dF_l5bddZ{p}p
zD>@+)>uB+~!|2D!jPvP-NW%~AJY3Qo^zdj7aynKhR9#4%=eCM|#8)2J`Ig%G$2FaB
z!!Zy)D1H}(&doGh@c8Y1WP}h?KAj4=W3;GeajK|qx)P1NDM5@O3a;5l%Z<@1ogGht
zQQH|CJaf%EoTZnBEnW}!nb7i;$t>`60Yy;No$76~i{pTjXE1Aey+^V0d`Ji)VA&Wg
z?D+g_k$#}^Y?Y0TiQU-O0b`Uhkhwi`fA53_&ptm>i1#&heVzNek_5D#t5^J0Z|;&D
z6DyZaMWj9|+%-S@HZr0^J3m#ku$j1;K=%qVm2%<~coAcHv~I{)qieYna3Z%{-m^S6
zlY=54)dj5vV*L=^M!&RIPMDUQIrkG>cZrk2i|IA9kvyIjXW(V?d%Rd`tNl_dd7Q?#
zRoVS0)9Q^Tle*1gr1cce;Jzp#^S*GM_^5%0I&%a``SvSBKiGN)tL(?M^u-H5i8^O@
ziQ)8eCq*D2YwBB2jo}(YOv?Sm3VPpj31t>RgA&_4zlUDnOVTdtTJ`ZPblk1<Gxl>4
z&FtIK(v?Hv%5qxvnrmTP-f|M}=gQwtws9Rq+d$1SU(Rw-#}>tCiW7Ju*0XA@4PWBh
z%RNwB7L4s}Y+;crIQK%1Pemjz-n}bc$#SWij3K5Y5tQr_hk$|^MZ(`U!wsY?L~zE5
zI50QHcB&AVb;7_C-QCX`J-5g|xe0$P78#Jlf9i6%hoe51!X}u2Gg(_w(qcbTKc?r(
zP1e!O=Xid|S&(dAyf6NJ<fueZY!tpK7<h<DajbaUVp8^=iOEz{@)q@m=*H~nUMmYi
zRhJ_~G0!`vpK~TZ%gCJ1@#$IF)(6O%fm>e}UzTFRz;#h6vV|lPAKv)KK^v~D;*-Ue
zl^?|wmMLgBp(}?`vRz3RX9><dHf2u0Az}<Bt(IfYXdUa0(U-`psF!|_A~i-CSQu~5
zmj87;@bs?!3r&;2-DHK{?k$w@7R0z5Y@<tlRg)N0pe^m484mdJzR3v_fC0mtA*aGT
zA=N-EiB6FU2v`R&8^4CZd}czGb@Le8vO4eI<sOVn+R#UYVaH@X^vm?&XF(;DVf%qt
zjC8zDwp?-Y6i~(ZFbq227)kiED#o@Zw<{t~t$WugiOhGveWyB}`!QMKw&;6xreJ?Q
zG5G=(bd7uRv|b=mhDzk4&!xbRe|n)=WkAGz_8LnbHWs8O<A7c1p-y3*6T87Esy~8H
z#8^LesHB*L#t{A=0Kq^$ziGcE_H_o#RZlIj`+D93`Ilq&EVf|k+T!I2vyEY@@LO(q
z(zdl}MO<gw>j@V8c5>}-pwy$XiVDgTPn?3#nHAVh8S?z|>#$Dpn^rMGWsU1gg@2Q|
z&zzLSDU=xhDW^=eiY8B_ZoO1MZrirs+;h)TC_{Hzt8Xf=oMYN9O?QrpY+4qcd+scA
z&N;_gg-#p=ER~JVKD*ZZ=tqk>6>G*SU8%&q{PKBL`Rmfiw5PS@z4tCP<P=z)I@XxU
z3d%EDGhcY&oJ_a&`W(w*_wIw{cfWhZD&c7zPlX>{rt%>pT4B>F|B5RXz%sMN{O5lz
zf&y+|Dm)dW?|=WP=JeC2#Zqpuz2iWG=bv9|9(-__RpPR&{CW~DFDf<1jT>gUFdZt`
zb9GeKKKtw%Smdq(gF{x4$bk*+dGPB973Hj(oD%-(t53opiztO{c1f5oD$S{=zwf@+
ztm2sYb@_ALou9F>9#UyLfBr;s>810`F~^L|o=E3aT9g0lzdnLN9IkvSFz95ox$e3%
ztfG|9C7%z!%&8pZ^Y@N-oM09BL3DxQ_MUr|n3rDK0GU`%u_V%vTC8_|3d(zW#Issy
zo_cDf`R#9C!2ps5YXQ#q$?)v6XW}_MfNL*9sk#@>O)zkP>G>SUp9)kemCrwawpHf3
zvI^u&|7~h&wbR;DR$31S3gy58Jsm22l#YjQDdwxgD&#-@@w1_j9+VD<BepGz7OgbD
z{`JfDm|@ul@OeM;%o&i2O9piR+nvhk4IB2DM;>|84r1Ul;1tD7(9b{ptOryevrV|*
zf;n~&Nm_Q^mEE!4-F)+tcEAGT<Fm*7eesKz+v)x}n1XE)2mMfSoyqYY4piWCe$6!t
z?cf^Dh)`HJ2FYC=JINj?uy>$gZbZSBpW~>7c2r)N__wv0rj};Yip_OAnO|1l0o9!T
zK?>~hEB|D3L<RPHZ+{h2h(m$x92MBGz~&U<Nm+97Lr8cu8=G;a#TTyM9J{b?jc=~=
zjbOIXWrJFcmrV^DTXZyz;8_X3j9r%E>?s4|<97_MLU<T2dZuOL<3+duNT(JJ>OgFP
z%z@m&)!hHGVduvart9*}XWQq%aOA)jUK2V#htc4>@JzKHBd6*xoh`?tkxdfi=47C(
z<Z>u11JYzCgXJ&~9Z!}?rhHJ=d^=KULb-qQmoI0qz@7vP?D>gWV6$9o(V%QvFrerl
zr$F&CIS8H6I@TEk>@62QH>@*sOeZ)Bj`{RFiV=>V{O#-Jt<V(Wr(g>4TmSN0tam!_
zDA^cKI1c(*>FozT>*?Jeit6h4Av{xDr<ab;6+arXzK{*`#d@Da-R)ftdVco~jPRWQ
zwMQ_tIX%6V!4%@pyD7wb9(B%lx^iG;b>(f_ToeS$!a0Uxo#v@Vbn+W9A`*|6t>Y*n
zw0~7*5ZYSC!S!0h_qsTGvv=)1<6s*T?$<H>L1^n(XM#2|oj$K|gspQ{zC70_%C;6i
zK)D9?-JS*Z&DYNFoI;!__s+2l>@$u77n>_0`lF7SDzEjnjbS`-9A#m<&(XDfFSVfW
z&>Z?<UXAj=E5W{>_)h4}^Ub>Y$ci22lXopMhYw}9z-BwkwuOC$MzlB0XqR0*<vX13
zO4ehZOrj3QVVBP^%q#Pp$<^Xn`R~V9nwuV2)nyIr%drOb_danZrVy_?YTaeVlc}C@
zy87Na#>?lF&nx>_4X%&Hii>kWA0j9pmKpnW<>a<N0eI36pECv{)Kd<1lv9Y8f(3TC
z{iS>?FFt$zH1k0@vcFtm%aHA~Qwp+dStt3dv0u)x{{4(Q7w6JrJ#R)k!*>JYWVPox
zwl8(4>+GMi{p)plHV^6d30qbqXWOu_+)Td;^|=l8lzqQ8Skl-GBM|dXi{vEx!_2#7
z!#_AC7q-QGcIa3>mJP~*&sJ~RV3z6X3~OL>mWOlg6yo#E;Lryqrp*swEs*B`*rfV)
zkd5sU`w@(@H!=mejQ5xF?1ZJ&GJG_<e+4X|4mO+Vuw*-b`UrEvF(b@GSjY9Y-o%qc
zp+unY1e!2y_|`30^ZH=$y@%~2jbA2Cnq-yN;r?^srORK6Z2J1Q0{aONpb|U${>mTU
z_%T(OU;XL@^S}dd*jxCnVx8k}eIAD$-&efG&4r_~=IpbNHSc-P$ySNTcd+lMKzjM*
z4N%Uku!<>$adGp#;w)U<eS9CWp1b5-dQ=KseDQHmqQzB^pdW%h9M@$2;~!tQ)|@U*
zre)u0NT2c3G0&OuO^edYFQ0EML~Xz<f2j1m^UlSVr<*dH=We>`LUYC$(`~+@Zl}Wf
zjyqmTS#!E@l#>>rRMee(@)Y~NcGvQ=N1x<SqglLolT}{R`Zvs<X*gD{>{x!m7SxG(
z^Cp?;(?@~zYCAZ=$u)7}F!S+`pJOfB!|dJl@SciWDo|NZSYN_ZHoNqxYoTK~v7Wek
z<l>=~FsHcZ;Db1h@zJ{bw%cB?O60K8n)QO|y86NR?C&C^9d@*ezvPnRps<{Y>BYM$
zIK$!mV=9Vi@%-qcE36f8n7PX@)8@u}2G2#;w8xAYVinL-W@mR5P*iR;Kl|CU*3z40
z!EN5WiH0*OaJc|l{%3Nm$5cvlAO{C61g;Vm%^NW&WC^U8ciZyd`4J=P&5bvn*Et)*
z(QN^G;;3UUBF}F^ncQk+U|C&y>G5`|dX_r}6VUOoBnLXm8s=k9;tv@z4_95a07~Ww
zICFHGRMh_FH!oY(Y_FWl6kotQsM9VS<>0e--F2s13ut$qd$une@bNf)zdKp^ynpU<
z7u#o+Q=Vs%DI5pG(<1$@yI!?S!q*rl+ebRe;o2zf!;aIQbDDlyeKTKi94-F;`JeAa
zU!&RbWjY)r#P<B@r&nVT%XT}3JL?{w6Q)lWRTdmj@UC~AXb1h!O1yB6GhJ{U?0D0N
z_eg<VZc~W2V|wroRIb)mj+AIYMQ*Z%oVn2O0~MXn5yMBC3a{d24&Qgr$^tvTWk&_}
zdv0H1HiXx}Mx*`87tc3Su{K5OgF_yZWkdBAMgXnn`XBgk#|DgT*AR3Zr(*q&p_oR6
zVQdqMYkScEufaMgTlO9@TMsmwZTNW2j|_ZO)SxjPj%npbpmCdmdrzA-$_&MEu8wK(
zUA7g|s%(SOX%9MC>-IL7Ef^TXv?w2C8i%QvN7P~ZAy}u6!H3^TSSw>3Mum|Rujk^}
zP#<V)GwV?X2f>3MR~Q99_|3)oJp2$x$MgIs#X&mjF&)!Jd_Y=_wRQP%o{ox+DZ^@E
zF<%Fn>dg#HdoyN8P2N7lxN_Ks7yo9k*@<lnEKAqyZN${w!OM?f21D)<D7%SJl8yzN
z$ru1JX*kxgL7S8-8=JA1e^`GgN9HNSlU)Pbeh@>sY+;?jbST>}ebs7oSUIJ+b1a`}
z;4pP$y%~(UIR;bau#C_@&Y26M!gJGLcA%cJ^TB_?=S!EGF)w3UrmsJ}!5l%Iac(Z2
zkuSYxz8Pho4O^n_684leIA?1*y!;5sIyk_Nd<+H6;gIFRSofnNY3mK@%qrC1HE%W9
zSa_`l@lV1i+hg$Ia41Im&A`+(6Hs@o966*_ib0wG?a1@|uRaW*p*9JR%=4X7i2Egx
zR6ZyRpWk(8JDO4h2s&&VHUVwNU~~xCxn_Mnh<dXLa_>OL{7npO;zwTB2G*j|jz?Z*
zj~!~paVj)C<3mul(zU&JIbd6{6X|V6+Pm>AQCYlpU%I}qO`L}IW!mULW(w9xn}%nK
z_fQFW0Bz_S7zrN?Eee$Dly3sUGfX}l`B0{>Wtc+zW;=!WB+Ib3Y3)Z|xYl-a=m$}z
zXFZ)x`WfcAkyr<Nrfp;Llkapo!4c~d9TkwPcOEjk@V?xNGTVhUrFJv~?=rr<_%0g+
zW$_G@(`2mu#$`!{qYYv?=GwR5>yA^1-*DG5)9fj*e|J6Bz{Y6xpxvN?>xfwexz=Hg
zr}cPO?tol;U$Xp$p}ro2I>CA|5q+F#=r;sPFPo)4_lF!eF^`-Q4teZa;vE>=T4nBA
zyxIKwE8F;sJ(a)H@Z8_@j=5$urcC4Sjqp3bf2Z&!{-`}=%7|V)OXSb8*?_*)R&1P(
ze+^`BVcAduTpuy8%1p*PVFDCOlkmZ2I<{&%NP<%IeOJm;i1Qg?`@0SwAA)`x>rW;6
za|6v}wD&AOI_@}FWh2Ut3i(Z#j&U814`D5DHnl9rLAZ~7<tg~+G<V`qyxXzFRS%1g
z=~1@zyV>_?L_hct`oNfS4l=N-#rG1Q>+$$lH4**b$#|}h9W%(vn`=8m5!ssZcgIRt
zY`C_L`I>?@Y%KcwK_1zE<NLDJtU+6|e&+$R0&VtzW+;yt2mXwQCCii%m=Y3onZFs>
zH|F3gzH9U0SRQ=$QPzzp2lh8MpkKZN{wA<tGWBSGhtis7Xbsp5!f&jg%>&My<zStl
z0{fmNol}T|IbAjKbbe?JY?c>)8tlPGOV%a+Zdtjf(H!o4pON1Ll+_gUi^s$Ai!x0@
z{o){iq#RkXA9~Ak>&^FH*v!#+{Bac&6zb21Pn=-BaPe#_N50A@sk|s&5y+K5puFA!
z1<?M^HVtk2fEqInme(U-c^&t|bgt<ZBRpHz{yT*@e=j|V4gI|mP+<S`r%z$J-krA3
zz&<A54YUOI6<SneaXR62J*0A_!&*+#3NX_&wzMQ%wCGJJkl(OzGHyD)_xRo(I<(eK
zJxfbo#>E*+cjNuF6Ypv&G~K(M-%_a1yz$2Kp-7&TF|8ox^j|7sX~|5*5G@G9X|u09
z62Ezd;#&r%g=X5!GnJWCN;wyF13EBW^GK_}nmwCJvZ!}pg|#aE!4Dp@(@?u9lxb=H
z<u6|W<<dCZ%bMTSQE|llQgKUVW>~RCIp6lSndYL4=2&arOpLq6mb;IfXnjjdQchd#
zI58%^XITcEs+)>hDp9HAbyHji3gtlY&oU%Kr*Px_oLc_R|NMf~bm}~p$&ooN*B^Rl
znY9KE>PJvd$c&buRMv9(XgW?ay=~h*t9*0HJjzO|-H(6#EXyoVq8-iKrO$k@-ar2M
z3Je-pA5zW+`Q~)ZWAPn@*3L|W`J&Q~d1t=Eb(2b1THJDwMId8zx=f#Sg3kp9j!?ns
zo;z3H$ePn+4<1}+>pj<krgc20)aEl1PMcGTb9#F2qg{o0_~B*NqSh%anI}$jPX(&8
zuqDGlhJXR%V}5S9;cPpA!_{djqMvwTxt-O3Q@v9j)*C9rufP5*JDoer?Pynx*&ZH;
zGI^c-KFaG<ijT&(B9<o|pLMo1Ec0NB^k5bPuFD-fn?YI(OIaxs75MLe|7o_o{B&ti
z&vmYOKjRLr<9ETqMw{>8T+kjdJ#GPCh7AVvP=S5fWhcPj_1Wj3lkb23Q9I?g>ywh_
zKmYT4%#<n6_Yq1K$LY%#<F`NSJ?mq-E>j+3rwZp9<gCloOSA5AU2LY|92qd}IdjIr
zlKo^m8-UNig>`%0!>P)@^rhdT{bPRddjR$O_~R$qHjs7A)dfBWqwzhDGaFDorcIXN
z!5)m49J%zu2R?8rrjZ|)%8N@gxjld5T!3vxtR$qsM&Y-%wV)BgkI=>j{1EI~sY#I)
z*Hvixt7U;r401UVpaT0^Od-AoA8Jw#A3he_HL&Y#1D-mcrvrXadUf-D^B+&IGpnJT
zs7B*<7#+L|rVKNGcIH$w3mplDX+mdV`+)}Y!lwP^m#=O$Pi}aNT8Tgkw%eVD_Ny3=
z@iFAtKUrvw#fv#tM+Na(bS|I3niY?(*<+sExE~K;dNfrkKBmjo7#)Bsj~Qw%!W1dz
z%^YP$Lt#M6z+4>KYO`|JVe?<lt~ZO2$3gf=+k$+b3i<!`;#p9P!&(-8ew=!3+d=d6
z`n~4PWjo9UC@@obXM&W8FHR^1E|@&jym$TtbIBa6kAaR+rVr7)B^SpJhy38Y6m{hB
z)w|6D>)(QcA3u^HTNgQ8q<G=a8#~xsHD|QB1fxr*<3nOSIuZF&BDrMTp4o4LjNif<
z*#F_Jf&HmdF@?BY1KU*=cUSNCR2FPRhwrJid(9$@5`1dQfq)Ck!gpkaIx-)1^gJj5
zu9`RA%w-3Xb%q&;=~(st_odC|_NCjbq7kTV(KAIHT1U%key9sSyhBmhrr;woolA~y
z6&~}$p8LqlI@y3{>|9L8`{m2#KslLPoo4*bwW7na4IQG_b{sN~tlVXOzH&EBYSVRw
ziCa|ogHGLSeBAxmf(hn=S!3*A3|dq5WykCVYhZg8*!bxE{AbRxYhW`>AsnB**I{+^
z|DIe2MMHyW1M)_c^G8mYfVw!&Dv9_Zd<Qzri&2js!#Yh*Z`f~EV&IT%@4`Yt=z9E^
zylVPLb2-|X)2EKGgLqiQd*f_dhYu1Dy}8p`N8Ps(@1%xMeQ}W^p>a?;et7<P^UmYP
znwj|cU5E9XmSSqLkKMc6jDUO`cku=k2{)fN-CT0~c&jvUxA$hx=Z}K`tOfS#SOc36
zARV6ho8LzZ>|^XV@!(|7|Gtj4_TQdZYu2HiWx3Gt*<{l~pL^CbJ;TiVm9s{fPo6!)
zTL1O7Tx{arAI`wB2k)g7s2{iC-TW}#%?(s4N7Vtm<lm0^a>?v5<|K@opE9z}jKVYT
z6k=ZV8NJN=!^^juPv5=F?DrMem_qzStkpzIYrf0h0KeN_*<yYTD?4aMkB$#8>+uf%
z&~c;9HTZElcl=;Gh=Yv#n&bQV_E$EU+we}{j0vn8;d;UR?7~NkZ3ojoN_lQD>h2`G
zr)!XD{+<ZG1N?f%Vy(r8xVzB5vX<+NAV)_zccFdRgm=^3Z)`U|UbYJzKRn;APnU`h
zsbNafVd%enXx><J#c^1taO@B>4E-BBD?ut;-#oAu*gt{=HU@U39O~#@H^KrNBiz}q
zD28L(Ml0Dxuu}QUN7tHZ7<@uE3g!N8w4ondILTTRv!3ijJI!~)19-3B`uYxYfX^e|
z_eVKCPl)*6E<4t||IF#=>*xM}k;fi<VB3f~eBT>8%wy~Jn%C`;M~oYr!_^rz*O>g!
z+_C0bC?u(LXT8a#onbW1Nb?n}Q~3Ygx5Cu&`M|lM;PI(5rkab;hH=0PmFJrv(=y0%
z$E#b+1FQBz7AFjDpyQ3m^JKiQ-hbRUa{+!UoH%)y?GM}Qz5K(rZUcV9E!ld&+`D|I
zd2kJ6h$;u9#Q}oX;hlda-uYC-v;Ub(k&t7iDa5IiKL-lz|GH+r9f-_XFWAp`5$}cD
zz~+_Bc$brfhX)nPoIY`gc^`hWTs(WU8IL~TfS48@xuT;r{I4O)_ZF=+%lLaY+}J~|
z5g45OO$_L{_T<T_wlLQ;6kY^MjR2dZHW<R}*}V%Nr_pEabZ*V-2~c3u0H&NAo0<M=
zfxQR0zaLwmxGHYBR$xlsRd!umCkK@jT*sNy8oPH7>*&;}Be5>@1gl(fVY#Q(9~B>&
z9M{Ob_10&*#LsZF)*FT2g)?T1wic9BPEz?ur3lx5<r>iI*6m6u)_4!CEvd}C_S#dd
zr6<otIVxioFW!i0nP0c-Ij0qGTt~MW(=LyI++(f6i_;!6JQe3$ADdGwQ|ZNXbj&-I
zY_vfA;0I5$O5DupR@il$zy0m|t)kN@qo{oP>%YFlOq<53Zdvowj%iRK^~x(7%-we{
zg%T*O7-gTIa{kGmoME`8t}Bc5-Mv5JZ~B*C-iRrXUplIcC_fdcEE`&Qj=<nLPS;J#
zSt=VhZhXti?$%1qBn{VU{_{USAMIE4w<5~Gb*HIh<n-h8(=oqPuF<NOYlw5YYs$d$
zEFabpuGh?UdOL5i^=H9?W1vJk#ZKSs%HU|(Dj@rQYnjhM4&lKPRMJzi%c;4k6l8yu
z^3!^k*2a7;sI=vD)J|3^@5t#BpEw)JyYWb_YdMG0=RI6o{Kr3Dl(L5Ab>@M}Qw|oO
zGI|!ivr)0jx<|$J+O<2e?ZLgf?X%$GWZIkofm75+bFRSD+%Lh>dS#~)IJlcE9Xp<(
zl9I}Jmb24gF@4s-DN{y;l#-QDzQ1M_*lxfF!=G}>6v(nLqawL8Mf1JyJ&3il(``#|
zzVj;L;DCW?Fm&Vd!a*iDWv7+rn%J8+zXc^O*FsMl|L{)Aecg4ZVGzNw@O5R+`oRGT
zVWltQ<hflKxappAvSS_N+UDW*f_zvO#~(i#ad-U<#lZ*P_{P1dI_T=d*S~%>2IdT-
zXb~=$_2AAsU&R{J8*TZ9>ndlN;9A_YE@!*z29IptzTdXZT%(>-!n^6-nJxz!oQdCX
z?|a{=w)`Rzwx`+7tXj3peC=!Zqy`D_0=c;QKzTTbfYao2APT1mCl}_KWx;luOkJM2
zXBaB`X$>A`m`#@6Ut^h^itUtGQT0fHT~1So(~6`KQ~fkH97Y97j}jxjs;U9D5#evI
zW+=E>QTRA%3h`KVOMyHBR)Gx*>@}VRcCZF^G74;dka`j?>aRfAux#hy;6nnHW-~{a
zzrSp*IRgzUNA5p~PRB1^-eMkJ2gNn+$?A9+7Ddl~^b{zt<F75tr-xRv4=>wpe)z&B
ze4Od3_{*B6y@(I<(}rX1(~0=7H5S&R17m;u;K!gP=umzAfj7-#RPIq3f_$Ed4_yCp
z%?b9S?dE-l&Cj3T0EOjtD3`i^fR2<6jl)f+Of)x~cdWH;4ZpY|?}@lT3%kYW41IUe
zI;)_q#1DY3uh@tic^Y|{1LgG%rypZZfs$(;UbvjfFyD?ZMOr);N`cL)#t(<f(aHYz
z$Jdz$pukSOmE&ZN5x|P|JLgU}=b{5U0UcOc$+#DLwkTL<{^gNX=3CFKN2bxq%0Eu4
zP<Z>yk>*F&Ewt8>+2qN5tpm#SBjzOxxcCXIU7m+RbYDYESy5+BMg9Nu+0)EvP<V~O
z51gb1K4g|Xzh^T}AzlOpHm7oa3D5DT?t0y9z<>hVJ;yWiCuk==3D$fz*P^q1^Hb~1
zElYOr*L*w>_%ZgEr%g7WK7R%V@o+k$p1BtEW4N~TuU^_@ZpQ$PUHI4#$1T7I`H!78
z$y@-1#mlH~fA$+pRp+$3nBC$(p;W%|l*xF$U}0{1N^xS14#T#Fa(<0dh<8z78)K&s
zKXG!G@8Z!hMPFg_@B%*k|HJQJH?JbCTwNVxyzZn4=5OD5ynTM-v@3DbVE(C0dI~>I
z?t^l8(dxZscUTeET|^LOHb&cC0tNd=&zOdpDB@3vM#b5~%Xef{VBdNDY4&}&7;9TS
z3kBB?U)*H&!jjbfV#+2{=R4}-ryXN1hJtAdf6!(NUxL>-Q0U(=*x_4Gt;3H347y7y
z189pkef~T%uv4*_3yNmQ$oIjoUfFD(-Lwz0&R_ss969+g;Dg{7&zx*7!=SG@=-W|w
z*q2>q1vaM;Kf|1Z4;gO6exdWoXPeg7_rA8(eBpP?@O2L#1;Kw5mLGZVNt4V!UOF4^
z;!5)-KFr;Z0SLEa;L65*e8#$w6r-hnc*QJpE!KN<A0xWC!(L{zZFuHi!$+k%7H>0u
zSg{L(Flb#C&&|L`!uOv%!Ca0(PP4IwA!lof^A6yLGA_~c7}~h&f3Y|i4TU7fpx)nf
z;Y{=4Q>WOOX%@k{=6}%6Jc-{Yt!PKOa~xPS2On%cf6+{<>>Wu<jP7oVxK0_jVsPMn
zD63yA*=7d`cKaJC6xUf8M0DkXapv3^qwMdc0qB=^p9B6>i&kKB3h^I9fld4KhcM9g
zR=meKP`N2i#EAE@sl&`?FPUwQ$IL%&MwpPuy*}jcps(Nay7|`^w`5fE4@6()(+ek<
zzkBzIsHgGzAH5bz5r{y-1Za7^cQ-yr@7-g+7qU^O!OO@|qs^$%qwz61_cdJ;7P;5a
zv&qS3fz9_2+tBV@zQc8wUVg5leKUVR8x{;q>@H5A=(d~8IPw*a@!Wm)tJs#L6jto>
zTyez-b}DUW71^EKv>N@+cOI~oevFfj@r=cw%rnlIVW-2UqK*o~uw$80>2%j!OYBs(
z;r<rW{`99WfI_9Kk|}&YJ3N~=@3ofKRAPtgk+YQLl+Uzm^qEna#kF7WzyA#=c<?(r
z<fu^k$Vc90*Tc0Puh2DjNagoezH+zo+W5y}a3`k_pE!{gvYjA0L(l^Ap@)`3p|`cm
z^uc@=(wg-{A3Cjb`t9zYg;b*6fuG0lWh(G~^rI)trcHYxYako<fie$+-?5f4Hx4*(
zE{LzfS_D(^&NYykkF;fe5R&Cer9anL&i0Ly%HW^<>^Uf-x22ScEc5Z>hgpRvr!~%G
z&C0T()%GKgEVqhHCm;8;M5Qus)F>KDb!8re>#&0xXz@;koU0>LgdTU?cxz2Pe?A6;
zLV4TOVTPO4JD_yjWUc$1^)Z<+?AzWp!+i9kXC0Mhz$}P~%J<uDd)`jj?v#AYFV{Gy
z^6%VpgEiNkOfG%uk~qkM&*m$yY_tk>cWuIi;jqF!$0{5B=gEb+<(8+qSU$V+Oq<UI
z<2d1jNwDA@YX_Gw9xCmrpyo^k4C7?tJzT4rvj#W?FvC-E$ub`{tRB%H&7GFLKm6en
z82rFNJZZ-|M+GbgY_QxGE}R1Yz%K0`(|P^%Ep{q!PBre7w~U3yoT8tKcFq!!$%V^+
zaeeo@57}w5UEO5c!nM9R4ZWNGKT|y6b9{cj@P*%`>Ivhe;`;NSzr->OhYcSy9A`b?
z02{V-VaI9rPdaG|W?h(@nz<k>JKMLVOE<xyez_gM5q@qth=Oa*pM3H(L>b*j=70@O
z_kZiH&!B&C^x!y_HwPDSZEY&ieaCu1tMMnESlMN;4rSrMgU^5d?Y2Dq`|{!M%7QU6
zpR=)@hb{fQQWjTVaUta@@E~=BFGK|lpU<kmW<8J7KF1B05;d6v=Bojg*NwEkhQf%A
zJ*WENC`ooU2Mnk+jZF>a(7^+p9n!!8JJuB9ar2n(aCt(dDa7*xvNM`s3h`Glg}A>4
z_WAgjo6HpAPJ#VpD6m(A6xa)o8D{<#A23dbh3fYoU1`3<Y2Ns8-0Og5B!K#Q0<4e!
z=h_o79bLSS;w@0%JhNt(xg8&&e)k$Wll;Qu6s&B;aNT|s;<G?M8xfj|GhFz`50xz#
zz5j^?<IQKzpJk^(=SSu496tuVN(J^GSD43F;f33&z&>Td5cAb5kGE5)f9JmC=I%Ff
zZ2;0_`gpMivcOA`*f+<Axi0M+&*MY($4{SV{_KL;R!QKpg)6Zfv$qFUHuvCz^I!jQ
zWnhiYa-bM=l*4UIkLQ_)+Zcw9Vc_PD8*KjZl3C^=C~WO|fc9*D{_tI4rx3sWCG&$7
z>5j`Vd~m!G7TAA-HL%$^%ZFo`ZGdIT-Agu`|MSd7%(;ROxKKW)oU1e3+X4a;SFgBt
zj`a!k<}4_3KC*DKxf&mc<0$u;{<j}qY5p%1jrQ4KK`|XBXQRg-?lToQWn#K6T<RF%
zQw|r=Sv&XGQRcrt3<WkiUAZ{gDs92633t4(!Ti(H;6o+5UvCh<bxscVJXt@kP)=BV
zU3dH#^U>3%+G%TZk)f2rS_O8dDa2ucJtK#KHoeVLHOhbrq(6fKdo!jxZDoA{=d0$A
zHJ?6jrs;sf`m1=(pXbaIuI{+DhGAJYtZTgH%E<|Za34E)qWK)!ZcYVnn@V@Bn?0jy
zh8#~~Wcq(CT5BF%iRZU-TGG_5jL_CyoN;Npc82F+y?E`y@mBfx-6uCd#?Cu2XyQMw
zoo}wd^Hrar6P=1bhXX!;b_((P-!%77f$g4s$Pr8--g^Z$#d#hI?7#c{Yi22?%+1x+
zLB<cAINtpImGdyB3(Ilj8bL&O7XKCu_WGYER-1o)7Vl=h<M_^UW#Zlkoo&9oz-zAk
zbMdk~q13v1&M5OG$jB+8gLSlBfXsWRz`g+%*!~paw|)2&vk5=k|N7B2<{+lfwGATE
zW>e$jq2JXT9Htzv<N5d~>+l6LZ9U4R4yWQN(gFJ#|L^fN<~vYg8!ivPf?@`^ccsr`
z_gshN<n!agq|SHV$4fl4f8c=vd*c_*H+4BYOH7ta65Is~@y|U1`8cS6eJCp%qCTq5
z;GC^5p?XC*kjJUxhnUZtKH0qgw5fPKxQ86q`k2c&g*f_a)&d(d))dnMn+ojvUfX8=
z`nRuB17NGh5bUp-H^F@Kopa5Tt9O|%KD5eg#UQAR?}AKafoJQMtLB>bpxpRjDvo2d
zW1sE`ta<-;crS6#P?u+y@z{3)_KcD@u%8irKQRpcPsi*EU&V8M1^QTV(r48?jOXYv
zSeJ(t*dy=(@e^lHMW1Afc^noVpSgb(p102a4HL5RbjpHV7kTNYV;Fp_{lZyOY~N;L
z4g+3;2=Mse`QY*r|AK+9zr?^+>-fyjA$t2v@%it}zhz_FX@uzvX4`JGjWZ_=HQz+P
z{6c(Gh2?;~-NqB~hl2?-O(D+ptk0h{+I;&RbIdO>Xz|-mZiFn5fpSp}Zgkx7(Y58|
zgy97^Fh2>_*nf(Fo$talu5kx-1~+H_%Q9_1fBPSPx5V7K6u<Snq}|=0nl;jV<N7nP
zd>daJaw5=o3E1DoP+o7@ya^w%xfBBOg6D4lz9CMSG#Ltw;r5%NeKKu-fsaKnv!na|
z8eGGD?s>@?4x!z9Ju)1^Fw!pJm@1o#;Da!4MPrmTTuPu&I2VV@=)kIKQ_sFeKXMj(
z#hb|SL)g;m^akhCAbfxNO4p;f;kw?H-AAjspa1+hyB4!k?oi=!{q<*C<x))zm6@yr
zfs2yi*=N^S>+7(hiPLpcp~>l%D=PS+Nyp>j3-*<mrglHpgnki9tF2v>WK^V|b=I++
z3ZnG=X*sIQAO5h|D%YqSVjg&n>l0HsMa58HshYmGYuXjo`ul(W=V@4Db2|UvDj9Gc
z+>0+hzH9JKAFnT2vIz>hMXBq2W@w$xDVV9C3_B`p@3`Y-YxPZem=42J5lW@)JKuT2
zQA+Q09>X#5zVsu@E56@_?s~q${NfkSnFk+y18D{`y)b<$A}_#q$@9<WwB;iio0ZFP
zaB=Y&=RmR*=mXNqp4OpGA<Hm~lWP;xGT60O>3f4C_V;Lc{n%qGaDoFm&;<iWDl$1G
zG%cVxm<Jt{$UM=4m6oiW64|9Q9)s+zz4la0dp*U<*j086F5LQ-`RPxevFzPFtS9H5
zJKGM_7%^h7O_Ne~WyUl3p4rr5t+(&L|MgU#mF#Gx%)tt@Ts`_e{Jg@n)ek<n90Md?
zOUX(GRGxnL!)Ic!z%-oC{#+cu_d6=UUwdscl*^n>9NjpS8K0?Zt~uE%n<*uolc51K
z<#C__t;fTTipUu=#-P0Cpq`G0#Wx38uy2*=*<d|D?vda=7;bUhaMpL%XJ@_Tbn&xi
zkBd+M2hSB1>6|j(Io?mJ^Ghx{-YU??jmsuOB`l8*KzaTO1~07Xe2$q<EOAv?Y5wl-
z-q|_e!^NIykEH{*?=v^w{G_!ac5NJIgm~Zk7D9P^Oe$X9?HKwoF-=+`fAy<>u;s^c
zpkkPVb3XF_vv(eVQWe)5pN9%cM|!VPr1vILR4{^qqM%|$MeLf0MuS+uB*ywT_HK;9
zUJ&e3rT32X-aAOM<p2HV?(E*(cX!|JV;@L-qmSLackYxk_0D(BcP?)g2GV1n=Q;dK
zKlj`i>!!iw5Bmgp|KShyEjxs}9-FWCJ40pPBLCYnZCb7kXdA$P{E)kT9rQeN{5$VV
z6^OkgYRu9fPSHCqa~!}m6!u<!h)qSV#tt$D=3504(jz~{E{Z<aUVEC9j7saF=zf%0
z7WAjsv@^>*_Br>2cb)zU>p^F=6-o~Q3AzXLnJo+K|1}@j2juhmXZgO)lLzn`tLqcO
zf<u5<Xz%?|ocp3ol*n}dcr9SX>>%^W@_+7YE`XihCI!ImDGTfsu@>0!ft_Ojwq}KK
z%eI<90<ahQ0PLos5bx8hqF%NRm<L8J^X4WuS1?bxLy%Yzgx2G8WohxS4#c6MbC1PF
zb-y`?`E1?>^A8>T&X<L#Jy7AMn-6tp*IAaJ)da{C(WW@BUW5R=CyMHOEb(<HYiCB6
z5zh(CxJA}}x5<whem+@95(UaIYwb>1^F;&L^<=etL9+^GuBf^{23Q9qAuZ8rRw;gx
zsU!MTX!bbJUnN=16}`8?!qpvOkttcLY<&gba^Tjme6f`L%i(K$=+sr_?~_)0zT(`Z
zO*u5>R4!e>w9tWaL4i$Jcaq_1(ZPR(A_Sl=m)tY@%E4VJmHW+J4NaqpdU5vO<Njd;
zl8R@64S;=@0Br4o!cCjqtPm-_fuA0=)V#Ss)U<vJXm0ASZW+l%{id9>3E3dv{$aTQ
zmC2%jU$F%+N|AIu9O%>Pub)dUO_K003V`<3G*L1^7pnt5cRomZPNY!IFUQZcZ@d&z
zo|UDUY!Cj)UfJq@J5G4cMWtxSvbhd6?-LtB!4xRO0ZE}0erv)C^W60H0;$wdTwTdP
z=~j^<LmSbP=g*5ZlxNL3UWegZ<Ui9E!o{bbY2VPH#>wW%3mRBTaV{N80Wk+si2s^l
zrrg%XRFUR1agIdr0PJn%`cLPZHIla_uT`5glhxB1I(S~VYNr{75{>H&QvP5bDqkqC
z@MO_?iBf5^d>hXKxIkei(_E8b9yp`AxmX}Y39$<#+g2QWzcYEId2#X@GaZ^y<_#BU
zZxYg6+ClxXa0octAz$9(<a$94HNBV!QSl1OQgE-|(jN<K0odJ|R8$fX%}JMhc0L(@
z`0dSPfvvu6EwCE_U>|;~PLi_0A3vIFmH=Rfg=s4q8Mv+jj2Sm>bS(^EcRf|~xq0<?
zdIruet$8nuTw)%ZuqwI?jBWj2sZ;?|8~?KcWxz|gyjp9)*R1iHH*FdHqESEUQM-h>
zwR3GT02H?PQ=+1h9zL-9eY(K>=%WxXptbr=fgaCHTxGUt-LiccB^mYIcA9UBh%Sq6
zh5yyb>Nk6tYhq+(MX5e=SDQ-ami9H$Mmg`AC?`Lm=Kf&jI`i(F^>$4N^^)-1&Md!K
za$&C!=}V$SRwGq`LR@2+XOJ%e(a;Y3v|Ty$ox{O*#md2$O(w9vfvM!*lUA5}N3HPo
zOeF4t+6W4_M%vR<(msoFu*aIIIc7HfUJ7*4^Nb3jgSfw+S>5zKt$MheF1cgn(^_CZ
zKS}Zk&0iE;YdIF!?PP&{v;b`8EnG-=7TEKx5*eMKklMY2)`)JkOPPnhS!~w1@k6<F
zKipaP9UEwGDR;ip{Zw<7sQTSONNhC!Z~ELp%?%@UaC!MKm4kUcs^3tkyr!$(7s`pA
zlirIG$87bR8LUt2%>lDrS5x#d>R(THt7XpAVQ-L2uGswdsXa!<0N8~zUOJyDcN^NT
zJtw8tee_T2Pmp_<j@l0wVvW+ZZBo1)59sb#AY569tASQZY1|-YqFXwiq6^-^9k;Ry
zWH(ob-v6Gs(!4IK^l%+{keiZIJ?+hFX<T7BPr$9fRP8?(NpZ~r=>D}^<vK+anop^H
zn#&@>UrhUO%s){0FO{pMVe)}JL@Xp4v^GXT%?Zt`6f<>-=96r+_rrpj@+f{Wxy@;Y
z8zGHd++nPj<=hO`Z{$Z~`!m(}Zz@IPyE@e{XRE&)Vh@?7z`4ob;q%P@O<QM`G-H#+
zk9xm`rOo|1q@_>uUv9{iz_BWUgZmGd^&8~#dY6<^!vX6@(YO{WsN)3YD)n*3pbd7w
zM!vdh_O^?|#dz`jaYFHIR9)wj5F3SS_MZNs9RBaz^M|N=`oZsT*?kf1E%`asd|f~w
z*Z+E8e_y0X5z!w5%pjDh`kSUdZ~oVp-&EM4cY&8)8fz_(9T)(xF>qj4OC{^BB`P%W
z0Z6cp<(!3X49Hf$emPlCceLPUs58Y)1~3GG<^A`kTVRNYqy@MO$OFZ5<}^G7k2l^J
zZzfGz9^TIYZeiKkqeml4Lmf9$dn3a#s}3JNS3vk|bzyH%K%I?+=7kqF6}5HdwgbEx
zHR=cR!3WdB$+Mqo+qSwaLtFVky$ow@3<$Yej8uN`ng^*H&j7*(56E!xhxG+(@h6}B
z)><IDgz&F?#u=wt@Eq#yWCHabJ$kXFA@|3}ni+r8jT=|;Uyk|?0QkZSV+8E#-A93b
z!g~yUbODpS`ZH$Y1+bQbX2OK!QP!jY>CkF->QqPa<?YfR^2ma9wFR*K`IZ!0T)%#&
z8|bTDoBN-N{_fF$Kl99}DER?gUwLI~fx+b!J-XcZnVvRnr3Jn>hR*ZpA6V?t-T=6s
ztBC05qVwff9)Z`Z1%l5LAidGn&!vN+1b}pGfO;F?7?9ru#uAq8J$s53Rc-)0r4DyD
z(PxIGkY|iRBkkT9pdEkUf%7aF9V|bct4&!hrHRQ=blzAyqig_pPoK#ukROZzk38~a
z6!km)*KfVGlWE?(N)~m=BbJ4=RtbQ1aAk!J0^k|u17wp`V1B}aA7vkBvAth==I&w$
z!V*1bjv!439^-Q!nGt@`z8cGa+#}#e+Lg)PjPp+f2p_F6puKu1K(pVkUpuvFW&qiX
zk|7LdpLzD#(P9&TIl-$da>aTbf9jM4Y%{hGyFB<LJ{Iw~K7ihy`JHPh3ZZ<`tXWlJ
z`5(!l?~Ip1`a4H|MY7BzE`W_d;|czM4ZubN*w$B#?`h0L<c18e&fdEhr}16`{$6}w
z1AL_8fH)$}-}g5F?2F|Cdoe!2T#$9t%)8qQG(yWOtHsVel>Dc)z!rf0p#00t+oD4i
zHgL+Oln#Ng;M*X79P8ZyHOcC5q_GZK8<h_o5a|+Q;Ac(n+XZH<5J32Y=(VqtHDNy;
zY~u4SQQ%;8z7BK-eZIhq5^!U!5qLmlC?x+F7u79ky2~dUbl@jx0mv%=hav(P<6n5A
zwA}D}{GJYs*Fd4@4?4hltH-J43|Xu|%a<r{d-l3rX2@5I%t*i^4&4+AslLT^=(kaI
z!q;eJ9q@LoUDC7`wK^21SQC=6=5kpt{%Up!u>8+-`M~@^G~|G){-DGC!J-Jh`m|~S
zqGC5%{)mz9hjg&~!o(Hk&B?1R01>$(kU{`OUr`3P(_vS!{KpHZ&>|G7<^gS>wn-BY
ztNQmuFF8>FeAL0C5;{i~$4_0@(E8_2SJxu@{gT5!-ko9oxll3)g~F2K4Wbtwa8BK{
z^ke$%!{012{{)<M3p&F?2Z8{;-E_Elx+ob>RzHTm%!MDN#kg4;=&|ZE6Gh`}+d3pO
z7YEAjS1xX13TktmRoE(e{hc}l<S@{K9=vS;Nm882@?SDYJ}iH9Wz@Agq5~b$yRj*w
z17F)-ipT+-|9P(Po9p-#F0Zl3SV<-@KGw@R^EuJjemG;D*}@^R3%XffR!~>U61K5e
z7NE{#GooyCZZDs;>a+NsAH8I&8M&MdFtrE;DQ{d|zl<5uy}tF;>MkD(pU7l`4*Tm5
z{!Y$^{Ou|q*tZ;cAJ|x6-z4kR6;d*At%Gm+A}8T@>KC&^hxT<!<Tu^rBeS9WrPKZ#
z0<qpQ_Z`w0TdVqfD*D)W<o6Uy{czCOXej20Kk4APtL8r9XB9|yvF87KKAL03ODQN5
z4XftB{wm8K<O8>~HojP%d41+r;{b+N<kNC93NuWGu|aiuu4i3KAx?izN1#oy2E+&U
zr?C{`qANeIQTep&Pc}%if9%q2W|zh=HA9>XMf+X9_0<L}mq*XtXue&&6TKoEPSnjl
zVG8l|0NAwsM>E%$A>S@Dn>6=?g9_D1;KVhpE1P-(a9Eq5n0EX0{h9~%2&h}AKKG#9
zEo{}=9u*|eK8?zl8?<iMDP1UGzoAckD^Tk;_2I3!({Msi8LDYsTdw(sHME)llgk>H
zGp(zfYyo_`4l6G6_ih~$O_gu8cO|nq+5<%O8`6`N@ziTtSQZ!B`dA3`jqPfa9Xe<U
zy$-0KMCNvsS3Z#+y1z|Y6E$Yr7CX@l=uzL4)_TS8p>Of8TLbtNSs!%ptwXx1qI^$g
z|BdLjs!YQ~CH&^N<z_yhr$6XW;Rdx4zIad8o(rpPz%kaeJ!&K7naSEeJf%H|RhaR$
zMP-4P1JAB)x`^_f`O7P}Kk?u8_qFmZ*7(IK+C%upkwC^_{aTo|I_Tjg{b&UG`xYJ2
z3>P>#;FASmCXAD`22|v*QFFjzlqFqIzbdJ=Yb2$;s@j+0bJPW_?suvk*&o9!1N%-3
z9o%L0K0!0@(K<9;{-=3An2SOr6Q2t@*uA@DMQiO_QvEhxXkKRw?3O%X#~G!4#QU;L
z-W|0EF{YaU*895GGu5(SXhIG_7T6f`S;bVf&o1p*ewJL$kZYW?YLzriq_kLAdpE`i
zt@o1-jD9@yll*A!l53*XW@W5vB4yG_YwX{;*OhW<5j!`zD#rU{+)dPh-s_WAna{P?
z->-LrNXRgCOXe2?R5i<XR6ye?Z>T-!PdoLFfbZ(B<o032{LSV^wPiS{z-Lw~YKELA
z*~n7GC+%<)_l3i#Wm^AY0c_@2_Px9V?9#dmgU)%?PSzMLW$NhNjq<RUL^cPsZ`dmf
zqv7&r{+9lB1M+ht6beeNm&kwl04X*VSN}?8FX5g|^hy5TiT3?*DMwF=eK%45l?&%H
zFJ02uG?jv_uQ-X4%)OT@f#0kIHmqA~wr<&~@#7oo2}*0nz5Zb?<-30j_<y^2HkKkM
zipM>>7YxPYyFJ&8NuOe+63?w8{`4p`Wp{yd8}<hT7th^`_nB<Gc+Y2bJ=ygA@%-Oo
z)8`&%!1ysQUmj44H@h*C!xGgwvL44icQk(6ex)`7)B)JszFlX33S8$sX2_89t)E}p
zKWrz9UrNAUz-NFa+Yi+X@db@Fa^uF8buH58UA~!9%MjqV!o2?agfM`OJORPaJ-3ko
zXo?;Mq>y2)T(O>>GiRM-?4^o@E5I{!*bZ#+?gak}I0ZfRn{Q6A5lH8d_C0OdR2Qvx
ziva6Nxi$ImfCaC=|9)vWQO;4Y7^l3H6}ilrv({RN<8Rv^Si1qf05hRrcEFN9j{6>8
zW-f;J8804tj?X@uZ9!1#>H?s7-@d1rs#TF~@&Rjd+G(5iFiz(9?Y-lUGc6$JPl+-S
zunPX{vpJ$dUJx4p+S2O10|f=B(~|+MKB<EM06+jqL_t(Dy_Aeqws`vKk-}X(OHler
zJGsQcBAId!BC^OM`tM~b@A$A~Kk{fGRvB0Z;?ZBbyxb>w{r&G>TA%6eT49|@Uvf!H
zfp=AODKgoSH)8>14B8#gpK%16LFdk=%Hp+-rRa9)#C}G~?%lzN3<_@)2LM4y3*Zt3
z4J@(k`JFEjEU&RR-!44C1$<5S?hVjbMG7Usb5U%*-m}_ciZUK07bw(;L;g^{-*e9&
zByT9ZBXxKAF*dRGN5O;m1p10&D1f#)*?{b^poYrb)s?Z;y?Z0mty=>N9tW459P)_z
zdXf3$lUddn&-GU<j<33^O(v!USEktXV3rs<bh1@iab@^XXTR59-`>)wyKC-wKB+gL
zG4$nb+>tIy7Ps8e$@-cPdd{DQ|2zHYsi(eG|KA+G%z6WjJyi6uwEkYC4BHPLd~lex
z0B5Wam~VRaY+`BbnX3~8%1j%t)>+LbpUkxFKphxI0NCDKE9OY;ZJd0wO7HOEaejE7
z{)#8c#@>qyU~`uSasu(E^uGpR`|B3_o%;V?0qk50?AYFLc)tO#FBm$@tn^ce>o5%;
z*p;O1m9E1Be<>}1Eg#rX6yikC1ZOR7L<wty7S5(R{5V@yn02H@Rz~1yJ{`(19rD`5
zB)3l!EMVR~9h_9qAzEGecgU+vU82Cj&RYU#A0D|>KJ|PD;aX&`s8`xtDu0-bMKObw
zaOCyO%awE~|55=jUoY5T9vrs{6xrKDb5o1X&21`~+s~|(C6La6R3-{>(jbH;@G3eu
z?ANN2X(M2i!zC=9U7Og$WfhC^OkK6jd?e8LRUHT!SswbsxdOi*(ZO9wSp{bmnB=#}
zTI>1`XPdEp-{U%vtEL08%NmzA7c{OAq6$_a5lbjg9?HhS^%z;NJRtdQ<iL}5B`A6&
zfl|-vkf?<Y{*pNe_a03AFR34p41#=Mr;9?IJg{u;`pRUp#92~PEVXxMf%4}!DrZ^=
z)VBw{UgL4s2IRuQ{3<D0;Lq{J3Gxe#;)E7t8)$1C_>Jt_%+xPe)Vr48U*`XiL~Z|!
z{$<R_A2yk5Ka%egXQ8UJn#mgH^_%6xMwUMe9I~{}DIvyDY<k>xo4^D`3E#pBaD-?z
z9}$J`8ogxr{T>2r;$yapD1`87O&=qD@81low*e*Q0x8k_YvM}t%?ioJ6-cof$#VY*
zf$#No;Cn0u>bX%sOudJ`wN_~1vWG0Nt73soBn59@!x#J0`v*C>w(-XcMe#@VJr(#B
z@YkadpC6$RcQHvwf7N6``$x@PEu>*vP9UIP34<O(f7vgKsJT+Qd3y9x^Tkpf9;n@I
zTakv84gTD&in+FRRVlY*eeld9BlLoB^=FH$1)?hspC_uH-k~vdmXt#9&+WM#;0dp{
zvl;$Dpy(IsbMI=-Tg3XsZGQ+^V0YDgl|Bk_W9eHT@>7VDujKIZMfC+>r-wouz<8(1
zkCK*^FZHsWAv7>Tb<L+W@};j9nrCLLH-|#k!WP(2h-+TqvBL>GzCkav&wTrXc^Ty$
zH=nCd))okQhv<1b)fNz`Huth}8HQ#i6?v;bis7Oreoo-avTZ&POi^oUeBUkl`E%ss
zIFTC#))oNlYh{7GUqD;bn+yF=;8D*yCC%w_f6z{Twkm7w^v!gth4*Wt*0(PgY&63J
za*bV~K8#|73tH!1@7=(3kpkNZIsnbwsa(Y0)+f`0C+*l8>(EB~n%BoHF}HrbBnrUp
zto8iKYg(D|Iuv34c2aO--K$|JJu;Ap-QFXa+Rqld%)qaIFmtrl?W27Nd`Q9Y`uf^u
zi5aG*=0NsqUM@C<-4_5LuFyI%RN&w9Gu7wx+Z9kysyVN2DRc7~HO(nW0_tAMzm0zb
zDa5ts{j^_mSzsT13h@zA%pCCXe7h&H5>wmf)7*ZLHV8FiK`A#~RIQklM2eXbIv6Ud
z{RR7K+m4FE{DtN3Zta=qGYu;gHD%?KI;%Ye$`<37Zk4jkI`f+Lvr&_m(!aM&MRTs^
zbNUTs@%I}#2%zo~`HlP6#8u`E`G&UZ6P2Lc+B?53AI%-5_yB`Zyg*kE0<gpJNlSBJ
zZ!tcdul?|;Vvi`Vx)DO7#dcMu$u&~g`AT!mi&NI<ogs3#N8BPh;CoV(+ZQygXnxW@
zGEtyRfa1@NU1r{xzSgYOUeeCPp|YTO^IWH@rn7tu7Ss6RvR7%Bs@M?puC!poF7uj{
z3_lQv9iGQSX}l@6lGD`x60vpw0*5KYeHPd@7i5YCjn;wNr9jm|!2YQ^bSx?5A`YHx
z+p8YTQ^*NR^3SCN_M#LX_eo*dy|v_39Ax@j&-$je4!n84Nk`y)14Yu0X0J2PPhV%&
zZlXPuwtU4JS&k17Xnw2a6cjenlVdtcots9k1hOFkb?KGMmz!NXq!7&Hk!tYoH4{C1
zwBjWylkPFsa8{H*O^T~oWOr0=11OX$S1tr#V=4E!{<QY_{B`DZ0aNJ)SZNL(Jlw7c
zl#8}1qV;p&z-}SyW!8$s-4<|W)27|l+R%H4_Nb}xJM95g(KC2(ddJDv?{f3nYvV;%
z>#aV#SK_A^pto}6U;umP&b<~)ov(LKH?#mTFTJ$61<&Jwb^bP>Jjen6&IHn7jf5T-
zieA5UrB`;X-+K=TBpfeV<+b6Y$QO&ujvec%4*TWndV~4?`z5jfjre=!dB7ihF8Auy
z)G7(MbBAz30bhb;77#J!63;=cj^!x6boos{F3p-%wp7+ZesR56!T&;ujD;`O`^d`m
znL&fjwS532LEFA3_9@21VkxVA`spl7fedJee6jk5V!la}DtaV7ZFqEg8P=yVfGSqy
z4gjQ|TyaI~C_vJaSsq#EFEy{e`h7SV@`n1EKH||PhlCQ2k@*XCzy~xIwJsnx@~u!I
znjHXn;I|ptZD;ij09w6zNej3(YE)TCB;I3sW%qJ|;Q(LQ_)vDCkoIw1y3`9<a!2o<
z@tF;P4e;v%mcG!P$5Pw7`|i`9uXjyMj*Hft|NFmh!+xk4?{(^wReGI5iU+~+qrLfa
zZA3rE?>9;z9@rZEoGXj_#4k5>|M=sXVTBati1u<3N?Q2ORxgY*PaP-44NrN)qrT>7
zBEu@T;6vL#wq1Vi(=Q9!;Q8lA+wsox1j>N_>xK<0XdHOg+_UVgp`M4O0rmx#4}Gt8
z?K0N#Ijgpaq!s!0?z>avk9uKvuA>YnZon4dR|4^ueU<PC=6F{|?%_Y%qyA=IjV*g5
zKlhDt<2C>~cltK-2iIV2yYRwh8c)>~HQ6!DoRN4<arp*+vOe_GU#Z~RzN-lVY+t;n
zufJyin}&gA{C}N7oGzbVhY(qJ{iv>y`y+t_!T>hZdvqj%7QUHxbubmJpN-Hp?@n6&
z^#JT`0<afEd|<nj$&Pa95aiL0RV}4r1#L<ThP1$9OG96<lcXl||7vmrC}c5@7l;c5
z?E~_KHdOwS-GOX=T|cL0adY=Mr&^!Uq!lj}^5ifC>+k;v2p{y-QVZVLJjq*Hep7C2
zrF05~MH*C|gaEc4sHsDhC(dtR8q4zfBpo2CNO9}N!IC}jowmk2HhOvFaEgqz=^1@>
zQ*)|l#8vEsfiivf!xqy^R*{zekcCtE<`?k%$2OH^X;;-06)3`icf3FvSW|u_iuOOt
zPu~Vni@QNvSobc)D*CKivS!H|<cgPv|2lPo_WEz{$Ora9S+a+M=L5TfD8yxf&B5AX
z1p3XVI@rDD-Pv9$$WbCFA#1QF&#hz5kT0DRwZJAiY@>~^s2`^T+RHwfZ+^1$l959F
zsa+NGCn-MURlNApjl>W7#&e^z*}ql?(0lv<w%YI28`~OuF6KnoBTLbiFHbgGps93y
zOb6H(=^(SWX!TE%Wf>H)@#Mn>jiED30RR1>P96{Ka3~bjJ*e<+|3e+m;x4<waS0xw
zPsHF+h<_UmVC%qiqI_Um06SC`>P{bXvyKC!5GL20NWFO8L&Qn8r#K4nexeW$`@nW`
zP~sdEKdXLRAH^NXJKkJFIe0mkD{$(XkLQ^s7B~(yh+Z<zs#VI|a(Xp^h$SqotIHzx
zlP5L_6dmyXEc4|u9kMwPQ8xfZAM9Gkw2+o|5iv-xX$|ENl?;7muat)tZQ5n-|7?Mo
zv`!$l6BIE;CvJUUtN*3zAd5C~6yl%BiWkt{2F*XO^b&w=Uo5;!hx;${Xr{m=pD+Nh
zAOB>I`Rl|r<{+Iz!2;N_YW@op;wK)y1vbfkA>jB)(Ff1ksPRo(k(Ju!ZvwCTwbP4)
zBDuBH79!>!uPPY=Bv+aHwLWcE+lPbtQ{SfL&7B=<m{NL44pk$3+kK%=LLu$}*gnNG
zA}EbNwya<-YgI+c7x@#;HO#~NG!M^TzuVlVK5UgAT>n!hookgaFZC81hW47CL|qAv
z&Ck4Am);P7ecLzMr}%&}0PM&6wzO*kV=SHFpz`^%`sKf8X^ar)x5}feloc+NQpz2j
zPO<BMJUVB_8GVNF{p42*%u_RTAgeZW0ha(b%OCNL>U-=BkEIX{U_&9^;-~}ITLdJI
zkY&-GANc@mgiH4bwc%wAN}FCfh^!@$v9!i3A$IH&l?Jq*`pUW=8;2<VK%4K>KH_09
zBfSpztT~kg(0wl<g{D8XsbVgYiy<h@<Mmh4V0`1td*EmD&HD=l)Y9iEi1s5_HYsB+
z6Wc@M^0KmxGttoR;sV&nL3O#YRYlY9^y-!oi8{pF*O1l$F<PNi_?Q6SIWbHGcekow
zZj}#5_C|>U>-R({{obv0ZZ3NxS9X=Be%XTNUNN7bM1%!+JULt0s6fl*0^^S#&oyI~
z`IL7gn=W$aaDBUKrlUaYpX2P8k*&4B)?Rq9l(<yzNU>EG-V<K$QO`7kiAe8ov{StD
zkk%%((>r31diuMiW{2X2gUb2wr43BGI_2WFQ#fu4UszLj>u_?5*gBR7%-<r5meNw}
zEvt9kN>UImEDQg5{UU{Ia--%-;HXPr+tw{+!@6~vn_}O68E2YrG>7;`T)4c<S#Ema
zA~RD^yzm2|FF6@6QJMVtx_FG?cmOu{m<zKCM=M@d<%+#eIkCX5P@#fu0bqZuKbXv-
zK_*=|>VdK@UAo2GfB&ap%T>Us`t>VVfTds(Mi!AuxHfd3+@Yr?(_4uvCNx>f0<`)8
z?6z%dSc}zi<s2nDab3W8?2th3ab~6%sa!vT9=TJeTC#?2VExs)*oi+q>-pF?4y=m4
z?)e4AavczrAW>R+Xe>ejbH4m?js>r_YJI1BQ#PoBi%QuL+G8xW3l~QHGg<lAY*PLQ
z;B@lj<pMxwhb?m12b7a>5ddE8+Vs=NUOt%&0Qz5kd91CE8@d3bSmZje(w`1(gas@X
z!~ns_ig@+pmZX3G_HiwO{b@wLYoDSshIihXB87)FVOnM^f&pkzI*1N=q?oWg%Cet+
zU$tsUOT7$`muv{y=LYo;;=T6T_oC2VY4=>n2mlnH>R5{VfyQL=ro3avE-@oU%n4i3
z!ZLv0a)8TN+Vs(*e-N9%#Hj07y7%wjN!g{=*5o~I+)^nlj0=Ou)CJ%27hl|5^5=~p
zbHF+opYP+wEe`hy#sIY0SfIPHn)%7hSCHr2x$7*AIF{rD%772)Yp-o5B?J$^&io`-
zj+f>11OtV)vx27YQGP7AJ?rDF#~{|v_ul)lrI4rW^x>1FoO;`BXILNBv3YZU&6;iI
z#TUof{_5&Wf5Gy<b?a)ek0kYb{`~dki6_3X@^k$PrU|U6GxvSHZfX3_oVi*)xJQTk
z18s^WKFkLm13}hV(*{8RHWt_@5j1XGMTx}jZL&@`doeh#0>Cbp<>R>kc2*)%FFNo$
zr4Y}xz|N|3q;@|{A?^nvn1JpUfSr>R;yR466ygBqnqd8bL+sb3ec!Zd@i1*{*3-Q|
zh4|l@)TfBXd5{i*R>&fP37Jq#K*W9L*EeU%a*M;HWb@m&OBBHZ?!=qtCB(m?(EH^2
z7N(^RaX3(l7dT9ry>4fSLfm(lqzB3fXdK*G6g9FoNd%_*)21%-WYuu(d$Y|_4!N9M
zHOK#}x9SG1xfaqy;W(A|$x%zp;O|#@G}*{Uo8VrjmNfm(I7NqLMG`$8b&vDM7iG0~
zKWS+(B5>T*Sw8L_?bB3r)+cIFPFB+P@xXs5Vq6OG76QxvaCj8rq%WcpGw_qy=7ZTA
zz1Kn&d#hf&Zk5lx{z|`uHaWygCP4fCl4#=}8ow&iMtZPYg+k_;zO6#a8p))TIS$es
zC6MQ4QGjldCA<rb1U9{i1$GVq*jS0aE5(W11&&&CYBE)OH<iWov%MOKcD7I$^K#{h
z|CwPvk%i1(Ml3X&b(rl!^Wyo<{W=(Jo<_@kmv&BlGS_U7#&9MIaWzsq(XQg2z<ZhW
z4*K0lpdY#Kl+F9il}NlinxpWc{qTA9;~o(6tP30#Pt&}2liFy7=G>p%4SiM?sPy9b
zlC1#ac!$lt==LiXcKF?WRG{Ws`Np<EW4X2%Gj45D*<30hF_BdnK1-)-4LJ8D`D+us
zeK=^+c(TW-rgyVSNhxw58@90XHScwNWs2T1d@oC)d3{c_-rYo9e`F}c>zeL*!C@Xg
zQUP%C$zgLXfPE-5m!7BsK+AoPl(8)3t-mv6qhR!j=Vi6`z}OX$H9_OMsOaX$_H7}+
z7|Y766K1``lKY=y^l~bn=i!#p^@ei6@X`hKO?CC3cxh3NZx@O}Tx++VLY%&RrQCiD
z6diV9QB%cx-d{q-=1)4<_*w@NSH3&fT4&n(sz*W1l{4>XZ>>uG&m61YPmuM+n`4(|
z0<gQ3%5NUKN&vRN!JvbZ?ADid9w!>J9&gQzfQKZea{S}WQ_LUa7NDFgZjza6NDEr)
z`-aUk@8cpx^M(!Te+}f8Vvv-YPOmQCy+?4&c?HP+ziVY#-uOlB8Fk1O4gwHAkdk3*
zvAB4yLA*N*D_1r8wZEQG_d@ZI7n<tsnQ;0^d?9TB?BQB}{S@NvQ55U0mgW3SqIPC~
zs-H=XHTQg0pOk%y`pqBSnPZL<U7wpDDylw@_HJlSQ#+rSq;)Lgd%0MW&VPG`StoWE
z8xWY*%{$vwHMe!B5!ccdIRIdf7K_l8Z>tZ-fXi3NQul#wr<y`KOp02l<5ZBc?2r}B
zD-%|jN2R<IK5|ic&Z=6>{8jT-742^lg@fwv&waVT&h!3w?c`qQ&r))2TB)dtnOv(=
zp4qaHIOBtPMj8WAjrd^ansUQb&FW<Z&7-stROGP~;;i7m=-uLNv5=gl_YB@L5O-FA
z^p@(t{f-Z2n+dCTMfSt0?=7t>S&L}auyh71CFRTfgW>_}2j!qMCsXEhmL)f>TnXfq
z1bDaJv3;9a4~Q*SO$4sxFHk^WQh9UYNgiuNPURnzE^|*Z@BHbAhf9lsalnqCvK`CU
zurBm04B1n>t3O<}B)1+M$$A6Ij34jOD?6%f01s%6@hR-A1`oGP^fCTC-3?}SCkyX8
z(>;I}z}uH!%9tSZ-U!eH;Dx1dks<~3T+oV;M^w#I1n?~unEkWelPN0;6a$oX=upe<
z!DtWiNk>3#SXBeIW8F-kevKMcvcMP?kHd%05fJ~A^7JyLOy#Ah*QLv;qRp<7G;nsf
z<%Mb)|Kt-UEDMiy<l4A#MN5C2YWhEm^$knpfB$<@RR3tvqKc&>C++B<J;l<E=A(~h
zhJ83AD}dwk&TD8Xu$_#epNs!%i&%ZAld}XRW~owzEiJR>OWgmg_MY<D`>%gZl$E`w
zY`|TB#D)zknzN;_h<|lJ;8e@u&&B^8>h6ag8g8xh-Q4N<%WfkI?Xua*2e~sg=|=>*
z6|#<H7+~z+FB|G^H>N4WA9OBy{`t-HWO&DG)5+=|`Y>Y>>t~l23MN-y-PY1p$MdC)
zV#mM#J=t2^JCL6-c-w8AtPCP^KvpI2?yrz5#=rdK6Ditw05)KBMHzVf=}(<QEDyoy
z(Enj|0K|0x&>dw0tng#Oe-|hA6Z-VK?)or%nLdn9aePRZEn75v-F?ANHHstezdy~o
zBj7r128H@%m$j1ba1<wk<};T<@aHzUyrLYmDd`~xEc-9MxP|qH&NY9a{!qDp`sr-j
z=ebAOJ9n;YDd^)VAo$bZJ7*FL7Oc05FwmmY_qbL{=brdtKjoA%q#XNL4*eG6pXsK*
z%KD3`a7>Q50CtQ(GXI1sV4oJ?y`hb1h6cs)|MM2u$e}<!P5N1Mpv>~hy%J9XYeeyU
zp(w<c%P&nh0ASyR1-2+Sb6|epfNq?8U=I>Vwa^c=QUFc#gJhL(ZAXvFEnKEFePI$_
zsRNiVMC*rNIO@j)jDKk804oW!=U^_`fW^pl??}^6^|1-@pbo#D=vmM7Y*b$THNHka
zhbH*I9xO0()M7oa4UL_Tb%1_#gK}n2mzpWPfRWxx`4zf<#3Jio&Zb9sw12p#V>Q!X
z79;un78Bfy30QhOEUW3qMD=R@X(J2uhwGY1bMefYmLfY@@o;dC@6Dm|WBD^Q>k2#z
zJ@IYl*0VI^$);r=j`I&`<T!=6Wc$I4wU(ZeA(5ef=*KPOFOLJeR093&9m)6RFBeO;
z?wD472#cR+p6=Jil#-@nc849$VS(LG7TCLd7T8Ti@%h?Kr-|w^zJ+5db@0dKaP~q`
ztIx64*P#~Gf&5e5YMBdVv7fB(b<$*R->sLmi(j2$rU;Y_2ldORyDGh=m2|k8j$RW^
zImf<G%0LTja<o-AjCyEnb(G)D53g+**ZRv}F%)U0=_U7i0obe6PBIF!p{x{lZtq;n
z^by7TiIQ>B{Z5^DihOF{@&0VH*r&WxR2EEkwW(xo=~BlWP&)<%`u_$!^M=?f9v`_p
z%8IL){7{a)zLfxy<PX<K4`0!J-<j!ImAOoz5C^~xSzsq!H}`zNr(HjE;Qo-T>4y1z
zU<-JESrn}23BW#*-4lA~Cx*>2f19{A$^!dR0oeD7c_JMY;t>E_Z6!I|bxto5`9<G6
zuYVg;SnE}?vgO6{!<lQ%2NPDAUpO>U+tGf<i3;x_fkHL(auN^NRF(p;H=6$PfxSmS
zn;k>CR#;Yl|LN1jw5){1L<&LqqC|1_e`ad|i73mQpzGiCZ*5r#05^{HkXLiX8vy{e
zwZOhQWPzQ=jUGyJ_k23f`u+`f4DBI$$vw;M+S61oUeE)Kk`yfcW0>gYpBl5=%#qc+
z9bj}g^}Fk}XL~^YuoF?wCz;I=i?&Yt&qgmyit2MVa}L{Gf1}&^2bKSKg73XV>HUn9
z1M{do{f{S8Ip#Jjmq*C*U_h(|wi0MBg@J!w(Oim$DC<#(^oAXpTb`7{<MWf(dSjd9
z)o=PVDR1spzbP-jjq&D&ppw9{@U$qmU!1sFmi(c$NQW;y1wh{^Wz1TdYl1ANk;9K#
zyG9AXzRC|^Q)PjvUtiMDG?6t8hcd|o(t~+ttmc|4-kK4mJXS+<=&QXNTGvPM(j(i|
z0-EppWUd*3)wC}RIK8R`oQr9Hmt=A5e7g}vqH!zC(5Y+8ZpqjNO|VxsuV8K$6GkPm
zSOm$M`49^6SOA-Mjm8y=n0I?OGQ~8PCmWE>b^!@5N<r#zxt|F$nn<*#S3lW2dqD$J
z2Ddl4Ay)#q5=dDBSRZ1Ey=BW5^V6Y_!jvY8qQ&I;p;T#EixvnUTcn)$Z~Or)unYZ8
z0PILp(8mAKAD`)8qa*ob?;D^oma*S_vmgvk9gw2&HP^J2Z{`}|YpJD9J^yR{{*o5f
zw}8A@BJb3?$?sSNO)Wb*L;u+TQsFz;-3MI4Pc0VVSeyn2F_6t&V*%qOOE!hAJ&Ajw
z*kQ`a4IGxLfK6DG1L}Deu&zCl+c)_0rVdl5t`ev>Lm(w!ifXKJ+`M^J>k}4hS>pHu
z<+$L2CYC<f0gL|oIrSYdedy3BqVV=)gKVMhKI4o!0yXO^+tlh0B{tUFUwt(%1j18y
zBRceEqO%UNtj(kTHBY{;7ur6+J14TLR;`qj6PgGoNvtgB<U$_w6ZR+@#MbiOd($lK
zy93Mt#BaUz4Ebh`wZQhQb>Dq=nn3)e*1F#TssOSj#Pm{4imGke)Q|<Mj63A23Q!MU
z)83d(O+J(tYvC7P9Ao=5Wd>w~G9KU(z%M7m&Yk<LvcQ-zKiGO8Yk=b$ZfGC!N1s}K
z7`srRPoBIY4EECJ|MHh}EEwx#8C$1i%eGkH7;AajjP_$ZKm2eHjgf+}(X;%$a^*Jj
z=RbcK_D7FQQ3$#7&MqM<;^2A#_yS<l|6O1%x#^}30$oenvbpQ=KFN#cv3z!cafAZI
zRadnR0MPL}y;top!@4Elc>-WR7Wln;H;bC5T$;&zLW6$OP5%kE1!-5PP)uwJtxV0D
z)W@HS-kmhZLtze>PCF3j6Ie~3aYjA;ikx}dn{obO?=o@Xa<Kr+weBKlcVxi4di(8X
z2D%Q2lT;cmGv}VIzv}vns&J;Xa{=s3B8k30r-=4}?E%=uik@sT+yO}RBmXnMaxJja
z+mu*fpFecASrr3d&$z3-Eb|Hm9%3AB2{<$wr^B54hA%XW@C)bbvu!ICF@Kl8=DKP0
zeHL3@4rZXZq#SO7Wm5$$rPaibeIUVaEHxe&K3{(6)<x=}YaGU3-lC$VqzgI-B1!Vf
z<O5r4V$p(mP5)Mv&8@N&$;x8dUo`5nMVq}B$1OKcjS|=(3RD|Z&&vejJakSSOTXxU
zCiXi3_A@#tzHf{UlPp-F7y@THRV`|6(xG2FZM+gs-;orl-!8o*Y*Lw^6z4fM75Kcw
z>N8!3&eZ#W7S&)1aUFhkt*3=AbVwH|UP|ATg$$;Cv7ohEz*v+42Y6BDd$NIrc*NZG
z=0*WAd!6;F9zI>b{{LOu%2YVa1G|v|u&)t-z1Ig|H_HJEab$8(K=sHaTg=_E`d{x3
zLM4jovd(<r!ur;id<vqmB1Ara6ovhpv!D?TwVTquyF)c|ZTlLQ!aiA9AI*65%Z;+Y
zuJ_=#GV4sE6%k!c{jQ@9j6b}tRp23fY`vo>#ASi);JV6sX?jR(7u}^WlgMH^Ha!lq
z7mBsxP61Uj*X)wdbFcsB)Aa#jJ@`}i24Vk2vClFu%$IMQKg)&0ge5wVS6yt-9PyV<
z)vclpbo|K%`XK<tb?@q61S?G@8UhsJz3Q2+vcOIkg*au0f)xNemO}hxOCer9<u$}B
zZ1(;=Q7^uae+qydLm}Q!6yj;}fzA5$hI|p<KWdpdfZGGLu{H);s7^9pU);zv6DvXx
z5EN7+<hNZfL{P_iJ;qxTSPLpBJzQnP3uFm}cwhOzKIj9mi)wEB=&I(Xp(yI&t+hdE
zGCmh--veXG4=}-m`V9nSF9}rYEMK|w!&K(>pnQ(@wX7(_@qwLe`z2Es*7}QIm|(`P
z_fh)m+CR=d#at{JGknaZ5@@HbI>2~X{<Pm>pXJ84`a?Zg?!R|=Q&CyViagbYCY3^@
zF>WA*IO*E~!9T8o?jpbAT_+B~K5Q<icCvX<dkDQm+r5@|FS&mio0}x(&qWV?=ZCV+
zm4bi`HtP)Ys1(w0A%o&xc<?5dbe6Gz+gA!0ZSlr*&nno7K<oayS2Q!H*DUF=F=d%j
zBm%{rBqgf>a!K-oqwZEeY@|MOhg=_Yudg-F?Jpva*#nt<+Ut$X1YoOBvXmY#R+-v5
z98RUbA+lB43$%T5tc4cB*~z-&n@bv5H$jv)FeuxpdS5!nuV|yl`SRv5pUa1ReKFlI
zCC7a}AW`Cey=b$1ozFEJ0QW-=cb8x90Ub|~@9JQ1-vQX4%??qBhu$$bm~B_RxOw%$
z26DX+>Nf%T#$KYCc^CV3!6tL1!2K{Ai^@~6Y(X>Z(k7;Y4!?6lt^{%=kUa_f^7GFE
z^md!gn>GsI@R;IQQwtU@EUUzlI-t)Cz+^9{-_BD^XD9uREU+Vz2gLlh{!rLrcKvOI
zU1BVR>UfUalsHR70Fr?NyIP=?ewk|cQA~52H}96!>lU+Q$!2STiNz*KL7`0`UH?9w
z-+&=p^Ir#Oy8ZSuOqD9I`ou2_R;T!y28iFZX-|};A<t1a{8Iz;0Js43U;$ZJdu?cd
zGcwo%hZbF1!o>do5(44^>S2xT0(o#=0_f*2FCZ&c?Eu(bKS(|=mv(kPk&6d9?Gb2}
zp>u#=tfgkJUd>IdTB%!20tmkU{&ay^%fbLD`zd?@*R{qYvF#)#n71cd3vkyDs0(s(
z7QL>`0b1SJppzf{g73t4_mJUFKm8oFr=~sWD?v2M#9z8}leHY5J$sF9GnYPPahA-`
zkXMr8b$PjWq^|&q(?=XgPTZ_RhE=A3?wfu|x5aWAHUL0=SLdwL$zBY6xxe+6XZcLO
zN4C@z|KwPjr_#?U0oLx1KmLvVqzwqn5syFKD^OX$M$kWV=dQK1)j#P>5*bhqlocL)
z@H~yhM3mnyh3HQJ?7Q#&(6$G6kZF}F#myaeoT+Q<NrT3eT+FT5V*Q_Q+7vMgq)oBh
zMyVw0kfHwalcjLS|FjDL$0)}1>C?gj)9$+Wsk6q7J1uZdxv4w-7~1}h9czoFp{}AO
z2OLMIt8;JR05{z5Z<{A_r5phHmtEF65WwEOd!KpXg)!C#J@E-JYg~3&OIe_wqF>3C
z19{V@nPai)rcW_0@Bxn%da{<>wg89y4*+&&{Zal<rQ)QS3t-0)Oysg>fxUO{p6~>F
z{PF+$KCr2OE`S~B1?l((fPK+Bv(2&%k;xXTha9!Q*2D^ceZMTQ7dQ)S)&d>K^lw})
zX<FRFslT0q{PU3V4_c10%44q%y71Ze%G9+H09y;_gXh#1fL+P6PK%Q;2Zm6HdluN+
zJRnC4PjOMG_iI(z+ya$f41Hmu@=%>!n&{u2yh_&XKSTg*<$reVlIB?*F5%-XQQ&~;
zQ&~G*t+p}vDP;ws`~nI7+^UkfM&&A^H2j;fMCox)Rq_G_41ISCTj<VVeUNNC3+(>#
zojkz$z;=bmepciYyLN@^mO`?~?u&Wr&F!MX-R1+C+spUwlk!OlkYn>oAx#&RT4I6S
zUjX(FU)oJ%q4WAp>Ggq4J$4;9Wd12&>-CAN%r1><8}!jvl)eMbYhX&M-=r#e#>H!s
zR+<4{FV^X%HaiGIW2&Egr(fTxrm0*^l*L|7zv+J|wr7F;oyJe{hq_5sN*lQj0<b?y
z62Qhaz;tOw_m>6sDj$V-B^^8uk)pwQaSrq2m4!Bf`t!jr7n!kQkJuY(`@E|3jg8Bi
zy98jz^X1O7Q$=Tf>3eg`b{*>5(|0mbqkh<@v1u!R@~U3a0Xgs0OT7iKVX|-`Y=K?T
zDkr(ihx4hT%Ln$4W^k+pw(9?~ti6wx1@>QMf$b>7Pf{5#mGAj<Qiy+|HQ`D5DP5$x
zduWLoO9wIcw80N?HLH|FyCxIt*ySIh+aKOjSv^h{!vY({pIxy4_La>{g9LtUgUZGH
zg)+ij0=P!664hI%_^3`iC0gI^4a!)?lI+a!LHQl6Yk75$`-UtuVtrtDkq_+0(rkfE
zb1vGv$K3ShG&6UXUiw^@kY&nS-A^^0<%2GTp&b?(K;yGg%zOCzm0Ij|fJeIO5Tyku
z{##a`O~r(gP7pqvaSnWmYk{4Hye<@g{ham?*|)$xARpN41#B*efQgh#eQ13DR;HE|
zOWomL8j2fE0x)=%)~7Bn>cHmsP^Y0YNttuJlv8R7P)#*3zL#ly-zkuGvgoPp{HQcb
z7tU*G#QVvA`XRmR#0KV}Kp)t;QAifkUx+3f06P^qP@~;CRIfec+h_n=VDFb&^Kh3G
zFQ72+&Fr=2@^|Fwz$y5s9<@sqFdwU}OUfcX)j7&CvaH@=hJ3w9j5pqa67y>-DI;B_
z_m*06<q$NUnB#*e#8LD)twssCfv6`X*~IS{vGt_wr>)p#x;`gX0;e=VBpv!q@7vr|
zlFz%`kSl>)31nXa96ImVvCaC#cAyONV<Caq<>Wg#UtX+DL-TI-v-!<D$<hIZc&-KZ
zZ>C9XL7+Nzew-aRMSr~SzH<%WP%44-7t73f^VV6g5=&~VssT*Bkad675C{hVAAm$G
zB}<paRUW-4egKSzg8bWWrBJK2-M<D?HpUl#3x8NZ*Qim-`l7B}`DAN7m^UxyXz|nc
z-vQKn^UVp?e>h_~t3bN=sfBhqH83=*l(W8%l{gmRSWY7YK-C5f%G<IO6T@YyVeQ%-
z=8ZQd7%XfZK#HuftOWp$4fLnypC4l>oXOuG+@8&~Rm^Q0m3!{GKVEEF0M|F)+$k=o
zN;}~{7|Z0jGFWkWlZltEXQ@n`Xd~yVyh)SF_B&Sl$+jKT^5e%ZwcsM@Q`R0m8p-0l
zp8d|XoC;V9zx?tz>+hT~O+Bd-fHMjfsRpdw2M-<|1^VaDpHF~z*MyeTD0e*b%*Zg!
zJTk5>w?kO#rs@NnG6O{4d+#TfMS<VUD=;$vo)(hAba0Sy#{6{_z^4DZKwr7{-gCt2
zfIJe<E3Cu&_8qj=)cDbMfqvbsTf-17yPM<OU1R5lS6>}xSqHdE|HBIZ^wVp~B|=qQ
z$$H+1q!#&xMfJ@$54C-faY4F?EU>q2+hZ-qH|re01sE7Ey|krV9#qz)<jTRkF?8ru
zJAS-=B}hXGvy6``u4tX(iW#o#-X{Qd4q9NNIlf;N{(XdK0Gr`(LN0*)`)K+!*U|#m
zvcO&zV}b4Yz|J8GacvYc0ob}$t8_tgNBiogw`d`M<}~F92Nr$1-3je^u);W~IfVlx
z40xTz1(Ue90g+~y;yg8SiFtdrtVKi50k%CTYv(IkR<st+E^cgfI1KUt>>r{i#Ea`d
zs$VMr?9?o<Nq@Hh?4h!j8~E9x$N{88*|}B;^ZbPk1*m6zxP*n=b+SquyIj_hv;+0h
z2lPliYnCwAwW(stYC$R@P!wxw0-F|EEM8r0q#XK(_BbF-+rPgv%{;SEiU6TdMAoJ0
z1h8GYv7huC6b1SGNxsJR3UsCa;u|_h9xTS+35dDxn<es%=xkv{(U~>?dy@d{c>w@6
zKCn~YP{!8R{igqJ(dOnJ(MXS2woM1bp*vE#c<PMm<~nJB9+cK9a&dtuq3&e+6u0bS
z5|lsQtTj7KM}Z(hCOyDR?c3#);^vl4r<hivNez;#%j387xk(ge^&a{z6M)^TauNCK
zjaV7`W#Z-$>gW@ltyVamZa|B4m;4;Nc_~r7Jqq#d0<h;r1K1iPLwYt4fL%7w8ZBO$
z_8=PoTka3W%IbMfXudu{<+@Sy)dK`zpC}+AsGYFT{6@ZvGx-<Qf%TLto0_WGpuGA~
z2ME`_CkyOA02>Qz6e(N>jaRFzuhSk_3TC1Z*ZN>7#2Z!!<MCOicK9(Ig?LT^*v#$A
z<!a=)?|v}<nk`x^Cc985se_q5axc-NVL4Oo<N~t3JHc8}^NiOeuyAm(5ATzHjt}g3
z0CqzG*vX6qR|;ye25R0<=bGU#0fefBs>u@quzLVtV}b1+I7UAc0@wn;%B9hInO9R>
zZT$6%MP?m9b10lH`spDTH8gdyF}6tpFbxXkr{vyb0+!wMK{ZSPDf>OvrKY*0X~mT0
zFITRk@tF|7jv6d<NZ0#(-<gE$<;?eb)F^IVl2QT|*r^7R^(e#zV1MKTuyq(!w`^ha
z`BlxN(2-_EtWY7b!hLtzYV)UIQk-;3n_BZP7L&w%T~D<>(~0V86_y4U_(4#e{d)*5
zua^?$qazo~mCZV{L+KF;s6KbJt!(}%Hy1&N!KA@@GD^%c(G=nl0J~0#7TA=9ePPWZ
z-vk5L)pU68DrZX>`Dc&H^3v#~=I`>K9cC**Hc}LO_*}V((4IaU&22OAi!YL6{BJn{
zU?ZFPvf^p~#CT)!`}&v0_O!k&Ol2tz<c3@c<VxW1B=GC6znE>?wwX;EHX5w5nK#*|
z6)9TOlr3MuoR}|t?}mq0(!+n6c`cCzHud(e74dRRcf5G{r0+Tw*zf4?fZA5)Sxl~3
z1lEot6%SBj$dKXIq7wNdUq{7UwrplFGG5UEbJniiX|PblKPup}3w&o?=bQ!W(|md-
zalo33L*4)+ehXcmsZ*z{^{ebr4oA{-&$-Wt%rXR&uCjnF7U0ejn)JOt_k{O}oB_XD
zwybI?)3L}TURDABzK=cjwFQNP;!+-hKPl+T`}Aok8s|z$1EF3jsr?&~h4&o3swYia
zA!UL?wr;dBzJD*jyj2)nOq4@(dKuQD{@G_oikkS4ZAYF3kOaWa3?yQWzRv=x3m0yP
zlC3AB_-&e~OquUx$eO&g^~IYNJvCJJ-+i}Gz~%-~k?*jUytH|ck95jdq{vC4nl5GA
zt!~|N0#E@<6PNp+|NPF<h<kMu9;AMc^>zF9H8aV;-$ydn%exZY2JvEf*Jn$)lJ?<@
z+@*~o$e;gwc2e!_Po2Jk1u;IC0rUx6L;0hB|Mu3R+8^J2XB;DA`kn*LsY9nuayxio
z(?Dz8M0I3-nK^Tflx@BVgWrdA*56tTVt}VX?ghN`WSlW|g?Z(bap5_M@?$l8!wnq*
z%i-z}`OL7s%wK)=d#jwob<)O)`GN}?2U=4zmra_q+`4CA4kB<3+J4GhqXx_eyqyP!
zM6y7EWYC~b!~K<c5+B%5f>)I*CD-Pp%smtuUU}tvI~Q_|v@2IGF1fYRXGvX&lp|5T
zyLRmplLA!aD<wCN36HUH)>)_8zhsp=C_^eCxJ&@p=jgAh{=E5`b`6@lb1ks5kTyaK
zq7W}yw3zK=Sv-DB-_EtbPH*!CQiuzH1;DPP1L;%_gp(~9hmn>-Tvon|oCUV7)wBS1
zHFI&3iWdA!Hoe1*$8?0{@qQhmtrOMYuDu7$kMbw8c*|b1YWG25Xup`92Y!)j>tA)Y
zmJz0r&Lc<27;r2Q_`lk3ZKU`QhA70bR*ok_4udjNh>P|_8_yeBS2i~bz;*y}GIgXL
zSYUeqwhjUXINP9!_N-bZ1z<NYMWq48#G5Dp*pJrX-6OKNgC011pvYrXSc^%Y2Bl4B
zQ3jUKX1s|03h5B7xabfI3D6*U>KtUjktpBn-jf(DfSr>R;z4Dj&#2p4I^6Bjw%eyZ
zvqODno_q{%{_$tCS&9IMbkImUBM-`%M~fd~+@OQi>7t%eX9_2Mx&Z8_L<w0j9Tej7
zVd<wQ4^xPfb|lMi-r4y=e_JMC>+UzFn9<uK+J^Lx0Rk~Q*DaHnmYQ;qNI+PtaoF#h
zr4eg8$*DoPLgp^{VLw~czFGC*fFzIjC8}mEuwyC2b+9n=kEfaH((<NHAO`9~KOxw<
z={vtX(A9nS?y>%Jk3Mxso77}Nm_q!M`DU%=wVy(L01ENJAqw#zYqWT2Q&wkzJx-vp
z1F-YzdjCeH&D}j38tBAhyBcLj$>!r(>&>5J4IZ9blu-lu^!`_`My9+h>{AJ(xmRgg
z06UgKTz;u7g?O5P7iDEFupjzD6yozYMga-{*xls=n>a@*<Ws$#_;hxR4{T#Dk)PLe
zQHWC)+Hw4{t>#{_Ml8e%+h0jo5~LV&g8<*gWeb^VB@5b@t&{cQT}+gGjuxEtAl_P-
zjJ$aE+eMoKDa1dN1$M*oMeN#>Oj<6^I{9M1_v1O{^QA!mc3A;-P@=im$L1$W4zG{N
zY=LckU;|)Fam0TBnB7`3E3tQYcGP0?j4Y+MYai)CpBlx?Lzgr)g|j&hX@k|WczJBt
zJoC;Xu`B>`(k|NL+^fCP-Cb(Snl`z)(3R<EeOdtf`APWSh!jEVTi?cIObuC-v!}3~
zj(NoOJ;obCaB~aS{ntXjON%nEN3%*^+(_z4ePd3?-OJ~)7P<37&CSjlR)BA90oY#)
zz&<IP?U$$kRAY~n9sVYY_?P6bI=m)GCf779ZO+y{qJY+^z+megct_$9^}G%}|0BBV
z@0afguTd~m^le(+{N>#GkUoV)>{6jTaVf+Frj8F;V5c&dQ5Nc0D;B_(JiolGv1umP
zJVARJntO}t@E6UyAIUO4OiiM(_0T!B%$2RHh#cw2DcP{+U8?=VXLB}~!CC{In<5TO
zPnQMuD;Lx^1>Jjx6KE2j3t|Cm+Er`HbOG3vaslilB$yi|R{~i}fH`LC*3BM>EoE!M
zulg)jqPQ#x%W2=n-kZU9YzJ!a`?bLSP_>6qF*{s0IoIy=y&uF5H+b-fFaX59(*up$
znv-?jm`uPD5^!?ZusIgI#@drWTHM1b7i(^SMJR*;a#?@<;lKa=Rp8KiOUVtmO@Nvk
zAI;sm$qhylpkXAvJObnoSc_l48>oY^BHbZ&G4h2*+YeS!PQWF=7(hAd@5&HKGx80&
zQQxPZ{w@rDM((oJ^OS@Ed7#~O0a-L{TG;}4Sb)aLz{N}V=d;g_GP7o_vB|1vnp2K9
z9Xix9=bhIuNdPapL^Efuw&1-VG^}2|l>AS33iqSP6B(kD-({(kp_BK^6~K$%$N`xV
zqQ85ogl@&$^FQM~*Q{9;)3axj01(|vC-Se}i!{gLZ+*dn^=6gW=jb2w6R45dJJE+-
z`%pG$u(^KbnYB%qF7@O?oUsu+FBAW7fBV9w@A9Po0vMvakduKi<5@nB3R7)Uj|L4Y
zn120Ir+Gf8ccN+2R>_ijRM_$vP?T|buJ-FEo*3+x9?+Kg;L9)PS&Dwzkms(w_B7L~
zRkf5AIH(ss)Ug1j{}P}vr@gy&t#6fLg5*kD@Z7)tHBrxb#$t^34jpQmi!W}WHYKg#
zcJ!wQ<HG;_-*>i;aGm;K#f(xy^XAocCHh^LegxI_ci%0tKB##HP@nR`#89>@)a6;{
z%X6_|k=(ZZ;~!tyd5n9^U5y)8wDSsz3jToHrcGNZ#iD6eUfe?_0Qi0To~E&q?EWG{
zK=?-U@y9dGqD31leK~T(-#jwFvOK2%Y=3<s-#OntUw`HG7rmB5;&|V30qj_b#QbK0
z*uQ_je4L6x+-^cWAK1A7c8s8Me`YBG0PLQ!z+Mq+fh`}{m2%1lwkEW3It(8q0DEC3
z0K2*X><UK=z}8Tb<&Mg;bc>lGe|g`FYVdo3a^lrvN9-LEW}2^92x2b!e`@pbVAnd<
z2X-O=o5P@4(jt9GG`ypiI0|tsRI<Rv2X@L7;;PIpfvA}QY$E`>gnSY=v{utBhKYim
zUt~EuX~lN)#>ACoggY?yKSUb*L5(EfwVr%hSJz)<y)={*NLp32f7JzK@M=MR>9BtE
zXYc<`8^*K1zFrjKcgY7fZJ9;2{YfVLPJTbD&#c(C&y4wDlbI%vf4=(6Lj9H&gwH}z
zIof{V?;IpeNpvmda(nr}ep3D`4+p>|Whlh26NUJ0ABA{RQF*_1V|ppXu`-{udb=6&
z!CW(4Ku}cRJxcd&9Vt|(%HiKcMv$&EA`5>!I!@@tlBlFrrDOqful%t0l)vUw4v+2g
zN8ul;6i6ZdYlfL}TN_hF)`&4(AjeAS0PO1p8n21*fqggt_Cqli*kJ&>y8vvxl(BjB
z2SAd|V#|4Z>KZfT8_^@VHrDlSQnL9+_qwK}4jfVm%vXE$;*-e-HWcClmAgqH>Lk4k
zrXx`P71F5AWPz>uz*=B8!UFq9wW_9H9a=nLePFLO2Sc(xN%>!b4{TX37LlfUY#U@p
z4%Fy?4oAl7P~eg8mYCTa@ENVE{;JZ{jCsBA)j2u8X()>7%9_(k3Lveh^`K#eqNbGE
zmSSdS{E{?2u=~je_K#r-aVdOBaq1%rV5ew-t@2}mJ@At`W<)#-Y{@sK3ui}sNA(%c
z2li4?h%fOJ&_bn2H-JsDKdX%%|9rlAUM^LB)R-WEHD6z^lzFgs)5v^p@1*w6E&+g#
zeKya$FiUJVZr;!wbA#yeAGxrRWy?t|#iJ8D5Wqh4tNEy33)AkDlJ+3h_Bl7UJTXSb
z8oxtBFE$>JIMJk9I|pEo5P&^E3UFZsNXfX40PL?s<$Y2-AUV~{eDRq^cFND%Eg#M?
zUo6-hT><LKUV@5->Yel+!pRSTxbI5pL}l1V3U&j#iSF7eT#|4kP@cp9Ha@V6q%?<l
zYt+xC=HOrp>_h<eracGE?V>^dX3>@ixUc^5!o@ml7d7XJI+%^FL8LC}{)PhY(rrJQ
zVNwbhA~zZ+D-xJv+NBe~CQ(^nPrpV0b}ofD(#Q?D5=e^#e$~PL)@_^3wyoRD!2=%6
zFY{C3B1Pn<x1=dpD0O!YX(?D56Jb95eFE69-h8G%t)ssw02^P(Kbt92R$Bk4P-+tZ
zB~Cl7h6N>@<)pn4_n!>;LSAaV{(6B~BbGMSY8L>&vGhIl)UpC>ipUZ(uLVBbeg7xm
z&ah##%oktG4+ArRKv!MW)|SVW&;LwT-+9m4uk*L<It$vSHFD#|U9u+KYSyg-tlVXR
zA3x|q*#X<nKD(|3nz0lQ3IK5<M=mlSemGq+^b8?fJiGtX>JLDS{6>xXLBP>L3x-l2
zu0bQ+sZ;H2fyXYNR6em32dJg4l#g~e>#RB!hy|2QCE$M-%Tg>_9e~YTg;gn(#F^VE
zk9?tTvH&8Mqr@d(sf|x?lp5T%RLT(>mt;$nC}^pa{cRc>FZMT<y*u^JxkXBVfRy-;
zrXN#QEKW%)wtTcLplRX4Cq)7BE^YeUV~>4fR<G6>awy<)`KqhhSjua6E&ZRK-({@r
zKlos}1&+xBC4sZgmOC{mETkeEp2Z~qK<q~!%?Ja)0H)VmbDC+>rh1(8GNcf&Qxy9X
zE!asPqu-&xF>qj4Yb_irQ}Rj_0IOepb-bA|W0h4dATE^Px7^anvL_rb_Y^^aw)ptt
zS(=~bTLlE-{G#{ZtFCITdF7PA_KQsiC6pygHd)$yTrLn8%bpc;>r7V6+$Y~*!)9Ao
z3Y49|90%A7fL*FoVO`1ITLLZ;-VwWLXeilqyL72*6@5^`@dxO}KmUBL=Bpp9{JBSe
zXx_Z40PHsUj@*(dM;?v)r8X_b2K|V6uu`SswqN6hA=NhZmoxS6$n-+}mCo|fxBxbd
z7_ZPAitbC&8Y#;X{Kg|7gJ=tE83O%ID8#XdIg$s&zwxec*u?|DzF0o67f1QP{;w>s
z+nb8g`eMwbJ6H>B0oV_IzR1keLT+b9ZDbvVcrQ_i9|`MStc6xdOLEMTt>!~n1TGY?
zgB2pa@;Eb#3Qi2Z4Rh3e{|^D!56B1h6<PYgHlh&UWrhd@8VP`{Nt(GGpLW-_It&VN
z`M@4HB4Qmunq357KPL*$tbAbmt8qXZfElZIn&-zaH(z3*27(X}pbchO!sl%vZC1<5
zasn{BkpPRPm5N$_diky9S-`_-xI~le*W&h%cW0P?E{yoVEv${q4f2~j;GDW?>&N7^
zQ!hFb1nv))U)V9rx0zl0<abFz=I}ml;9eGer-Kv$9+yQN0Cv1}E^hjv8)P#|AkWR8
zEU=b4E`B3fQ@weUD8vuby4$74Cw@uCOX=fpI$WIr13^v(^}Xr>u<z<r!}JtgaSlvN
zX)iq+z74V@uRr*^XlhJ#gvqzGu~gYvbcwW;31yjSYO9;%KX`>K|2gI$RFqH6hsDg0
zJs;S6%wSRaj+WK4^MQSWDs-a&r2*ZgEvrp_?E1mly-t>tZ%$rq9v{6dZwt}u002M$
zNkl<Z0#@kyrS;30L83%2E@g{U!;jMZ?Dy_W^Y!A$OPsV?%yT`p(QaBn0CtFtBGrUr
z;?t&3h~oqM(|JA%Y%Lg`5A5<sbbkUp^pl^>HGiGBChP<IM3u2m*avoc0Bpp764D_}
za$m{)!5dOQ7zdTJUXp?WYXTBKLH!Ppy-M+uOf#(qwdKq9RQ=T`SxEi~Q?nd&7TC9`
z4{zP07YYWCCcL8RFCWP|w1Ez-5>3HD<y@=7gh8SZ|11E&?ymk=AOPSBO6RwH?F0dk
zZ;l1PUK#~pcPW+MJbG0NQ(iBj93-YA;2z*nfy~d&SZ@x6(zHq~vM#>wf<|hu?B|1D
z;-|h?V4j+~F0z)Ye_f+?dg9_H$BYH`+HHGH;}@rB)8ZX8Sh;J@I7|R`+ZrB#osPCh
zH4WyrEqae0E^u;y_7Bdw7TW6i0<ecmNhE*h;5pT_6U8Ur?NYY8{OxIG@+vX@Fs^ej
z<Wu?=)hli86M$Vpb4P5~r#yi^u+`T}$Qt+)F_zR$k?nyrErob2fGuU8FD`9lnpG_k
zSY!5q>v!%m*S|l<Ow^t;tev3ul-Dn5D3+VDDX}yKW|G214x^T7{fh;#^J^|YP5uU7
zxFB^O*fjS%9a438YJ9}%pEBtXcA9)(SIz~nQ)r>wsJRlzS^_`)^pn}KeY=3Rtzi%w
zE71Z43z|}8%9w%$3TB<iZ#V|KjGP3p6P-o-#d751u^<0V+pHBjLkI?0{N+26-vJ{5
zLq7UwW*8J=-`S%_V{^_q^}^s3TSV5&%ppVI@Lu!QTN4G4Ziw2mLd6Zx*GLRlSY7*{
z4@wU@YSO`4(4Dma{M>QJSxTcqT;Mx%nKEB-@4VC(pmLk`dz&}!wpP@O7jL#ewkscb
z*Q{CET9?+URW?&%lpWyagAb<3-}YjGLcd1UEw30{ptA1Sv#FKoYp;DTOWbYY_ioC9
z-(CRF8Z|uTI+u5*%zPJuWahga{u&hT(MP|sbp>?dK49vZawiA%Z)yN`hGe(Lg7?FR
z&j|zY>_N^sr@sBmoL(M*nCnD6JjGf$ySf74U4Q-Q@^PLVWqY`S3BHiF>pMgXl10$t
zu73b*VmS<uy?(u??10Z~u921tfP^R^v~61>4C=ajd;<DHb-h}~BrfpGl~=Y-8o*Az
z?EbNtM^>^MENF@^cn2Jxpfjb5FK%wywX2yWU>lvUH@>mnoib&4cnkvi-+ud9){58F
zFZwe>N()=$_k6U^54Yxj?XB;-ubceY1Ky?<_KHb!$dD1%pR{XZ6kP7P=MR=Kfosua
z_tJ-TVd&7w){5ML_l#NmaASo`-;EbsdH~PUr?0X~0nU#$eWrHpG6M431zO`XcVVqO
zWy*@M<ulgM7hQCC05<u){`z<+4J{8Bn{=@zXH3VAKl(21$Gpni&hsc<(06<GY^LAQ
z^H(_0OkeVtufCdVY2i@@LT<<h%j+wzY!#A0Dr3&yZ!={UzCa#s&In($U$Cg^uD{^-
zG=JQ<05*{ji0e@PhjIbz6SQ#WQi%KeMDF*jB>;eZ!O&S|<whMwyP(6gqemfrWGt`&
zQWglv`AoEY@5w*jN>Mt84~N)1U_GRbqXm@(C>PdZSU}Hk_;paXh>};|nUA(<GrUh=
zN%-JRiZuAZzPx26`y7*EP@oKnD8y6sfz8rrDZ~X14II8mD|qMyM)_lbjSuXi_{$3_
zOu#kL1IV8(EAi3t-!ggaZu9L%DI8#Jt3yqz;2S&JlQ<RTPxP^U<hS??<y}+YLQPR7
zL%SYN9sxy2>JsHVpv5xC2R0Vicgq5sO;9=ll;8k3Z%M|ZmTSXE{V4!JWw8B*O^W&p
z*6C%WaKiyU{-KFR|A9Dkw>Bwz)NlNakq!!RWn(R{uayP%9-jqvGx;lj^(I+hr;$P&
z3+(Upvh?srbImMK?MH>UjqHTwVpHOv;E$&`%BzA7j|ZK3in*X+c{}ccp873aqibvw
zfL-sQZ>5Ip17&4_Jy{mm)w1aiHam4ZUS^O0*!^UIy;c<Wu@vIj1F$_G*yDUYu$dG4
z%epNdg*Y40bpj1vpS0RMIVKvwzCwx#_i5fN>Zof|sA$*+wx2>gWP#nav3y{siTakI
zy<7+QmO>l=TQ6&N46A>>ENa>FtOa%o1wYKF9)<WEQHZa}Y=PZP{!uj##T|;-%$xKk
zY1}Vdv)znavBL~sxl>@AW;NCX|9t3ApSe-t1kFE}Rz2CYsUqbjfpwKci{&5b{zqcJ
zI{^DySzzziOS{cW$pT<|6yilwo)^|?EgL8b@ey$;#4{^l#Fp!~{@b&_j!z+8UI$F+
zP#B~gnO7bWmD}?(HkgB<Hpi7f|0WvS=QlL}J&bi21?8u{T4<h}v^KKlD($}Y%9yA6
zG&KkHVtp)xb=q?`dM<`STo$#T_S2z{eBY&`EwUpAhOYI2tv&Go?H}wOhJ-Y(p%DKn
z#{g^{R7!LI;y0(7nd|f(;-3)gwBgMikvjP!a5eV6iy0@LDtd_k><7*hfUP}hEE@!J
z2n4XX5uy;UB@66S=2c3vN3Is)QHVFKTHF;n>XR%1u>0s8Z7%M7SbzxXFR%9zfPHEi
zlTXT7yoVf#U@644{(UAT2b8#+1vUp+IYA+w>jV2pv}$f1zefoGV(;9!-E3UHK_J;L
z-gZjvjfxd7CZGop&-c!9j0uR@FuMSDyffxl5&Smean3POe{=Mg9;MybdI$Yx`C+Y!
zA3VRVLx)Zg7_ux3Rs+srDSG<pwe)CY4GT{1I9cT_mY>(z0+#kDKQBY9YXLN|<aV^Z
zvE_(O4`Aq@{~T$7FP9F$(E|^hYpTheNAxBnHnmiKQ%1m-UqyGm$qXMp*V3Ltl}x^X
zJOvBpRsCyPfX6QbmkYjku|S5F+pY6}p8&&9M7M2Q-GbAMVf+Ap{`ov>Ey}w#=@Sp@
z&z?OShsu&wTX{Lh%FLA~tIx+C`^qX9cmVc)t@X3>TN?{_B~B|tmdV?!mHhbeOT%Cp
z&3@&TZLCGEYr9A~9;k{hWc-G^`=xZ&f)8RSkNy2L**c~>UZlKk(1;a7`zqS_i!bKM
z8hd}ZoU|dV0$7{}`HA-|u)mH1uxazFuWoArzerw*zolNsJfaU@X2y?SBLCY|omEuS
z?-%X|Mi`})5-AapZs`UAX^`$gx?@ObZ~zGr=?<lS)X)t>cS|?O3`h(;G-v*2t#fg1
zxR?tTeCJ*7ckgF^o@eh@VB^?NqF_3>PUj<{W=9DDJKY~uakx9h$G@w90Kb0of~6_7
z3-Px3_!K`(pY0ZDWgSlp0p@*H+>7d4;(OJOs1Ke2K|uwBn&a&H<UgBn0LvPgw5+YE
zTy@9XZ9H>EJz|r|BH#UR_BYG_%6zF~-o7fULKr*R>Th8jEzv3Wo@IlZvoouk>A9m(
z&mp5t?kboVfHSIxPUE_gH${;Lno;{JH<KE;&0DD}IQ?WhA+qc1bV;;0zl7G(?8psD
z*u9EZVL4J~s=WEpH{yCWT&uei+g2A6Qvr|npHyl!a_;J#3}@LCuNFmrs>BPbx~#=;
zfjV~~*BA3@!GpZ+AD@B;{Lj$@;29p~tm95%FedzD!pG##wL;s&4}S@fXJ4_L4C$Aj
zk6TW^%<$6(X=^2kyYwWPI1KgGP%PA#w7d3(0q7S*h{5z#INp16fi}lT{81je>zC3r
z1iTtpM!>c?E{pMUy-}S>EU`b5PHnrvA5MZjN^b5JtP%Iu%8}Gh0>CjH7ttx?YWYt9
zvOFObte>cj0>HM^@SHb*+gc2oQb0l(lI<WYlaT~^o2rI=Ewp#+_0W=yAjOim)cYr`
zW@ImDAZ?88#<%gtBbo2(RQGzrU!@|SvuH$^O)|91Q4S@MoDyuvl21Z>8EPs+wp9a{
zj17nAe>}x%>aa`-5$fy)4invU>Xc$5-f5v5{?i$Cc&x_S=2m{I9P3Tb$hOAtD>ZW!
zd(mW23t;?uqD(}Zf$k}oKjlf9@ec7<1*yEQ=U;Et3d1`FVu~N;+dc%H4b1$fMD<JO
z`=|;TYj)_rFJxQQfQIK#Ut<h;Gk*7?aEGOLh7fF9M1saz&;w@T#yz=_q*%bz?WmT*
zvG7$Goaholro?n6!~x^iii)Aa!aMoD;4AP+rkx_g!{Q1_h?@I*P}{T$z`F;}u;aHC
zSKwt;Zy`zQi%`sui4xDMO@ifZ>k44_v*D-I7t?%-%Ifq{_n%+jsk7AbfZ()a#Sg@w
z_E=+idZy_@2Q7$m<g21&fb$YF#Ya3T(wpvI80rk*U5DfC9RSuY7kGwsz_)}401u@6
zUAW0656Kp!UvSXofIV_4-%{Er0G!ecE9YB;cf0!)xP2BeB%njpMQ_wKVz2ZRCsy{o
zjJ+}7cKY1_5!Cl}$bAgJiKRE5E2Lfi&mAM8oZqKkGs+lX%mLc_^yKu_#BJ)eRc6xR
zmrfCRylhE8W@uzz<8mz@(YH~!SNPBM>UHNf&6Ndzt5Vy_)K5Dh{Eiu)J*r8>P(Q3N
z2j=z%@&E4T=r(;6PC84koFpt$k2HjqXF?MbL!|P^uc9|R1-PWfhE0fKf{p+SbJ)hp
zM0219gre+Ac#7ATF0zVqLv3W$s>_lTui32v{q9c{1{dqhTMZ0kDvKQ}M%G&0y#{k>
z?@)PWo~WsyK&M^gB;xpuHpduJ01!!gr@g22nVBu^UEeGXheJkU5#p01zvX%AjlWI?
z{<ezx(cy=LyR8^YtNLc8z*urEMLAJwY&*y4-_b!#f_wl1zr*^Ixt1xaw{9KNm8-0X
zyEE#T&#&gSv0_Rp`*As#P|Zk26^M$V!N^H+AbQmn|1C$+uFRgMP1Q=mc@!m?&<n;>
z&UipecFI4We}9rGOcV(I;IlhPcVgqHk0#Mk#E0V&;v0su2_=&^W2wuG?MpCH`my!r
zNZfJ)M7<AQ?9XJjVM(hDXRq(1HVS&}>-{0`_Z!CvK#ym)&Nc8__n|n?S0|$eewh(*
z#eozo4Wm(>AUj#=@8#k#7Y2*Sg!nzqg2Ws!mg)S2a`Xv4{(*9`$ImYd7lde~ickCW
z;wL+)kkK^Wn2JZsf2xlm4i763YJ=eMO7&i*$qo!bT2HILMs69{WV;R$-0|^p%Z(i~
zFr;Vx;q{$L%)5H`)W>)uqb51g1L-wyaY*M%)BLGF=%15yABAx$(1VgZ<Be)xG_!H=
zQrGM9HIbd87ctgkpYGnD0d`pi00?G!J`NnIq<f4L&lFtU*jY&Uk(lkwV96<X$?^A}
z$3aGOJkN_^jmLrD>pU4#xrf~Q>hR8ACO{^!Ki74f_F2EtWMrOS&@P3MeDUUpr#A${
z><00qO>_0^fkpq_1S**-iFYnV20Usa7i;Go#_uKZJ}UA80MycJ>&z_d5C5$P`Fk`;
z<plskh(cDI+{ybi8VIOrO6aYZXD@t~3bczw^$hpkqSbfVUB?3p*MEsm$p_=qjJ`7l
zT)s23I+({AKru|_b?W1tVAt4r(%zsa=R`~u0S9XZB<7fnqGGSRpJTxqGeCcV{v8vb
zhF>HoPCWN=F$mj%{(TN@^MW{^Uz}p-D<h2e^Y8v_UE7{sRXK<5S!Nh^?wJGdIZ^pe
zr9nQ5MwxIb|KG^Woc!F~Z%mQ)fuTb?g&<~>C71Z%Ma@-5<*WrZCX7{Ou?IJ~nV({T
z^;1;jeg(Hqs(bN{PTECPw`*`s%=wp4EqwXJ_i8%Z__L7>yL}rWb0w<|oyShGqjnm}
zwr&Ie5FWD?(yz=p>?N1Z6MTv@Z7CGq_u)BqwPG?ifgB-+ug1;G{&s3VDwQ>4y0}gx
zg?|XWG<jWOW;@T~a`QCH^;2~kYYCKmX;CxvbgS1YJ6{eyQCRIOrqg6#5h?4h6(xni
z_^0eT-y8+;<yzrDui^>-P%1dIU-V_bcK`XOD1L<jOo*GUD83VUUe?{(L!q44jFL3}
z`|Ulv5S!SctmB^aO1}Q0S|xya`)j<DOnC;go55>1V#6=v28HoMn6X+j6$@%9MT^LB
zksTX#nkYNtkhJS`<bg)+yI$tvG$=BdDUfXu0?BoKT@nR%T{Vm?>H3+ghONUZb0Nyl
z_JOUQ%G@+-+ZF)-lev$iyG&cnIhe%?K(a03p&KUyGH4SAj5}$a*)jC+UF1hCkAKt~
ze-mqb@ArG_F^YWv*qjY;OE*<*+#i>A&75uR?$3L<)><fXmx$P`Hv$$78@Y9?{2ZNK
zce!?3W&C$?zhwTQ;@OR<+|j$rd6TJn7b<P#5GYq4;T*RA<~v9XV(id@egVU{8Ab}i
z+4zoTx`gsbIyK0)dIB*%=nRd^5?2mr)7%f88G`Y>T~nyd10k&)4(2&~>#A5{9efDW
z$lsM>&mJgzd9{*MZ$s}G#Ij=0_}g#erws+%d;Zh?3S->q0M2(QF~`JW8ktjvTx081
zi6+Hx%Lv((8+Z`rdsn|Ema(H3DK}SvV1y8Q3H$>BjQ05SH%S!+)_N|23${(8y(P}b
zT+3vnBw-1QKxiED=N{>e)Fod$t9&|5b&6?w*yLyPtcIRhf74<c(LF$G_A->1^!@Gs
z%a5^R${P-MeX0p%JAAI9{SM#hki7F?W}z1dQ=;Fwllr!U7@>bltq&eTq~q^r&=shT
zMV{_H$17hS!43fDzV*v_N58;r`Wj&DU#EjT9jC$&RF5&fT;xz={0HozSqXr%g@y`a
z&)v4EyEjsOn%`1c42Lm5lBCDGBz{Ar!GEKpof-h@gnuR@1oP$&XFM7RlR*+dt8vGd
z&Xn%n_)RAO0<femQdnNeB*_k^=Ku%-nvDH<x%QSDE9_bNEI@7%r157^+xfi}LO>?c
zFcZrhkbqTk>`X%q2v6Z&ci^ODL_auUO>Onz?|t&{#|lAXrlW@RoyHdrqqn4f+cV^v
zEKh$kuI0((8c7fLfuh>CqUVYnV)vg<cZ<@Ba2glg%d)_9G4gmu^WUEo6a<iY$QCIR
zG@XC`>)tq4g%;oLVy#lh#J>+rgNy*jLO``rie6@m<*f25cWZgF6>bR`JDQ+c`Hm&#
z0>znzSku7b#k#A0r8b5qu66weo%d~xRbH;Vqx!o~zNdY&6-l5U8(bpk%-{#3+^}*m
zG=hr#2E&~iQZ=2g{%;!nVj&9J;R6j?ZS~>5CAb&6-k5R+e!vE|oaa?Z$`9A8=a_vN
z_{R0+XUXA-&q17NKiHI2*UVE3g5P56l6%Eu5vygpfP0rO6OQMAO2H1JN`1cg#a+Db
zC5P9#zpDe;n7A@2py|Yb>dvnT6-yuR4cJ|2P%U6$e&;4P5LXs%N7clv%STR_0=-4i
z+XPNPA{}F=mR<I(lJVYDZwFiE_+RMWpcIzplAcYzqzK1kx$Sn1p1pNPjD?E^*A!|o
zdoBI;J7c4`a+%Egv;pajq0D0nMz5Ayyw3JLx4fG$cbBKWk9NVo8EIVSstXqLzO&o;
zN-lix681r-{3W69z({J3mn%VhB2yzXaH~LgvegH&|LzTQzuVpEYm~#yeqLueA)poq
z{FuW5v)=%MF4O6QSIU$k7-g=h#UHNLK~P^wj#t_l2~4>wBp8=>1B#zAI>)8?eMz?r
zX|8Q(wp3dP6jiv1pJKd?d29Rt96%)jtwj{KVoM{~VjwIh3VP}7&>loO>%KscEW4+n
zR@bFOd&b`C?=aSia8DwKpU(^&2vV4?%Vy>jXt^X7>c0nIm`2XwBTc&My#$aX3@cOt
ztIdt1u!))dao?Dh=zpQ8x!=~fBV_L|R|lnfEn-?@=?BmbH7i@MSod+H@lc3=KFZ@R
zQq876C2)>!O3H-2)a7wkYod;F7yotocWITK_fTOK%pemohR?#*<YnwUTPgb-qKM2W
zvWI?Fx>d0zB?<}5EdSehD{X@hHlG>3fsa?3uq?6ImtEfab}a7vB5`Gl*`5CjYz;{0
zeXYfIjm<KaINcp6XPfp+OuRzMN?XTId{I8hK#d(*v1P}|EEr5zJ#QJ@+H_LcdfVBj
zF-&6@%9+tI)Yw1)_5Ii3S7WTvzf`BYJbHl*(~pr|xpu7L0NBCGi^d&a1H{f>{8XhI
zH3ckS_U2qkc4%$vnPMm95q)=t6zLM3N+sjWAJR|6gaC)@Li?vXzX<T{#1t06Wim#k
z=CGHEG09Y1np7B0b$y}ct@gx%a5+7DsSig;Me1Ct^<iKR04I{SS2B~d6YFyhNyxnr
z)&fX??imrhLqA&Ve=38ngr<>&=}J{+Hi#&FHiO!&6`Y*a;Fxn<Yf&=j)oXH6{9RyF
zq@ZWaUwn?FOJtXJETGKu1|Pmg`UG73FI8``{CS9kCgAcZeaKk#`v0ia_gD1Q@VSny
zI85)Ct%4AWH_tW$WY%U<{(kmwzL0FwdWpkjV$`1Iw>xc96otq<c+%1v?R-0GnL8$g
z;+``5Cetr}=zipEy124hIL=FqIWU1nai7vT!fF|-4#ay_Y*>4dmHfcZ954;?koz6o
zF~|qM*Z?1;>m8!dx*mjzNS9jPKdE@{3VSXxXA_wV>N)KU@iG!_v$p-}JuN!2f-t}V
zDNa!kJN|(vHq4T<(Q@YG@!gojr;=!l8<N!7e$pS?daWq@H|YF3uTD9x1&^uYg&gt^
zpMJxCZ%}U))rWB9n+gkBVlncg=pS{!<;WrcmM|0MC5FujgRu7Ax3i%AF%5wNcIRDr
z`_d7u%|xXHqBr1rbf{0|snZL8jd3fV51+=0doE?|b&{msK-E>=WhM_JZ{$L=$^Y(x
z;Dmr#WDa!&#F6(stDO>AllC-rwV|=2&0vjp8IFyR(*jkHB7Ez#pIYpx2|Ml61Dsi!
zn6YQ1DXbOGEbvEiXCJFy9L$#ag+L4ETnq&<DG?m}mi-P`JA?0ga@s9tv!bmTq)CIm
zV?_u$Es*1uu<3Kr%@32o>F5VS><$<KMaLI`Lri0w!mo@KrBACm&Fo}68jClw@kkmm
z^d^?Hl@nehntFc<)oK1YkYC&E-%6dm^-QtG2w<^)gk4TAH?n*kr6HM4h5Hd}62rE>
z%boPX`A0t0lWxm$^9rzWzE_sqh$ixATMLDO`_@}>gM`hFs^?3AHR4CYYvz1NrCjgx
z^~s)A>GqQ`y5DIVTfISalVTXD=6mBqZ2rVMQ~{quSMj+YQh;C=Np%`5F6S&M#lf7j
zH&+q5w<huO`1{z*y@L1onbxl$SYX_~Nzc#b((yA*ms4zoq{t_A8ag8S5tT~020J*y
zF6$jd!?9ZyHil3}#lRDeTzBr{aWU;*M?WYe*Q#a5v&<s;r__yoLOhFE;8g6E&2o_$
zn4vRI5&CG=^_HgQYsxYX#}x+BWpL||m8NPl02~{HVWu)Z7O+=~QxEN51XS=o7RPg9
z;>hlhQi<=^og1xf+kjy+Icga?-xGr)l9;_0g2Ks#_a;GkSs3d{ipX|s%NMRw$)jZ9
zht9vxRC+OkdyX?8Z@OIcw*(657Rv!`FD9q&VA9V-eUAbY7C0{JzdSnvFJ_&~iJfd|
zT^X70nB!SSgkt0^3i7#r%LyCC!q4T%;(qGtDW|DT3ndsc-q)-C`(^?V!{?pKo1$2t
z3o$J(+2vgs38t{8DQm<KtTW3V|G#s(G{B=PiOBvZXSyOo$LxwwK)zv}{8HMvp<06l
zM)x?K$hd;M;qi{Jx16EGw|0JNYWH@3-j>VN#^IYo<R0ImwGat>u1i9v3A5P;2RtCQ
zi~VHy0B8&Vd$acH9-A6@Mq<$V_$#<NR5r^iT%KJ^TF=O$yGjXEcdQj*jqPWa)-)Uo
zFHo2_VfhXEB9x%(nQBzjWF}doye^Ttn<xm~634)s$AMtRNaz1%)T&hZ)?K3Qh-X|M
zy(2h=3r$ldFufC--EIyD?8+lVOgb1V;DZ6^M|O%UCBy-lzZLM{9SVR;Gy0HK4iHQ4
zkYhR-_~-%so*&m5hY~O=swXgSx9#~cGh^K`&)g+Rz@Z(4DflOgn~OJjL_RRM3}CWw
zH2-f&ZdwHac%yN5v>n)qYEq;8Hf31XKitAoI+qm+w{PG!9QPba`}^{lqvQsS*J$=H
zT{#}Q+S<C@_@z&#34jYh7$i9g{Q^@-Ct@&Aep_zL*c>D4a~tEGoc=eR;Gnzd^Tnc<
z?c(DNT)oJ7_f{RR{28?mN=W5NK%1~dAr({S;+$@z%14MT>c)BuUh9xibDBWwsy0c{
z^(hLSOneZIv6<<PsrDqvFaZvL2=3F+d%swWf2?(02pMzvYS_x>>ECGo@<1DTh8!;X
zg+b_Hw`MiF_{}V_O$9)VQZ$*}iY~=NnXsI-c0s$z_Ad8;s<?8|BIjxg#IKC!;k~A*
zDAFM<dA6HgoYqmn?)VUWz<)RKFxLUPg-hO$ZH*f9qz7{C6o^SBc}yvN+R3Z<y^{Wm
zY-{*)JW(*+qj$9*;#ODdweva!`DnK7EFY6R%I`P!3|jJVQxVqyEcU-u9|Oztpyb5&
zKu(xIXMmq!0KSvNyLbYr^x*F%4mmWn-3?GR^V`+3iy-%PdWU^l##@&+A9JPo&m_w7
z_W(}&Kx0cSYI8cF0mq4XGr%SL|L*UxyLxOGJw11XWbSNUl4&v8*@A7y2y9G+d`QGp
z?@54uo;k_}>1a|b5(ozfQN3P;b7?mFz<#eZ#eLu?Zyx^{u#zs|SmC!{@r|$JbWPGv
zvBbQy4F?tzb;5J<@9gm)8<b}v-#maU@|pgKVqj?3dLcTUz-9VGX3u+F?l%hy*q`mS
z_!M*Wo2Lc6nzAfd3DO}ye*94Xf`a}~crh6G_?VTw3Yl1~MIWq5_@j^RHEPgxBK+%j
zHtLzK9j?1(zTvuHNRA;oPP%ZeexA+}m*{{-sZ^$rh?9fff3y~6WljL(#Ia{k-N`Iw
zscj_<7RM1ZHJs_5G<?(=f^^~h%2TDq>r2-@xSl9$Q-{v|DILi($stnl#MoKa&kvWe
zhIh6re!d$FJzJWt`FVX}sMOdtK({=>)+hYv(;~fdb&u9~Pkj=1H2e|g%D$K|qD?OE
z<2#fuHnNOPDk#M(bgYdPE%P!>3A5Rg=wFm-*%!?3w@y>^uUE5WC|ZWBgFWfhnclCy
zXmlW-v-y`^oSWl-1bRompYzX^1Q{{C<!l_{bpZSCWDVZ(<o(ne+{3QWT@Ma!*6p|<
z&jr!146$W{2zJ9}3IcN7)=->+v_^kwDzq-W4>K%@HbbunmFH}giG1fXTgV{98V&Ol
zRqTAl-Ra@DyWOJ7uPs}!A6HQimgJF@t8N4UQ?Nfovo%RP9<PopmmdUVCk>q^swj*k
zI*iyKoJa1w`N}nNa5u!=A-v2sVfub<X3;htHf-EjwN;bT&K*lF)YDSo6RpPAba%?@
zJ&8U4>xoBE*-ddyM)awxTEp|Aq2^Tk(A5H*sz%Sld2P5q&~i<c=qFeLwwmj^1^*k!
zn8?v4rTJ|9^jLC0bi+#pz0)S;4%aYv$ST;pMoPiX6!^^}B6wDW$!5GiMR?ATE_}3q
zm`sGvXmq+(+ziE|jn}q?Da9)CW(esvB_6s!kIo>rlOW}THgA`vP2c@~ae*rawFq+W
zSvk|nuP+`=S2`OQX{#PeVmmh@w)2u)O&PrJs7D{|?r=X0E8`uo@%m9Lh<Q7q<9OKa
zMPT3LF&|MY>OC)kcaU)PQufT+@@*g}%Wwqg$PzccBZ&;a6BmYH%5JEzsfpy}QY4}%
zH5J@s@UQUfO;b}Fs$1*$qv4_h52nn(GmTmYrlu2jHnJ7QcEyaxOG?2_f_#Y$#A|`#
zW8W6~h~)NVSB$_t7Xs#C&zosQT`Fni@DC)5EGVM2)@&L#LpNDr`FUtVu%Z{kcq+N@
zcNyhp`PCc1ueADFVR2BYlFzjiXEBoJ!vSZl#B)p<);{yQ-fCm$sVU4($vGDD|A>Aj
z)R<W0SWqm?WA>b1akMS0p6}}LJNplF`bZJHznYcr8PkI03{~>ij;f}q;X5^TU$CZU
z<Da;^bf${)w$>~cFanzs<6P>QFa5F$jT3>j%E}~F!M&@W)2=njVV#R}n4{lj5M+)w
zDtPm)OX45Bh14~Tq{a%Q6YWpd!h`BR50ux#Bk)g%pq>mwJ|-}D2OY)2$l0!JTMbSF
zrZ2GI>NTWqL$$X$)n4Ci^XIX`?E;*ZL;Sng^=UsxSh~<Qj)?uZ=AKv!O+~ul9~`AA
z2ndZ#fX9;@S;?w>a5!P^J!d9S^B5XNgRO}~$c<r5#FeYJRQ#_L`XhMfhoASvli1#Y
zT4zzjaL1!m%osk6-9|9aRyrEjWREC}8GqjJxz&U%9w*H8E6!y@9(E>I?}_*`F`+by
zM!!d+Z&>dzf8Ycy9*!W{`yk$#LhXawUD-Qs82kazD^HGln2m>jV2)VC&Ws1JWjyY?
z&Wr`jX}$4OykC2SGQfBwd5UeRi_13Zfas2nh4P1&dUtIk)m0L00^+gu*xwJ0A(f4$
zRKK`pT1g^zdc4wr)Bz6D&?{V36c1)|b>@SB(8WU(db$e+55fxnu$2bPT)C9>J3YTG
z8xtCkbw3gf*XqpPtnpv5{-sJ!z`>~lp25omwlXD^l6s^AsqKq&!D5tK?R#e2fg3_2
zD!X@`r$dINt`rB<tH0{~&Q&7QyvHEaPls<yGq?KI*4?e=6gU0MiTyDmgc#-&yz}k&
z!qppTP^G=6@}fhZVqw$C^TQK7`4H5H#Wp!|<w~csP5+OHqK+6^p)-Ph4ITX<R1Jr|
zg1r>eMTCbs$3=f^`Q-%}TLLl!?@pr=6xV8oKcLmsV>DVKio7bcIU=)MIF(AHetQg}
zT+&ZU?edv~$aVL~IK|b={xIjH{qzi1gHm$lV{+&i%VUtRfcx8_&}arDma_m{kGwaz
zXQIXU!MDp~qpr6We3I8%-~5S+_LP~~XY$^}o?V@q6r0!vah}kr3+;vDdVTwStii0I
zdrQH&1SW^6q&U_2@tBWukBfdbt!~qWI;Um)Q1&+xxJ4;?g4Fun`svcyZG7UatsE%*
zVBX-Am>g<<yuxd_VNHtH^xsmn%C~acF2)4vCTXTO6+7Xf)+XN{k})fd36I840)IXo
zO+?ZJ2Yi{I;K~VBV*2}ADW-gfojk2vWn5^~GlF36XA1pl>f?DaH(0Zmiw0Tj)vQ|(
zbUmV|;9aE=^Ur79)S+Vj-lY*ld=z%w6OYgGyZw4-u|u!N|2sa7=1GNXwjHv<8_&&$
z<7~S2brB9`i0=4P%P(-|%XQOsiTHU^%}}^^A>E2<RYOW=-JF+SO`Ko!vL_umThT@X
zrZsW_Elh4>t-=!-_v@)|-r9MIfAKW`pvDZ;KE`dvL=@8A22C$zME~UuI&pKvIV|rn
z4cbeIeO!T*KSGrnubsl38a*@V1;J2zUb-quBDQ1V0=xN29FEj@p0(F_sGLMD*T#5=
z+ZkP7s9+l^fk|W%FB=%3E*$EmCBpN#>`D(pueUniLgrTl=4mV=2=`t5^d#9=bG{t^
zsGcNx>$WEPGG_ixSD!XHL8j%;f1UB+#TtF_;jaG?eFE!uMG-n+S`?m{$CrdLlsdE?
z43#3VGCSHuTJ^C|LY#)yMy=4mvy&n*1($1cx*8P`@dgZA54j5%4UL;BIN<Ce<ST?n
z;COdmsQ9uq(=LWZp5MC-pqER1KX%vjLNj@Zkv<x`k>i=s_y3N37pB=@32Qwwd2j@2
zJQm)8SU$X?n+?X_M55R%B^}v;loc<8yTqrex)L35kS`48R=%vj7K;>tEs9(SBHllc
z5PL|%7G_3?q2oW0??)YfI_J$Izb-nCjd71g7JwVo;_V5-w$LNsz)jeA5tHcgf7BlX
zh0IQa9?)oBkf0*1f7f_u;LGxMURNbQfhAJzm`9E!D;K|iz}i8i9>R_Va1=fp#N8w~
zgb-ywQs%)Q1WlrDk)_^v2bh02u%}6cJZT%q%f0E!le_VBlr*ynj!<t*i*KtP6~9{g
zTnlxMz=}z5!@ny#i!^MUJhGe-OItOU+^@07H2rtd|1Q_d4!wuOT;cAc7^()YzW0}A
zL6UN7NCXD)WH9q1a%v$#y%Yrrj*MJBZr&Wa^+>F@3A=`CSXhj5tD|hWsAs4ve`})=
z|Glinut3R6`ng#a@21;f_*p>AU8}gG5vsPBS_EPEJ`a-ut4~7|w&YiL%tM$e98rCL
zaHahGtvMjzGr9-yA~A!Qc#j6#;sLslJxL<e>^Pfv4ePt*&DxJBAy*DwR8_YaDEFRV
z*D2sCD-=<_xk+)*)~?#yrfD9D?avN+!SvUj5cR?mg1JQx`?bLn{WJ@{529MH)rX<J
zhmCNTD9620pVUl8bE6PY;aNMFz%=Ud^uHe7lG0LiTO+yFy_CPE7NrEwkm}msv*Az?
z{3EGfxy=BVSJcg-6K~40Urq(0N9ex*aaMsi*90NfXkXP<3)3Cj1T!^-3$g`N!TkHe
z4IzJYrfojFD6I1x`2x+4%R7)8gCr5&!B;O%@?p6aFwp^f9(=s#)=2b{W`)mxRuHn4
zyY?&QyH1jXrjvPKi3pyFFrmU0W(lqMRz`c>$$&t%YVnC=s$OWE_cAWt0g@MV*=0Bp
zn!O0UVxS^hXgg8w30a@vnN57WamE!fZ#((OpT~&XoqXa)l}#jeg8)ySGP?ds#A_3I
z`^}4o1oAUS;CJHQVx{4dKXz{LdFA=RXg$J(U~7x!oWO<XVBPWuUa{J+paIThI}Bbl
z+lhZX$R~ghfjOSt7WVuud>SP!kKINp7}~lXaL%ox+T7r@X=qWjMR}i2Vmnk4<%`T_
zN+w%iJZbN3&XU-d?tGHGQzVd+xxBc}EWPi@#?+eU+T%}jGqd@n*~?hccHDs_jIvuE
z6`?#K$o8!Bb~Ul0(DK;HwRL4Ybga4LK3dyI(F`(HJ-1Z<Ogz7-iF^j;N)J0-vo-Ux
zuFQw>C0hsJ*5<~!_-Tk>u(raoDiED7x$+-@p&*hRpCKmCcyajcyRtMxj~L~#+oMc7
z<gK@up?ZsQF}e4ArE_&FpDm@(JX5k+yp$x)MZ0kNll4|nPdkf_W2lxU$EEraab>{e
zV)0!4&Dq364Fq~-I!d1vmUIR_Tco8}ShM5Rkk8Xr_=MY{Ror){#GH`5Z~2rD+~UOV
z&pQbY^%_twq}z`k-Y=_?TpCB2@ZR#3xdoh04SCbq7JszP$y(0}$8Fij=DkiZ{MI@@
z(I=9$*e`-=##fEOQ)PSXAxle&3l}R1DN>JeDk0bb#3X9aJ>9WJ_JSkNt?ysbO$$H1
zPZufkGPA6cd>Hpp7sRv@yzo{!6M61xpy*&f&ba;Kp=<7w7<VFJLd4GJLgVS4ZiTZ*
zDb*Ui#@Mbq`#t}MYsO8{Oy_smr!Td))l5HEOxR!S6I;3UQ^_5!OpRrT5Ub0ix3wVZ
zPlF%P5$*kcaXhdi7GQFNZMMLQm3QbND-5$?HpdHjgbU!|Y;iBf&YGJnB}wWL>%M)q
zx(=EA%1&&3wSvT5a9kr)P%7LuXXnrI&%kq1_uAO1q0vE@`!e9RE`4#Ik4t2zAM-w5
zf$$%;XUy{gCy*h;EgndkgG#4|tmg4H*H-5&-QS54BkDU+DAYNeMB<MScgIy|pdRdi
z%kAhehFenHVl`(8^?jhxD);*ON}fL~0lPIi1HAnaDck0?FPP6*^WVpsIhqt4@1Qca
z3Ve4Ka4Gx<tItF*rBI6rf_Fd<$Hokpv6mqXYXc_OP&pH>-o~w>yUY{FXt=f>vzYl+
zR=bO*5M(}G&h6e;=%B=U@hed$b#8YSg9;Dcm;~qrS1)&DOeb(MowWZ)woI~Vlw}8L
ztX|t@R1RWaQ7%C)?n^KHK$~LWO!&#77%6dRBC+J8xMvs7%qwD#N4c)t_vT3t$Nitr
z?t{1cW||um3W|N4t&gs!^NXm>9$tI=mEIX37(EG`={)9Zc{Lo}+YAKrdf<>Pq!LUd
zVpH!0g*}J3H<J{vRE@5KT^^9CwbMunuO}u!>-hV57H-Zj^al38htTr)_Vzz31cg=W
zahN6Lmcix%NVrW<6=vI7N5nN$=!#?GrsvgTaq9Q?1KmdAxuKB3iEytQ{YSnr4tKqT
zQ2_0!PtKoZ6R3z?267`Jbwb$Wn+^Mlv`|kT)3`0goXeSEPC!xL3j+)#!PRRKdRM{7
z3-b1^tSMtOo`wqJ3YLc%R|kGWcP8PVa2Yml?6k)~z>iOrca>lr>)(l4A`}*KhoHjS
zKStgK(0tGA7!G)(S}T6R9ifUHZd_!_zjXe!0FFKyo{<GE58DL5axkw$o1FE;r<liW
z=CChI<IF=JI`)%dK`P?*4-!$X!n*FR1|d3t9Cr;0L-FFYD$S+NWWYoi%W_})MFW4C
z6v*2$z%%gh1U+hw1m|HTDK(sGYUp>jUk2w2^m={>v?g63T8EqX2K)Iysl<Sl1Hzi2
z7vFcdHTGS(#M+qkW`KVsyx*8i6hjyu-jgl#)M8$K9#s*kKa6LZNYG>p*@x}S4q~Gb
zatzzk*thiB`S2VV{2&r13SY6&+2N@9fMitWNG)^gQr~Atv;!Yy&>7_D$HwYelqGa5
z4Vl>kss*p0YnoGMR7Tc|D&BWKOuVQaB=of^+*tj$lwdj<uR(7GqATKKKN2W|2kp|D
zXq0w6;gJ!{0^YuCVku|E5k&3nPf=Z9i>f@{?uoEkzo}GfAql3!B>$dy(yQsqQ8%J^
z*9Yp@CdMzu!sBlLJ-c7pS`C2h<^CcUM7{H$Oz*aTw;SEY*3lM>T!;EbVN;9m$%;D>
zISts87l}4`9e`}YP-5ZcxI(c<79C!nrDo>T2YRUO>)`gZ54@K3%@<HV1nJs5m77|i
zg)ik<JL&EAE9aaPQqXJ8?%N)SCU|`Z!Z1!$X^E*i#>?3a!L=+ETkZ_wXq=SrIV4TV
zBsNe$SUBwPP3G;39X%4P=PZ)gZGnGGG3jHVwTdzJtF9j1IMBFl8!=uD`P_eVtO_g?
zNP)tzNWB~aoffSVS@~a>#7H0Jhd-D^W<D}_%yaIOQDvGvd}8q|#}qsTwhUjcWSKg=
z-?IYb>7Ewjwqo(_d&Hh<_hYU(A=kQM*FO3N#k$8m#2S1U$8Zhg{`7S&W!;NHf7@{%
zRqOZUx~lO5C0`3>FN)Nb^To}<$zPLJ%>|7DXIcca%WQ`+!_jTt&clY^qa5@VKZqsj
z^l`4ksC69FGwJ|yNi+G}$7{FOODi<@u2$+XcEdR+Y3NBLePumGQkj16UzLuPmivcp
zN`-Fgo6tG7OubT_rk|voLX=!4uAHm$mLX!dpRc?2{zdQPzext-itBBn#H>0$_fa(I
z<k*|?iOo;1XyDM0Uu9{erEIPs-03V$Pa1>n1<(4a@ecz!0vas>t4*>VvncxaOy+72
zKe=BGNU7Hhi|G$IMLMmo<%P}-RwY~rZxnv>AvrhIe5N{LZWMP*-u)JSz1ewQ*h6UQ
z8)<k_V`hJ%&U6_;$Vz#e^sREH^p0BKzewGOYUiI!@mdy2Bi3fxG#fI8tduj`UW3<(
zi+XJAock7)yMA-sxhpT5Xm|@EFFsDzSZg0u-~PI@c!7D)b*GRG0MPn(pGZ65`H>oq
zohBxRZNY5c9(;F|uIKIld5d_*qJ=p}7;qMeu62&cp&R_+9IZE8mr^|9&E|~JCzGK{
z$D%vn82$;GjJbbnxs=O$T)w1l2c+)z)RrrKv-U)oDvQN0r>$x0?6lIKqv9~l=*V2g
zdbUR6j_17w`b3@8lo$1fBDw#6nSVfi1BH29Kl;trPmMGnKd|-_=OuTx+tX<p;M6D;
z?y>4&$RmKjj9O>wjLrvu+geO)E$mu&z5W=u>S5X8q*C5F!bNR~+)@uV+i5TNX>mt_
z>6t;6ekoV?md&=hDczmL^n%~fsKF*K9B&12vtM->?aG4n*}P+$?<?^5D^^$;ptvXb
zsb*-PBm!eJN(1_fBH%~Q=V=z2?A49hhVDr44*W0$SYST<JEmjFZYq_$n-+(lfA3%}
zfs(9+ZNS+8yQ1ZPOLls6bp{Kul-ceuz2@s3XwIAIcCs1h6%}w(R=gHFGn<7U9IB?$
zJ8zE_L5VAQwp{0ARY{*1j_)r19&@$Aj?`FT?IN&XpDs5Dz(ffcyaXp2rOZoe-o9?O
z85+1cR!S-cnfanV_^DGZJdgfxtCxOxx1sJWu3^^Dw{kw}#BWQhK=y8$nxJq%pqR#K
zoOb2e@R;4Nj4=D##_PNy>|0w6btsOY@VjoTkGG;tB#&!Ua#w;$efvo&4E0%cR<+I7
zDE?8|;)MYQBgXrqtS$tB=gmpJv2I>g)}xOo;`okkC33>3QDW#Tl1>roONW4CB;g|C
zhwIRETpw93&pwqa^d|YnmRxJs?v%fw&%l4JcTeX35a;Gv(xpesFxbrq{s4)-ehgaE
zgcwZBy&RlkVP}$B-t{ox)_{)~*anQ|yw3}#RimHtE;Ei09%uem$JL8YO=r~vquKKC
z-HYKJL|nc4McF_xAsp{XsPCsxLbOer{*~0zPqDnBFXy*an_fv?B-o$(q=<gq-)3&}
z`Mzht$@1UL(;9l?rW*6F)7z)Tw)vg?LU=xn#c-}D9KpYs=W_78t>eYK^nW_l6$1Sd
zNAuo92fA*XsMYX{H;*d$;Euys83bTdDY=8F;M<2C<OEM&->(_{%-QpTu6^qHOD$MH
z``?l^1h&CZBx}~p4&1i;NfTD^?wB<qPQFPW5={^t*Km65_i|4>_HH`czP28tE`G)w
zB^MaM@XD(RW+O0hB>ssL?u+Wf@(;yBhubz8C-{wTQGEx~N`~P1Aj+*eIGvPZ#kFgj
zYX$Vps6-@aLnMH7X^^$4QE&)@f^~R*@?U(oHiK-@7f%Y+rPA-ej4pT-X{KLe6s1y6
zd|p_2^}UWLdW?SK`t-aluC#e(3OCZ*#XcxRhIqi@0S*&P{P~~nTsQf`uY7o=2)<J>
zoDQR5s#)g4y@9W%1aW0U@28U^WkOBw(e<#5yr0*LX#I5R&GctWy;gNLgO8a5m%OQ4
zts&ZZzCO)=H)Gk``lS?#C(B;`8R#NTjX$y{>TTOgU&2=}l?|03NO@c&H2O`WlZWoy
zCr^AOi#>Hhbp57Cy+k>*o$~NLn3r}!<6YIU%hzw8=$AUun|I2)5QApUjO&rCBeE>{
z!O?#8=3S7)wHOW^>E^wrH3Y7z=^s2=lf4{jPu>o@Gxf6rkRzJ`c!qS4_gV*B8*qI0
z{;&+A6`n~U*kK2E3Px@2sfGadCbg@G{g43B?u#9sx)Z@W$zgpPWyL2S@)VvPzp7lu
zbNWmlWB#AiSAPdKSg_UgsTv{_WEL4lwEnz98mj#GKwIGvp<LJzxGdL<=NfgTSMu^-
zI;(a{nr-d{BC^7l;GS3W;4)0VgoxH?5mLHRz{H+hqNCh`q$2;BG+=1NmyPgczPg#0
zN}c-j134*=`W4?n$UZBf_Y7f`_o>_;EhzXnSCMJV#JfF{k$eqtG0oW~4xrFtP+^}6
zx)&{Ct*SoTSo24H5HoKOYYe=lm@8;>C{t(?tGXR*Q7xX%#$&?uu<pqDd@Ox1^hg@7
z%xk2j{cHz0Dr&dVl)s~DR%cUO@FN^p60(!Es?)$LywkF5Bbq~wu+V#Mmb~&;1A4k0
z->DrcA(gm@djn4kSKgevqBy~&$?L=or!4`PMFgp^rQFwO(;4M|dQNNeY|KOa`maCi
ztJ$YgNs*?|XxEm{MO#@H8hsKH_~l0R<bEy<98R00!7QmluGf>wd)M3U+0I!mH+%nb
z&X`-JnO<8LzsQ~kkr!C?h#Fx}NS+}o>Z8F&mS9?~p>aXk5lh#j{b=W}?m1m>KeL!4
z(VgqM%cAa9du=0eeer>Jbf=MCqgcQRaemCe#pZntGH~OdHZuM4aIb<VEl0>zF!>>o
znH#(Hj}XvwUt59bdkL`yjgun~*R#g!5cjRqVh#SNkfHU7cYtMqDBF{9v7jmWNO|wB
z2dbND#JH&~`J@?L>EFShz~_Te^9&Li1oUg9&p+{<$@y*f#~)s|5eXZFors|MQO|!V
zqD%88)e6nr<HBF;E@gAU7yDlhhF$bx$vJ`ABHeyWng!%y!e8&F$i*e(njpS(^TbkZ
zsqOkdA|GGI7t2ZzyiIc?pC|{}j5Gv_6Ql|I_%2-h6qTy)$D*lwgEZNAxa)YhAT2-8
zWj)bm-ElnBdHaXa<8wZqlU`V-^aGjD#wLv2J^T{I-RH$uVi!6##ph*7<Ycr>UK8F#
zwjAH7sslC;6M<1)an5fXx!;Tp^kIs8OuN=qf9|wWa=1z)<bagDD@ZV%`Tl?&$=(-5
zCi-t(KoPCnod0dew$YaZ9+5GTV6$eyW1CmLKXLfNOKk8JjN5NE70zP5;bS1aJGLiC
zS(>@$6_d}RIA(CY3jzW7JFWSzedF+)I$(@Pp_*WGi3q%dHN%PLW8*E%8(im?C-g-o
z(r?DJiCu^tdp8>_*(!LxWwYn~Zn~^W+xOcZ$D68D$ToTAXNQyM`|f{{f8l*W|JrvD
z4f>Rs&fm1WVtt;7uDVEqqQ{OvR@N3dT*OC~&efzxlx_?UA?r|Kvrn!tu3peG$ampd
z>vEziN^?B)o{s#MD=$Q-%i}R4kw~y&YBTj4zV+QxuUdlaU=0mpXS$4<Y0{NK=*|gd
zhB2$Jka@KGo6vv!x*)Q$vJW4geT>KBZRl{$o@`_;@;l~wN$FnAKcL*&Wiwx?SVCI9
zm#isAYJUpN?kTQ4<t12t#1*0l-M+1l6tMKPo_tf+S<LhKM%UnGvi&%2;DUF6A<o+`
zMZ!7sv@Y?`g0r$X-={sg%JK1|tpPar*CYho0~MAW?!ZNlVAfG)XEci89;^T-2yBTD
ze`XQ$E1z5wn0?jS@o7KL_oF?dLT#u$&u178U__XAJewC1`*Ag%!sUkg5_{7YxWaKp
zfRyuog+n9!f{xxr3i5;MiO_7_+60-Md_6&I@HZJHOC%$0@YP@bv*j~TCO`2WE@g50
z%t>yFeYRH%2SzsIuf^>0y&^h6z=?c9vsqV7;<dU0P!Mv{D522UR3;?6G6y>ouU8-8
z`gH^)kKi<VWFi}{IB66AZA%S7+XL044ECWq5a{QGRU(61i7-}-)Z~~h)Q5Gs7_xuU
zrxM99=1o3+;wp@P;F?e9iq}FC>BS;p@!Qq;L5HofDh&HJa$=NkqEHK*8!iIN`~_qv
zC)8Ih^=%Y}Z&VZ-kB&`ze!cibe4cdFcx$P13+94o`E`?~uDJmVIwjvdy^V?2=m7|d
zhnsA}CALa(Kxq?}qCdIm=4fw9&z1Q)3B;zVPkQ~?M_t#=EcCjKB&qWL-2EHmLfD8;
zEah$22tsu;ebN#-BR!-n$Sti0#I|cO0$-Wae1gs{U-g3o<(4KGWD-RT^+5R%y!bW(
z6&@M39TxX|9du?68NGF!#6m6+ekTl5*&ms-;ez|K?29RD@3e0)@&q{!Q9vHc@GIuy
zt7xWrvV`+~?M0#uz6dG%biR({q1-y$q_SGMmKzJ<XqVtoTf1drg`EWQY{oE-=C||v
z_m@87(fxl}0M%#<g1cs_;d^W2JiteQUAWH*B*LrLQ@))?!*7JGGtO^mV7DpYJbMQ*
zJ~X18lj7z~G?4cj6?F8vi3CM+bKf^U+2{L!L%tBuDE2Bka(!n*Hj8ge0km|2E_7#j
zNbcDf7nj0wn(G=k&O3J{u<>>H$Hs42XzM;hV(maU@{@W}aGY*x_C0XM^AP(pTU5x?
zKiJr9LEAcD+kY{a+ojFEz8{h{{a0F^BWW9ofXQlN7F~A{_3tmxu5PP_OIr4FDpG9Y
z(}tPT!wC?B84P?Zu<BRoxpH~w$CW;`{wOIgFHjLo*@ORQJUbxwd_l9+XFG1dP`*Zm
zef8%sI@LVK(KJEZV$e1kpUV{!bQ169&j$~AbE3KV6RG8{k=mwXQ<yoDaZ++sG%J*>
z%H<~ccfU$%j!kw|GPP*gTZJvL<ms19sG*+Hj<42Y@W0aV3V&<1<uC@0_L37{N%GI0
zk?o)R5O0ZnItKnT`$;b%q8BR%)z>HZSi~e~O3Dx8z$kkJc&iLA%SkF<UyO`jo*vZl
zA7*Xqlw8YU7K{^%T>VS`kK22Ax=&)&T|0gB3qHP1T8(pab3SVfemP{WwhUMm=ey&*
z<Kw<)v+!UqpoXzrRAe^G<*Ve#v0u)cKTeJ>j-B2tGn1!{F4wF1IA0~7uMA+}X1QK1
zSl-HfTU})PWLm^QJY%&Sd^*IcNyDvPG~mhA{MW7QZ}FCdvfJK?Q4zU7JJ}V80~VL#
zhu{%||MvLuGNi+UAJG{^oN`-gk7XeF<#XN}AO0I-R&I3lolCa8->Bh4PGN&o;2N{K
zgx3$RRnx_rm6$y%AEbLG@O9Ab^86<6<}WLWvnEH!*q-x2`_Z%x;ZItj5<W)`{$oz7
zEh9N{MfV%Y`v<j(3thyj%xP64>zsyMU-&j?PLm=}Opk?qokKW$A1$x<O!Gcx7OPix
zjNMMF5G!W7CYRf$>h)Pm_<r#v2I?W0K!KWC6Svg+6RRU-KeO#WAr$G-oj`S0XHAC<
z6U)uBt0B+mVt$$(ojk2%3ZmT9zor}VoPlcM2r`m3v;yv#RYFaL$QC>pj$AKh&5N;b
z@zac*sVZblOSCArPK>~{i3ffx{T;FDZjQ7`Ed0nsY9RZ;e*B?he;y+Ou<U-g>iV)U
zmb+irhtTXTvS)TMGtKd$3q7BvVVB@Y|AdGxWy5D==zdH;{-3>((3sOn))3UU0naJg
z!ZfN!V*fA;xQxs|Td~0C!*Cjs2bdlDbj3$u+Zegwo^iExn(wEEh$Ityd$^0%Y2b{d
zjpluCOd90uyJD^yZoK?RRPjM~$#5gj$4v`QnvSpE-((Vf78s~L%YWIleZZi%XPit<
z!Kt_^N39D6=1d5RxbWF>Np+3;^PtA>A)k~vjKIOpo<0U81brjX=mKyqt?;+&mh&{+
zT=oDN#1|fz_;HkT<!z7E*~;*rK?MbuD3RMBH&t#iMPKe%2S+ExF<$5&NVEudKZ{BW
z;xA2mV#-gD_bC!zaR!?j@4(Etd!_E`jjh4XNO$(5@4C3SW%i<y!8dsG0Jm^Ku?WP@
zc80#D#p4V{4#L@y<XeiNWafMJJyiXzu6;sXcM$zyf6>pVu)D?TFPvmd#XUx6(g7@d
z+^0TXsCw$ZF?bfB?Y4UQ^hbY4x?hG%jHwaqRVV5I3lR`-urp)C!(h#U-#{&f$*y89
z*c4LT(K54HdmNwgCUTm`HB^gr4X=1orQW3c<jBB4+$<*D<d=sg5&6P=H?}`H$5a?s
zFGTx)?_At<)i7+)9oZE))R!4@)(nFyY?j}dm9pJii%|=CL7OwspWCFFt3YnT6vI>a
zkzRbZL@0!z{3g#bBXDB7Jq_LR2tS;YvMJ}C_~e%@oHNQ|5gOP(=<-rr>jFfc{Ny3k
zGL!P5iI2xz?|IdZSpLVC16|~r*h^1K(iHPTBz&xh%eHr#$^S)*k@!mP53PzFj=vAi
z;^6oMi{l|=)+S{^JR;5cM@?NTO2DD_g<9ET#1nM?BnC)wYtG*KtV58AtWstx^=S%Z
zEBM1>Rc6L^LHCGHt|7{AIE~cIzuR`l4bWQ=g;6`ODYI7H8$P?ve{xNF>JrK4!NI{1
zOl<u(fau3(nU^GcWa7C&57#n9Vk<^xcb=hq82Vxs247QijQC1}ImQ^J{}nPKD~n6!
zz!;}+_t1AO&b*F^mDpGrGc>$<9sA>LjEhiJGAGRGx<7GdEGEV+`v)~p^8yD|wvoK8
zqIk~d3?{UY`q%YC%PCXleLtc-JzU;f=9=7^=jrmp2bLACU(7Fwjv-!gz{Ab-?c0|>
z&T$Bw9I|1j*7YY6-R*L2fz+bl{`~fUNIDm{idOtkQ~liinYjRD(u0jKF6XgPshWkr
zTC?Jv;I3?dlt~o5-Y*PMgQV8f-?o(vC1m2GtE1C1L3%Pgd>X}bvg!9J9N-0<RiP^0
z&y@D|-{ulxq9TLl-P=czd0^VnYp&`TG#Bo?w*hX050Oqb&Pee*@5KnJsDgKY5o=T-
zqN^DhyZhNiG{1`Kt9rWUglbWOUg#K56Wr;=fQ0vhskj|8gLQ*h_PtqP#2sSb&^Y7^
z(<s|4AT^S2U*bo0j?05EDX@6uLYm>Y!$-DC@CqS3NFZ12%4m>vEb6_*>sqxvO`XR>
zw=T6FWFfPK1r3jlvfT-!B;qaJT7>~QJ!SC*9G#z&ElhM_-*T<fEF3(E?}1+ZsiJWy
zCJp^MOLYFg>w<7N<5ykf&((f)vUrF+y!;|N<<aNpHG~g$A?Xaa=r#5mP4;KU|HgDf
z$9z7bj}l9d`{;;!yk0-7^OG~U4qP>D@Uw}#sjp3dpJV2X5S2dFW8)ZmjI>BO<C#h=
zepGm9NZxskcvB|+t03}(Z)Yrs)>B``eYI;GM+Lw478N|p_Q!ByG!?JWHITg)-<|vM
zi)H|-!B2Zpq0CV7w!i@AG+h-pR?O<7(EgGyMl1W6QXvos7O|gFaWtePl58SHFiDoa
z$UXPc#4SXqHwl_Okzdy_Rxj#lP$^_VvKJKyoBC_y<|Muut{H793_73k*~sxL^Qss}
z|A4^-WeO>g5|SXg6@y}p?y>@QG4dLn6?$(2?o2%c(G^r`wo7^`JQ$f>s3)TEu(zzo
zCZ99W@)_Jp#b4A||KKte7H!p0SVYWa&Q`tG5ip)5FjxORIc4u9NV_`l$H%W{*2IeI
zpo(_m`r%5cA1M!*;D7-(QhWLF%k*zz*6;iq<XUWyfEO*9MSZ@a=~ThXg5y*YLqsq)
z``V+2{Z94J<Z_|)2FvFhmTs9R;$_Q?n+gH-$<SM0?mV3HZ_GsftY)S<;*>dG^GX7D
z^XDs|?YEc6`cc1OpQEqHLVe_7bNxLK#8#BB^afZoRW=`*#@rE~-+IV$*c*(pFx~-B
zjx&vm-w&jb&J=sx{VR*_dW~gv`_Q06Ja=}+BW){cRV?3xjL1C9KFjsV8|eQK_Lo#u
zO!tMW^U;bxn#z~&w3=Nv(bKq#4|YoGWw#c`5d-#Ab`_>xC9ZS<ZUxsH;mJcKNf$LE
z3d>l8L)ivH9R5^CrPNx*k4i+6;D9e2O~@R`43X7fmKhte<m8j>r2bsORUDlAt|K`b
zRbj0I1~IlCiWbAy>PMW`PpWlCVgHA=w*aeZ+x~{7JEc=Vx&)*f2?>#uZYk;R*dWq~
zNOwp`Zo0d>RJu!Y6Vl<g@Q9v!@A3cMH@@fD4|}oJ+H=i0;y1>abIh>}iMT9-h~>3B
zMtNwmt6tvE1hQCyWQ#mG6;oquv+U;Yj}O|{xjM5d)dV+U;4cfV*i%H{*Gk*!M0um{
zxZ+M)@>>|3yCBTIi7?KsY|esCG1yHPzv|kL$l;cps`R%Y)Jz6*81d`+Re4wR52!bN
z7J!$n7)%<a=z{OXFI!B(x+?ZRyr_@gNcd80kRrNcc7!ttHhgXP9vLd9lNY=xc9qoq
z(i9ZFx=4nBH>sT7BEKRw%K&CO925LHUdi3p;r6mwz&S^wp=)C~G1=BqrF(rHS^e}F
z#2?M#+G>s#=kX;9E^rgi6>YNk3BoRA4d(U2ro|wGX6T|s2fV#l;GD35ojuO%smhxB
ziSlZ%d=GVezlpFOjfTjK<o!*yQCJUWb)nEJnlfGN48pB-2QUM#!w4vtopvaYfa`3`
z0%L-qd%3+UI5Fs5(E8fxWPSfYb%sExJrC9-2$}k%vr#jNY9x3PW;X|ppuF4?ABSs(
zP3awr9}UaAWvT`$iewYra=ewxz(Tfv@|!2ESMOz|5}p_&47vAkJ{IG5C+j;T7h|iL
z)L>v|YhpA|Xve*tNl;#DYUI$B<4HFl=Yc%o74oYuAXRwa(80LjYZ;1>%n|4kP6OYe
zgWCD^x!%o)z!S9+?}J`Mf21bC3V*`+>Xc0oM)Zm8W&ioo6h$2=IZXkZ{Wq!x*I`ke
z?@lM~>gPN8Ep7zG{p9UBmj+;!SJA~$nbH~~>f({!z{=zvUsLYpJ$#{aL=~p+mg7(4
z+U<$q8)C`GJ5wgA)VEWtd&!k>Cx$hq_2p5}Ne=qJ#8;sV9M{YlpDW7QmxbJ<;C{Be
z_IV?0@kx_(k=pAK78A_bEN2NVE071mhPEm3oUg}$XGILZY2ij5DMV<;#zC%&8==Zn
znpi%mtRd8NOeq4-^f3avozKA@zvUO(BcKGu_Y2$d1bOPh%Lh^oCCfq_$X{0GFOh86
zkwEeAFkEiur+f2X!d^3$$nCmP^O@urs2nssKf@k$^xR}CdaaFS$K2q%y;iGAUvMEA
zBm|GV<^bQi4YdasJgb8hNkaK>$OiOwW%Qc=rKdbw{S?8XGwarchPu(rMsy66Pvj?h
zf-@%?oT=-&N=_?VopD(|2lC;|K6mX2i$~b*e9^hW8HWLVrsUjp?{6rg-<gWBNm$f?
z7jY&H9Q5CcNXbo>bLk`w1S9V_0;|nivlQzM2FWn|brV63ZaDoy-g4hBR0jePI>$Zb
zXO{OeED>DE=6G|G2uJEuHeaKMt=k?CN|vL*90<LM>FXx=ASL_8O2PU>Y(@~Bs!chZ
z-V14}L+Fi?<yCU!I&BHlV~do27s(B$2#ZUPthz+3#+mq=uZ8A;rZgp@Ul#Zpo-lPq
za~;`wE%=+(iAssXx!IAsZbdEGYIvr54Vk#7Nwew<?mYvhx%vi8iwmZB!5XH-iqBU@
zgQJ0^Wo`_*3uGX9%pw2+6TK8JK|H`YJY+;wsD2u`apW~a<SzAfwis^OLfHedT)(Q@
z;4)7AIIWhEt>;@>4eO05Tq>jGKtsu-^Tdef<-YZLzb*AV45Vap55K`@{P6n%3RfZ{
zd0J{;nlCo;4wPB>=8!Sr7+kXXf&+S!hoJ=iG7_A#DoaNK&D4Na{oK{ov4jZ|Bug$@
z_zPB_^W(jCd{l@fd~2wRi*|R#!<-(E{XMy1;Rb|Z?WWGF5P#Ioq6h>++qX@wK%zP^
ze2jBzV|?{%u!1N^aH;)6TrRwdzR@CtoRuJiH@kq0Y5jSc(g3LPi(T^@xT*zZG;NO!
zWzgVI54mAoIKo1^3b!Rn>%As!)LpZ^Iy0Q6^DENO)}U!TV1d{Kx}nZNRUhA>wk&Q}
zBDoB;`77SbsST`S)H=C&iz24(h!pFX7bKQxo~Q>ah)dHJZ#kt%Z!K~X_rlKA`S!qT
zr1EreEt?o!xqjuG(R|!$*Q<=Nc!v>tz;YJX%5RTq*<8m}Vy>5`Lb_B46h+xRT``_U
zu21PxJNSH^#vfLpD={6^w3?GB>2Or{#8~6#dJhV*V<lziGE<eoZvb*uON+w-Q8$qb
zy1ozsFOgSyeSMTRUT7HCYlpsI^mZ3RVo#cfJNuJ!Q~>iPLqcpug-SNRd7rrP=P6SR
zZZ&N-8gmLlXYaks{(-A=i(T!p@f_xIywCC(GAB$_#s-3l?}R*ET`b<oyA0o_8309n
zL!+Cgv#Bl0TDV$|VklZ2@0IDKjOjailoxB=iO-~Nit8>f+ROE{$vP5>&@TO6(Fy|E
zmL+aqa!%z|)slTlU|A*v+j?|Lo|l{5cz>E0!Px_%%ha3(HP=S&Am{T$Z9&kzqVmhg
z=#it_ijFHpXohR^qpyskyRd$Owp*iTqs8m8{W^l6k10H?JX7Pl&N-eFOP4!)^=Z|`
zeQrXvrB6@iT!NiqQZ}_|vwfu1zMsx=%-79qc%zf0V}B3TBSAB^7v7OR6?$QUn{zsM
zDpPi?gr3^%*7woXxZY>C62s~;3wvUYijGk)wHLIpfAJy7PE(?M$RjJQ6jhRAD%Rh8
zOtzh4ST2zI3fiOOVA?})bw$VF+llSbU~XH0SGP#*RuXNh>&%e<&?bhuTJ&zik!0f0
zFi^*WF*^dU;iM9;OQZtc2hwQ#rBPsf<vQ4Dy;9WFYnRaW9n=PjX_9G@*=+p$V16CB
zz0g79dIW!Jxj&pMsog$QBb~O-^?~z<!s7(O{v_E;;8LZ(%B)~A%cRe@tnvbq8f?L|
ze8W0+y`Uhg{1ELWS9xSh=cwIvDGgF2-A->4RuPD1VPGgnJx2F*kPP1BdC_g=J}Uk)
ziaBVI9|K0$GRFtvs{Uk@Gp|0f@yYp_YykoT7}Aj(=HTgYn<S3k{_H_aoE_YJ3oX@m
z$dgvnqtm@(UFWKXNs#>Pyux0k9NoPtk&NoZ6!WST1jR<$^v&0<%%kUZnAVo>p~}+a
zL(3cj4TqZcYssbzE}?KLTnx~6d@k2ptO18*BCK~lfG5;|8T_#&2PoK)N$4AeXN@6Y
zOEL6S{5yIfCqznW%A!Br@puC6Tx5?7oVkV6U%kG#5{+J1j%GVkPm;`4+MY;A%fqJQ
z9R2-fM|9nhu0OM8GfNe+GX5eF(p=ZG-C;uXlzqy)+sbDyLzmI=h-XDDGL0W=)=ePd
zG`e8K&goQw>Pd`s#I99*{e|=ZagW7A!g%=FbWhTj4}wDMJueZKEqymnoa>%%*!2a9
zbbc180c(dwd_hy_svW@(+2<cIk>{J-j(52%*I|!)9(n4LmTExL!nx<XzQoEp@oqw)
zV&)40rno?b{(_(Ux3{+QFFY3yU>U|}wK-_)PdXkEv&#AeAoxlNi6B@k`#(;&$I-^g
z`C6v9>sXG_)^IXFU}E=XpwOSvgQJAa+lyuz?U>1bRQme;(VO@u`(k>Z(*s}XN|cq6
zISD-cDwdp%ZAC#(v;F$A0C{s+RPo!T?3;LpCF_vWXh#s~u=1#z7ZnpYP877Ose|Dt
zun;c4q0J12_$eI;165Infw;d0lRTmn`%L61ZI}(YUZ!cA*ViZDz3Xnvv`)d5$@io5
zzRr9&6Sz8L%I5oQ1Kn}yDLXH<3`E{-Ql^E(&Lulcz*Sr>qzXTbLa{rng)D>)yfx4m
z+^!irZhNd_aHi9K<(D_rq(j2uHfvR1^KA&^=3&#4T>Pw6NN15AvV3!y;xowQm~G=P
z`<A%XiRONxOIL;fON}ZmLbYPV{l(btxwaGY7|0OEFF|{b^D`Uz`r{>w`CgQ>V*DuH
z9?9pv*R5yz=HbA^1bsO6ZW8j&v%L45W?2?IBQ3*dxs`3#i+jp!B?R!FDkKC=m5>i6
zJAk+OaqrL0FXLoB^~ak`hVdXMqD*PQiUWdUm>Nu<dBA<*fpmIJ)?wyC^BZb*l7vqX
zB1j0pZ1?l)>1Zpe;9*)4uQUl@GHE?5F4qB{&AvIQ!y!L~ABK02t-T4_TkkeS|M+Q^
zfhK->I^5v~jTMPoe{q5ZwOAH@J-Diek9;yeBcmE;sH8uq0&8NI1hrwh_bk};M4(E?
zvF+-Zq{f+e63hSLn4WO2K!{@}uu6LxVJflxvV-$*r!s4Bl#>U6Yo<pM)zE{?p9Gl8
z2?Wv+uVQQ|&~NMzUT`p{>m}5GnSE>kn(8~Hby+Z6x4iEPS-mvnMonM!5Qia=a^Tb&
zl~>D{D_m|7<f2({Ls-;38dttdxIe1cEy`uEz{Ta$;E|)OU(E}ybweMPWdOGi<p#C3
z%*;DC5UZ87yeW%w&?g^5P^JZItHL%(vb1y0EYPsBojwCmej#@iDxMQnxN^3dzc;Km
zyI5*NF0UCN`k?$=H}I56_&rU{^E5xvht-ROBM5zNs2HcY>gFjB{Ll<Uj5^GX@*T4Q
z(6x`AFZ7JYo*EzjHTpWime>`rr6*APGunVZq{f8@qC|l-?`o@cl+}MQgo2-mO!)x1
z^8G!VfM*BZEutk({QHClB;K7z#p1?T8`?S%5XHKLoMY-aJmn>E#wSC8gSDCk$U86_
z0#%MB&m?d+5^o6n$<If!VvM>YiVEvZhba0#3KP-fyjym&IADH4o#652sCH|erkG*j
zlihR0Gc0)J@!ee_*RZMA=9N5exnp8fpt-vFWeQ#l7TU+!D|~!%I;ZkV0x5}Y4wd;;
zO0uU^)}>`1(PjX}JES6`j~yQ}d*N~<c==p%;HqYbhofL%<Ar0)LzfcX^@%fNCoQLh
z#WtznDYyFEyO8B&<$_rmN2@?7q++Q37&OQ^LX>nYF|q1sCErQUT>o)kWAUD5wo9wa
z!wtrDlGbQi{!S-}-n>VvzDVo6P#7w7*6l8jW)S+@1rntAQM+~`T3anotwt79Qv_0A
zEjb#hT*JYniU~u9QT*FUFN$_&sw*5o9D3GAb-~!KXb!1X=t8!j-Qb;Br}YyljWVx8
zqQl(os@{wA%M(Yd6g!|sxKMPHa!Y6myNV~ITYI#zDAam*x31eal_*|lBX&J%g6_$a
z23Ch5cZOYHZ=LxOiUXnb^Pq1CUd?9_%YqSM>X2Cz!6UNng^`T8sCi5!X?F+x^2VC&
zlx|nDezB%n_$rvbGRIAt`c<wOC?^U8{XQ9YK^{!6u?Bow`7_HBwt3kJ!x3=vO`fl5
z{rWXaQj13bH-5Fti!C7un3P$<msBu(C0LEZ-rb86j|!9>l=LTn@|A^VvFfYq3ZD8q
zPlIFf*F%8ab||Uw2C@3$s_csti}c{6sX$A6pVSv-65L3vY{IXhJyM^8=ExmVy*GH&
z_1}hRfWng==->*Y#<8C^Rrl>&1d(oQ;6&(2J<a098B}Y*{(z(#1W9QaV}P|};PEC{
zhr8a5LoZDc^9+{87l1qA8lXafJMjZvW25CLqQjlThinPRJxL)IG)spM0{*fJfjAQX
z2<Qx6lIh0lg6ODCfFaf%m5DEg&{gpa8E0ZaaOuG~)LQ7_g9})U&gN%Y#$=vW?=&`}
zRe3|+Z9aDP*5!nyYG5HTH*1V<ssfWE%xg4F&3<2JKiV=>hdgQO0R@+ZFJ<qbs=p$K
z4xP%?9~DfqV#Uf!ufHD~gik)*{<xLzsGl<uBCBw6MP_w{=8B5Q9Npj2yHQoT6O6wp
zy}-ja228tmxu7@;40krJN)KpV;1_gAwlTS|Kp`C(-^jaC^4N&u&rrORW#@E454B<j
z`?GW$F-NNEV{mSejg?2~&KiHdmo<=^xvPr+cY>_mS6-f$K`WHL*Or`W+TM66G<%u_
zKVOfxWc0PW#-&ja$kTbl3+7LNNPt^RbS-m9^4d09({h=(Sry;KKWjJTwe{ge3z>Wd
z#!jj0J0*dMThP7bWX2Jt;%bG&YD>_io}T4i@*`nmgJdofP8Yv}UD{KQ2{JS|uF1dv
zm(f~>nHt`8m>P?4M%&_@m2}6NZ6HJH0i6juJ(wN_()}HTd;FG`g&pM)V&+zo41BK!
zA>H>usguX}MiGa_XJ*L6o=ZwgM$y(EOMs0g9(-`lRYzl#OEEysDmvVWYZ`RY991d`
zkosr@diI7y?h!G0LbIna$Sq;G#1hiK=)+a0_uY1F>B|GY@#Bsw?MYb1J*h>m8~pci
z9x(UjKzXtzuUjAO9#HsD7cQ4bBz?3x(!C5N%4zY22hbP|B}~qr=W27+BjwZ(2nSzC
zaEGTD6|Gv4eI+4+g*ypVBa1yqjj?g#%A{HujxRXJca`c6m2f?@BsZ{{BK6O_oT(Ys
z_7Z0u$AY9KK=Z!V(B~GFIui$+gZLsiuCY~-wnuQzI4G919FXyHThj0A;neHA?$*m!
zTn1|{v~Y2&e(r)e<Mp$z-AXoZ`1=+`OhKe7`rKU}yKGNaFnH;}^wYD@>iXs7d)>oB
zT&K?il)5Jlwnl76Hl5<1Uk#xEZ<dWWqld1Hxrb0wS@@=`penXLb9P716|R}{%uOaW
zq}#$I)XPw1OJPaQm>`B!#c^LMo2Do;3qGgQ(d@hi{~7~TMNV;4Hh^S2lpl<$@>uy?
z+?d4J5X<a9dd*X|!GWjKa<&Y|;zFfQ(e9~nrt+fF!BZg>T`ASAjVw<UO=VvB0InFD
ziUfVH(uhfuyn!_j{TsjBPal>ezl^5Q@adn)i^WY?o|4QiGJQCTW35;2QyhA}n+|no
zI`eY0CuuFaEo!%-3QpJboBVzsqr>}I^^#=T>u=+1k=U_A#c$gx1!L1rNg3)0NqiVI
zH89RmFHgK)&F~|zDwN803T-FsHuA!U;`1rMS2geMe~_Hg=GHPIAhnJ3eLu*?SmD_j
zL!Kr~yD~v#QTm}}h)83O7pNC@USy%EpLW^->>k?NKI<NAe8AF?mbNh%5FE%;L>+~^
zqd1qJOjS{bGpQl3whUpJv-JM{g`Fe7AydLl&w#}7#<9(;s6JxLysO2~<RUYCJtTK@
zX+!xP2`b72zrleg&aC|l+m)^R$?PhrFvNJqw$xtjSO7CAw1B|gJOG<|A1=@#jpcRO
z_NfgEn~y)gq=genz`OnyXR_D?BR0J?B=8{s18UouBc#Qzq#2i+f$4fEI<Q)29<XWX
zYmEP>ozp1dI00Plqyg06rv`+HoafJg?38QF5U)D!@HSW`etm`ZT5v@z$&u)U&8Ly8
z)Ax|1-Y-r*l-wV#No(w{yXawxR26A>+g6Vt3ozdA<Jyc@oADD}Hf1<Ml3Aeea99tN
zi;0cU#d57wf!)qTdKc$@4aj6jU0&entFzs;RN|35xAKv&(71+j^jD@V{>vA1j$vbM
z-NFo`Yia|sW$J-%D=CB%5o1S->y!In)TWx7qmg%PZ++)86Cic=cI$!_Um>G5Sdu)l
zHDV`#U{Letb~tl%OCOD?7aNO&;*4LBY^xzokEF~Vp8+){PUE)IKV1xdpf>Sa@DkJs
z(tF&I4B9}#olJl}pAB?N_33aEyr1R9(K&ohz@%g&oSjy?{&k+Ph;X4(1-CjP$;zZF
zzlFsgx?blc99J4rvb8=@Vf%MKUZ8R-;_+*Ywn7*WWLul_wNb&rsrGt(B7eyf|GIof
zSJ^~F7t2lUptk`|U-t`RWK{Hh46XfD=TSgFimua;g2TKmFrB;K>b0p#t<TH@q-j4!
zjX0Y>au{p~kIw-ahuS7gWm0VA0lUk2P+w&_DSv9wx{v~2r5i&^atjZ0aF-fsWSRsB
zi)ZC|@``G4l;cP+Tu!ah!wLsgfjMeRz>k9lwi=key(kRt6MiMo^mgq16iPhIWpi5U
zv-@0-7`&8LAWs!qL!aO6JS;=(T2LMB%KZ?T!;5p?V8olMk~j{r?^Ct1>`|qn&VYG*
zX8QfYSNY5Rmn69gZp>*c9q&vuQ7mWmc1R|%n86xMx?F3R_G~#hZY1-uWbtq|+)Yq|
z#a}aIb(OzXTuryY44T96$G0!Bj1pNm_~5qEgLg?Lq41YUP>*fVku48XVJu=+ypLl+
zvboW)U%#aR_Qzwj6*?){{#8nM%l5pq>>#D*gbKlKU>SF}*P4&K@T9gtzDGD3r8KUt
z1>l#~VlNiS;Y~)6!LKDsTW1BudPW^=_<7u`&7L7Q-W;=&O>K9ymu8gl-dwIZeDs_)
z5J(jmPf?3I;bpPzEqnj$*}ffnO|~BVclV~P?zNRz7M-{_t-9-T?K*+EvFt&Cm4S)1
z_ur`z$gh?#-VcNkG8R!L^4i|(Rcm&~$$9WKPLg7`Z!cD!M$Nn0@$#tjrS8@F_({nI
z)mzIk+e6sl3^`7EHHFl~`8`-D`R?IUL2Si`^k6E#{LoPzVZHqmWqjV%f&s=w^u5S#
zwqfG+7v<cK=G%da2)hoYK$00e)81EP$HwiaBHb!U^2`Hs;F0TPia8@S{rBX!a>><h
z_i?6-ObfY$Jx#N@)A9w)sBF>H?*(X&_zN{?F*?BDOk_P$gViZiE<NGn`#hZ8iCkD4
zHGI$UV(wo5s09YZz`deV+hfNI12*C&5mcLL6F_}Y4>tS|Q<ky?vsY{5C$`RS%h6vO
zodq9jO3W8R?*F-ubY*=Bg=Mritl_Pw)MqOEw_t}CQoC1`6NEROuxkgDHRK7t=g+%Y
zXU*`6fN~<>m85q%SheC)@J{B&H`_gA>yHKuW{9X765E$qh_@*<O@!d#PMAxW4Br|u
zc=_nQc&azwY9yZ{ci?*_qdvCaz@{DK??Mv*M_3Ny$WiETlC?+X?ooSws-(_y$;@}f
z4nBNjwiwGcH<$nxf%x*8&KQ0p?X25-N58&#p6AkyJ<;L@Go6$)L}bWJnbIRi8P{^e
zJK_K{%Xy8mdl@0!FJR2R)m0x%1M3uq-A`kc7$16B$suJxmO8#T%^gyInhumn*nz#V
z9pth&K4l$}l9q2k*KVC1TN@UAhwzvyu07ylEuK7vG|q~RD0W%Ws$7)!m3>lYR?=C9
za0{ZV^MHLIqXl@U)o7~2Kl3D`KQM`vB(W8mc=tjO^$RbG^~Y1&k=HOG8=GmYMAk#u
za8=D{YETh<!26;XPN21A9CTkY;1MWja7S#tv=+31T~F`#WOk!<^?8?UMMT~@wg(x1
z5l`Mw-&c=ce`4sigt2DwqQ<28bt{@0$rsYy4{$7{?(194!;eb#g<h>tMwGX_fvz{l
zjU_<SUplvtQ?<p1SJi;y8V}M$V1iz$^Ss`Gn>et{MmgBFv%zra?pFrNFeDk7CS*HO
zkjeR<OY|1xyXF<F7}VYzmFGKC41FL_fBwz4{c-`mX3xWWi0`QU)L^a@ZdmBz(glP?
zU-Ldiwj%I-`XgZ%1q8Y0z8j{Q^Z812wGYpCTnTllS;$e=5>qnU`u+T$(0Cu)a<KRF
zf)iMPdLECRa}sLq^||6dxM<wK;Mgz{no>%D;n#Q^9Uga{%&9eK1Qx#<*G6sg8)GNb
zZnQ(#9{5Zi&7XU0X$2%-91bkaF?{uo+AvYiSPq35iv_mgueW)vq(z?cD}0zaC7rMY
zI2h{i#z6jD>+QA`e3tc0yI2O+o)J)Q;U(tBF7%={{?i@=S`U9=b(2{9;XXU=2)1r9
zQSxK=9v3L}V-*5rUbiFRyq2y@D&uyi8Kil%`Rgr%jMGBT{^hY1kB(f|s=@%=cDWiv
z(!kT8QqTF8WqRJNPBur{g9NAq){crtxhKLSVI9dMqzCg7Z}B(h+vm)!rP2A9Lasi;
z5f7-{95%V=8IcLxdyZ>kcyn9`RN%O9{xEep6eo})_ZI2}`Tb&^4w5g`I*0FeNb@dO
zOFlkx)g21Z8ah2-x!5qbC~}Ee^_tY_FMI;ea=_va1MjEJTNh}KK~uE2mjtOun0MQ)
z4HlLM4W9O<*(XvysvM^DL?8`lZSV<ja_^^^Li=KLQ?6#s*8SZuiW`AzO5uek!m~za
zIbH1_WzVMa5#<HQyymMtAJc%1HgeQmLBvD7a8Y#OtE&eMkmbezTDseeJsIee7lESA
z_7p=~E0?5mK`WP9qOjs3urj_a@(q_aR~z>WwcMMYu-o#1>LEfC@=p|?B~-9py}may
zwz0pRq`ar3zwx@rI{6{W&eazV{(k2Sjvy8lCGy?1FP|R^pRFKQz%F}c-wZu<ju^ou
z-9Th<oD3cMOd22e5{|DT&brH!8tgd(pG}Wd0b-0|iKjJ0oZWQ-5zA^GDK0HRGsg%+
zJvq$s?j&m^XGqf^(^GMgu@AP~t-axGp2{7=B-c~vdkIVx9V1o5dMr}Dj#BXM6L-^6
zX_r90gzXJN{(h`*_FU&2E1JM#uv@<)!n|)58jpXsHJ^ho^^I*4`KlDRE*b|^kAg_w
zv;F}=g_rkVMKKSxX2?J(GIXvK>(4qKY(42(kDMS|kyd?UUx1kIxxEx?OjfAL+fBD&
zF$=R@X7aRvSzG-W)H&T_ZMBF^Qv5-GYd1%D5d~jI#ye4ksI3U&rnEh_kLz3WIwyF~
z433-Jo<(eCUgT7Iu}=UaeVb*qw{}8G)Ig+m%Go?7(PEBwY07%iB>uakZDQzEoXA11
z#lkwqqLo2-?{+?8ajAEFU^ddUoKeL!4@6kLsV1(W#5?R==8f~(bXGmr_3pdFni^e!
zIuAnx2SZ_WdqIM?r^xdb>~NaHheDlv2f5v%JmIb-(jSuO(Y~_@@jGXs*(JzfV7ZnU
zT)M37r`<PBu?k?(o>&s%Fr9O5H)p9bz(gxaTL%w7-av^4It7ZM_!|<ne;1w`Jec=7
z(92Dv?-!Oa7M<9_N8?N*pBmKl`3SQ)Ap~-IJoC{1{oE7mJQxy}Jc8jOhaxc@Rh~C!
z#E&K(L3Ve}hB0~KeGm7~znb<dE1kO`scRh2GU?_xcU_Wgrwz|r!mLdq5W0AU@a`_!
zflgbBz&-QGYbVy9$V>F*m9<v9$C_c*dD{hM28lV#Vr?DOWNXXNJX$-493lot5kiJe
z@H*kF9|<3>i4YGVbS{p)yjIXi3=L})2(e3A1B&4YVc;1s#8nb}t-yz%GjTw^rx;{U
zRXW`t$KSjDT2_w*>34m!);7hq*cd1yE<~l4Rc_F0a<wlo3Ehbep=xj;LG09I)n2)N
z<ogk3p88CnC)=w|Ak?bYe$Tgn-=oowGY&PGzlP7UQ6>L0?zr8#bpzJMugVDVaw|hO
z4>hlM8wBa`#9wXLf`~5F63yRt-L(eEiXiuI+D*5=szz83C4Rjbr1{t&iMc<RzTou&
zUA@9{-Tl4Y)mBmF+E4el6Ud^jHtiaD7fxHa<3xDf1gA{dwk+9FV?4XHU;@6&H|QOD
zq@%$@Ik(iS`e;9Iv|~Qf?FA*)exbIBIXQ<tk^WdJM&tQ*fLiCvE_%=Qm{F~(-o_DC
z+Jw-yFNCSTdhanrUu2EQ6XMkRnA5A(<l%aw<z|jfUpBjA1W?;R5p(f63u*m<3HfVi
zq?Yn%xBJ4Lk+(^}K_bkwBKZ3Y*L%wI+{k@xE;>VcUj58z)ehMchs)AeL=-I>Mrv0>
z1H3NuU<XOH3tAaa%6GmKzK{ET!}Bmr=ec<El5##)QV2&5=Z3xLZ-~$$L8M}57&xhq
zUX&slOl2$s^;xpkdSHUZUMfdF2(=t~#)`7m@)l|1!SlMuwZoM$!;w5O4rJ7Ib6v^8
z>Fp~OdUXCX&uv&&RD>hzo0+&I)Cgl@KeEpt8@=qOtoha{OJexR(yu@9sE%WwH3mjE
zDI+5<HyXX%R{9J7bNNp(ao8@a!jG)_wt|y-kWd()H1KA+7!7;hV&dUUquNuD`;{ew
zn^UbAbq^AnX|=q|nBS<(avB*^Ai!t3Oyn59pRDS(eWWbVb&wW1Xc?~>8MwjT_zIRW
z|2@K{3b$a$a+HJB=X(63{KIKF{z{z$jnuap1O!JTFUz4w&wB5LR+p-UvrCcA(r+3v
zxG}xV7SpI%LSBJAOxrktTVad1AF&N`WsxT9!*SINaBk~)AyO?|(;*sX6W|(rJ=;BC
z)uw&=Fqi>j=qa&wSwm_99>j^|nGjlh%BkGq5k-R4bVdGjZv(??@pYDF<nFRb@Mc}d
zJ9Ve=e#&j4@IJ~T9opk7do?1o=Z)Lp6LHnC`4d~nDS2F%sEZb`NtDXKnxp4Ft86KY
zE9s<U7cQPqX&%djXUg*fOR%3($MQ#@2Kr?Ly6=fjTNU1O?Iz(#7lIuJv-{M&<_4jV
zkkqFmR_jse!h$*NWvfZdxC3PP&y_rm{X!d4k3J14?ZO1uisWnHm4PSxV6+CmX|uOd
z8|U^kY)pxAc%F6+y-e_<HPZq%cD=w_JRStuB*M(75(hVgEnr`L_~fEefdhq==LLI`
zb>xQIfIes6ug?FLdtreGu@S#Pa^7|2AV;vFo<2y#8TKsf!`dY88EO37$EHao{Y!nb
zN=hhtFyrCK3jtV5Q~2+Tsa-RPXu($18gX%C<Om-q^B$J)!S)>LyPea-m_6?7qnJ{R
z$L3R1R-nKUF$?WkkJk?<ZufLqra}H7VMP;ZM)uJpu_*;CSk!Gc+uHmZMaf1b+-3Yt
z<yVKew~w<fNxtgDRwhU=zZ29@lh>_>@v$|cZz<|r%QudNZSW{7Il(@bSW`>pj2jp|
zo&E6j6{E%rebTVKIn5~Mo_V8B&PpZey)+;TS%{i6QEyDhZtf#UQ|~|}B{U@_F;tWA
zQxIu~EJWpd+gn``&*4G+$5x@Ynu>M*3`aN9WI)}(`zXoBsMumSiS#s_aJ*aFebXQZ
zx54!XalTq@BVN5(iMu>vQvnUWeKd2PD#<WpRl&G)!!~HJ%YX$vNEGwNVhZ~sI_{HO
z=K=WGQ~92XM<ib(;(c;_K5))azG=B(LBl+x8(;oK`?v#j_$nf5QKuAP7W4ZiP?%Lc
z?n?$FycVgb{I8$?`Wq4~GK;(;=QZk$-iiBGWqt7!&)Y3EYe4U7q?2cq1kqFNoddBD
z{Qf;dQ!Ldc^~Ki@iBxk8-{LDZs5|5Hho*aaJTjhrjpQ&;K0Ng3n2i*@U02TBVcv~u
za5ER<jZ(b<Q}R_XyzwFVQ3<|>Ky#s82|g=|5HWwo%)YolR-ziplRRjC<^1i8mi1Xl
z<m_i6c$3*W%?v4sKag(oKG=xFfRxL4istB3SfvT#x^S~2FkV3tuWwX({AxCr9kUn1
zgeGlD>I5~)OZey8fzE6T0#@nfa}|pAE}P1#ND@ED^>e)GdsvGuASUqm=S$rkPuynJ
zP4ri3w~xZ%1+kYqY|ySR1ui@Wc6-anV^kAMVmgw0cVKzF1S_m^s(UWmX@<W)*nVL(
zYOkc<4_!Y!8!7~K@@yQMe*n1|D4Y%T>rp$})f{S8oO<N^GXqmK<MBCYgGE_<NsJMD
z7JQYj3^#JWtsfSmhLQ!IlRa|kd~pOMkqwTQe@W6wv*Vf*H=At{@ZQsR=n5(}A=j3r
zzoo?AKG7_I;&6Q4eL+I<c7f$;HfTZaz)9{_CPD&45~iBGs?R%Fbst+ySj%qtZ+&Lu
z(7T-87yuj0QP8E`L#D(GxR?7WeS80%BBdH3EiGIX6S!bi6C3L#NC*3Kp4eYc`eXo|
z<g&{t>|s!ReiY`UD~#OCyN_kR<a@Gy21*%-*R^^!@#Wd#PQ>&UEL~UYMQn=4palw7
zeD{K$@x10STMqCZndwdt<6*fwnZCF8($n!$38VWtuqEwZRUNT_nhC!?bo9_ya%R)%
zP3EIPv8u>1NQE86_>7P*WVg4_8A3Zg0rrTP937=Ko1L6Du(<3@_GvqPR?M+kfC*K1
z4}PHy`?a$EcI_pOy!wcj9&8NK(H^8^F;C<+ajDw=7{Q32%?J%lliYqV@u7x1uKVNG
z6Be4Q`8ZuT*)34}x7dJ;y%qOrPJh12M^>7T0gY1}#E5H5^ERnkYLM8jN{Yj?d|bk^
zU^3vH7Qji<vA0kfP<wKltR!_(>R`}K-8_c-w*lRDn+!TUqs5vY&iQG8LZ>iFvkP|j
z=UgIanR-A%LV{Yn<uJHShlq|^U`i4MJ%j?_s%U3@p;888)W{6@ocPj%KE%*<%KBcv
zN!icZ-Zyc(uZ+CYkipdW8sL$;Q$BX^4{7+s41K)6R%(0Qy-xgmm>p6SfXf*@T+vbm
zvTi>2p2~kmh~%=oo8%sg!*5_Nc)s#HY$}9&<zl11i+6X1i(^5!>TRw5OSN%q85<3~
zi6XA-v5>z>*{5gFD#P&1;C?pfs+?SM)IQ8v=StjJ^4Vf<0iHsJtmCQrGsSm0!_*oE
zy`t^%qAE)wP~3j(u+=Gat=Fx)HOsY*bL>odE)AId>8Pz%v`=h(Z25~_S#+vV`zJ_Q
zJJ{)-*v|A~0u4qmrpit*w;wunqm=%53_I*8LUnm?x5twEH)3XuGNdY$hH;OKXGtbG
zdd=uLif=?^4bJNxBmeF0m5_}wr+%>K+<_RA|7|BwZ#$tP(1cRaMiVo!t55R&X6`|a
zaLw-65o>nK$@qvn9P}XZMBKF#9Od*+VFXP}_~vifUul$@2fEqK!yo0+GY%RK1)$+C
z_O@^<cxUZD={)|hFP{6=84C;2!%{TxESJpkjf;hq=BVKe>A2EQXuiXm4@moSKzz)y
z3B94s!%(`=VngA51y;I!8px2g#Y463GG>@quM}VO@G~Y<6Y0qEh2z!qnhrcY_$Z>n
z5gS;#+x(O*1DF^<b``3T_MPk}=7$Fot?CEQfU0g59ADy9`A-vKk{5zrOgR%|8@+yu
zS07GYSW?fOZ!^tp@lu}nORqLhe$u9=$o=tjROEff1uhmO=3g-NS4U++?lsdVehNgf
zLw%{57jiFV@_Xwred+Om3N9Nd{twB{vw}S{86jy_-VZe`cBBVeXeH65Hz?`jV^)=6
z|0VmNcqDND>ylXhQ}0)JR`cd2h0`WNEjDX8DQ_H%A-*2A?kFxgj8=({9}CWKatEGR
z8%a7_7`a!QztZG<_t|OS^M}=c=~p@(k~&<~ddpCzbzX-zaWn7Pxq!a$JHO1ILMmd1
zy{HAXAtj8}d{P=r2j^^QX5rHYti9gwsUZF2z=Nm0M}|JgE=y^i7YXOEc#~IFv!&gI
zEy)wdBPC)h+p+O0d2010IlQrh`k=WA11}QB@}o)E@$ssE0o)*w&HIti{Be9a*h$Ei
zL!m`~p~(a2Y!Tn#D(f@SQ1XqY2v)!aAfM4pE5n(rY(iGO_Y>j}fl2P7`|DqlZ`_4%
zz{(+@alMILt|X6!@6OFPe}{G87|^+^7iN?wJ0uKgt9UDbCH8RO(aZgLmk1lyR}_PY
zjuPn|v9%8^x)_WpbIBZv86S|If^^0{7qtBa3;)*U?Y|-qp_TNd+M+4O;7O9g7Tm8^
zl>}huNdMv-9Ut^31ny27SnnqXi~|l#v~VZzbYK^0P`C`koQQv+N%&R*dkvxLPbK+R
zr2Fe+UOg<ZUL-XK5u+^Mzq}`m%^U`OkMPcJiJ-%t6;220;a^G8#;W7ct$t_$%3gt9
z<-RF@2j`5Y=3Dlckp060K$H(!Sax!NQmdJY{*;k9T%lm{Hhzm8W_4N(xSs^3lY~a2
zz$q$5Sj2Dc-+ur~#+fW@K{sR?++7bh0EPS9=5GAy<CG*~Gzd!Rh8J-utmsO8yC8K|
zitz$($~XP5-za3ko_W3s({s6!T%JI0PDC=kJcY2H)(5k&0tJ`2rkjPcfHzHsON_cn
zhO2RCvu_Zoo+O6EX<YP38G}d(V<`WQsQ<zkkzQ<lZqpIynFT#<WjKsmbhU>AVDB#=
z>pQpZ%m_GRZh;tGnJtlJJCKgrqkg@S`tGx}x>&`J{YTXNXB_-L2II)-8=R?$?h<bA
zYdJT!VdT2yzG2qq(~CEYRrn}3l97ryIu)VA*-YT;#}&FDk)y6-BX6Dj$P})h3ix-|
z>+U!tQUqAH#tPApm7ts>J!5XSq4AzM$if!=F)arCaG9pX6jO@xxYdTv+v?T;CSm|;
z?MaGL1Y6d^GM)cU|4;0sRColvGjFZb++;qb?x5d|;zb^%2O?)K<6b#M@^;+VlFOOH
zuE)7U2S#YmME(FLs>14y5(Mw$mBj5%+z%}*#TRr8=D>IUR^+2UKZy#q5L(MYA3y$@
zYqc7$8STuf?d@Kd%Zk*tB}3Wyc*AcQ3iefg0qMpnK610W=bb!$A#%mHJVDsJOiDlv
z$CWRY!4AbWWnxdEOBi<YFCy|cV;m$xO4U*fStOZsLKQ{Bi~F`|=!x&0Oq&!trDEog
za(TMkjj3W%9vHX5{<_GNZ(W#H?;u{jw0Cv2SB>x}B{lw?NGBJ`>`Aq1F-MI%BALS_
zA?;4QkB?4pw>=-%&Ee(5HZn?gM*bllF$XQi+Qw13#PS55EJIPPW#XjN_VzNs0em0Q
z&+)IzT8i1e^{M}>rMqJ-K-iEs_v+h|dZ6o{Zux(4-0M!vn6c9|;dfaXnnu5){MJ1v
zhQiGh;02M&(sQuFdn@-~S?2g&SSa>LgBrcN$;v&*H|7=Si~-P!v)|3P{38=E!+mik
zM|tRD@8DX*8xcU4`Fn6DEPxGFhLK1_?lyjGp)wpDWB2S%zWzZr(~Geqq3eskH1sk5
z#0t(Q4}Q4D7CB|eDFsYmlxN!Kf5z7Xtn?qkpx5}6@(z|%Z@d3$a6LmCn8Ah0jQR(7
z`)N5f<<lA9oZZ~q9zKsak5GGBcplrBGj_im2bAX&$wzflna7Un(Qqc6ZBOTnr?@5?
zL-&VM0iBF9%|I>pg=dcH;@@w=@FqQzf|s56UNxK%nxs8ZnX+KeJ)xHIc!M)(UTT(m
zN438T+K=N;^w2x&sG9nhm!THOk_vAYtKKsLz7wgMKwPobo3-r-d_%%$ikjV*h<g~;
z-MGWq+p&N{wi2KYZ;WWoPw02)KV#SFuvurU-n8dR5i3gQHzWTB01apD1MNC%!q^gi
z-5}#1%~JdyG;5#m{?;)#p<<dlUJu`vA_`TggH>VwN%wysGsqFyQW5Ky;R4kl0;U>&
z3i~NgW|~?dn8EL#3GvZZDebNn0HByE@9*JGKHHk$W+vv&Gm+ub`K<%_G?9uRQWI<z
zl%4$6STR(3uxtjND*PXNlt%_t&|-ZL&N*ubchbOXKr}c+P*-7|r#du6Lgjq<C5{|_
zzFa36t9tQ&cgnJw14y96eh<wbCbzk>@pyf3Rjk30L52GFpnr}2r_7Nc^#r&^EIKCM
zja@|k)`lZ#&GAU%O*&<zs>=Ma`#s|6e0Y-yW1rc6-||!n(3g`7QjdW>ue5C0=a(`;
zfNoN&Brt^BJ>fsZZ1t%K@(xMs=Ht6&K*_i28;f#~tUi(dz7XX=@>ouTaN-WI{{}O^
zba;ya2!^*`4HnG)L<t2BU>wDSdzF-GKn19fS5c9Fh7U&Iw?>@>9lsn;XxQ7xKET!2
zFxXlKq?3tTVEAXH$AislvHlUb=7~<1_x1Nvx!Di6G*q;8RB$YmO{loDElDR?K8s;n
zIL~3$mg7HS_IQ?!e?PUN_lpVu$~Mz&o&S7e0V7b8>iV+(+B&lx`x|69uVsBk4yR;X
zaitTsShH6$rZHz(wZQh{`|K9^(ve>r{15tmQoU7U2JrQDCub-1k9g}gfcxio>;JLm
z0QfUM!ni*-TRYHui<y^Lv3u@-<3JTIL8k-Fknl%o2?E@Ps=VJ8QbZe84A7owM`TJk
zXYJ35+M8Se(0^hKHI&!WwwEQk6RLlceBfWu6#&=uNZLWwsCE>(bII(-5|`sFC$8Qa
zE4r+=2pgL8UikVJbOn48mjA0$e%n9^AxTumc$`Vmu<Tl9{N-$o5{8&R4(8>PbUx%A
zu+X~x4WQ?AHpw%cLzMZRcI?PGq(7$DnAHBeQ&lsB01<+vL8Ov7jwj=LbqU%*aq<EA
z5C16jkyKE%n8R4f^7rp3(-RD+Dz8BSF>Nn&z4fPZ8Tvmu%n67p-N1BR9Usl_TQ+Z%
zR=3&c?NlSA+;P)f4+#q}rmq>4CH^P)c?wFQ;z{|{OyvDr$K9SH_5QY|OB4`7756;^
z|B-yFDZbi48>Anxf%)gZ0KZ2Bz!Fs3tqTeO4%|@`&VT9jFR1u=&_bZ0Ux52dj&Pzx
zfc-7Cn5-{xKWpm(jwXr68q@E={TmrO#nKs2b`B5im0!GgkzE5=YvVU=VUad~Wo;Co
z{K)<>k-@0xsHiOcf9Irv&Q$*iWI=DwrxsCm1c)i1lWaItn+sgqPcl3DsR|pYV<7D7
z{EY&*D#$wmAQkq1>(uXm6+)3dDE~0q`DQ<?Zt?#!tNVX#%=?Yy@g{-wgSTMEiR>1Z
zxjNO}e`n1vfPcUM#Q;uvoM)rS^$PPCr}LMnVuF&9UIVtSVsk(mc_-OzBL998XTIr{
zgzdF*0W!70AGt09TSkdl4T!E@h2Dta+l@HRz2hG<+BLlA<8K+1OS_Na=>NM@HDviM
zCLLfuK<7Ht6K8;!WYAsp<DbVu4E<SAKk%3h;fVFbRn%=|mgNV=yjOg*`@^x*uU-DH
zPL<6K<+Y&8$|3JOV$Uv58-LGjp)EY-xSp;2Dn|UbGxF~pdGPeX9LZ!W&+SIUpUx4;
z6VOn8y@#@oNHz51^t1Dy{uZHs3zj3PL0jtjrU@~bx9b@{KkuNm!XHH6hI>F=PA$v3
zNc^8%uR6JOJO7Iu<1Bb5LQo&eonk3G0K8*-L`SOuNxaF*(k~Coe>MK^2di<nx*nIS
z^h*eO<N#eC@Z9>Q2<W=wal!r{d#?|8IW<JT{5uf}q`yIxSl`EnSH1xBao*F>_;aBC
z4lAz@y1t=dHabRC5Z9n-1s7Yn$++a>w+<$W2d_v!3<dcx|M1qll|GV$+&XYd)FFR3
zWpDua7dF1o9&tF62NTw48xQ{16Ce*zb|~{;gYWp_bn90+<&~8)=i^LHVJmyyG4&UB
z?+)k$ZzB-hlzj1@^idf2VKHfDn8KLAp|vtcv|8nWht-H!>~k2GrSx+m@9yV+=O!H&
zv>hXu1F(}CI6oJ0_(Ss!I|qByyj2oH1XG*kfOVb!;#BMLZRMahxPTBp=o*^^=7G$=
z6v$BDiN)O@e=&3lGcN*~zwj+@yiaL2N+0otb7^@HgI;o&pSe`b$H^7Ae%$}+RLrik
zB%HIlp^|>!JNlGsV>+4gu4A;g|8Z_8?$+NY{>05p1Z-r)RI$G!G{_I@SD?CJ3U^<L
z@G~kD;A^6~tTFs?Af4iX%o^sS_5L-R3(Orp|0}?hZ8iYE24<rD;i@db+mcoV(Z2@&
z?+2Lj>1<43tv+aelE{cVT|`{H4T9aCXt*N**8AI*3iTiP2(m=#f#%mT3}L<NrUEG-
zB5HxU!D)<e&cSNn5Zb#d|G^0?pwRD}INl`IQ~F;Xy@(JHNu^+fP$vQrCADTL!>>U6
zH&*;q=a-@_4B+kd0M`|GjpXdS-$RQe2Apt*l^3)_QPB1BD&@$3c0LGiyH7<%`rGJp
z0KYv#t%#0#4_zPqO4<0&4#(_mPl>rxeu-KnF-YMy=&q$_WRy+XdILl*#X#gDn3IzI
zBwZuRJxBK`F<pw%|KLbHfKEW{C~6t{gufeCFu8JW9PZILsUaT2YAt{BEvrvu=itgX
z<i9$FTpY*%CS)trSJFpdi(Lm+<YdcrRT=IC?mslL=Z%OHP)opTp!F*B|6$;myDP{H
zZZF)I&Hca-{O>UQOWuM+Xl@Ofsf=3pZ&UMt^wxdbUJ~=H)(*;D(NMYsma3FKe{B2x
zvwsr_Am02PKFM!vKm0vfV0eH_rlD`@UvtuBHBSJPO3eKadHgLwz{<#DAfq!IfW>;O
zyl+1*{O?&9Ch$?e^2cAX-b)$)m`#M7F3bQMXY8YstL2OQOKJY2D_u`PMK6AxRd{-9
z+M=A9`(**sR@nBi_>ZLQ5b;Icp*da>YbafBiGV7J`X$+jn1F%4C!RzPw3VxF_Uw=5
z9f*GmTD}1Df|VDl);~!|#{;MZN<_oSOJG@K_ame9pJB=F))8PSlTo{S0KmxKYT*(w
z1qR&uN=SJ{cSX4);_Xb@Clqs-Y-O4M(c`!essXkcTx4F`10mFpCIA1zH2i}BxwRI~
zk1~Bc1aT(ofmJ&y`d%l%VR>1Zukf4wzX<==K_u?&!!?kydVbwPeFwEoIt0`6e=XO4
zL&|&o>xTb4c=!OWY8rIMxWD`cz*tcf<gFh@!8s3>stx_;kA9v}VEUV6mj*rL{iP-{
z@qY<+oNeenZjqajy1tf{<$`?CADs_iNr?AuDLJLU1@~^=^_AbP+alT+yGoC;qp75b
zt$fG7UoQ0Bf#Ji*p<DN*q3h=&?$iIDr|15cfx^l6@>5b@Rotn|z*tEX>IwyHcZez$
zZ0o$jXk}sMo~EMO#OKPZiLzbVNhe&1NL5iBT%6K2KN&T5NXF?sq{f?Oo4usCq@+{b
zd6Z4wtfi`24$~%jmw8Wzp;C5Nh)2m0L+Z0QDQfVX*;h9j*gmN^#p`XkpsxXx>4<dD
zA-@0orErLeD6StTfK$=&`<6&OL^@E4-c$TmPZgMY#21Mnh5thn>4>ZB&=0{zeJ!tD
zD}WcG8=G<RN{C=7z~TUzAy|mMzSxscBBNU9&@%cKvOmxM@#~3x0fwa&;@|r6{I)O8
zMesW<qgz-)DuGvQ{Bbb11Fa1Oo>v>1o0Idq%NbY-(EN{us4qTZPg2)iRPO(7{-0<g
zDuVQ{T}VS*4QiQp+m9#VwPO%o`_zDZ4~iO$a?f#4>28?+I3T*0K;LzG7{ogj|00G0
zYzQwjFD2Lo)FM;}>s)G0l&w%#lQ*W6k2C!E&rJH!NfBuY5u}|c%_B|gJL;zEWA<Qy
ziijrRlr_#%3SQ}Na4#pO{WGI~4piX`^8tudt13(-h(YV@wQ}$eCjIDr`Xogp6|@#7
zI8=upQ9b=|bfl%rhv@h9Uo8FmSkqrpVWL6TX99$+$1>Z*y=6IY9povN{%LbOcerLy
zJdM|ZCVk56^g5Dgay9}VeS^@t!4sM?)3Pn5F@0G=C4TXpDhxjl*M<21OiX%2+l27~
z(5@~k3mr)VyK702_Wb4ic0-8icxAW#`SixQ-M7BQ#;RY+FhS<OW|6khJ<8~RIt^Tb
zG!7NGa_M-W#+G=cl1r7K&RYOinojvjmZuzd+0>g(+`LH|ZyQGaKWdmv6=o*<O{%F_
zx~F3T_=NJMyEN<nihSx6m_aQpFW^44{bs*Nh8VzUF;&t4S%S|kEZWBzW4|x{lODr?
z?jh~Ovp=FBD-r^uK>G|T-%*mz{8j?TUJ3XeGJr^-)QME@uKb;ezc@>mgGd0a<)x25
zKK<E{{!Y&VFm9z?F9ioV_-mJd{IekZkUAh8@h7YQ@eAZ9E&|;5@FKCAQo$1|p5LX|
za;ECpSG~9lm5)kF&E8(-GgU>=%msMlAvS{+y*QwqTOIWXld(hvhFq#z0;hJTGXMCX
zE5tnbDsp5&K(RpD;)_Mq^HB@lmxGv-4@>%E7-=VTJ&L|<j&~<-7`{l11~dYP4c1;<
zA0pr?7m>`P+7Xh!EExg-#XHKJgFBgpl6b%md*=TMX=v}2%axLnOmh7dw}rsM8%y9#
z+r<EybYen{UmiTqVS{QH`=XddfTtV!;SOYQlwi+DBNXp12UIw#FiE7FQQCeTp&;b4
zj4#x#9n<rTk0+>IiQxaBQ}3))=V<=j>qpO{0!1PacXS?rF&IHQg<2^1;mesa#P5M^
zOzE0DcUmrj)x!-dpDaYxeGr%MX1TDvTkc$)F%g{+E%P!gw3zZ8u1-#L^uhhbbDAoX
z+YX(FwTvp~u08BRHYE{RC!n1C-fsZaFUYc5uAZ23hYkvv0Kx0U)td2iv3l;qR$0$6
zKpIa|fui=EY2N3DYil3`vPdAQo`HKwO5!;VC@L7}swKc1!f**ru&kt@1^pf2KvyE^
za+^wrtvlg|gJH*Q4u>Pw;jd1e(vO~~v(@MNC~r=bRk?8szx*Nq@bbZTETH1f>#!^5
zTBhlS1!2i(Mm0&69i`{q6P&#eV%G1}<0QXYjtU1PTir}wEl4S?g1Vj>Kp4>M6aA&O
zfWuY<@~7_Zb;6?zs`)+dp9|3+930$uxMu<cpuP_XtgK06pz14XaJ!#0)h|&F5g|Z~
ztnl`l)^t0m{|wE!^=N)yoR^v1Q{SStOh?De*gbVK`&L|tVNXZ}>7R5D#w`^nUmNFR
zsY;UZHD}Du6_jF6daHc<g;hOzuw_(Wa|!x(9R=Gddw$*hWc%;Z*8RE)n?P;aR%NC<
z_I2l`R@wN==k{7Beln~iFV_bWsK~v~GD!HH2gGv=nYF5-Tm2DaRa8`Xq|X)@gAQn`
zOVR3EYBAZWjal=)*+YeZ-_!kQ#sHrOpE9**CtE&E$g69v)}Q)o6S)#hQh9<c9&USh
z9Lj?=ckHD;U4;Ytk1aC#8`zPFVRsBqM|`RQ9x9R7VI20Vma)a+rOX}b+25X_1CjFD
zU%a7I_yAZ2!PW8e!}&y1=z0@JLN~C>liFMtqt@XGN>wu47kAJ>$AY+zvh%scxdb0y
zh4%+w{@)@osJQuZkDrj+IystNxhtrHr_ud53V1d7l+SP<(N{0-5?pfu;svtC&6&${
zw;w29cIWc^Z=t-~z&rS=ix%f6>Yl^>Mj3R;jfkE2P|OLxiL{CbZJgnBJ^eio1lToH
z#2u0MVG2p#(}BYOkHIAhXgjm{*w3V}=%Nu}$6AV!&~^|meOq$ixezV0!6?R`KPvq6
zEjmd+BCyVEtYvu50lFPxg4zQy`p*tlj1O0Pm8uF?5`AH(e4$`gqhwj$jb%H44+JsR
zK7RlNAJ7i~3BR+kvh(J{JHPwLfUAPlqYPV<G|GZ=c}^oox7#Q%IQzcF!<71{`q0*|
zKon>nnH%~s#Z9MQ(c$&gU<8I`TFp(<Kq<Tz;e6@ZVz6?V{wc+j{a0~Gsa&LkZOvX)
z(yjAL?<E$ygXkea^zzp^H^^U;6a=tXrm_t^rV+>|StGO%?=O2?>*pca3FG`wsvqOR
zA%MP;?-r3QGXy+67O>1n7%s^0waBYi`JRWpw1?Xz%|uof^;Iszf~QY&<x|@1r4_;!
zM=}BPC2|i|413Z93oA5uKAmV%S3G2+5___<T(?psM5G3Bzk74Zh}QV?35@->9=j(D
z2Xo^O7>H=PWOw`6YWgn1^^<4OAIixm^D$evWJ~{loV`_4TmSO~TD%l3w78ZarMLxm
zE!tAt-7U0ukmBwdoC2jqf;$9vrx4ts6!+ja-~YP5r~7ta&&pXV>zvP-*|TTQ9{YbA
zF-mMT@llaUX_2Xn4cQjXER7d)m;Z&A8$!BBBh1BlDxuL1;_M7}>maGpH@O!+E5~2G
zJ*GcqLcLlQg$`SG1JNr+HqgwZ+uHA7Ooq4Ah-6dGJ8r#3F;%rZ(VdA7zJF}wYat=$
zQ)H0G#w*#^gRfB!|J_TO5>PkNB@;Ln1>w=c?WIK(p#PNr2PY*f>0l>cRrCnqpAw1T
zU2_@j-``$)8GB!xw?HjkSNeHJCIXLKtymAw+IEIV4!-xAU+j+YnD@S1sIiKkYA9ht
zKC+~}Twk?+^HnZU)c>Cwy5r=U&UYB)SL=FG%b)`nX&<H!63{)My(uvKFqO0BiK_uY
zetKg+)-TG1FSlQyJPE2U`06@8S{9TqLpg)3g#Go!n_8m{XA4VG#+e(wqkG&{6Q2i2
zR3>?rI0oLD#eAky9ZZ{5M*8ur!qx1x-xXB;cetX2Q4dcjuc=b@%<TO(lc@27#-zKj
zu6@O6;kSI3Z@#g(Yxz5iXH@&78=eiBcLh5B>#-!%(H4IMGYeIj@gMRJLg9(OGabws
zU!mBW5}*V<uG*&)IHqH}T_KyzS-~X#Z)C|Qq=)gNRuLRZ19GGOi{o-HQIed0>r|xy
zSG{E7XvFN_Xn9rn-JYMv2Rz8IYM1MVd@L$tMV>8V`h0W`V|oMJX)U}%bB#!Sbh&jl
z!u9x>SUTu4kAJlaLy&5B#zwClIq%P8!^v{_r`Pa7fk(u9Z@Y6=Z}&U-a99ajLYcy6
zEpmS+N=9g8IHp`a8J#3Z*lL3QR^;U&XZm*-$H9fKm1;y|IJVoG>lfT81Is8$RNuL*
z70?2yvfOms-Hvk<Ui|Z&5ta}sx`*C<&Z)#7VmUe<bio_A=XOZj<$|>Lyn{5AIP^rE
zKpYx-8#Xxo2mZga6Zp>c`sJxejwaq=?Szr0paDASSV5C_lVF&nm6Z@qVVkiY=v^={
z_JKp$czc46lwBWy+`U=Y+uwf&F#~~WL@y_0!UnI`M!S@Z2eWg?FcNJ%c=p}p@_#ZQ
zQN8b8mjwvD13TJb-PC>e(oP6`m%^Po0y_6KZuw&O9?yic*p&@i3Rz!hjOfeD(FvkH
zxV2#Kw$MKLRX-$czJp2u;%R2}?izohXm;LSScDV%*_(EXVzlxGEM!E&ju-QPwFMYa
zmK9LIyFO^<{okw8_s}K4b0mVBc=JKS2^^1j>2oD4)L(yRpOE86zg{!~E$jiH1wltR
z?B>J663%o%Mt=|jjW<_6Uv9q%LWT(B$5oo}&-Z^12>Ed#Uvec(*EEt!&{YzjMl|B;
zc*a`Bv>CJ$ljdM(n9aRLRc2}^;h9{$dQhv%IekN($3rFIZ-`GT&K8bOokYy6mT{g$
zw?kX+TU1yGN~(LN-ZcwueYmEHk8wYQb2DIB(eL((2n#O`ZZJ8noL~q`c%T1(9%s-@
zj08eVh1rx<R8&fYUtaqE^le~c!<s-#4fj4?>s;HO0C7RbkieEv;i!MOrv<5QZ)m&S
zg?vU{afj@C9?YpaH}8-nhsRP5CE7Jrmnwla`+Q2ZZx-eP==36}M|ynB?=5o6OQ(yK
zAwIt}3cB^o-m<8-m%dDS^B-&_H6Y%!Q5%`E7~r0y>XYn=Z@!S)vx$7g2&lXr2#TYP
zm@ba4>eS-E8IY@PYmqekjETlx`fgp0jxT5h8{MPAudpgdRd)Y$fI0$pjEd@;LIhv^
zH{&3eKll#TtiVOQ*%OO>*6UaMq+uyFiq=JqN=X6FV|P=DxIcG&d3Y_42601Hr#8{B
zjv{8O!P1MJc76`x&B@CEXyYAB+`#b){6DO^?ntPCdha;-A;S|nvOrG>bWX_;@f>b&
zTEgv;2gQa^&NqJ0sbs5nm?yDsN3KcdLb@fKthU!!4iU-xTue%YI?R=4&X#EotPWo*
zPP_C)y{!S?d&nRbNLfsZ<YRwG1!HC&FYT=|D9PMuu0qX%&q`GabI}D<HIpX+3(7>?
z;^M1=^9@c0W%~8MfSHa}O7iMC!hZ-;Vn*v6W`EMF#P25`swycZULDMj%vBhSg`srx
z-1kC239^jx$dR?gkdTl-hZ7agwGJ>QWl7JS2xvQkUWxs++U>zy#Rqo%I<9XS6{oo`
z9p)=j`=Uve7i!|vea%-()oEpzcmsBu?Pi$F+QfXX$=LMppNwtv9RB>W`u9^QnaLk@
zF<EX<d2T9R;)!LeKbE(NE1+Qv;@qO-b8}41JvG{5+~}N>2;}GgW15$iW<3qqP<uee
zKIg?R_t!_Nx%3#gTl7}Owbro`>eo1P#BT2@A1rce<EsN085wt8{^pXMVu_NhMsfs>
zCV#S4^unii&@svI#UCc@NvsclOIsSI^lws047-s978Z>=mui|h>-`Ec%5YGy(nLR}
zL5TxX76d)A@IeDW4yHXJ2-^(8lk_~M-mXY=254&#j0WYu!f+1^nWTDy1pF^L!=+3`
zn0h;hEu|k~a698D;8&yl-46+J7lV>8byH0=5o)GhF!A;-9m(@;Ya9rl5Bev>3)9#q
zjTaqWe%*t0t$~2i9G?dnp@}26SYTeSw3|sk)K#7Bu;GA}{|7ZS&_NL|I$_9A5Wc>1
z;V_$#!)*~Z3IVEnoUA(W1u5a{%Q4b4KdBZbg|O7yArH^{bQAw(`U0mmQ*+IptKxKH
zo^8RI>U|^75F=>8Dm_$A{9KHv{W3_UCUfx@ZkxI{jraCz22Cz<uRL@_Rp*oa;Mb&i
zHXyd_umy6nmfd-L&__GK0B6hJcXoe%Y3ZUjd%u28$fpwZf=(4$-;GA^?9gI=F<0|e
z6gTJ`XG<T0?4M>+bOHFPnl@kAE!LIg$?m!*-Oj2QH#lZ4e1(ZRn&#q`12OGE$tmrZ
z)5^t~rD|py<2NizjcfiN^8D}07V1owx+V)`R2KpJ$nB!Gr+^YRnS79jdy8?q&vc`U
zNwUNYJ}}`8v-em{k|7g^W6{mBd;HhqWuZeKGs=9<BP8nrVW_qdZV-FTTjF#Oxa~_6
zv0Z7@`YO`#_ak-Xa*V?FEK2DJdTjq)C83OTQf6~ynn<TC*=r0^L(}v6Z{$iH6gsb2
zBLCyOIMkSW%dEiYA_Z}#>*C-3{g}bzg0BaJ!*+#^vx4G~CiUN6mjzK%Q(JgU!=(~$
zqLq=*)zd!dXuv-uPNw!5{b@M}r~BbU^iZIkgO{zuuVb&~Z_Z2Cc7AQ|CzLo)Hv?8n
z7v?}#AMZ)?$(dbH@89x%yUXkgZPvqWdTETAp*vri!Z6W01d-kq)G|;|S%)jly<Un5
z#^sKZrmoIC%&F}PwS$6rCM5S0VE)?|a+zRLS_yXAH7R5P1J_URz?Ji$57zXMFR&y>
z8ABm)P#=yy4LahdU!@j@u60&Akq~c^TY>b`+IO{_uegJ$r-R$^g`EElJNpY;K+V>y
z!8_Ku<UgUu5pSwN!-F<K7v|LN-N&M(<%et6>}kP|5{q92nR7*G#RzCoqPYjS-ckKD
z3C+_ViW%}yrG-z-gm^Kb8uVrmy6BJ|@t;7=rmtJUuig%xG~8;Ri$B429M?+TY{a?5
zP_i{$d>t6t_k@=yusKoUmWuIo@RkUEZd$X+w5eEOPbb53Bn`qN-n<N2Ar^;@q?m;r
z6E#g7MFextyxQ@n+HI!x%!$$cyvKIXbR6UBH(rK+aP~&6{0$p(3AuHQM|$#^I|Z&k
zVvp~9<7O!>**mQZd_(mCxTvbOGdgjjD^t$v&#ay;YO%}LfDbQ;P=3=PwhQUSdN~WD
z5q9=BMQxp5xQ7W;gsHy0b;HjpGlJ&Qksd^`M>~}YqQ@B=!^{`Ndla16$l|}r;39Hk
z2kuouuZC{oyNffNw#KQxS2lo7T5axyeUIEsmI`o3A-7hu$(m`)m8=kni#mF^p#9c4
z#B5L3FT5#h>?8sBt6YkcI9I?(>_2pOCLg%PGV})X_rC5I<<V+^21oKT!W~;jpOLN_
zkjf~>wny9H`N-H_GrO;IsRB_f#k>jlDs>9iYRQ5IS9H<c7;=*SyTePhV%5N<{Z9fM
zup%Am(s3)#q?Y}c`B-uGPnF49LNm392cIm>8o?L&RR$D6`&pJEey1h887NPk;{p#$
z<dD}a$MfAml@*d32}`Jd95qsdc;?q6L6PYT*?;~sVk;NU`kQ?KnZm6G6X?Ux_r#J$
z8dp3Z@3)ql-DBtYhg5eR)~c;)6)s*ppJwqSJUGu+=Fq$8r6FOQw8zn+mCC2qnGN4n
z&?(K9#z?xQpwfb5NO@n-8M$H;PhMWY(}KDWGZWJ=|GXhXGLtIL;-RB5uHXEle?BsK
zbzUzz6w8F;KluPT!x7<&D<^FiaVGb<-mWozn~}^hX<s9B9^>JdaQsLEBx#Zdx6#MF
zwzcx0ReEw~=z#^Hq1i)QM-C?lnP(%Ad~FGvE)?cyYidvzbOIOko-i<J+r^ARiZdzS
zL9#Y5$iJsfeb6+E-%AB@R^f<{Yz(KTlX=hJ>y((DeJn`Msm@7{#EM2z6G^aLM8cyX
z26-rXy!4KlASgtfo9dZ%G)z}kPEJ|q@Zr5qNqI_tZ;V6&^6)V-t53EexGzwu-zE0O
z#m4r87os`~1s8!yJji?c-Yt9iql_QRVjtG_5R73yDs4}mB`5j(ZTdkYPXE3v+;Tvw
zj)BcyxMC+}mW7sGYjar1d`CAjo}Ky62U#V@W}e-Z+GM)ZrF(qMw5=N2^YT@&hKIBO
z2My1c{&!z<qPRadek}hI#-_DZ=og~BUI+3?w^AS21XwKupf??gu^AMM?Xc0ETX2dS
z4u4$nLTI6gpH-h~5gE;OYNXAy;oSR{NA=k9`hJqQ#~1omCx*uq@5dg7bNubYfMPo<
z4C7l$;Pq#eCMbT6z*~Rp%=y=d+EwshGLmbs3z}bCaGSZ9qll~Q(^ZfCt<wlfC5dpP
z4BCCMgMPQ6<M-x1FRGi{S^p1yWtp-W#?qm-45i7Ervdwg@2InM{04sPMGsqTe;pJC
z&0*gDZFC|RhrX^-T^T;_mQyXJI6uW|ROx_-@5drbI7ieSnGF@n)$kVcXJeea_SVVH
zsqd=bl<)?!dyO9Z9aXV2(N}A4%u^_XqJLn;QHA63ZKNc=yD)qHw!(vR>Z>u~sEkE-
z;1eT<HHkIW&9@!flR-?Ysg^|V0KR<w2OD_wm;dKO@(N{hNFe)rEbX6>^kj;+<^Yl-
z$!gqC{W+W7L~&@|vq!25AjoKZA(E4<l!|!MTiMio2D#*tl3@M54r}W)tR2gWV4PZ5
zS;eF*Oni=|K(s^s({4OO0<>7Kw@UpN5^iKkII~IinRw08^Lac<drpLk5&_=vOzQ$t
zWmU(YhfODmhr~+XYiK9=T+20>#2HvgTC>A5SOXsX>tFT@Q&LEw-6xLde;n{h9m_D^
zTnhJ^@YghKJjILV@2zepJx*-;K`mnTZ=j>cSBqPHy$KnX*B`-T3^%CJIWW(e4u%@6
zy@wXp?2>B@(Ot2WV(ntxGFn;(MUi`dfC9vU^;{Y4c+};pGI$ao9(mG8%!<amn;7A(
z%D8QSZNbsGmi8hB&{;^7{>Kg7J<KR$>qt?8<6t3uvJsxEtMhOSNQsv=`9=4~bz3y$
zh_OmFf>u^!Ew<W6#IM=iIX6mR;<~$K-H^8$R5vRR`T6O!``1y4oG~LgjH_}GNCn>V
z9Fh0aCWb6VkBy;6d-SS-MlkA5$np;7#4Seomp;wWpG^0~vv^7=-SBL6wuLk%tV!C1
z(qFx4sq%kV6>GD&#v(o4c{r^)jwMG#9>$Ww?)2L46CA^`9f5}~Ze+B#kYXvVJ)koG
zL#}xy=mYfVSoU5p#BD*$WosF-^%Kvridhnu3by?1q>H$uqN)$GoNsi8oW<#lz%?|7
zSt~YKAJMKuy|=vl^FM;&BPJ@K1}Oz)4#)r8`o>bLV$r|mC(FCQ2`ds*$kvaG`=<4e
z+o@f?PK+0;H^zHfqL4~4%#nHViTy<VDZn59yM5-O$YGz2ys`MKQow|wO7hnt{?AGN
z0DmWH9uB3yNFq@$DLJ_|AyI&RSW$ID#{~Q*p`dL-g4HW0;OQ20^+yMbPr-SuwBZs-
zKgNU!-bdo4xWEEY`a|j<l8q<aG$Ts9D<=N8#huyLU>YTr^Z2ZthH^j_T54)9Cx7+r
z-uOgLt{AggZ`uzL{-5M5I!ImzFHuR=Ga6;4o>JQ@{+8G!x>hxL!2ZuK^4bMJjKu9W
zalx$vW`D5?8xDU8IoUuKz_V0t!FKxo*Y8Rm@gL;#vE)@NLA!Hn0ngxC??4a_dlr8q
zB^ulN5gUM>x1~>S?&6Z)@p#_+Y;Vb}OV*pli8h|5Dm$P1t9cXPPIXie-@wy)FabsF
zp#Wav^6`xlKr`4631Q_K7W-|e<_O%j5^vrE-7Ki#FfT`j%jU7`ad*!<f!uq=+J8v9
zW`6IMgXEKH+}@~L9;jLd=fP?$=~j&~5+VE4dVjxd3D&_psgFJQ|I<_3;Z~r4_w`{3
zpr|V^xnL4wy*b?O&;L|r&iFny2z`r^{`^sh*4+Qg%*@Ov>~Hy@8x~`P_!h9IZ5^P6
z`@|j}+OVlr_)zhN=(}%q!cboht<G^urh}##>TSr>=H_h7y-V}g*UETU4_X*zv&~X%
zhJP9w>V9dx4k02ncUQNzk9YfY*ZNju=fFzDQs>{N1-#q1b$_Al9*nsf+sn7m!Ou9`
zwvF4YAf@Q6={k7~wn7V@WaUCK^B3rgmsh?!rTo5@Qzei6U$(jPqHx}@sWkqM=+?Y1
z@+l_UM>o9VE8|!(W6f#>dv^+Aa%9uM%vYw5?9SQ$+1|CZTNj}m#NE@Jk$nr1Q%K4C
zHcfW4{m-Mm*Z`d2)BW!EN(3Q?fFZhJ^F@9e&=-v)Ek5&F-ck<(S5)HVv{zTdSDEb$
z$OG2-*D_j8?6*aR3_mV%Ys(ew=`1c~Lx&z$C>2ZZw7^@xB#Kp-9fR{ml@hC_IN;37
zcBc<y6ay4&rq1Zt8!ibfGTv-V*FQBsnM;DW@rIp7q(<R}2=Qm#-3v->qzU32uQmaK
zdQiz!WVibt!3WQwI{sFxow#6KTnT)CyhSU;gj)p$&N&^_Mb}Oee#;kk9@K1xNo>0p
zwRnImrYbl+E!!eTs)k2ei1*rTN8F{0@rdc5sG{sqh@%PH^7p4s1N?PA+JSeXTmJSm
zMRbG%g%*fwv6<2W+`o!`AeoN<<L{^gC02hss~*~n5yabB_>t2Tk98Czt<9#z5n=OW
zaCIfq=f~D^3it`0BfHLa0p|55M_rW#Pu&jV7R_Ck1!jmRZQ^VZ+$#)UUW_=dk2@j$
zu~=X*A~*K;P#;2D;1|u_!BXNpPd(LIW{2^vdC}zS1RBNF(YHQk+3i=^VTfa)M-i$^
z#8x(?L5F^e5mQw^>MP?ySt`v|n0_D3Bl7cG3+w#E_>=Lqm>dL|rof@dHf?wgpKY;}
zC4u6g-qH+xevzDPChL~e2fxza34Xl+52$dz%*L^i+){JmwR^NhMg9c!M9lFD>&i9u
zo?YjDvV*o9{O^}>YLU~WcB(RWyq&~kMMB4qfw%J$WJNlTO35-_H-i1P<L!d;>j)~W
zg(9htZ%$uSpS~&KyeBy4Ie$^X>gRVI1iO2^7P9c(*J>P-?F+34BeUUJi0=A`DCL(P
zDxBMIE`gmtydt}r45+ZM{7uS2I|wh<CrF%9P3DWujbp}MH%+(@7-+oVDvLPLF~2(J
zAzZ8MP_@rcjn4MPJ+n2PmlYF>IGGWfh4qW}HvD!yh#uP^V~wJuvvyHTtx7vgV1y7T
zQoQ^)N4(UB4WNike)TtbhuhgC#^pxt?;0LGZKRV)G7fM=`gQzc7NwB!lUu4|w_$u!
zLDO`JHb!%!4TDg`dFSJZaOU@06d-<fuS{C<H(At9EPn#g+0w|_ZlN;+0?>7|99Me8
znB4Nqfwk8$F8y*bC^t7MNjU4~Va`T~)_2T7?(dz^Z!tRGB3iBZcME*0u}@yyaorOh
zM6F@(W=7z#I8}S^M$Og#u&Vet`Lg!}eL21k?6qTgSHvapB*gef7avxf(&zHyx4~S}
z$l~l-V<<7pv%DW~iP`bp5f35~T|C>9TAn<DcZcy+?{`QFXLvU7TcwQv2kG9}2V>DI
z{nso*FR9?^tV@#TIWO#_aop?w&8Ggt{)4CV`HkBKg5o1uD@$6L;E+n*ET5f9?XqF#
zu4rJ;Hp7s^ouQhVngD=u2uL(}u{Qx^>Ss5z+nOQe8H)^`CbXik=#N3VuI^7{+;XoW
z2`oyd<sUx)OzDiLV)km!CjrkFRYUH_ib-h{ZL^B_I|>kr-DGBUf!!7*ibc@5;(3B~
zmt8L8sa?;RsPa8sTZNqq&}`hV%v)tTMHciEc4K>f7f$PcnmiMu?x&=^$dk}kf;*ZQ
zbdm^dyfSybd)g2I@RkLsCNBcgyyNYN2#8u@DS-mk-QIb-ZDmd>V-}E75wWb&5^_&S
z<9+6R_V(G5OQ(}xp}~0u2uiJ+8Nmn+OU^G{b57a=xayH49UcqxczS_XzRi<MET;8-
z+uIknF>ILEG4}3r<-R6=^d_L-V=;K%1{6s=Ke$`?X9KER@mz_~fo6oj28Og~OW=q~
zg-gc){wz)DACz5cRzjDo_s@Te;&J5#)H`}w-pQ*3`!7YlU!uM+Y6cYtB!{c!d40Nv
zSQV3WRG{4(#v|$Fgdj-`R0#FZ%4=TZ*F{FM?1_amf(cZjiWfv`dc%@iS!klz`x-%L
z7W(=Wm>>4qKa-`fYQ+px=EkhR3syN<CpHVIYTe}IR=WYr{30DD9WvtC?Mm0CpDZ{@
zHm^;m7{Aa3eL=w~tA|3D-&!8U<{W&y&_B~v8Ee)iz36F)qL&I-N?|Rsc220*MPCfY
z==&=ElBkL9%iCzaS=*L;E!F0F-=ryxR{~7vN=ze1!MZV79nOt0*_rPk9@I%-GKJ4b
z8U$b}SN=r{Wq%&8$#AtY?yK@I>$>!)k#2&3&r~QRBMnFC&kt=<bB?Q=ty9)b8VVq8
zjrDGG@49q!wx0swnCCfrg6M95#asYOY30g?fLr^r!@zzP(gQdv3jx~1ttajpT5WY8
z5Zr~dnxB4*13Sk%T8Bdg#qgOreECa<-0lGfi8zBjoqDVjqs#B!qS;O*WQ#|kx0}#%
z;Ma_`B;zy9xXHG(8i;v6bngny`EN*bDdlq43S{Du0W9LAJvjWZUR4j)vXb;Tm~uC|
zgS5y)TI>zmzti)f^%3mR+?)T~u;B5~j(Ru>uh-SJi4IH?e#+gS(LMf`$+i=?K|?60
zsdJRV?G7B0Pk8p#NKE1PVquiww0{S@Q9RF7$D`5Avy=-&zfP*a+AcfM9*&B=FQ(|9
z+FpFSGk!twaVsS1Glu@cY?JJj8GhmwxP6OG($7{p8~r-nn0T`v%cGFI3Pj8~r@;C$
zNonw;bLK3VN<ZM98<Q6ItYTYG7Y~gvatTNA<5xSnZv}XNGidNlT3GznXdW$t{94S@
zv-{x_32|*$jlbywI)3HQTXw6)ZXwicTRUzqgl}EPt)Crcb34kN*uC3ah#h}KV#A*V
zD^>IxjEu8ew9`yA@%I2fqf}7LgEBTMM)5Lap}eoGqG(R<6Z{YB*nHa>XJOXOifHU`
zzh>9k<(r+4gL{5ewmDn3(U4(PBT7ZlU3M%fs^>PJ^fn=<oWt%g=C~<4`hSVECC*A(
z+ZGv8RS`nEKY}2c{up+TGy;|7D9&~-_FueYJ82v)p~TYoP$<qCWP5-Ex~w$Di`{`N
z)2%qtV0c?T7dP)6O{V2QO`~J_t%LbyhiFWR*q{|!Jtj9Le|GLTu6`(ueLT<OBpng?
zeSK$gb`Kjbz&;CxQ=Q}8-KhWeAm_<(pM0cW#2C6lkJ;d_K0ACM!RC2^?#tlU%5b#y
zg!*O@`ye=LH~!JHQf2LkZLS)=EPAHVS#myh_~!QH+z1qsvqU_+Va2D}jAk()_Lya%
zd5m&vIxb&#^9Xu@u=-mi$O+O^Vy`>8j{P-Ed$w|rXo%NYTU&B4#5BVgAj2*v&R|oX
zUJSukWL=(vN?<?;+=-pF9!RPvbhL%(J6k7|`)^3DdSfLLt#Xv`{eM1-yG6!ibsCK$
z7M+C`dn9}qM`(S3UC<y5OUJ>S7N63X>;~folr1ofYZe%cwwYseFdW=YNsgo0x8tc|
zy?*%-S@5u-8fvNzv}`m~7(Ad;@9F;`A_7!98=ZU^?$G`lPo;k`jpgyP-WDhrDwb<7
z_aMkj!zc8J6ugiHv?14PfRilv_Ew6ONiCQbh^!pr{nVyYLJnOzAFUrWF2rSw>NJxQ
z0h~%Ne9+dc=;#q=1GEr7qyvo^c-4E4T-E->w~eU)i0r*2=YiFd8a+=|r>jgPk{jDk
z#6Ck(x36y20!~g0VtPP{gg?lgh|*K~o<3<mO*smT`OR$CK`sFcK0B#8S#2w~XVlSH
zC&{qlPUpLl2ElXjtxq%QPqB7!*-wxP?Fw+-38?nP1=DUr96~&pG2T1xe9}*NEGYRU
zbLQCB3<9*7LMFkn0L_CXOmSv6b<NeWEdJU?G#X@U)3R<avq6LlZ;p0pq<Jsbp{|8J
zE@oiwQb+CRM#z2Fa+;O7OZmvbT&Y<G!`Ul?-ytoAMogGkTg>N@>nP5W3+<woZ1J)Q
z<vZlHB4BCkWz0YgA<%w?nURrPrQ)zRwW<2JJkkTgvDSXtjic&Ye`S3EqP07JT6=zK
zq%S!FmyC3D$D%DaNgZZz%vi1`9>tYh+ffG{jbjoW`Bk~Bb)bxOEO}iI)|HQ58|Pa%
z58$LI7_W>tE#R=Ww{<gXv|woUFOLH?77p9<(FGXLiGMWNnwFNz%wW)pQ*N4%D2GPz
zVI2FFA3aJApke)N#(NH6O5qWiW|efO%*Ti_2Aq^(pJi(;r9Fo4^wH{!{O#lN9hReH
z8_a~2bX{+VpTExA+S+yT@EZ>$N}y7|R!Ne$?5&hy`zhZt*!81Kj+-<KjrOTQub+K6
zPUNLFO|GW+O*hd_zCK#KfGKy01W7YZXI*IBFHLeB&j-S#@C|A^3csCkKJtqLYi~~h
zlG)6*)L^?$n7%(4r~c;!AR*AyFo7pfDVlG^$aR45N^hb2OF_;~U1BtgHpb;&>-GEL
z>b|~?J2>;hhb~fGaRYghKe0#>PfkJd;!9^3=+p1TX`(Ax(OQ`>^y@daF}S%<S&zf6
z!jcMNWI5lmRN8K1RWmQDf7^BgjX><j9#y|EoLL)P`E)3Rz3$mz%r>E_K-ndNGJ^xa
zxxkxaS>ZXO8d{xi?+}f>cJ$M{Fh)0HA3c8dyp|`wJ|OwZ6&UxlPH^0^scN9pAi>Rf
z?PQA++uqikjYYdD#4j=mP=dy0R!ikRI^#8ViI~&HcV0TbbSmqwp~`~w*c1;6+G<kh
zl~bvW!k)h?QerMtI-8J=O=>m3s&U4QNMd`9pzD*a3pbI6n{HJ2bF&$Z^w8?lr+F9c
z;fEBk^+ou8r{f2%9v|Mp%q8KuR3CoUk6Iao#oCS@c9<VD-8v>p6xcgen0gliCI`X@
z92s;9!MoZQ+XRyKBkZd+CaW*pM6mTEtRJe!B5!M^S1gdG<k|jKAHJ_K+s@;)mN#Ga
zSX0lrG_94|UbaerMNQ<t%YNrlYayFIanRAEX)Vb~8Mtr_mNBJGo-e6L^fW{l===Z-
z4Do3(;myE{!$*MWNQk9OC9oP_zo#>Q+@j=$dTAs`uD70HGqTMHgFdG{R&c&NJTEZE
z^y_b-kO-*7rRfxv5l<+(2}bwNS7+b_dlOI+hJDu<=E=j(c;~$7%liUyW!zk-1*d>J
zqRw(yi#$@yu^VT^JV4K8;$vG^bRSUbAw{J>0!Gkd7Gi(MwEv1@OmwZr)y_q2n-Gxw
z&AkRgg^|t9Y+L<Rx0GBt8n|SBWf#=-zErF?+-e{!AzbR*X|+yWd?}?rf{6_8@S-4S
zrPa;y2TM}%?ICG^E+cJM4eIO#?;^gOY3^~*ih+7ej|))cOQ*M)MRSQJc&irDnH(*>
zo&m*!=s$^q;vUu+pxEGO`|4@btQ`~AWwX47J~JcldqJ)i@~wgI)cZAld-SCpFCwvH
z)JMJQP)<w;rk|Tf3@F);UHyBYN9Bu&*8a7)kj7N}hTBi|?}e{=TWY#yi8;J1*BNG&
z_%rqcu4=47&PC`&{4TjW>$J`q0oL8`ZV6sGYCd=H<qDT}EC;wL>otBb>JJ<Z@a<>H
zXfnNBd4WDec>Y|hE<+w8prZUyt{4yFee9|>Qs+weGGS8i<wz_<>M)UOgKR*Y7OP_P
zZgMm_(rt7kOtElMeA5-8i7Eo8sBMHIGl$M<TN^$p4v{5av8Y8quFQ|mObzYGYN2Rj
z$l@og-mOC&G_QJ?B-wt}PZ>+%o`3T`k5O~q9TH`Zdx!=|d`1jNP>h+~w9Zob-|nQx
zVshSNqRr5q`3z8dh5)L_(*f%?d0=b4priX)f2XyMc6#yOU&^{|rO-w`5>NIHcT6xu
zPGHe^(c$QwRH6};s!Ke6Wm?eFPRa#H`d%N}@s$Ky$6?K`GoC3QS#atNx3Ne(YSfAr
z)MY4<@KW1{B;z&&$qJ5{xuZ!zVW3kZds9t=Wz3|mSwrh5r=^CgEsMO?>UFU!kvsvX
ztbpt-u@b2<@RJ_zLE8x)V6_=L$a;GPAbi1;cC1Q#>=y|<h=XwaTp9K98*lqN@!;8c
zZ$x6)?me@OK+=PDg7kA@Ge1gh*SmrlvB7-}I5)n%jt_{le$b0Z8_cxSqipd>C3CD-
z+8);*^gm<(U2^xC1x)NCkfP(Gr0i@eE)$n50kISC$+xDnwv(NmZ%sT%>DT`WzWz_O
z<vsFt|C=JJh^~2G5UgbGDyR;vKPH~tCYAM14fVF8;oX?{S2BrpZbcqhg#kss9j|AX
z0LrSLG(w>fn#7y0b=H!Ouq^rb$mJix`tawoqV1RwEOX)3+6-(|I&ooP1d=68Q^T_i
zH8ic)Zi(NzUea!s`U)h7`0yu186ufxy3VTn;>%_ANsqAI;sC2C)ifMQU8qwIAv2y-
z$)|(x-ONUrdq)&LHw>I#nIwf!_&naOR9<IWaUG@*g({JVik{+M6@yj-<BS%yT}cau
zWGWT3qjHDX^UtIuvU>9NGax6o*Cb&iiC_4mZ<~hX?a)qr3l+qg58C8k+m$TL<BTMd
zRW4?7Fi1soNCZ&Y+>#gG{wt#&oo?twEwUIV_jjMP-j(qM1IKKWMdem*(Maybe;X*}
zwl+1_RZ`|*^8`@MV;@)xC^pb*>Wu6@DyY(^j8zhBR&$d!V;o(ef9S&;OfzJ}WCBr{
z9K|TO2~p|0`N<R|ZjiBF;SJ7Gm$b@YoS&Yk8!=XKuD(z8qh!p;O~mHRl}5kAEM5B=
z0F$xLVeYPD<OYo#i<kPfm6QX!LylKwHTp45!ik#@Ekrnt@v4SQ?hVot3r0FBWZwl8
z0_^c;peY$=H!{(GZ8$EwCrb1}tfa_J%XzlJSJS(2R}CD8`^5V`>(7@9LYNdEjZyF8
zC%H;*$2(Q^7=F{&(!SBueds9?8kiam(aZ%_Qhj)CX(_&RHe?vN<1c#CFIjHqlP7L-
z%K#&Zl&jlISt40T>k`LdrrCnFPttNf5@F$3O<|F`tZ>E$!;DR_9X%}@-a0C_l<9rT
zAgQ8bC;qL-oS^;IeApD6UDC3gsC+DR;D!b^l|l1;H~N6;hlLx1a$-b)O3BnKB#`Z7
zN;L3^IWNf8mlX0j%^5>)S8aY?P`@V?b3imJ(29CWY>?Bn^F^v0a|amfoY8!=1Bc=V
z8rS%2UrDH|qN79*9-bp^i&~aqH2Sp(Ceq=h-xU(}2j+iN`t2OQ9ELz!ERit<rY=hH
zL2r-M53UDAYCl`P-NoZ33tzA#j{rt{6J7+fbIBY!m-pM_^z%9U{V4OpTuSq{n}P2A
zJ~J|T8gNbuJY$+BOn}Mz+{hf9XER}+FNL+A53{8}LKx(#VV~%Holh<zI(gt+Uh4CA
zlWHz8$K-7RtyaYg>)2dHqUNh4nOHh3PpG^^I5FJ~6?(==8uFXNqn;c0{K;$RMDF6*
zj#*CjI!Ud*v+2K0vjl;Uh{el`1I2S6in_IvezJ_PTc@<A5VuNP!`-iv0h^n^ffE^K
zwc$drt>)?P!)NUp8r|@Qv-}s%#G4!TS2e>xGh>nMfz<Q;&zGE@rjS6~ieCdCof2lU
zgKIi#z!Rm1KRGqk%f%@twO5;<Ww?i**E&0U#97M~u7{iu;@0n=70_PNXJC~Ujd+!v
z(qT-w#EZ4{3}WkFBK^!6z28qvlcnZE&aD~NJ!Y4L<FK=<TG9TUYU++Q<$o*Gd-GM|
zJXc`@ICNR>yPcQOI2tln-}%h*?n0&mkGs=8^sJsO!YU5U7y8;7-_N56Wkcv4+(-Js
zvbX(a8PMchW+BkJcs2-efjT6=arL$C-W)|2koB%nFxsCjOluQikn1qP!$DD5PHmJ>
zT^8sP-@=%FxG5gnu5_nFC8q3(EP4LF=DlQ~yDpL2z2=Qn4+y9j#V~$Cyd3{qgJv=F
zn<U96yOO_{amM!2ZLWW4mH6z7rt`QX1si0NUIC<CkdnEgxBr;Xf|b-ShG$h6n+FhS
zqX2rc;X0m+LTlFgPP=!IpqrnO!mjI*1@aYmzBkAGAs^}kGR4uXnCyId@PK4TD;hO5
z0f{W7B)n)B8?<9_CO-ScVJdtCHMcg`hYMjUUq94O1zdn{R?<UVJ(=Zd7m)I-ru(!h
z$&xwG=cmV7{-rU}{cy<Qaqo!g@;P~=*Bw;+`Qc(r=p^)&L4zZ=b)8aJ?JBvKB?TYD
z@f?w7X^#)xnF_l-a7e`oJO`{Vb@DmBY1;!7NIaa%Y=vis2%{aOQvB@=xuRg3{35)x
zwAmN^2uhly<+Dwg5+%5%slQpkL@Ha&PAwj5ahFA1&-zFq9<+x>zIje@+GvWoOV|CX
zF9^ggk-u)darC3H^LPj)H6}Da?!U_QJx)~15pFKR6x-X4KIyn0&)~74Ny$kI+`q_q
z1ZZ?f5zk;nNn%Uq_Y%QN)q2vu=_7q}vhMl?O)Bj9aiurd4UwY&@C;fYtulZR4Yif$
z<>BU=FK8kUi`(H~29m+BiI0h&bSFjaNi?%>nptM5t>v1b-gss47;o<+c$qBEGCHV$
zG9N67NlbTT?j1-T(57dLB~0BIP|z>#=Nl5^PTi>@7rH~aEbe=CMz(=>zj_s>!WQY&
zg}bV(mwG==Q!!CACyaTID-BAUyArG924FppJO|${@eHnYQiTFAy9NNNEV=VPTlBVT
z?Jz|^Qy#p0)6Xas=@Z)0I0GD`TcN*nAO7La%<lm!IaGtxwGb3B7)qZoadr(Q4ROxR
z(Ow?PO<HGTm?SNkzFMIEtK)?T)DC;Y_}X9=b!pVuwzutkM&H|${v}0HO^3N`foT1P
z$(?8h<Tw<=+u3jZJ9Y3B0<{|Y!n&XI1>`N=CtK|$+Js}6Hvjf>5LabiOMT=WgJm{h
z1GU|N)jqy%QQ&(;%#!Wx5LzO^wP3j+f<4EiKDx0-rGkowViyLxNDR-CF0VH^0op{l
zwPtQXapGtqvytHvy|$}(<RnEZmu_QUd5gpAShf(M20Qm!zC0w)udA)bUV(sv;lo_i
z2NR@Dt7$##Gvu<iFDp8&y$0ytN=I4HfP%nSB*;~-+M5{|aNEFhIAQqH3U6~%-EU1L
z{Xdm#-z0T2zb3wa1XeD5)7W)hjuvP$X^^s5#LVhH!5F~|=zx;1h90~k5&RCjmT@1j
zu$DrDWZjDsZ>kik_xzGaJ0>OkiQ60GKl3}7FMof1m`>ROs^QXMl5m0EPjyGRvm3Yw
zJ9VqexsE(5!B;^IeW2yA<b)w^h3hyhq^40QW(f68Ftr|f=qJJB@2p7UV;6n|kuJnJ
z1h#d&_5T3rDXmwEj`Ttuc<21__GbGJmu4$vhg5aR2}J;}oC5V%Iek(XsWWBF05x<!
zWQ*=VK=&?lm9w9;9J0>TYcQmBbRcx)nMMZfu!ar%(=_2s?_2)6L!tcp`%nA8>YGAq
zZ0jv;ytq<Z!oy#7WK}vn6g`}-U0gU8+>dv6SQ3L@2s)}y_LEo|zbYf8<F&-RuK<e-
zd86sCXG7}NlwxXc8mr#N8&eAfwWF=20VPd4aZs1WS<EMan3BQg*crI^*3#J^P`Lpb
z8$L{(E7+H62j1|?;!<J{hKix+M`>?C)rJB2uAeVPOpEC*=@*xG%!i11a^kc_9-Ic=
zwdx?sXC^i{+l{(zB0UxKA6{}_Y2q!Wt*i?QiUP{t(l(-93T(Rh0}DqFl~=VoA`fIM
zS(9}<d(;7crzD;KYRv2QLA^0VctWjU8<l6uR(L=j#GPP()EDX2jj(X{4a)Yn9f=Re
zdNb?yT>X{(h*O3K&_uXMx(u1;qdVK+BIYsX(@E0qH2FolYo~gbC0oS(Ly4oJ7(hLs
zUE1Fe&3s~>gazmA7fMOq24cUro^~34>bl?{E7g0!7`$u)sf_VZ@@sO0^AU=gzo8)p
zH&gFDUr3}a0(Nb7?32B7DY%HSS4TU(!s0STfl^tak`3|fX<DL00>F%5?JoF8EL!<e
zJ3_mvdzaH$p65cHXrXcvyZ0y|$mx?wsuQTPH1s6JDH6GSp=fvTrWMLP`?jepO+0gH
zK05s)ED-z+=bt&1z}17wyM<t0;P1&x1cs`^I#5>b5C}6ww?I+#52$vf^mDX{!J+aq
zZh|H*d#Uv}0>d;q0s>kM`zw1N#;-(vvUVgp5q(hhy^IyeNS_Q4mF4Qxv$gm8FgmLq
zs)X1_jCHSZ!J`ids6Tajdi>^`-<dVX;_<<^K3x0iV9N05f5ocACm^cQ?IQ5<-Qf)%
zkd}ayJ`(AbDo&vL07jsbDBe&D+bzCL-=`JnOkAEUqcN^8sx`6vzo9_$z!%urv+cg8
zjq_Ct1I^%K#PDea#I#P=+b9Jsd@<YJlynC~s=Xk@{4P>)Of4xRO_$-dymV^EkBSCc
z4=cJ%<<40?RaI5{5~F&1tz>+Da~TZ#^d#ILVZTHs(XWvD0GIN#Bs##<xS;c9hJHU%
z4N*_;RPk`kx4h@)@^7{*tgY@wNq*W9$p5oq!S0?7o+v9)AHT1Tj1SLGJ+x1RNnefH
zed=+kg=ag=xi)=g`3U6Z?o43VJH_M&uNmaY#_U9xPHpQa_iu!=VJ)Q-?peshD07}g
zF=>qF0%20p?Et#(sz^kRBYQu#{vvSSYq0MRZJx&rUiJf>tQ}=!0}Nf@(FY5sH;D<>
zeA_jg`+^i_7ExMtEg5tavv|WNHZCxHgBWOGw#!RN*;fBBd4El_B!GSTpRnV(^}ACv
zC5LVIYDKE^nwf$3Zl=w{&Nlc?rSoRhH+I5-l5;k34y(y7FbzL~yV@L5Ozw8geG(R?
zVYkmKmk;=F2-=sqZisE+)0Zw`_uV?~Xwc%#R^ktNL$-i7;lNB$+#`np+3*YHOxop-
zsrf-iw*@9&&I*=4nueSB0@4(oTa=8StlG<qJn*nN1n}6xuF&;F5z7Eu**U%_Msn4V
z*V)(oa-(l`XR|}(7UuV0jx;6>s;{%E;!P$A{eLD`vq*QQszGoVsq0^_Ww(7c&WWx=
zCa*W^;u&j9*s3*SY;*KN6myJL=8)%!$?s%1wObZaZ_@g^3-!#U+5|J+cNFuhWEsOz
zqrrw=xQrNr{I(`BVIF4}$@7CA_BL=Hskj4!xda0#VGpHWQK6~|#B<!IBZ_gNv)cvx
zXG+SALkZ-NK?DPy$vfDN`X1vD8Yv$JJr=KDf!upbb>&a{0|>^BXN83L*~xl%zQ2Ka
z!Lc>su7hkBmMj@!iC&r_8*V*d<D$ra<V-;tRT1{C=tWUuWfyv-YN17;FhdZVO~}lw
zSl)W}*ogF~yYr|Q;zI76Sl~4x6T5Cc-sObt(Tpe?r|8%-Bj%an<i3|XdbTak2pWB@
zjoH8F(3Xc;QJ>BvBXge~OAg@C#%?XZi~Sz&B*S%U486@#Hq5vjRrQFRw9dbc!yCPU
zO>LdpbbiXcU2BQKa4iYF7Wyf}PX4m<cb$p`b{dM}DEyZtMVm2}^iJ@?DL1y}YIrl@
zc&ciM?d(GPWEu@_K<-y!Lc4s<WA0N27e43|mLm%e>JA`+oZ+QZc;@gy)|a0CWc$xp
z+hDoNEwW2hk+g&t;vtC3oxxl<o;X0sqjkh=hQm?8zc$`9ChI%Kq14GlkNvCkqzk(I
zN(UHG_bIZ>CD0|}I(jcDkYXe8&uX`&or}Ih#~&R98l&G%#WsZNWwH(S<6Qjn3!npQ
zLkJ<1W8#zb=|;F!vO3<-LCgc9ADzVnxTmnLb!hMoH_#03RU-1hIK`<U^PML2MI?^t
zUrFu0r_weTf!X(QG}qgu>a-=FH<>U!d_5~KfG5GeaqQ>E?;NGg^J-BZE6Eo}xRMOv
z88BxuR?FW(v3RTt#bz_i9+yFni*@xpR6lZx75}2$w=~ovXXlhO@El9Jn1=btGD=To
z4U@~Wnyu3f&liZG2!!*kG|u{kIwJl!k~<9I_HeQR4$dUtdj_}JoYmQd_~QNy8dGVC
z3{R7Wmu7tEZ~s7{ok)X+G-c~=7rs<gTks=yD_W+XeRlR!ZjLNfZD*ZE@BP8<%i{f$
z`{A7cU%QZdQGh7$w%hSR#um6&I5z3W4crmB>Ni9|ecJ1nBdRws<(=C@c$CrxmVD^M
zj`_ZIvh*j7w9=V+Dy&bVW_O-91*YqGYkSP({SeOIaht{IxxR#IDGq+l#d1uV0xJHi
zbrM44FA{#kV>@x>5c3;S6T&y2#{FK2Dr+vC?evYI?&kpKYMfAyaLh=D%extweqh5-
z=Z+P?j9uYc1J0Vo!?>8!pjTFee>8`TO6xgx@Xhm_h?4Wa#6w5n^A(PAtcHCyrr!5#
z{UN?0vtiNsY1UhzM|p>o&pp5^Ovgzs{j&zsrZs%X7tEdjUL}`CoD@}fjL6|$J#-Z+
zWZ@r%?ozmf`Xet;NDeBwww*Vz#1^mCMTG;Nrmj>F2byl_VkOL<Ayzg<4d9E})sa`C
z1Y4(#$_*o2O&>q*HT4dZYxo)5oB0b`N>Yv%Tl6e?jeK=_X)V3{&~LG|0}O_|n>?v<
z`64H5K`beTP}~b6AGVr%6NnNp;_6>doAs}^;cdiK8e;dtF!;ZbX-;%OdFuZ(4~;HB
zicB`@6RnWnU0HNZiV-2LtGXG98-I(@JvPed?E#!79fqsECt@;x^@e`wRB;4j;tkIc
z)Wl9O08JAnv-#|j&Zb6v4G&86th0WMb7&M@r|K9d@*(_f3~cVJw0@^8BMUa^qg|vT
zuP?*jxYeWLSqNLiUS0D(iz&MnhmBK75{NSHXB*Zm9s0`bhC8XGs{J~P8^iem@LE-m
zn3(veLPcR$X6DcH<4M5dO(glF2XUV3kU~=0tVe0TD7LNg;)NnRh^Tr4)Bdw&Q97VT
zH5~4Kf%>S?TK{{+j^`Phn2C&N-v06PJ%p)F#%`%~W?hJzo?M;X(Y|!a7f`4a=8h%<
zl^k<OLZl(AMTw^T;{EH;LQ~$9|3(QQIu%HGSu*F<=)CehZNC2-?S6PRvQMQ_tEL!d
z(0Ozyu?j-~ph}Cpgf<N#4de|uuBb|$`n2nquQt^lB1FKAng=U$a)NetRR3?bY8A{N
zv|;pRj;9jr#@j!&kzselEa@4ty*+<BwodF!fuw?o;2VDk=atlo@I=r0-jPeG>N`iB
zvgg)?M#60hRA(sl>kP~BNKRYy_*T%~x&H-@SRq6+=$Nsci6qeQmIb|^J)MdfRK`H`
z0A16|qQQXcyLQk(Q(uz2dyKBEoA&PbU@+at@zK&7Yu6$fj%?QrE3^pZqzR;^A?1tb
z8ccdDeZC<rmu|cE!;Bf*3-Oyi>03@eI{jt~IyKr1hJej;?dexIf`^`6AGg;(Q$(LB
zSR!Wm&(FUm!2JtJT+E~#S65c@8FdO2uf#Z*oVzv;X%!}tn<u-^vDiYcumn82!_R~!
zK~-d$96!S_4<ht0I-ONgZA|CxfF;3r#7z0IU$L;oFYi32aXUN~I6y?YoP6!Kr+d+T
zwih+*+*A_mMOfdGo}jGj)CxH-o1Zx^lm@zYKWNqv!3_gxz99!2gYamYso~BxXj}K<
z#w<%^Qe*P_I6w=O2X3}eCT%wX@}}0v*q$+z9aCxW2gRuLg4Fw?VWXN<obp~GyT*R=
z)8t3UU?`UY`4BL|>o8cfsGX@dllw!QIjB4kBiyV#saQlM3ivZjj(B;~!m>YLtQi6|
zJJa3O057%aDn8NUD@ZcARX;X-y0}PzsiE(mUu&@!`+{6iAy7i~;nc8Lf?3QJbR4W?
z)1XbwEb>*Hnpa0%guLYpnA#ZrTpKc|6%WJH6ilC&%%<9z2WMn$*=l~NnHhT8kG6d}
ze5MZrK+!S@Y5m1iBbX_&Coo^mO+9NLCrFvFeG^DS5GhNpwY>stqD_9OjDPA;_knf{
zxq0#Oi8A-NT;?$c`*rQoJO@U*o!dEzxBgbwl(7s;N+o$sLxDyMM^TBG!ScW$=ptcd
zHy?;CsoH>$!?8SEi8l+TcX+b~nS;O!dvws|_ha)c|Ec;g3E|KuNtJURbO1Dy-pXu_
z>gM!_%1w#bE_#ef=?|+V;k?v<m_ClsQC=xmvt$<_KE_|STE_X-FSFLJi-hzd$L5Yd
z2{Nvl_HS~W&mHn2hvW|$s3qpd51N$sr^Jtgy;|fKV#)6pcf?kHIFV7(q7Lb5+~`(T
zz;Rvg2b+2F=5W1&QLa9bK1O8ifVpiV`ue|{W|c1E4!i8=L^lkclE1I<mslj*>+aAX
zhP<KaePE%Yg*#%Jn7a?v!o3<uw^=`fP<NHSZk+xN9b*(KK)v%coyU9AiA(a@vmoc0
zmM?hC<ki7T#9Ri5{%jn6=z4>px(#YViJbi&^_H<Sj*7U35Dp6`$uMqi3pISjyENdZ
z?NzR_{L_EtF9r{FC-b3S(ALrnDSs*idRs@bYmgSmsNL&Jv->d}7qaEIl|gu1AKd>q
zh(yX-Q88I{&4L43@Snn`MHn?;PJf-HJ4_*!^ZTx+u^;c*D4++R3$uLK`6rwQ-i(U?
zJf=ZUR|XukB8#3?u0cPU#vl1d<GpyN>PUT*8@SNbVswpb;2%ot<su9~(W;rFn$$5C
zgN_Bd@a@sW|Lo{aBC<O5pV3g?3zV68b>+n;w9|o>z683%8=J|s{R`o0W3YrbwWFF|
ziiT_0Jdf*6T6Ith%aF;krI_n_;8YfqkfW+#vLa)q!e@R%RsGHG+f(E`O>Ed@i7I6D
zCOU)yOpnw<KQ2fsn8#Br1LWD=K4vPOek#fUhycBG3sbT6rv)>f*ByDs%+QIN{+odz
zl%Q0s(S)RJJLF6H{dMs{x-UEpv?+9_GUfN+`Gs3ADy>d?3Z1RPVLradL7DEkT`PY8
zw1U<SSUC$T!9Fp)>U7<ZwvdLLxVLS#mKTYfv}q<KC3Uk$MGgpC=?OS5{|$)0ns_K1
zqh^^L6$GM9h6?~0;gZctJ~Jua`7jT)3McbH7{idW{W#XU9JFGylf<TyW#lj^FONnX
zQoslCCyq$4Hn?3{drFIRWhRyF+y@~4TFS!kglD#Qp(fTwNx>P?ekDvYk_qarzfh{)
z!$#;qayAGH5*P1yB`%QYFuC(3?7>SN`|wm_w|j^wAu(P?`6~R|<Y#`IH|w#Ib*tWL
zNz32x`S?t=IR;yl(^Rw_mpI$|Ifx3Lh3<(WhRx<1`9tXW`H!4VMpCR+)4=Clqi=2_
z#8cS|A~6a^^2x9MeLIS2P=w0SdHOs=l$e@5J$D$<=drKs6FSEPYW)f@PrSYnqYUN4
z9=!Q~s5;B2sQ$RyOGt>clF}v8-Hn8lf&$V=cbAkyBi&unQbRX0NOuZDH%K=#FvN5I
z&%O7>eSK!lf;H>>;=A|$>^+A?77caXcb6s#3t<@0ok%wQp_EDVs7?7XNh}w}zD{z>
zIWgQk90!ZUh_1oOvJCe?ul?-LD`z0O#J8X0BF5TjNsaC-f{&OsHgLvFj2>AL<dA1k
z9aFl4xR{i4J6v|2BU?$@!=mwHg7O>N)hE~68#tLV?P+>!H_8abE?L|oMH!PPdAHfr
zbm%56_E7-s`-R;L25w3s7?;LZ4yXG%9<u~uUJo>o1S+~9$E$5i6A1^owSbeX*B7_0
zS16=m^Riku3l(~%C@?y*`jL6Z7q{q{Fu8yaE+pO=)OY_xCi}^KeMz@e*xTKY+<tZ8
z@ABe;9v&7edpXQcab|Y$^VU9!3Y5j(Y^Pr|o|2|9xU9<}pVr;3o1TK8B_x(c=+VBo
z(z+ck7B}ckW4R8g55al}UsG)XDx9N+w>&5meNsPODTQ-wu!Gp2J>|a-B=<{f!l3h3
zpUG6HXM15LtBl<nyA7Fw8vTqhz0t?859=vCNa3b)3@^<7R);IUaCjCawkhPtcHIA#
zaSoRfYqh<wVz%4+&+kFbkJ@cGWMV?^bMGn483(P2h5!ALzo_(W|4<n6VK>|c+)xI3
zI2!;!Qa_;M8TdzL7k3(VCbHskXj#X56a=ywmS(cz>Cqh%F7|VaE41K(3~UZ3F=jzU
z2y*XKbXdf?#z$@UITP4FXudxT;#F-;w~~)?Y#ofb#z^|RvOCvdVD|?kCWnhy(30J=
zmS+`jo+{30dO95+Z_jt6)=o%}9Phz4GTH#ub(YA>TF}e$H^US4<eb!BY-SbH)`_||
zp{zUSVBU+-v=LsX@A8;F@Znw8_oa9J;eMXHm+e^?m)FKJa{hzKjn173wfPpR5nEZF
zUQX!*6D4jZs2P&hN2s?cmS9#|v{mxi*6MpafLZXFPyW{>EdpYH$zKt;W))>YQMro1
zd_Fqjbfl}8JCMymP09SiZQW}yogk2gxz`K18)<DdcCO}&a(QmVU3jwcS$gDSU?}CU
z+DQ|*1})FHzq3Az8CAqqt{8ZA?o>Uk+G%VUHW(~UVajj{TWd^Xv}X22n}LN(o)I=I
z=OG;%?T%hq8|eJ$U%^Yn3W7n-$FqJMX5OCmS_Z>w%TH7GS;?9I(e((ZRSfazH1b#;
z8rGNx{^YMuCunvmc&u^EpoRCm5mw7e>p9JE4$xni2M7wl5{!dF^p{bv$#KD6i6B<_
zC7zEVzV^{KmmP9pwdh~B_+i=r%Zc%eB2o0>{MAmo+n)NRZ*T~Yfb~9(a60*}$tIr!
zQN@*tD(<Ihk9r9{szmpV5CC2^6i+7?t<AP(k;z!fO4=CXv+YmRbiC9ol=%m+Z{e`)
z8^jO~VhXzWo^@je+iE7c`yW;{N0)oDy&S7g>Nr_yuIWrUPDuZV<ApfeP{^9!l)-hq
z({0)gY{I5P9Wu)_&yx!5dV(|0Z0$a`4a3A?46^ZDy`y9CQI{I$7omyhz)i(F)#A&@
zBC#e}a-81i$0UNOa($%rQI*<F?iq%9Lx3~AT6_MJyHoz@S*_W?i^K86>t%i!npwj(
zcWgbURrezS3o-%Q<nuMPi%i$}m;bB0R5qO-CFnhILGJ4yPG^#biO1>8XXUL<J9Vf3
zJ^_Y3S4_|wLRepS2x0;1g&-IW?+Fps)==f$8Gf&v^6tw7^N{RqHE_uOf<xN;FC0-r
zoI1#L)@>P>H5;xlEqCSwKutOPxx?!7hu*Y)yt$97u%K3*oUUZLJ}Fam7V<<GqiB7P
z$ya-2oe&}wi7@;2YE<YEQf{|X#<OsI|M7S1^L#H(JJ!xjj8)}9zfO>oYXO2Zofu-e
zU6VnQ`n{WduR_<L3bQKH1jUUR5hU@}Y+W&qf;~9~BpLjKr3u02H;yoD0WFq5!owD`
zlV3)c<6iVb1zkPb+hWVdor&4jy{`oBCniXS!_V?;L6I1*AOolfEuCfo$!P|{a{{c6
z=7JR@>fnpf{QH`lzCRmlwa;K7naz*w2|4l5$NDsYfOK>GT-wC!e3k$<BV@^bQ;3@Y
zrOv>lfQ7y7-aL+){HMGTMidnB{7gaj%71;Cv9(b@IHJM2Y~h}7FpmYI$9ZV~os8^v
zH~k-kOir=D=$+!P6f@O`iL+a%XV$oa)!~=sMDYM+0K>YrTuVFf(@hxUV#{Q}-GP+{
zd1I`r=GkG|x-w6`Hyb<5&5~^>o;cZPI)}7GRN23D;4Vo@l6E-n)CHGD?$N9tCUqUn
z|Cv9qw1s@F;?ddUp{+aolV_=A9cYZn(TeB6)<f#WFRW?B)BPwt8p$4dOBEpX4=sxV
zW^6$ZejknP5`9bPCt3F@Wo^rN?l{p-ih&SkMU}Z<0gdR~Aw81j8Sbf4mqumyweNH>
zvdjl{Z|c$K+2Pf#N)~xcAA%rj;@6}i&8;^z*3k-zpGkAdPixF%k!<O@VFWY```pWr
zxz{ZXI$jA{$!6|#`WNqneuF;Bc0vPLE*H`XERXHtZBdbErE-hDp#9s^$Nt8Ed=I_<
zjElrEH4tRKq>wb*HHn^C$@S72w|?XAIxCbZK89d>uyq;XleJCl=$HQ+!j-MVp{=m$
z#u+j7t6-t`G4rR-*d=k=u&}+i_^w@r#d$tU*ZZe1>^){M2Jwk^-}j|xTxK~b<F{ra
zdza4vslS=BV0u$%V*p4|HMv-qda8%CRYZy$g6;HF((b-ck%|g#*vD3u;s>m-b8S)9
z=d3x#lG1Xw&(^TVT{*CmcIzoHuASJ|ja(0OV;xzhkGyO}h2Un!L&wFTZ~E`=h9XBS
zuITO0YN@at9=F6y{7b*&up{dK)&&2JcI>e<6G9WD0)L_%>IY+7o#P&(62SbnqS+S$
zq#gT8RyfE`SAy>X^T`7^_&aTEC9|QYx5~5E6h&@R?e#n0rB@>CiVx3-I^}o+Dy%py
z{}m;8Nuw_5z~R@bsZshQ0l4>*p|hnxEWo)5EB;Ig&8HJ*d!yD2p+Fli{#&<|r~zgF
zl45T1&zq5P?bQ44d2%-F?CR(<&|M+XY~DT3j{A2~mqQK<Zkg0-)Ee2sQ`U8f%M2%s
zn`PY}v}Ui2)q=(-$>Mw#O#lOxWsXf^X7Ah6OD-bRtdLD$dTlYi-u_b1KZ_=AHnsEN
zW{F8eL?kKIxka~S=GXF36ytzRS2p7vUFFgxmu^0}n=PIGMRG3?*Q1JQ@&<dQ-%vVr
z(D9o&pz)N%IVD0iFm{!RJYitde))EPE>>1$YTqLsDe4UqnXd|Q)Etz8_T0jJRl_N4
zlv2}18TQ}p+4x1&Q5Tx<GixogME0`XpiZA_=|jaAugca3L`#irv#7D{Gr6rYL8yoS
zW;CP(D4qb8MHA2(v#&ATU^DZ>oPUW?ZC1-}1%v1`@bNMacSF^O9$VFFdoZD-s$Lzj
z#{_mD&*03PnHN?|*Q#G>avFxV9WFLjoSUjcV7*E3&X6XtP9^iPk(Q0&&#<lkWnKMm
zz<iHG$H*w0O$+Q@%-xb*n-d#^@qYv6-h|^$W6r=wzl$hkq0NCB{9Rx4(g!}DDK$~G
z_bLA<`2%~)c_e-5-1TF%bPYhe!%C@yw+yP$G{?ml(`)Usz~&z-Ol0n=FB7-%;tl?S
zy@^hmuih@s*;=!os)}v0{JwRRNM~g>H4sU)shu*%OE$Yc3BI^%@ZhyXA<4FvlBet}
z@(JpA$w7G67{kboo#}Ecikdkwcf?b#un_PlM{0vtg_ac~WVl*txlJ7}U~tt4g72>h
z^cSC7=7ow<i(|}$>b{{Gt8+PzN5Qdwf^II2>4HmIkEL6L3328jS6O#wEK%knB-Rf3
zP6L5|#LopL*^7L-l~BhV)eSn-s2?ngkIXYAz4h`}Rsv+q$gwsnP_g2b(ClK+Zy3G;
z!{EfYe2_mKiNnfToYG7jqvUQKC+^X!->9NBg&f-{fFGZBltw0t67puz#d1_JRoDpU
z@=svjZn6UHYX-xH5|UY*Ez*(nzpxId=gas)nT8I(uQ&}`cV}@F0V_!CqtiJuzKj6z
zgkA`z!@Fir4c^TAM`YhsNDtirMWSbkMBT>S-4Al_1#iD@i8LOZ1@buiS8k(3_+Ev>
zsfo5OpZWv6O9lxIKEC2?NFpc@OI<<eN1X>8v`O%OPl@YbOfVVnP7Mv(#59?C$1~lc
zm;Lfz{97thqZSNy!Cu*(=CWvCLTE_!1|Qwy@w*Hq)TIEfA2Gl=bq`%|2}?-cUobFA
zoX6$8!s07}wUQmUXE^@MV3o*6v-6apitR@7tjT8o;?Rx<)9Q4+SpE%~QGUo4nlH6#
zgcUOwFw@FC^5%Db?uB>rhjL3ZdiprZW=n#V?~$@ynFIr5;sM>3<f^Dkwj5e`&)ium
zW0#j}s~dWj87CE_*qqekFh32<E(V5X_1Qy(1DN<H)*k(>BKv(>u6yUd>aa~c(NOVe
zSRN07^Z($Y$i*A!*y~1`Glafog~YL6+KvY2-MSa3F65p}bP?_T4Bmj>6-4PH+C!d9
z1zD~%qySbUub)6eu0Ndt@g}2j-re^MKYk~>lG{z1@3Tqacl-&W=Prbe_;>?s8+pic
zuC!rx$Im;>l$VPX=RB@lyNJLbl4_46GYbw?MLCl#Bw}z-<f2doEsL@+>`$VqleLNO
z9(C_o4J}y8lNbi@e%8oCpcssYXOu-tOZ9K0=B$^FUMj<sL{1c}P0TK<#6Fg+gvSnf
zfgLYdtdEgsV27|6uO1%DrrxQ57Y`G1+1w-Uqe~dk=?pl{B^I*}LwpzKY7WItS)!&+
z+%lTpkOpx+(}a7g#~W2jRo!YUq(CZV`vu(4xk8st{;wjiB!5&Tp#4<O<xw;BP67i`
z!fX;<_R9F{Z<bl-`}1$?L!#*}$uhn6ax#559Ar%R?@hu3+W$$8T_GpGyacS7jcm3h
zGtOHU--p~}*17=#ND19%<u{jJcr#2W%0cOj^ua&`LI_{tX)^QK*oP!6D?P?_AOh5t
z;DE*4-l<XLd?#-y+Odb%Q4+?ZrlO^iEuZH<LB%!MDnyN>ZTz-Bk;&W*40UZAK>JoO
z+gXec;c%tLOG7pq>>+^NLvB0_TBRHMK(odltnT22(WOE&*u3l-MQ6Bj>SN(yOiW4y
zW6ZS)$^-#{$E&6#$4g>Z!F2sZ8IOjVquq;~+mBoR>rhaCX`Prc67T%(X2~h0zkw;c
zTp1WK%y|I(vlxoG1R$?u9AGnSKI3L*f|&@Ctc+x`*l_Do+)Sf5o6GJ%O5~RQ05)dZ
zih^<9(NX}XRYLVi*GnI6)Sh>q?H}d}YjVW4pOfL1`p38$+3ef%BPEy;vrJ@M4OdxN
zIqapCR1u9&6W0<=56WLd&wjYk5`PE0bKd9mM2THo_eJBx<<`Idg8CL`|9?<FuJFIS
zIPhf6H%I5(zD;~r39Y>a0rf;992ZeTqYB%-Hbr!KzN=^;7jCVXZm(d@zIx&N2B5cA
z(D}ik$GM4+X#KZ&Zs!>gK!>Cu5qSFkiO&XS>_b0}kn_)Qr`p@EJ@eMjl8P{Auu@+(
zR<uk$@itWNI)R4dIl*+ZKiSwcWWx^gYqiHaVRpG;kKwG7y-fb>fo)`rLH;y2`o=$?
zM<~S87Pp1$O6$mMGVYd<iBOAFu0(0qb(%B$$JHABdy7KPB{$l~fVIZl&lsQ_Qk=Pj
zK{9m%668Sf()O;)<b6;2H&jQ#UV0cQJcq%f7Cd~;lqk(tUa{_WL6}~fr!3Vj39bmb
z60UiVdkcxmQta^Q-pwP5hu+*m6&8zCL>|A}-*<@CQHe13>R9m5t^O?q<t#DxHuv~>
zFTn#t)ffqI?E5Q8GNGq`KT;p8y#4kK^w;%G$edkL>JX0GWNc2g=^#R}{cj4yaDh3B
zX4u60;)|KA-|lyx8?iWS7lKV3OgMjlB7^;h?mD(B_=2RLoDs>_NKlQ6xzBD_f}yyi
z)mn6=SDV|Q-}rMs3`f0RkfRvdACd#YVh$q#U+d8vNn0|yn+Porh{045N7}kD2s(*K
z^?#O=18yLX9vCr(k%xbzN&#a8V@`(%W834Xn?GlvQ;K=;Ds_tg<R&c`Y|e)Q$Mh&;
zc<T6-GUXh{{)Ocx`it1kL2Xp+6jP75A1{B+TIj;c4egiU=iyMiBebP-0wD_6Mx*%V
zH|(PgPF#D(mJU4T-n=l4uXnaFS7LGt`^4UJdWUF~C*m+>!Z)bxo!LuZiuKVYC)9z-
z^GhKMW-u!zcc!zO;y%XHdG;Yfp6_C(UrN>KHC9_14?g(>u2MFBErK?SrhB*0s(|O<
zn&FLl*FZvl;ps$boRGOS1xvn~r2kj9)@uJ>3n9JsFMz9uXhxl7Tn7-!PgqIEEI5>5
zcw|klgp}()MAR-|a;z3?k;to>xaq?19nLNbLffBT@6g=Ke2eWW&$$BivoBc}fhcw-
zRP{K`S_{Imc)zlO5iV?Zvvz1pz@>xA=FW#a&rFt<2Bua3l6WQW!i|c=CV-hOqkF!^
zwO{X|_V)7kZHj{`ME>{$SoAas9+sam_pUGq?p3)7og{f#IwW(Yho7RU!&SU;(xS)S
zD);YvV%&Ol_$YjEo3~aK|D>uh<Sc>WMOA}u%{lL`5^Jqx%sT3n_u!h)vCj<tR1yn4
zra{wFwz9rxA4yL0D6;{W`_dcov{dqIsT^552%dpH@_|ZS0x1uoLki8|PKCh<jDeTb
z<oA2t1J!9~T`_x-UTa(a7*3)E{S3%12~k$_d%II)9Ce2MX!Kt8W3nIx%yGd%9GHc^
zk{7Re(|3Lc81FBxW;7`CI!>IJhop(BL0~JxmMtPu%D4%{K$G3pr>F)1a6~7N)Co&e
zEhd2jaRT!}8Mxc-d$);MJv^P6`G+nEmVKp*q<~~)BVgo9G0#ZHDi$oU=(VgU=_bWu
zqhdVwfP{M~Ugjm8l`J^jK}+^+Fj9+Mw)$OxNCVl^)Z^encKkgfNDbGTXh1bwc;=N4
zB;er>v*_0W$m3nbIkws@S-SQ}Du<fh(9jSH(^nu-%tNdpn+`CVie@H6{6*Ras()a6
zJ0{LDarO?B3|vfxcyr_0e>>ZZ;-VnR7^-ET4d^14Y}t=mqD1(ucKGy%cma$<J_*a5
zCBI`dqxQr3RWF*`?4r6DE9@hgVAL#vy>?YE01w-LFRHfLj)tUqN@pJH9vB#)<juOC
ztTq%YH->bHwdMZz1EFJ~GlAtiZW5JD+`x=J{#VxhcvWR()&L%$?G486CK&)1-}45w
z2yYVpR<)=lyJ3#lnzwRDUNPvn<mmm}+uQX8uoiNSj!xC^kOsC|Z~4*kd~$xoPjQD$
z4Aa$o;%2$8DL*tPRk8|mO>g%2ZK!P2(G=5w8mWHFb=iXkSECuwU{Q%(Vpwx4;m$sr
zYTm;p|1aK(O!gp;lA+b6J_X{s&l7YcF*7?ydxT22%cl$q>=865i3oHsG4}8WRNp~3
zn+2o4uN)__w273jja%z$OANN1$%%TdVm~qP8{%F2cq#u*ig!6(?>{$0UYh>AsF!kw
zr%7mnl!nE{9TV=;9)}ILBYOvi|1631#O2r5^Ch}}#ifK(r$V4@P<{aYydjA}Nya&-
zK?93tqJnjz;|>UwI=K&`kznkcp#1qJyy)#n1hdc6l2%yZrp}a^yL(1qT`Kz8g1e!r
z+8&oUb1*j^*44vss+;a?8q|6y&(DdX(1U+twpNq6#hEna)t$D9)9l`j&$(8Dfpm>$
zKq|azNb4WH$RNP9ap-vt5GW!j#guN#Hg6s3cGw*i9iasZQh?Z|loWZE(rNOSb0O>;
zFq+-aG@>!QpbqwYLT|vP>7u_^!+(#>bt9EYP>O%V1*nHH`1Xi2*Fbstw+3Gc;<WOQ
zAGul%&YW((vGfUDNu>#;*@?|XdAq%+FBPv?-@5PSN*&sCz}#-B>H1w*og1XFTao`j
zI+_!~Lx%?$X0Sd+*~-SIeSYd>tVP|~Mn!_fch`K$jesQgq!T=0Gxr)wg3Y)!$e-06
zN7i41Ix)}%2i~iEmQ5}p3w(U;^*Riy;Dc>t6uL#mz#@%?e8is2Z5C%dF+}s?cskv*
z)C7RHhK5Y@y(W<RfE{%o{|kx?O!!;N+;CrP=PBNodX@fJQ_=hbr|WCa;2k3Y62DB}
zZ$xVBj_5~u4%*m((2k!r9Iv-!>;=mrT*SG~M{$4P^`Q1EA$S7dd$@HE5ffGd?+PgI
z$?!q9XM!H|ZgDahnhE17&W#tGLO2LnVIboW%h}639%U^uxQYF2BuMq?rr&eA%jmDF
zS@=T$A5IisIElWFpp#=hr6k*Xg;Avhsz=_R2Gd2qJHnm3_}SgJ-?!J_=8YxnEX4YZ
z78b6f`KD>i(1Z#|f@@9p#7D_H=FEk4{9mBoSnI;i)y8X=4wC)E%NSH6=5Nd_b~dxk
zsTyY6JJC=n^(MQ968M9$&*M<j3HDA)PE+lZSOP+#NbI2v#HE(p15eW)4yLh?wgl{7
zRxc_UR!4+Itpur8{O_X!zONuJd3Ryyb}KWTUH7-YO45lGEk1B<KOeq&<2MTt!6$<?
zg-AYGhwEKT{_y<6^W{HhMzR6Al9AJD2P=-qz&SVVZBKVP9dsZl?!MXod{lQ2vfq(u
zD8B2^Wqfjlcz4o1GxABi#AK)2=JP%;(cThzW&*XFDSzzPsNi8EgYg_P+G%Y|L@poh
zvsPldwd8$bb#ke+b#9o&yvCHFz@blR|9I0z8#jv{a8O<vjiqhL7RS-tiRZp8w$R7W
z{b6h!fba>Z5EX#I=BKe#V)x#>ppX!Bvx|(yCK24LR&oD3cRAD~)!pwBXgjq#rc&-n
zNhcj-;qrwsBMbso<;rS&n<r4K-pIw1nG&@mfTQ2-l$kxDbTWg<clknSG=onO(QML<
zDrW#>$o^?wrs-;VW;vDdSE&NfybI!Pc2a-i8Mr%l^jD!gzl+FP>h7xy9Zk)|W#7|I
znYG&Uj^k+f0^WXBW-sT=CXKRXCc)`UDSn~(3qMJ~vq;YKErdqMRt~`N-MWqm%##uS
z%P@jIZ#*bf&G>fTsa3c@m2Q;3?znbxDedADmqi~TC8Pt+341Z?SNi1Wh_N1qyXjb4
ztJpQ+Hpe2Q#O*MLJHy<4ipL&|lz{WWePsIFM&jr9{!e_k#>KTXZ>f&-^Y;^8pNU{1
zSTvdHdZK~}arjm6zHGg;)jB*gifU&#HR(gjR<V7?XmBj)`Y%7A4`ySYnlxJINP_VR
zWeWKyo1R%HrNzFP33Xx}pHQx*pa0S8i1$?f#dKH<PfzT1;fN_u@f%6&@!qNSAFsO5
z*Vz*$?4&A_pGKG8?(3SEMucKf5wVS|=PGOB%QKVS94RIoQ;`1q7d;fHR4Mk5>t)~3
zbGC3N5s|E8={QUauGPICpvE;(Xtnv6w5!@fnJyV9Ni{Iwk3SkmjV=;EhGSS!VAG5p
zwSieHogVhQ&P=z3#V`41paK0ar}pQws>^O6k-{j#QeLe5;DQz6XtJ|D_E(rcJ3YDX
zgC>*`=B=J-5W(1K`PCimbX1JGN^%y=<B3hnKVAJtGS|o-ssmB<U`R8@_MMh{xu<G-
zR3J=og!}~)VvwjA1f>{)Bs`Z>A;L&qip7ButvaAJCluRd7cP2t=_XYZ?r*#HjeXJL
zDeT<UzqD=omE^KNRX=FPojSKpHs{goF5S%(yeE(sfF|pR#m`_9yBr3>eK!)n)b`<d
z1+Fa(x(ILk=X4orIT{!>@#aI8cG`weL-+#XmzlPd5|7HVLo8vjU+^0B>?wV;4{_RY
zvXk}+Z&y7B?NkH;eU#~me0z6l%Q~-*ognC%7M$mmMPRkp-RyvX3QFQ0Wxvie+u6fu
z9!^^OvxBO7mlZHtWYV`qYJX#cqGl7g7DI~@z~@CJ&N(52<3T^VNqyrVUXjo3rCA78
zBpU5f%sK@NeT85&F@YBdxL23wt4qYgPGxVo2cDMz>KjW}O^up9!>Mz2Ey_+8hywOd
z2BJEqEI;Zwi(P&zV0Y};dgP`{K<C*dje8VmGD-)^nP;wv3s$1Qr-5atMg-?mU>)DG
z$b9@;gQ|6W@4e%iAGu?S@RWMOfnZD0K+3kzN2|haoY-T|)SQ2dUkO2P9X|^9voU=_
zJVJt<#{Rwb1zT^jcb!ierW4;^{}$B5&O*D$7jFHDx92z!*sQslho7|4<vFnx3ZNYH
zg*=nrId>8$=MZE(fcU|+30vKll*1C8l$hCUwM*IPx?3`Bt0RK8Ai1EN+Kj7>t{6r_
z^zw|gMXNHj%i9~nrPdm+m-)rfkt3c%R%BwH@%a|gZF?+VQgsSIhFr4s+v=#dms|rO
zKPX`Qbn3YF2q>Pptaubg(^`lymc4c&8$q*2yNn6u6jVJ3qe;w}&grP~+V!a@#dFa-
z_PC(-;hEsWMKjUf!`C3Kuj3AG<63Odm-@D^?`_u}`ds;%1qnZSJ!CNrvcdi-ktzr_
zB`KW+s>t*MhCm)psk!o0aX&T8dS&6|hM>J)8bGl@ei?myVx0^g)V?~9`s-($M6_0s
zoA+i$<l$=GeG<-+AKN{T^LIlo3-6P=$2B@`6T~S+-OGGVV&z;%{m0zInqBgIXWXQ=
zL?TUX6K;-|bT~WCWS!jIsFWmaZ>AyVch`px+q?{XW56>Ue`8<e!5@)Nfj8r*($xpw
zx6)uQP*1i09}A#=CDc5LfPs+_9gDMX#qh8gqgWc}VQUhVL%X4yqvf^NPjoD70{Y&S
z6(w2Jd&pW&YrWIj_|driT%9kcz2&N}&iF%hXX~28UazAjiH{V(u6LZP{94{aU*Krn
zDZ41EVAe*yKo_=!GWNB>=iG+sKvndfYi9La`Vj%RK_2ceC$zSOb|(O;dnP@6#*}a(
z%I-e3P$cHoHo|F|4(YR(v)2eDT(W-bp8$Od<9n}_{3L@<ZMIcNX*X%S*FmoYixz+n
zb*|Zi**J?Twnj5V2k-Kpyw|phLPz`1g(!~Wo6e<utp7<pT+Oi3lk(=u5v99m%nN}9
zr42=-p#?t){>X$nQ!VnXcwBwG5_4(oFA~?=vhE&e?2rqX(opgkott6#mqu_qh)t*-
zFxYQ}HA_kF8z*sdChytq8EVK|(j!<D6OiiBVZFNZ&{=6wQ+Yp3S!(^5@@1foQ%J9L
z+K($0I-YE*Mh`BaoEtG0E>R+4Y3SYfTYgL-V1IN$VNKjSLzEG}=z7Z6CpcU3wfl4X
z&Rp>?#`)YZr0vNI_-NU}iB$HS6*1QfS{%~A#ec_d*NVjk(N{c=vszH5RY86Z^NbTt
zjzaQF3S8W+_CLXIWi3}|>wV%PrHv^C3ChJSY-bh%UDf*At^)n`G{k>DMN68kz3r8|
z{PDt7Z4$oKbh`e?ITk~LTYvc!MrO|UiGpHlhvB@_MrC=AH%ZK*70ZTyTz2)hFTFFv
zP0q-yo4-yQ8{YBWczlT_u_>8<@?fFugNj&8iseqv;xh=`&Ed*+Yf<Nu*|BCu(PJUM
z@02{3+K*&%!Ea%{yHO%1Ng2F!G^1IzJ@9g3jFwK?EVofIWy_%ugwK0F6Bewahh!GY
z!q<dQ&Lr&f3tBLpDQD{befwDf`ElN?jxO?<^yfUOHKA*A8kM%4y!TsE{k1U*h3)(v
zfl~g8^B*wv>afiLdDM&s@HK<+k6wW~pIe_;Xw7F5gx_q_&U~Os)n(nZ%*2!T7xpOk
zRqU#T$S?7^!a$F`$&Z24PzAAOQDBRPo5;RVbC=SYCTzG_b_>F?Z|!}1JgcRVITz<A
zxb*O8OU(8>KL;ymPlN1VlTF;(7Yuhtl#D6)htjzX<o3ZgHCIlj-~FPF5%<sRV)tmj
zwgua=qWT8$gLe{pCs=>8p*IQ9Rr`qRM-rjmeD@kbYOSo~jlL_B{|P<W5sm0&d;eNO
zB!?E9BneKE^+Bb?J-Sa_9ae3sBf(00Xx_v+QrYG#R*71^r?k@N-C0GPyn@%nS^kUv
zWn@X#fDU1HE|7YcJ4QheqexweIn`*qHgU4F_;8GnS(W%;bH}it{TX_C)+gF>E?jM?
zNoB|8<p*gxIhpaC>@Pk@GlPa8u6gMKG-fdc(QhQ_!SGR4lR%cEuzizxi|Hed`1hem
zPtC2#&+Bp7szr}t6G?#)6kP*bQwmd!`e|}vLM~6jwm(*PvP!6yhO4I0>_)_?ghh-2
zRCmN3v_GU}bqfysDy(w~2&lS-3BEGGR9O`HB(biBfTm*Hgl(hXZ)~YWjL@q#Yll`n
zbNJ}v)hV`M_y8e)zZB_7C$~tqieca;b*;LlWu}g=9ByBk(MwlW2{O!V57$^&MM}j*
z=0*@B=UGzTq)AAXfbs3eMhIYgKcUs7dtdzR`+r-ch5exW7=|^P!?(oo8oR$CnG=Du
zVSckp+Oxb-%FbNDfBu5GEvTHsP-Mz`KbJ3$$($WQ5*MyI82@vxkfiD>zfrDIGe#~{
zod(FD)cx{^O{Id+K=51&;kVSd3(3SW6L9N*bE_%4&05Tt1>}Ag3VAxAr-h1fo&X-c
zy@vQgHlB{nvWiY#Zt(LS;sGImkZmzt%n+}~KFX1s?7E)QH>$R#xCqX^zs!5eCjLZu
zAwBc@f?{kZ|6G;Q^s!ANPqG?F$*!UcQdAGB`TqjhZ6;CkiheLDw1^Y*J*#=6HsE!y
zEjO*6!P`fiv>jYyhNYu4VeqH$!XI}Zn>AK-Hq;Zj!xfk@=jU0{L8sb{mJyHXC(_|O
zt<2Ke311a+;2kyJs<{KW`yUbp2&xsSeTv{Jc#f^;!|w}`35wL`-ifIJFo0sOVRA{1
zZ>_)elqrWE-+{ik2`uSwiZhqv?X&yZ0ctk}g=nSQ_7cuFv=O!{ib!YcCvJ`2yinx!
z1fMhX{(Fs4r>4bOWsCGfIhVX3rXW~Yy4_@&S_LKDljW2a$7gDF$)Pz%Y$sVua3ngw
z#dU#CF29Y=1gKo(xGNlWwm(~(Ut>|b*u<|WAQ=e5{mB{DXW95%k{_sdqKxU}z*=l?
zbP@kH_d&?(`g=1dYd=U(GSimM?p;-lwwi~l(;<mtgrM$s6_uhYe)J$q&kdbwck)`;
zu>HuvsrfhGAm0y^5i*x-uf>9~%J-Q3BXJ5ukJ&gT#Nuey^#s_zCgF@_?HqO``TP|)
zq07j)Uu)QLtmXH-j}tK8;xbZD<~=va8q`n3c3IqayDbY)>e8X1WKi=bWbHbkUvrx7
zNAB~6Qu{{^V+0Xk{<&7R!I0ql9_!0@kf*hcKvRT!g%55s?CcR)r-opk^PF_<^U_uG
z`VpM`T?rAU?1GR<E=sUz_3)B^Yza)g_9iOB9OA$ILjR*J(r<|Q3^f~AP*Etdo)QGj
z)p+T0Z4(g&MAg^IPa1tA=&so3u5V;BJ}9p!QnR;L655C;&f)cQ>Q)Ju$M^<Le-kp@
z7j>$lf9G!a<(JRHVzvL)-ASpOXlvfj!CTG)UfPt))Q8ZfVS8CWeXs2sL)Y91RQfkJ
zJ|7P!^Nh&$67Q<J+^$GbJSKRST66V+8_iSR+0QQi<0yzuPoW1>mh&p&9;}bASy9dl
zptb60U@|;0@=dRXc!xlNiE!4ok_n4yga)8zes$H4$Fi`LeS>>sRlGt02!mc@#_b=_
zD1}&e$aymeqyl^$L~W>!mw(BWFVzUQPJQe=xQ;wg$9vu@ro@U=`Q0IGexN8t40@Q~
z!q_(iAJB~V{2hJI73-$q1`Byc{T(jszvOu`r-=w>zP$h_KJQJQY~ylmgMQ_sG4%eU
zs>^nbS>AtT!y*wkUTU!WYV(+p5aW>cKq~ftD{yeEf)i9Ff`uvuY*P{zvVK}e%^wti
zf)^IGh=KLJoL)PVH6~QCbOO?r&QJ|e;?3<qWP8#wF3lIHH_^U$nF=5L2i6gQ{L+~5
znD#>ZyZ@84|J4nMBb$>V8vFd17$|j0_M9i+Mz?*2+CrKwW>u|UI`J<PohKn}9{P70
zX-UiOP2Git``P7Vh^hLt`<5KP5lT+Bx9-2MjF7sBw6_whXgE7@*oTbEk*SoMwdmXX
zU~vkVsf{WrS8rT^qq_JsMj!n}$ac`nm~UC-m7Bvcl2HLHT-9`}Q@d3^Hv*rKdA^8^
zaIIr>o(TNCy!&mIP^HeX1S-E<9C<SrpHGq{wNqURSptSjKlWuXskX6a+!v=ZF2b=w
zyt|Mm?OMHTE_4Rx<^Jm+%<nOWEjK;@{Z`RA7>bu!HyC-k7sc83mU_jQfsbMhzT%Z<
zfH``D&RApt2yY#hqKvjpppKJ{OLgP4cnr0Dhvt>J9PiD@zRWYc^B$JEADAJmptdO;
zAmr~YWIh;AP&LVN%<G46J{1#hU=>~;$og(Ry~vrYydC7(-5^BqcU$cl<#0Ic6mTIk
zdTz31j?5beb`J1g5uO+VAc04)d;)q~_r~)x8a_+a-w#uZT^GY_hV#{iGfb?1CAhI$
zwi8i~Up(FykjH6@TK3a5O(DYb@j)%?qcXuOo=Y4aWDMfX8gfvAw?BUn!rt>JuPn5j
zi>dLldAxh@efp-Yvh88?fbqzvpKh*|P_dDw;PZW7y)#j}=S8Yk%4C47)>Wem&N7mC
zx4n6{+A}I`WVVEaRpVL7((;bZvxyW>t6Nz~-I5c}gGE`&eKo@@h3lX8?0xS`N6G1>
zs=r$V`5J~Swf$3g<%5jBBXTTL^~Z&XzBnOyAc-rz=MnL4ZuqMgrDTdS{wm&~t-;%v
zTxd<=om&a<R12O(;>)j7$&+#7r7?P?F?Zso=rQ4?UAT94B~riYln(^*u2wJ!%XQu?
z>PUj8pw)M~iO)kRdHl=U;lcYSKB|2RSVvj-2V3fEwJPeCf@m-OWGTdNcf{chsuf)#
zF3K{ie`bhb4{j{rA+M6f6OC!;BaI|jZtZmevdquqbiTC8F|?o2^Y;1f^55x$*7IZK
zlQo)ARqJ%EZ1crdI~A`}7YjfVLPbW0E3~DR4)Pm(P73Fpx%Lt*3#VW2j^Qu(*byX*
z1R9OBOSJ2m!6_ae(2&=xbw?l9-O5XBO!%HElXlsdtsWM$|LD>_kJuXf<Dq>1&4$cu
z%pHba=i-KHh9NS4xbkg36W1}fp=w~r>f*t8)XkVHnd>7VcX_RX%xJ>BP3(Q-T=<H4
zYIEScb~S`Qw01cm%4T&ep`CsvXyRpSYKx{M*f{zQN<zqUgm~noTP-G#ZynUu7haW;
z4fwaSCh)-oemprVk?&JvwZvTfC;24k+liW!s<YQmC|<QVEjTW@FA*o8E@nwIIKz}5
zJTU{x@STG8Z=%ut2IZuIM-#n2-b1cxLJxG(31~>pK`|2xW}VnB{9rntp7pnqkbhcu
zABtz7CxtF9*O9@?CoB5jIgYI}>mFm7uP*iW0+89N=w>z{wCg-{0Yjk9B9LJKbMJnf
zm;bvzG3R8vn8TPV>sRnU!7mXcKhx-9@1_>~^;=cc$tuy<zC5j1reKZW7?B<3T2vC1
zrQRK7JS?^9$5QI{iZ{k-{gJXZS);{0GW~>WpRgGStMDe?9IbfzD7K;?*0+J{FZmaP
z32TFrKUbMCh8lOe0|hiD=4%ioLR7B_;bj?}2Q5nV@q>L;(2#dMNM*EkHNu4%Tayvl
z!Lz&*;A>;w@8Z8sn1sd>!Tywn`<Y1P_G8{OQ`A4Ds-y=7ykP)KETdc+^b-+;yYsZ{
zn3Y5D$izJl@{2G$!@O(F`+!XXPDTj(?d~bk$J?X=yrN}-SkfiquR-V;W9@9G+SeAv
zwPvN`yi!a2c0?hqjHBLtAgz?moqhPnrs|rX6<VAVw0*Q8v_rI!w5d}q*N`XIp)!g-
z|ECAI1V}9fio#3iN=791J4Nr)mt0Bzw)Xv!tPaS(NzQms1L+n+bP@@Np0~pia2dDa
zpkU;nm?Q|MbK-ZVJ3P<IuG-}`=ddix`+iWKjO~DzE4`ZQG56JGM<$>lG|y+3Tth*V
z<)ZtIXy(Z6?XAsBaeL&%iD^&RZP%LvP^;|$C~&)M*RSnr$~Qe8vXiQ>DKGAw&k?6e
zu(^*ae@5lCoG{JEf@gLBh*mdxN3Ds?qrr$eYAGW~m4E{FAZW-(HKUM<20#9<^~NRn
z9eVgD&X#r5nhct_2?5vt;u$9+ToBbpN{XX-LodS&BG~>g=dr|{IK@KzNqTgisE6%O
zzdds16R!57y2{a%G;ItOE3D6?=+V&e$wbu_pv38&`AGq2@GvZf!s>xOlkSjXx5>sN
zt*ab1SYDNu$z%s{TN~Bi*yYmOStTI6xnj?0OLY365fNE)UE~1!f!ilz1U^w=Lk3Za
zw$^)|at5ZPQ~iB-iBjj#U*~Z$tkd5vM2{}WKLSrLPh~G(c?2*0sVQ<_V`Z4BcQ&qj
zGglGQ{)y6<gHjnyKgy{MY6-vpLb-f>G4`;ovk2>jm~lJcvX-fxq7}WaRop9C{QSEn
zZCsp+C_%${IP&>DzGNltP#Oyt2O_jI=8+9A3EPDbW?iS_K_4D)zEBB>OoT9vU*+4_
z_NFWgPJ|%;#a#jzJxea)Tdc1(1<a`z`^fCBJ-8T@gEvn$8p>(&ca*ueF_INktizr3
zgDxIva!+@BR>OvmT{M)Fi+HnC;wr<};%5X>+5|1a`R179j>iLci~dm}z)-aVZBfW|
z-lh;zA5m<N>dl_E^h0rrxtD3{v2zCviGB;6Lkx5C*Ry_$is(sS%)cR*@NI3^D7ihf
zQ+45*r5wcOrvmXjgUd(sU(VmLXo;7sF<%(_WmYm?{q3fo-ld==pd;F?GT19&H`rkw
zf?TIvZfly<;VD;6(y=7)vC77dzBjppXyYzxTxPXP6t66DK^hNPr$IdyEtF*kZbsE)
zp@?;U4rP-8wFvx6=3X%$cDZ{4$=gqOApa?qQ0ybgX=uPlNlo9uWbc1ct=lRFnMcfD
zWm_3!O@y=!@kl}b8(q^-#|7@HR7-y<{<|%+!WDlT!1a%oh5KhxNF|Z=y~&8U({~zJ
zu@Id)&7=1P{VkN!R`&k)3r29%9rV!%-IK-Z9OVMEQs;+d&kw2bd!Wp)$jeDLNk_AO
z#r!&<v7F!}oO0?=buE(8zXZ>FPLN}zSP?%S|LC`diJ9!ZA0cd_59_GD*^Ud+gZ4q$
z#3>Q`d@58eB2C@|M{f1?6KCLZ^8jF}ar7Hfbcqd!YnB1#MZWyGi#9@p4fAgnCvcBu
zKQlTv6(*~`<3xBUqI(h&Z!#gEk5PEHFP}beCj1G}4Z=UNxd3%)jbeHV-8KY#m?Lld
z&2I7%#S5qTKzb9h3BVn{HbfrDy@JOJOTP&uIZ5JvRQ3N>hIQmSyolK(k2}eT-V_Kp
z%<`G(e?VsX=;uEA(Mh`tSi_{ZO~vMA(b;-+OF9!dUUQW!1Nl2woi7P=RG^%^JY%yP
zs~J<=@)zVotIq-nV91lKGo2Xaqn|IXTn6p@UTLIrcaO*}EdS@BXF2Bf8b{Ec`MwmH
z5*}xkCXq9cM63sQ#TFHGAqXCCCjxVcL6Kjr8bWV_`QIxvm`;)W_k89jn$(X42mWR{
zdsfPe*Owc!1&olXDOI}lbE%jc{#yd!Xc<XBwx7^LqhJ?-{$Lil2)$8cZkU>}k<=o&
z{;&9`Q-Ry2OzIfK?8b@tVlYZt!m;btQiD5*#?>YL8HclZB<8Rj#=Za6NG5-@176`P
zD_JVs-4<*YtRuGSmOokd%z3}x2>bQu;yz%0C9ko5Fl*}gJz|_vE3jaTfvR9qYO{w+
zId_OYt(SqH)?~%YPvT-|Y}u0mv-`k}>U*ib$W_2Ts8>X4cWE7pb!=fb*Dk`_dE0m|
z9hW+{lI~--EUI+qy`BbYTSB>D_4#Y+Zy1op94Z+2Yic1olR<NPQHQ04w0MD%AFTh=
zT+CY<)9W7C+#4e-2V?(rsq@Dn;NBew4>^b5o{3Tuo1>utrDGlKPeOf$@N?B>nO)k(
zcgSyxk-B2G?Vh}LE8`9;ZIO0N9G7>Cy7GGv+Go-&k5g5IU;W2MwpO@rNv<jT<>k?k
z1U8&epKaTor!N!<7VcHM;}3@*3a3-m31Drk;*aM^xhJIUyJ?y6zo@7m{eY3vW0z_}
zDc%c40Wf|lhNlhiu7ID|agX4I6tF)lb|&y@C0zUQ0Q=uSL9&$ib+NB|_9)MD;>|^L
z2{nr7frES$|J0l4lZ*pihe2lQ03Wd^?;*1!i74sZblq!oPYm3nOki&IX0j{5caz?0
z?xB!~ky{8mv^4&oyXIMEi??LbwiJDoexz{M>7c_3=Zb!KIXgum?9{s6U&JevQp6#H
zlFzN2veP`r7}M{UJDS#mi4Y;ZA|TIv)vE&7$MKC(;@2YNFGfpdUbCtF4II?u@gLIj
zO5tb4u`#Kj3ZrWFp@7+_Mc^b_(x(}V`1sDZ%ke$QN$MoW`!tRQ8b2L)!zcgwI98c6
zp#%S+oQV}Toskln*f-gZD@m|Wi~d?3JMm_{7IW`mxdv6??cwIb_tw~~V>Er{-fMMY
zW~d}RNsWjNST%yh&t&hXA?JA%9JHds+)J7A`{6?($aO!?2u+L^3pu=f4-w$v4!3%t
zs=V-W$+{zK@zu#M$swfS7ku4;){o4$c|fD5=K9t)nlH>o?x!_8^)n6l&o2c7v~~_m
ziJMgH^G-mf5b65m1DgHbIHN{TNHhNGA7Fs@X&5!nYqav3|0QFX$N1AjJkDA_<rSC4
z@v2PfYl;s|9}V~+zt2{PVd(^rCoTB*UQAC}G%=tHn#tNkL%*Z{_!V$YTqvle<V<5;
z(IZ7{peHrzH9KjEy`F}WoP}T69II1KzjtH8^djaUbO`<j7jBxenmX;Rs5$R-BH8V(
z4DL4qZnLDrA;J`}mMgFc78z>TGrqqHtyBgb-U<43A5-1aDH~{K%k_28J=x~fDN5pD
ze_227-O+=y$UuhsV?TYqjdIEboMR?L?sDkC;tO6A(z2gFM6DeH79Jmwx1_&_G>K+r
z2|TR#$Md*^z7I|Q{=x=7G=5r4nqW%svgX*Tzb%ZDyzpSenMNn-D7q?8l{}xn2S@}y
z#H;Dtit1SN^%IgyNyk$y8)E39yA{e&8Sbdd&*aZPDN?1(kUas-yqb^Aa*v3wQh+($
zl|J6aZtxIa5<&akq)<6VP_!j~C5tnk&~GD@w8YPUCIGSa2ml81>dXD6f2BN*i=Doj
zMje3gVWLiA#9<B|0A)IhcCmay)1K??6Ozalvp5OCLJZN}Kl^4^Xc11R;aKCMhwRM*
z)V}6w_X9hu`xI{LQ6wA}4<RjAr2M-gw233j<7D}r|9t7FbWIfA93!1;7=4af={{-^
zzI_dfv5^&wt}vXGZ`cn%UD3<gH2w7A(fs1#VC3liR&sf`t<LGo*Sx4=Xxv#oS!$C4
z^qV9fh<!nV$;?A*`_VggD`g@T%Ye&KDU?m!(rJhDb*s-sY>pu%A=yF6`=B{L8#6}$
z=y!SeDWd?Fa<8i4liM~P)e6lHOJlvD<2z0LIMRfn#nHBdVwV%OTaA99w*{d}*FExO
zd92f|8tu%Q2E!+qIV|8}(*O-ZRNy`F>JRb$J1=NaVfS<Nxzyv7yVvCaAMv6z^jFLT
zj@RJ<=VFal`4X1q?k>EfrBoh(pwJjjXeTxE_vA|1<S%_;re~F!e+Kdo$OdM`hy`9c
zqbc;JhK>$w7r~Ohaef51;n?R$^Sfq;ncgV-Y7bS?Jlr)LBgEngyg0zzbUJNQe);K}
zk|I3H-?!C|5GRG|jO{7?ZhA)`1FGF#sfFxdKt+y*D*!^ng*DljvA#<1Q4<5QwO~5P
za-4&cdxdA7>&df_lIi$NV@K3vIMHdvdo1!2{3jtD52tQDQ(S=bifG8EGUBkiPJK$p
zjq&m^Rr=fO&z6@E5SJvlhF#XA9hldFa@-{2iHp8fzmc?!71N-A9SW2dG57ahFJqSC
zdp?LZzWpm6J*qPA=ss_q^M~}Tc}Sq9oH|>)AqrVlS?-<~8tQ(?ZMn}C-@1+I@ysDu
ziJmJ7?+25g3BI!$e_W|PQZDCL-qq_<H2uTw!0jf6j!(%PAVXa=O5)9Vz?1t6>nIu9
z1uu}>kzYs$_;c*m(vKXG<_fZZb6o`-f!h{*C+}oClPPk_T7NeAv}4BTp<mwXw5;{{
zDfr+ie28MO)9R?GC-NM-#@syB&VCiRij`pl0p=<%J;At55VPqj)y?@!uZ66n1=A1m
z3G4&cHd~LL@_~Wg!QKL_ul_q!b$AdSyWklb0{e?1u)p*X!cg6RS6Ut&VIOUSzI3l?
zZJBvf5O1!PtDQ{IWsR)n7)`4G3>^L8d4k$$cAd%;Ep4Ksr*Ff_A!}l1)a8uH0A{OF
zt2B|z|3Lo(V+&vN$L$;kfk?CH=H4sUN<$<beVu;9(Xiu0;RkIV=2sPVE)=lJsRe?c
zt^E}rC}6F_)WP6E3!BSeSSS>)E1d4t)UE^|Sgru;hG9{*y`wps%Ag(43VG5;YJ5d8
zZQVP4Gl(LC%Zfduvb0e_t~I=^A$@q;vK4!|%@9x_r=;@;Op{i&;-pSyl7w6?E-pIy
zuS~<KND@mPqW9o)`cAC?L(4nHMM>@g0FyCS@!4?n;NJtw*oC}AXgo_vExyP$0iXvU
zcjP61^#S^%M(JYo{df4ualm<$af{_xrc1^Pmjx~StLGVxkO$9ja+kB`UJbq?+&^~R
zM2+WWU#qQG)O3kjxaKs79f5<zfUnTX-0n-EOaVQH-&}Y2TxuSF2x!2zJfi!)g|bOC
zcCJX{&mT(ZL5Ex9MO(VHPunU2m6m#8IX4s)pf2^IWVk*>A~&$aj{nJqZy3|;0Fy`K
zK{lg^quc--eiR6l1!yi4!Q-ji4pWp((inZACBFx7T(x_xz6{Xo-@Zew_^8mtYw=U(
z9jb2vmc44=mkKC-#gJ3g{ruU*P~y)>btBO%e#cFXkJ=S?mY>KO!1d3{lRxPq{&XEu
zB!GZd!Fk5w`)QREN!!wU>2};jS3EKMMfQaf2L`VZo~1UbCP(raO?52k&&O+R`sTlp
zv4ngp?;KtdQdGoCG}$n5>;zORR$TB@?Bjg6#HX~ZiOV@v_xTj?&f-8gq>CQATFS;(
zL-ZGBLOj3t*nO*ZmW@uQ`1lI4RZs~fF5fquKnh6gWjUZQ_Y(RP3tHrn^Q;(N9TcH2
zU_9f^@OFJ_I=R`xO!?k}cU%)?{N|9FydGL*z|aAT=qP3t-#nr+oVa;!Onu$Z3#?ta
zSV#E3C}0FID_l<`+@bWFtRe?b#n?}1Iv{pwuy3NZivv83z#3n>#CD^tA!TA!r;62N
zgy|x>lvN(Jqhd_o@XP!z_V*@sg+}O_SOP<sf)C$Tyt0TWVE7TxSOhA*cTqp8&@;}N
zB4U4%2X)c0@~S0%cQ^*F{Yf#ldy+?H^uk~sEe_?@UZs29sL{T{@aGmTA=Ue-2JUDI
z2<C9}HC?A(`*;;a8^}<vO&ixGN_r{-J=RjQvn!Z^0;ba?I`!aXFuj8MW%hdn{t?3t
zT9$AzU|xPoDf#!B7gNHb)-QI3S(-f&v7^=GC3bCXm}#FfrtLWAdW%0Jet(i&?C4*&
zqQ~~MMm0H;!|QD~uj!CkEiwh|+HU1Jp(9*ldmn@h_}%tsX(HyeF@;vVp(y~|_b4oS
zlZ&ujS~0N;YIgG~iY!{&N0+xmi^S1NI`WX;nvjNiG%(=K+iuz#oe@K?aC+|?*8YLy
zGygv096IhKo+#{Mn>E-#>Q=VDGEnVhe2&*Y?!U)(`oYgpg1|5hx5X+^*ra;6vun=4
z8OXmA&to|$3g13PW^w<V%Uycsd8cn!cZub~Cr(tyX_;mt2MXj^X-K@l7NB!k0qTo@
z!$d9tY@<vCm3WcA9{c>8pGm0zozZ8sEbM#2R%ww_fleI7MH|wLyDq)c>tHdV`GSi-
zykFeyUhknHJIpYY(Zp)qKQ|a?=hB&AeSVJ)PGDpT{cE@EXVY6flS_a3CJP+TLnY?n
z8k4n2tk39~e);>GKwWZVbI$M_6hr_1?cUzJJ@F>Ne0+NPsX|ZEGr{=Z@vzqPCT0$K
z`o$9KQ;fuTWMAli_`D=n>mDLhx(;9x@0&ONA7}l4XgaH~Hn^^7Q=mvG-a>IJP>Q>|
zyS2rwxTa`uN^y5DR$PL+ySsaECrEGz{CVE*`cHF^qh#;3*Ua2=7dA+dUiTa<JCO<8
zSWZapW0+Vuf@5pfk+6j}=~S~$XXB_$F_fK;)9l5bKH=fw@qR_kf2B>)*9a8DdiZ(u
z2ye~3E38C1ii&ac7SasSACg5=8jB0iMBDYJb7)9Sz-puV7vRR`9#5}(CU{n-In**m
z|Cwo5xTB9fTV+V*2|o44VT@Io@`&IZCaPu&LdQ7U{VwF@=CmFeUIxK%7Sy+l*>lRN
zY0}9#?4gX?&g0TDjT+~e4(Ooh3p@SU4dr`O7Brmcw2`;0kK|6XfD*<ZSjDC+bc-cB
z392GT8*qmE2Y<{rueI5*u`_t`M&z6_#fKO%SWF$hl!`X4@w`!JjF#5I7tM7$6VUBS
z_rlBQ&nbKhKZpRAE+h?hW)`!yOUL~o+!0h#w<kv9uAM#xf#0rUL$6wXfyp7MXMzvo
zjV`+AR=;4fEARdM04;hbTZIS(^NJSt4#}cx-3IAJhPThXPA?&g)q@+E`Qs%F0)}Z^
z%k*N=r|To@zE?PD?^9G%uy($R8^;j(lW|h{T7j9vy9nsF42ii%Vb4eKI`{+<n9UQF
zK>o!xe{0z2gb&ncn9K{Ft@BR&58nJI`MY7eAh+|R+*5%;az_#d9(~6+jz6r33;EX6
z%6Jsj4Fqzs9dv(LHslmx)AXDbObEwW5~Ss>T20DD;YsbtNBU&h&L`f$a$_1xLL8E*
z%kzuybiFsTMKZ-snl;7!8V?}xykiY<8qf}Gx<PIj@HVaE87;UjO<>^5alA_!+J;KI
zHB7bPVGXc#P60ym3N=Y=D)w4m(1La@wNaNj`g;HLEsUSEI=h|iKh^{GTP3MedR}V?
zjKA*`0*jg~>6zU~v?30FDxpTap%U0TL`|WS0NPdlfb{v38N5y~>z~khIL+!R#*|k2
zogj!iX5+BT!fg;3X*h7ZpsZaC3J-IJSJ!$EJ!eX5#vaY1>%Jlx`GxrIsq0m)_W+AQ
zglkk>iPXYn^_8vf8op8U9(4^b?P6oJ%@fHkk{iI9f18?eJvFea<GR0cT~gq8-Fr-*
z<HD8~yVeHCGYx7J4|PW^=H{P*e7$d-Cb*cVD_@`Z(r{i^U)l{&$CH}NJ9FWl5?CEq
z75|r{a#PAGAilryE3@)OZe+ZOqw>|-p$<Eu;1SZ3*h3fm6{kNu6F;)ce{Mip*d=u1
zKE}pb@%<Hr+Vb@!!G29gvNd^R(V53w>xPfSrFiq9^56t#Pl`9qS6oe&;tTXxD=adQ
zGDv%ZR_wpkS|cO;9&m1YN6uVEh>H2d#C{oKEdjeS*U?^5NB!0V(%)T<-?8{ywMVcM
zOw-`%RbWklcG?M7M5^t$tKOFKQjybCLE#HSc{1IzulU>QYbV9RhI8Tp;C2!DmGCf3
z02_o-5k+gnekRw+OH<X^(kZ;;!+qn-)7c)Y_nrd@v!YP$tIZ3%aktG>+E}sF1?o|N
zW*hY1KP*WcA$!T$a&n>LpPT?}P*5F)a$^t@ohfyCK|(D+@`I?_8NOYwEX7$$Ev~>y
z-Je5<8;Xsws+IWs4-<m`#}1C9xJjRc2ScLzSY8-?K02MjbIGnqDDN-&oGuyu(tm(M
zfF=?}&9@~-OlF8?S8#f%#1O%5tM(2PfB6knhGerOH^_air~Bi(V(U^VLyECGIVnj4
zu#byTfL~Z&hL@xwNIJJ-!~CnHtU<Cd<PhG@2W16D{wxCK1NchaBT^o9pxyK4YQF_u
zX^}R<n8o6&?6oyW6t$ry;z+qgwZmymF$movMkJ}-@fSi>#UeL;NA`z%z3F$I>=Q`>
zcrHUsz&p%Tf>d*VIhcskq6sdaf;-M6iLd@mX8!6yC&EVhb+}#83i)d0KyL3Cy4D?l
z8d5&v>lEvcM3Re*i>|~FBS~^pH!*9GpJ7!ddNo8tUS#05+;4I#WRNqt?!FUcXiBbh
z<Y$$fca+7T+~OUZ=^3DB<@LC=mtq9Df<k9$tG(d1794R(8vpSeCEl&Lh8AmM%|*Z-
z^l`O*rY-q_LL#%pHQ!_XTY{&sp8@`>^_Q9II+8wa^h)lTVwl~6Q{%N>=Frv05afsH
zv+2AfTX4SbfMX&$%~&`%C+dK(KCWGYL{8*TtH+<C|GvzirP;@`KJJIaG;@N$%Ap4|
zqaPwu05jVU@7piMYfRiH5cxN`-dcnP3g<Kk^4M85OxtuI<6pG>6!Q4uG3gETS2C}?
z%TSa{!;wW2%ug!Z*}L-h86en3%6Ob{%2qPk_I;gObfA-x=5UF_J`c%f5Y91${^bQz
znRO%r&M>lHSEBxwX$H!Bdxl^0IW{^#PRSZsk~}ocGMlgUfLqQ+t-$FntR~yq$7EB~
zCVUj&*N@Zq$A3@gZD&Nm>4z_uw{wURYOWMt5ueDde<acRN7>aTgIr~W9}6<tzc_Nt
zj+Tv&W{;9y^rPN5Dm$~CWY`)ghGP?qN|lS!+4}+{gg6?oV+qdv@-cPry^T@-Qp$ED
z(eZQx2xuFV9#rH==@s;8EmYfg4m<yt%4`v-_TEX2;9xtYj}!B3aK7F4235OXmVuil
zEt#B2@PcUn(bu>M7l6{irT(j4tym5eX}17Yj%x9CfsX>o9K%N6A!S}(_}llSzHd5P
z{LK~^b$d|ViZnX2FAO<>djfYuMV#w50LQMc%`C?+Dtf%ms#zj^+%roP+*Z0StbVz@
z@~nqb5gTv3r(^+S)U8^5I$*b%Da<?5G|9(}f0EdDG3g*r`eAhtv}l;(6o7qv(9$@u
zLCA5_8+Z>oS9R>A7O6TZi_C+&BwRUtjAJx9MPye%+EL4i<rb{tqKo1Ackt^%?8>b?
z+1xn{<rWH8d(WUf{O_gL1%02QA$Yh;nQ%Ge&HgyvZy|aoK53DO?94i@8k8ervB_je
z1t(DHB*mll-MR!GR6#)eWcRzO&(VhA7ZB$kNWJ~HMFXY?T1&`6duT9#T4UcY1+;qj
z@^GEPLB&#@koy^bJIqSda}g7+G2BRybt$sE*-)A`f4XxiIu>p*r&+s9>`198O%!o0
zR_^!V$eu{)4-^%JXY(Qg(ZxR=!gm5Ky^XG3L!$1_#!MkHtZ+okC!L^oo%{=a=fS2v
zmaACO$CHA?-){Uf&tIejOLZd>C{2F8#&Oa0=`f1U>pR_rra>OqsY6|rtqPl8Heb%x
zrJruEMK*c!(*B@{fHh!w0tUGO(v4AgDE?gp1cbw=w-nJEhL+Cw94TJ`ie~UN+C0IV
zg^+sc5dv1-!yh31G<W(r@Awq0$$Go_fqI*D0ck$_5=ynftIt)OXVQ|9D(PIq`0{<j
zKad19eI1`i1q_8=uUZV7e&zCR2|Mi$%{^fv6Mg3CBS*G%^!Jps`$rVnz#zrC-;9by
zdN@u}1TA>4&hL8stx_JNn-K9FDOD?2${y4~j%a>?D*mz+Lg9(X5$oO;(8i>qGM2dW
zdJli=-tIaM5i%cgPVn{K9L>_NeZ7+J^hcnF5gK``e-VzG{uC@Fn<0Sj|G#aVvZ<Wv
zv0PT_`oJ^iF6&**k*M+0MK87&UM3fnv55B+C0Vp${Ky(BgXnW~UetSkGP7eiSP0;}
zB9iX+#WG2sbj|Yw^g=1tC7M_duR-ukI80?%P2qu$n;abz?ETR)E?mE9)8|$Joj%uY
zx#u{|aG^3&Z}>U&y%oo7^pUp(KCQ8MXCCvE**=aqTH+`b8e3%g4c25{cI*Jorq#A;
z+P`*$w~x4VW3y!UzbQLUYeu~(-ZiQ5=tAx*3c<lo(@s4=vxAEPIF4fr8w+`@Ip&rx
z{dI`+_|uH}V|lrI+Ac%P)Bl{ljY3t!iIzH}hB^=Cr4)CZcuPWKpMSYhSEaeLk-V%2
z-@NSN-dIfAOy#KO_6Iz1GpC}s4Kr784~`qYfSzE^rt5<<hYD})f9dB&edBVXjQgax
zjKT%WN*U;_uT=Tvd=VtC=2_S1OMli0d#qC|8}2D|ooZ?E>MNhKhy|P5q#qdaoBUoa
z(JQ0QPzgf}e@%Jo0rDoL^DUpl^Y6vby|kXnYX1IG2_Pwn$__=j1-L;@Swzd>a{X8M
zPF^!y_iI#*s0b{{%XFk>C^%m!l!6-bG_%o7S^8blXU!i$-^KF#IZAxz@r|0cX5T$3
zN!ka(Tz<*kX_*8p$g^yr+RwsAV#w>=%k>-`$@M7tjBm^7y?;fFH~K<!8>M9iulof4
zwDS>$>3T#$@`r1NxVMW6WyhxmEuK-Acr6kQ{Yjs&w=ZPUZaUt9lhq@4V;<o_uBHj|
z(-dNfYm_*J#x0<*-H?s9L!L4iK3E>Btd`h@aae|TS+EoDjgu@uGs@Gp&0AERSu{nj
z;ARYbasdQfiIG5e^8IkeQdfziq&p%R<rZRk-)|Sa{PmStKc=kO>TngBj?nF_177!%
z0Bcrck4rgjg;NbPUq=DS2MsAkCq9k#=j-BJ@e9nC0c*5o26^PY7)ls)eN<HiI_HFx
zTl}Z@XZjLIt=F+r*c(|*U714q>3z@l1-OTN>86H4rd1^AdC7G0gvS@lfGqMIk-r2w
zc~Xcl!;prF<+2dlG9-KWQcN{E>jd)R-r`+&GAIPiIBo$R6hh{NTU#aTA(Bp(_<U1b
zGIS@7{ud*Cq0(3dqPUoLWRxd4Kt!_%?kS_rDgSn`5G$hB^w&oAG<j9rB|rX}PC>OF
z{KUmZDRk#{=H1&p&Uv#LIGx((273kmQQl3e#J4ZsfH$bK7;?hj8#EJIve1NbjGi=N
zERX9AL#~L?M`y|xm$$XZFSTQ;(421k;Dq*0!`%hi#mpAZUut&*kkY(H=gbV=p+~Z)
z-Pvmg9~h@mv!&%nJBRvNlLXPz^>vtMW!8p*bYY7fPW0=8RcY%LroBQC@a_J##mzdW
zEa1d#0+3mUn8*+^cIRWzx&8&OEz8I#;rsaCoOeQ7cinApH{NvD!g)kfu(vMeDS)?m
zOLUb=+!$lR%J}SWsU)+Y<y04h4)fFsVX(j{<>IzaS^$S8Lh3KVPjWh|S)V0bpS3ur
zR<SmHB>4r#PRxvsMelEe^JSV>fMgf*RP;-i;w53Y8)OLOP2mvc$zrp)VoDQjM2c0r
zn0zqs1*A_BXJCIrSVvboTv^(m>svqzKXcg!Pf}UPy&6+}WdiW2G0xVXpHG!Si$+-u
z=@yp<6jK`7AW5I`&Av))=Up*GhMo4EFH0FTY<Sw9mk}zv%5{jv#?$ltlpuQTkBw75
zv$*J|xMnE&yN%u=))FM3u>rSRoWg$_LLR{=toa`B)*8t<iR{FX2Rp9cy`MIn;o0VE
zVvzW)M3{?z!zH}3?MEMw%jusm=EHlUP!yJjX)Q|KX$9FI-t>|`6a3oVV@aICt-Sp<
z7g!YV@DrTlX6T%8Z5)klhIU%K0Fbwre##LS@_TufaMJg_?61&kbO?g<vxgd?YHrRE
zrcqi@<#zuOD8RVK>)cPbeP=xu;$BCSz8X7;Gkl$*#(R;ENtp@8_ZRZKK9ulKI}gSr
z@l3dcYaHd19ApG_ge#}_)%#P(Z<eY2sj4D4hsmyx1Rn)Lbed#0@$o(>+sIixS30B7
zT<fhw^djZMj>{_F4Xf%NH!m7`@teiiLTNa6p^B{SyX<JN5ErX0O>!6+d6>7m3h~9i
zXe)i^3&ff1F6SA%?CL)c7i<Jg6jyg~oX?sImmD87SsF)mZ%0Nd^}B{+t<C4F%SXr6
z+RFiiipG{B1uHm9>7)xIU~UncRAV%yU&<MuxhlCUt$v2(bg}+4AZ*EQO9@TVf8X2~
zgjzbKmSHuY^L1|p&n2dxBgK6$!gL*fhky=#pIiF64Bh|^D&j39R|nE~L9^(Cg`@a+
z2TgwM{KMlRTr^)Mv*nPUNUft!S6l66lnHoBAWwF|tq35FF}@x|$`~$UmDP@b*4ruH
z3X|wS&l>l;%en%GD4w@elC8s$HJ#@t$DITZs8L6)1tr%pF_wflOX04eT*ASb4U#gN
zrI(S10Vj3V&)svhkPp_(wz3KVY($FY`ndv`)y{;AW*AgexG>rJ(VHvaBj7>Mh&5^h
z=Qf_m5$(W!m3!E*RgFl+ux`PiB5H@1vM*tYOT&oIQ%n$8a$!Bs27FBz3N(wDi+Q<>
z_qO%RJOvBIj&-gTM6@b(%tPlI-;E~-u@)h)JFmVLHopdR$&d^Ll=Mh#bTFRXypUxi
zSr5mK0YVX4ltSbSS@lwSlBG8(t=A0lqQDN!<`v0Nm!!RBX|gX;$Orz!llEqRl3i(2
z#9X-M$P=)KM>%nqB5(9L>YFm65JH#+#tgPt@ce}}R?NXCnTVKR8nZZ4QJu*pEh%wZ
zTnbmS^h8{XvjlP{-zK<B{hDNCrD}}--FZ@&KD3K`iT{8dpVhjjXCCtUA~?}5`aRHj
z>f%SfCaUYNMXNK1)5Q_@((6TS=?8EkEu?SDC_I#*>hE^PQ#b5ROR$+rKxm-565iIF
z`mkx_6nAkeMMfY&?ivjlhptg^u=>R@70(f}|5ffve3+fteVHdor+2+z{HTHK#=0mj
zWoW%`?<Xcnfbw#^?GCWsY&E0Hbm`i(8+Lz;W$Ey<q(0WHu^UMFFUo<%BucRqswCLS
z;p^Y}NDD%@O(?AK$SD-|YQ2RZg2L>nc$!lr5h|-Oezaa<EH>p{y;%cAVw>(!DwO=I
z@deF3ZJFlnH<6fmSjxL7&uwK6#w;(K%@+A6m~`YzTns+1@J6i7kJ#GuEw<2C-QoJT
zTV5jbtUIl(?Kxz(n#%~S5T}d!cAP9gcDpmzDIc_>ZOI7fwU+F3>8>8@(J|JkU@OBl
z%4FO=`;G?sz8@bV`AxF0mj3#Iiy)X|VEjOJR;5teidD@M4FQTJX{AsvL@&AcE|#^?
zBP312N_o5`5^g#dN%c1L%LLSB2M{ssNYt}pB7h6~RRh&;U{dG%uO%_eCgq|+2tVwH
zRZ)OwGs+cA0Xi4hZ@VsL{rk?LSI{*#OYLWJIK)rV;P&|w&<idc9gLvD`4e`h+HA=0
z;hvZ;h-1{EEjJ_!&HS|%UQu=wvd%JOaZ(1*)`Q)X26!UND<rMbzg*+8LO;&L8SOMY
zv6f>*g7ecK%zsh9WU)WhD!JlO77lPygwPSTPVf(gKH|@m>23M+fp=tbZ8R}LGq$NH
zc${+>8vi@K<OFV~7*G+fe<NXCAlp#xN~B-Nll&ARp%@wxqh#*m-G`rgCc=7!59~Xb
zTja-tGm=XZfBwlwP*@M?9|tA#N%8DyU^xY~`x!OCH!7qY#I*<N93nyo_}SB`<n-BU
zJW`kNrfB`Zi^#$VzifCl^3!7HOA<984!!*KGRE)TEqss_k{*EY{I=X0W2}L~X^4t}
zV*2vh*!`#=g&>ZIznIVuXYrb_F0DK1HYjxOH_|Sv>u($^>-*1G`Mkcl{Bp4jzrk?I
z&=I-!g}C4EKkkY0I#v#rJ4k=84o`57aE^0@o3N7;V_sc{e#yCxNI|{FQrRIVm7TY)
zx!AGMT!H^}lxeC8&*RU$$ed<3J0EzCKcsP5KDIzp>x@;ii8#zcPu5$3Hr~?*hef1W
z#*zL;UlQsPmA%JeEd}hvgfRX~;rtdWL9?q9qDpMI?#i$69RGS3Kv##8`kSu=pRW6t
zXoe~e+VKL&$c%l6&i;F(De=;B{kta@?BToXw*XbmKYyYk#b<s!uCgPoMXy9k(J;Wx
z=#zR<Gr~0X<$KgtK2zK7`>_*{Ix0OyXI5xo#)3Gn+lvD4bAk>8GMUd!yv%k^`RA^*
z_o*3{K&-KRdwKr-7TLbs@b!_`p*tMYA;w;}h0|>un7OB!Hh!vPk~bs#)hLdzjlUkp
zzujl21=3IpbQ9L&{mLW!@D;;a$Ua`Q_ao$dWr&pFUxV6I@KG{6zVN_rbrJr|XjCYm
z9AP(6(|dbh(BM||{_I$Q$M%zbhr~P<VH+1HKKUXcd%}sZljbj#YeH^&-3III!0r}D
zKicJ+<N|c7zrq(2Qo!XWPK?KdtCcbz8sOlv;4rT(WY^Yuml6|p#{MAI`NQJsK>W+>
zYjFx%j~{tel;HQm-4n4-$?C>J;|A3mFiqI~W#ztLIZghbfmv>^-PVL}bRqOL>wUq+
zsf{+7ty}egwUm`AVKcaWDe$GidX~>}nSJ&(m5T@NHw-{=`(KnKMd3w)f{gKOu?(Db
zTF9)u=Y}KiREV{7J(HXGXoz3lB@3jD&@G6tJROHKB<jN|V&NJ~H(Z6DdFsOY`1<2y
zaOaB3Z#qNZqTov6kg;|!g>>F11arl#;O0F#kCEx)Rmwms;YG=Ko75C+y_FjX%;sCF
zpMMf_^nYX_<abHJ@(~Wqh^%sBtbBooAfNVTMLLs|ObJuu|H)ZbR5~tNwfmyaj>d&i
zyuz_Vo$j1Up|x(~@fN4^0&~!s;}(%LKhxpn5Ud-0Z!wBA1%SrCzqH@-mIfIR^>jBG
z%3Uuiy8_*NkoBt5>;@#UTn#QywIIU`n<T?>76$-PD-z(88yPH^McW)-8$O-81Yj}$
z<M;cG#SJ5f@*%VP3L#_6Xk!$Z;7U6cxe?87Vo-N7+Sz<h_U<9?IS^>QQPl_5YI9aG
zYFSvPIibJvrJ9{FVJha>2y8+-H45u`QGXyHBsem7Y^0U}_rl~u%RZA4j}C4`XTrSf
zFyc;@QLtF6+(aK9?Cze6n;~D)%0+ELCMqJ77C0M9opM+u{*$UY4tPLP#)Jc$<`P23
z|Cs_T;d*=?xGrKLZA0Wc%-u50K~<-Gy4{}3PdIrXkk6Wkn$ZR49KF5Q@L#itXKZK)
z;d#>e#9Bd_ZMPR4dJUSAJoNjbUnl2sd!D4UAkcoxx7w@YWVNHDW|AkxOgw(;zxc|f
zzfYXE@$2QMG@M0fPq8^wjx;;+dw=2(1YJMpgE2Rf1rXF3#f!~;r2$AL;L3EJ)j<2E
zcTI~3nI~Td?)L=9FD3u@aT`QQf8P1(i@Hn@3J2Luiq}+6j{EYO(U=i`X7$@7Mw|={
zk-26cB-QLA<;>^Yy=gt~I7w!nrP?&31V2A}S4!)VqDScu5r!VdlE**8*A_rH6av`v
zeHR}O=jsVA?<>WDr{(5w5fcQ|REkN!-3Ua2K*%X~^)TaUh=lvLAF=X^67ns)Ve9+}
z?hcc&GLrO605rVOfCZB{$eBdj;eh-eMvZEy0h0Xrt?*!STjv!eKJx9PS-3EMFitbh
z)IA8L2&XH{b_O5yx$6z4$+65!Zf9UG0u^(1-=G=)r+@9^u7zS5zlIQ<Z?+JS9?n5-
zva6!2zvVs3QF4N!fB2zo{u#q0=90U4o2Aamgr#P2h`m;{ajoZFA}Jy3xaENsi5uz_
z3ccJJe0uKT6zPr?!JW}WG56PU@y1Qz%bhf7>ELS%1$4vnLwIP3@aQ8s@e)Va^I3YI
zJSSj=-X~>qkLksk_e@w>&GmS0{@<(=K4DRR<8yXh)Rj#M<Z`8Gu!IgQy#vV`j#xW4
zM+CGyzqJig?VdIwY4%%d&sddA2j25bYib%ux98>f{dSb=k(r1vns7Y$d8^i!k1aJ{
zI4aDM(E_M!pv@ThacMMR2=S#=uCit<@cB*|Z5#9;okk<9Pi#OYC3awKT2&<!Mj%TL
zH`L+p<L@&$r?iU<M`5ktuIwJub2edTvQ4=v=c&ix5PxcLV)Ygog`IX@X%bQWSUwa%
zw!iZ644<=3XXE`!^@TNb#DYY7%;*yIB$p0u^x^9r*KQUuGq_+t+y=}vJ(ln^rCZa~
z2-SDId&issIQ?iy1v(Ox9qnu30sHK|5l;uwD|)?>#^u2ThF&8bXji@@mf=VD$QlgW
z%69U&RXoiquH4gTAgEzak}h={cuM(IQuXD{1bYe?z-A6O*GukQG1G%a$}@L`d=@Jp
z?$#XkG-t`*DpiuDDXzqgDBbT(&)L>gM8=N-JpI`0!&(jK^uh0S!-q-JZM}tkW)OID
z6i4Ui3Mzs5V{@bJavryRy{?WDRcS9b&)*>m4{u(Lvkas2NOEi%POwfH5~j!uiU@8D
zBykhP^G17f>Jr}&=(EO8IzQWB&%Q}2+N%_CdF0^$J&u(G6o_wdC3WznxuZlgV4f~=
z8_luKyz1Ege-^;~)(14ZN^xBcncFrc+zToH{ZdEj%GFUbchPajwH?{WAY{Z?C(`>o
zLR9_dPhn$Z<x)7+T5<`nhf3+6)*YAP&><gsEF6zO6t|!4)I>)Zdqs>30c_+R=`7wf
z*`O3_yUyN6f<IR)-$jaEU8;tbu*O|eznDnV4vjQn?hwaY&YhY=7|8g%#6$EK0xeYD
z-89wp-+jzI>c@DnyqDULm5N@!fP8EGeS;M_BztmODTQF@Vr;aiaYE`G-<dr&Xj*iB
z@_95rsBEpCS^}KT<istYb;N)Eb<h)k8_u<y;77|C`x1eD++E!wn9tz5I#|0lv#Be7
zr#oshNhIhxZKSpMJ()p!K2JK3XIpdpaQsQ)i+O<ZH!if(19OSz#>ucQtlh$mx6u@d
z7yj!7LFGTCFrw);#?B4h&_z8b-T~M0^320kbeF=yOO(Hrc)G67KkY`00d_ni*m@<x
z{LZ0!bQgsk^=AHCs9Uu^&ONeKjrJ?FGJL4dY?<E2PF8phJ|0vf8N`r!X9Fzec?9`y
zbc=v#CKM_OM-JAnAq`ZPCT<%JdaXObF0R5Z+Kij=heKi9@%$!d!RPCX&dIG_GGi{P
zF-D1pU*fQ4RD4dcy^`!C{!)uSag%*$dp=`uNau3IaOH{$>Kw`(zm#nCutg^2=*Q9e
z7bGD_yGu=l`i*XcyN`^<qN%Yk9H$6$^XfaJF9JA+I21E6C2D2ov1E{F2kb{oy=p2U
zwKj59RW5cf2)k<2Gh*7o=svw@!j~abRr4xv3#@`MMjrXhZ(kuKf3jn49q0vx%hD#a
zIX#31`UAs9R53(AHP$?&f|(?OX}SsbHq7X~0_eF{zr50E^r}6h)f*n&uzz7n7S!go
zVr(~UXCP`gkKg!f5li!;^B4WnIjSqV*L)E^KgPa!v(Er%aof<x{Q_=jS>d#st!U?U
z%%B=3a1>PaLRK#Qjlxq{-M#e3Fyop7i9gcM*!xLhNqj#^D)ur1As+K5=J+aHi}QNz
zIPyj}9UwWg<}&I9pkL6Ft#rauH?%x&kETe2hu(jttI4B^CU9^hZtJ5bdnu<vMe0s3
zQM)P-I!4q%>wbiLN(H#AkyuhQHsGt7XCwX9zur{Gmn2c(-Kj-3qgQy4do6<%`@t-p
zYE3qIQuK9g-D#BBkTdlOr))^d&<ewnq<FS|(eR}HUx1RaKdZgBa}mbo&3*>J(`ZB7
z;%ci3v9On)wCfG-6^4{ui+7W(fLX|wLZJ(I1&Q9fS*4@HgczZ#(7`uCXvSC%KJehC
zr9$$p$c=bj`b)I6i$t;33i5Eq&^X1)qm_+hR?6<3;BH9g!+%tbCygTSU+?Zi_>M$M
z7){@wxn1<w@e!X8z#mnVUdT2~{^Or(EzWEjE$Vq9<MjxZdw+1zi58yB!A<6Wzm;Nj
zb5+&aO1sjY@jL7}BqblaW{<gLuq~I2SAQ!uT7>-rkN~U3Fx~K`Lx#n@QMM-IF;fa=
zzUj1@<mom2%p{?retw+H_j#)>H-K!9^ZL_@5IE7X+5}?PDo+WDM_ifgl^y$aP#gAX
zNF6Wk8;L*&5O9Q*I3RrvcM1sFpK`_Rt6fxv8LdCeRHf5j{Z?M_i@~3!+h(pODiMl4
zn;$V?aZ0R{XBci*GmadO<(Vkp=)4<={1_HBgd>@e_F$jU1je*EZwcG9p8h<^J%vqq
z99(&pv|{!2bAjXTzR!HceeP4i#3rtYq)6#1!0qIC;v)};Bxyp)ZK(90=(5^hI_A8H
zg8Dl;vw+Hs{EO`RwTYGMr}{CMyThfmCl&i*%!QSDybRzy(cc2mu@*45Okq3!Db8aA
zPZu1McSBS3en!TTre3v>V5DNeip!57Vdx-y_u(BPubHfo28D>O&laCG4p7~zAy(9K
z1xJ~7C~^MkvQ0vyPOE%=SaQ#KE@wg~qg$Dt%qRTvYIi6N@@g1kAIm828-W6;yy9Io
z=<}nwIoiiOw!Zg1f0}DN);i&uSoi{acXPSkc(y4p#AW`Q*}!9M=K<xWR#lpV6xyey
z5-DlMgoX_3Y#f+15lEcYqZ<>niEQ|Q6-gf&QBOTeh!)^6L+9hXEpZ1<W4!*ueCEh=
z*RA&U{o&NmI#bl-88^8i_1d4m#?|^a?pzKM`W_HgonU#>+Dn3;KSbwBTD%{r;!m}o
zn`+g5BS1Gkff#G|L28SrB`)`$L-T@kfmyL9f8uHo;;Ff>1@9VXd{^bAr0}-?jKOwy
zQIoPm8XUl5H`A%*B(yWs)em3Er`v@c`~|g8T`1i(_QXCbXhO}J+l+lrN?l8nD%fEv
z&G(hrc{gVX!x8HRHSe|%x0DJ)DgT`{3P`zZylICCB`S+jE;F^Gu00O$LULK9JCK`i
zaysNrv+cnfWW^q3nRVJEhb<NR-<T4aKB_26RVm|o__T8UAY_>q7YHRmj=B{cmq-7<
zc+ySks$q6jmoS%8gO^A70>#KX@tUkQdp+Pkw4#Ri+PHnn(r(fb&njK#g+_tD4GOjr
zpthC{k(;Ef7E2Nh;-w(8({pKKU_45bx3T3zR-(O!^md8JBUqU0XE0WbPlEdxCZfA7
zjcF?n=DLo}t8Y!$lr$>r*RZB)hJqDl<$|ZMqV4M+bKA9VxInF_xi^?wah`y+>pa14
z=azng|FTai>LSya8OElcb}F-<M|=b{r79m4zFH}rG59SF+OJOR*$UN9cgGIE-wiI5
zCIM_Zbf9JR*_4*ijt}IK^k86uF+JySP!~!g?J<DjXlm6!;KLQsuGL}LE=Uqv^k{yC
ztndyhlBwI&5RV#lDpU`qv*!Tmzh*9IljYB~vV~ZY5B__D%G0v~YFIOX_GU{PpHJs0
z|2W@gVETHGp`8d1cpAM3ZeYgP7^lGTeCZpC!0?AI$U;Z8*KxjhMOitE`SEbZ965zG
z3v3o@x?cXB^j#Q2ej)KpbBSOzKyR_knR8KIisxl8?RB$7>!>{%Nfg6%(=~C;O{lcj
zADW7Ugj;dlst|k<eRl5oB$+!sMzfdypi2t(Z7PM1I!oj-@F)|wjlyT=_x>Lct#~U4
z{SEZeqf&?0*T9D*k&Igrg$oMPUBND2LD;7Xze{Gq^=r(+7V-KQ**e1c#c)Fnef=&`
za|zI-xMbNoF&%}CyAV_<>3xenetWb1$$1E;If9V&cBQZm)#E)W-<0j-Hm6_$inF*%
z7y!rrKjVmIM-Bt_AMAt?xO=a1QZ$QBUy}6agclsY?{K<QZO9Pil($RU*&d4t9lIYI
z;N_5F*It7=L>zho(aG_Dp?4%qBd9c{#0;Dn0N3|o)5Q4i2}I#-_rmDJgI|Y;UZo)a
zIRtHoC4bP++<hT*K06wMx|44agLdGd{9+tO4D_cE(dTXoWVFkF*;bj7x7(wF#?mA|
zok+oxoEBpj#4Fbso0a@4fLPwmcPi^GuF&++>kPYL;MgSC8#<o#rO57Tr>iH7KI^c4
zIq%8>n#r?Yg|aJX#uI{Rg?&tDPq?j49>}@h!wy#V``uI&-HeO{V5>L`FqTQ%C7BFq
znJ#KpC4)ktFsgT0US<(mEKWN5aPS9>nkl*+vD<86)T3ldpjOp5w%*a8lR?~4z0+R=
z-p<{tVR#M2siV2G)$LV1gyDh9_co&00Wp0kZXCaph(^N}@WAJ=S{2R9JG75sQqAM6
zWu<A?ru323d^Q^AR;x~Y!Jq@@Jn{BMEKYAZJ>m0%hW*<56Fk6BrW97G1;0;xfQzEk
zQcuGvuGn!?<MqwqLZp3Sc{U?8Hx6w)RLh%+1)5pP9{{h<$q?12JEwLjZ4`D~euK?|
zt@=xjB-pVO@AnLtucWY)6c*cz%p4`|l5!Gn#510?d*om_4+nmHP~FskDUChLgufED
z_&w=zC8$-mBUF&bUaf!^Ctzq=HH?*WhH*_Yd)$dtPR~qPN^>1v&qlT?ulxKke*7Sb
zT^~em&*gS0z>6m&jV2{YW|CAn{-k&Z9)RVmw*Cyuuxi=vD3j~#_MQxxT^Htn@(GsC
zpFhQGlo<poXqJPN&P&nPfh)y^%!ppqk27ZW3X065gAb2)tqP3w;*R}E_mR|tqe?IJ
zaE1n8Ojn0(LQZ~50IfPx>8t_QaYiCUnhigaX>@uF*N{<44EW)}xt0m-blSkzx|feR
zclTVXtBuDk>nY%Qv=VcND_{J}&E$geadT#!{A!K(Poc}yn(9C{;+aRmdO}p?1fO)S
z!LrvZhD8KHYyF_rd`1QwVDmsIeU*`FrN2F0pQ6NcmKGOcWMhUo@qM@!9bG7;0LKnJ
zG9BpvyZ^b+0V}cl=_~c6hIRZTmmxfxk>b7Q#<~zBn+nb%K!<Hw)~;I(U8#{T<BLOA
zUIaeOsbr_Vv&ht~Kw{TS0sJy`VJjG-Vh5RblCKSCwmWMzDz6UcP$=YvlNxuY+=}jI
zc-3v7Z|6oaDZjUx7o(oP6{(*_$gX`9<c$t+rd}0d!F%|obE#d{mL(g(O>``OYsxbl
z$ftAJ(JS8gs;qPBd)ss0CxEOHXqLd0a^rt??J%vBOm%#;{u8uOpA!D8Tg-3EbK*aP
z9WR(@Ik}?#*}@;aRf?|vH6yEw1+IQFVcM+}w;}!b>EH*bY1K);s}B;kB-ijiKRyw@
zd0FV>KKrhS?AB8{{q82S@+O)u?RNg|n<#zt%;{`>bL(}ENX^hRQ__0!br6zW>;R!{
zr0O=pPa(7rH_rpO(|2;9$6Oe>LGJ0&^ISh4iPU6_cSZn~_xLenQsM%RO8N9^QyH%u
znh@;p02&xVBfD+gt`Bn%x(e)2KwgnIElFW2-nF%>pMK(p&EqdA@cNy}+!2|5InNtr
zQQ_zr)@@Z?t<TsP&v^2na%{ADIYGBm3I)HGcnNPK3Fi7|24J7!!7Y}*TK;=#ScVKZ
z+=s2UI)F$BF|c+X(RJ)w&U({4EoYWOqfJ)i?;Gt6!S<Ex{jGF6VSJ3P_a+T;=aX^)
z(t&3BMyV^Fq5F_m!Q#f%0@)LHd;&&QIitS7rpLRWfK@@RsuP(fjp1gU_i#0s4Y?E(
zQK|lm*|)cF;MVi=2O=$d1eSnsvA<h)graS!1V&ZH{=D^(=$1o5T$a<+GEb?Za>8;A
zO6iTesQwW|dtcNC1$ooU;YZ?V30dr$_zzKQI87n%mvOfG4C<`nDp#K$pfKQ=0u8nl
zTY5!mJ!GjmkG^`CP$R)pY-$@Si1T+2^(sXmzf3ex_$C+u^UVweH3#7gQQGvjj3MG4
z4eL#Rq&sDjBvHT5AHB6i?i#TLUhA39_8Os!@J7Q~{El2mq6Zw}<bi2}CtmD~fT(#;
z#|sPT<wf3>Xp`6oU_6_!tP{Nt?SGbhA(cS)!28(DeiqGoHG^U%WB6jl`Y6MOubMGK
zF+|~UbUuWo@+L%ow17e#YfqhLC*no6z2{u*s)$MHc=q+agWq-RyG#TC?Hez$4VS-x
zx2cTZ8$v?PRrmfqb?$5;|4hABD^*G2QpWXKxGdT#qd=p`N;8dMU?9uNpcZ=?K&!Go
zavk0|Q}mCp4UbmS#`LZY9`rXlonEhx6YbMSB>9nN!3@0f<LJd4)S}5)twz{lsHBtk
zz{6&@9d<K`8`qAr4;dZY%bh_GbP|*8^dKLt7%BjBCEOl?!G90QO=%4FseDYTQ-|J&
zQ1<>YsxnB6EXEu_GSjFnc@_{AHjx)6@3GUewpgP(i*{I<8~NeI{lmM!U<Uhw9X7E`
z3vozGBBwF}XU)aPed7(o)B_`XCQW>8iJFG=1BEwKNL-)fXb!bnU4du2&fE)c<qb=_
zy2D}${GZQXm0c@MTLh^Ta@+8nt`Ne7spzv;QdIP<YA^GggIhLc+O`+=kwXKXL{s-G
zU{*3F<1c1~*Yuy!Q>Tl%5h1CaD#ThT<Qt1*3!Rcb@ww6}@?-j`!yG<5+@8XfyqYg@
z*-Asp%89jjT+)c>#|mjdD-rj^+3?ENZ0@=;!76~DJt7-o4YmTLzLz`+8w`)Oq8{1F
zITXp)vabImMl6s-;9O|5dT+o&aS>F<shDE0|BbrpwJ<=ab8aQfV{;o(RNs<WuL#M?
zc?~JgqV=J=I#at91HPFf7~m>w!vKAM?-sQqS+A1SJiEX_vrgc^bl~CckM(Q)ae>!i
zNo%HOzn>bfIIPgL(-4I}wtupA*qw~~&WL%Dcs9Ax`PCaxnkW<Iui@T<dP@t>?2E4y
zX4H!OUQX;UTPt8y{sqWq*TS+eVad1q@pb{o2AwI^&2(WqS$B3z;3-)6N^MT0fqV<a
z#}kWHaaFqSf}Hu_S(=v6jc#+u<XOwekKTu|TDIPq{qUp<8L|2+m($+kndttc$Q06c
zlIyDb#Y(d{1?vyY<^Z(LAm}oDOjf4$_|-`qto^V43JWw%nj^L<M8-logke=J6JU8v
zroD5RQ!#GFEKRpSST1h@M_66@lMPm;KAM8eQa1=4rl^GctSAY1+X4v6*TdLVD)$Wd
z*#9JcSvEu3U3^#*Zf%A=xej@pMSnD)N=ZXI6{(qVUf)VXKaD8xhaG-uZbpZyIf_)x
zXKAC3^rO1fT@CR$ZrR*5?QaeaGP^`<qrUx^<`D!Om+D0C{D<f4m}|sWmoV_tNJoOt
z7cLNw<*p|{T;ef5F*h}x$1{+@ckpO-P=5o`U$<lbooIrHV}&=;k0$+<)JXYp4_y=F
zjvGnOUOMq>lD34(Hjq5=2saMDSg&QA>i%{wb(IP#1K;L-B&wJPD>OwkuPZ4Mk}I5_
zpMDCED`+}c&-9J%2^fg1P1bLuU43Pt;0l`d#X=VF5bMi(({WPK>Xh~lr#iM4r`DU^
z=midM%;0~H1vU75o!w1w`SgJvDCARUi-neBqAGlZz4Z&(>E{jNRTJF(ofXO&aTOcD
zAJ&Uw9OUNlw+n@S$mthj83*S)7(#+LPxgkUs>}-mAOnN*MaYCgJDO)oJ)mX$5cM<>
zGHzMof2Z!&rX2N}S{e<fQjFP#XT;5-SnDfiod)!ar6xzuf>82ixfPwtn2dugB2KVR
zvGPJ+g;O$X>gh_g{Fm7h!t*70x}4S~7V2>J;gK^t#0(A1f`Z#*uIdo^UVDKRrc*l_
zeMkT2W&T8+!?N!@zEbp)hQ|}$GQ>u4`94A8qGl{!vT2DR68OE(@>F{D{*zkUi^T4Y
zZJtF-hOA9(te64yjnY~q3`3@as==wqZgn8D%9(dj%QtcLxJq&@+l0w{BIVR5I@V(n
zRYqyC&?UZ)MP5WfTn6vt6;*}o5KGff<!&ZDLd=-Q=<_5qK-uLkGI1EDKcJ0Q*$K%!
z_hn{g;UdCi&*-{0WWt2ta=1R+-oJ%xRSP7F@8qH<ViIh2Wpgd~)f(sl$D$3)dD@yO
z-YrKLQPTG_>)mB~Dg7W|@$=yr-0Y>X6sgiwgEznjBOu?7pDW$2#KyMo5tw;w0Z{FB
z{?mPIQ-l0O_7l0mu+0rpL9LEV%a%%u(Y0TH{F6dy-$Qnv2R=QIPEA0xb-pdNNJ|3P
zX%`(a$%i-;ccEujoIb3DeQBmmI)6V6Hy4C(b+Eq5V*{tUCLrh!VI9xo!yK#?n$Ud0
zTXPc$p*kYefq~<x4ceOEO9qj?TAf<Cu|NzqT`*V_$P33uJ@*|{xtU__G_zW?f`yBK
zZ&pHcZdLBPD)<mHelCY<GmMf`3<>vjivG<7ifsLsqpc?lL$4!Ab}-}K{qyE*&7pl|
zgjlwW`O1J!J-4qj66!%sDcbGbxKJFYuKy1=4}3&;?Fx!1Qfy5N$fzvA27g`@I_Z+9
zBE;Vo5x=6?khCu1?c6tT_xD0_-flv=EW!vU(H?$(SUE)5NF%yl7$j&?<lqoe;bfXK
zSU|L+zWjB&mA-0^dh0nA`_!=JlXGtzb(L2G%hg2azNZELc0U}6X)DoYhH3XpOr?C`
z889Ppmp;}Y8XWQW@5cFuOt<)tk<PC7cl4_bcMD~FV_NBw%(Uq^TBvf@O8Jfd%i@cT
zi~nw}tl_R&R{iZgKFJGbGR#8D3k#$X;My`6xaJr_uR9#Vj$GokVqMhy0((l~v4xX+
z;3<BwAWAq|F{rD}oZu(2+tCVUylozjTH$cxR&m<sTI~)&GteSpXrQkMyg45&>(Uu%
zgka*VM6yY5wc}7p4q9`gZt#@X`>4^tMZzlXd8^Jo&1bYvB&%$6O1`kaE;~d6osj0y
z?NxbCBBrn;PB*EVD7xNj7G0+mvcFNuCPFG%rzRtxZKnX5Erdo&m#sM)O*p?2)5a4B
zb-j6mME(XN?*4%+qsa)ew2Zkxx7F~qwOT2zL)fiD@HVYl{xx)1qn`anFfH-9nQNnE
z$n0p}_`8Pi$O(VOrb|@P43qD9Mi68VJz{Nk)^$lt*WWiFDP&sg>30&(ky`@FBWsA8
ztDwznD`?E=UxEhtC(1Yd6A3b~GMkW2UoOe7T{twJIQEhTp34&13&W@W96_FHF$9qN
z?o0#Thij?paZJ0|$)slW=K;<E!z8d|=fl~c_GjQbXw7HQf-N`Cm}E$i0O^d@OaC<D
zD6LrCAq<=wC2TXGaaWENEEE!iq!fuaV-iQI8c?9l^G$>l`*?LBWnCkL6$ixSPI<$T
z^stGlxz5cxLRzbo9USQf=eY_y9%F5A0j#fXZ>s#>`#XvKc6^#_q$`<J@FGCWh$cPD
zpFX(<?xH*HO#ZtIqsQCrk$h?bpUq1mySasYOBLX;DxXLsTAM6lb=2|`7-aqs_g9w@
zH`xh2ad30q{d_q$Y*lu@Dkc^HK5J=66pR!|5C0=3;x_SmLC$W}2OXX9_16A8u}B?W
zqDd9cH<YsmrS0`y*KN`ac5#F49@KYW?Z2ns9)vxP%L|;*)DC}Bai<U|&TEy~cM%*(
zlGGvnjQD4HbfLv^S70w)T|f!h5P2jbn36HuCdQuj+_8PNDZy=R<%wW&PfCDzNp*Pc
zqFMFO00A*x$b5KLt({fhRgPk@zRky2_N3rk>4EFc&e}L8wG_zoK=t}T^IAYC^>{~}
zFE$0)Cht}t=dSi_rMW#&TDAS0yR4Gi|24EW2=(y-rGE57R1v*M_L;-1K<1S_nE`zX
zqn|+JubL`8i>K%cK+%4M&nyb0aB-m7b|WK>@}XPu+&|@faTeyAZC{6V0Pjf5wQ`?S
z_v`nAMG))V854!NQUA(9X-1<uF(V2vTY#ZGjF7K@3Qkz=_(866Rb1yiVPlLhM#myr
z(BaBg&K{pOHR8)R&JV%9w}&qdU_*PZ!0Vs*=(gtRPiqGQ%?Gl!`XPiLt@}i;GRW4&
zt+A*P*At~}*QWGD6q-I7W95*xkNUYi8p8uA0`IA%KRzC3h`ln+Q1z?N60`z~SX}Q+
zJccRy)W1=~E0bKevO)S;8|Nw`i>-zFwAUvyGxbw~1oL0d^atCn7cDzyw)9M5>&&{B
zgn>m^kRE)uUpMjJnZCVg{Xdr#3T0Qh-gmT9$a<@-btubw?wmK};lX0jXgMV>4+q(;
z^jYpb84Gm$=d$-AqW>j4(@sch>ki>k4LtGs1(68dj;>aT6)~dDF=km0W1hr6m}7CQ
zFeGS(s>^rM^T9`-85F}!!qbAz5x}Py2>v&JvYP<VoGCx}QBm(L);MJ(n&vmLxUV?-
zcw_dc<y%wVtg~=PGx;^nkTJiRUu%UO%%pv?54vB7vE*4hu?}d><WhpKCe}4;xNOx_
z-_dh4|HFLXIKt>LORCZ?<u@zzwo^?p+B{NR98%hq5e17Ky5gZ@64*%uV31tqp4MLl
zHGR|N1iw<!tUNnuZ0gVqFDdB?ZqOTU-x55XqntH>rnmmgpL9Aq=Z}jzyx7#wH#&gE
zDtA(FmE)s(id#cBL3@%lQDc=2R;dTarDr28o%u*{^(LA7ul-ljBI5hx2cc3~l^u0b
zXFZ&c?_mYa&#HHJv)YC0d#{|86)71{X$wg%=1DFwEEk!NTz#T}SG;fwAUpI2dYSSu
zji|d*OWcSZBRKiNm{t+-8t~|%Nk(82(0-v}B~MErRCF436;P*ZFYTqQQ!|j#sqaGs
zEfK0@tckvANcJ1IH+QP5$NI!otT#x3JiKt{CLSi;=@?I7R5u4c|5&DWVLnzjyWd%9
zIBn)Oh*Qjk>|UTv*I6^VNat<sNVe}J>L*><oGUV}O-=6Kv^-$DP?cp{(#)Q(ybI?Q
zCQAFCm|;b_S|=@>v$||zL|KjD)CN-{Ntx2e#jK8QOIup&aW-D2q-ut<(ySIG_O|xe
zz(8~9u!pgax~z-gG#g*+<w1+Yo0(>~=7TKb^@a*4<JQWwgtpY5KnuDNV<fbkz*G`q
zxq;-KpyrZt#8lAq-&V76PhNRs0zXn0UHTrf+AVE6>ve65cdkQ)eouuP?QfwR&|fs;
zA<cEz908*WpqSCR{mu9wTM5a%W!)31kk4Nkt9U9j^e!g`pElfamyNq0qf*D7%4MxO
zZsQxgo!g;Y^R0?`o^^WW@&+AyKY2th$=l?KpUT>#vx2B;w8|#g3|Fk8>{x5>0;F}J
zFN@#Kv0STk^SholY_^E7_8d3MOW1ZYi|YK{HetW(Nt)?OJ0}*r#t_>FFUY%Ff0^DW
zV<Ee#>D}0Z<<7BgVD|&*lI|4D_76!((RYJyk0WsABZ&qK4AT7UbJ-In*3nX-S%ai~
z`EtZBE5k4`)Arc{3~kw*PN>87VEy)DQp{X~jj6BEF5Va0yeoe>BRYP9M_^mML-2kV
zfY;&5*_dpc8N`)+GyU`(sKahm0)Agd1~sk-!RH)CL21)`xiGC!36~SiQ89lI$tdw|
z1^jPN_%0=;>C%k1ly>`1(N;nT9zU}CzxYubi|sNaW9L=hmn`7_tk+SCjO*Nm^+vuA
z=6C-QdTJN@#H0O@P+R^(^kgilq_Fr_{*Ri!ee-qlWTezoCq*yoQ$O_ozQ?_N1lm$_
zti7;D5TY)Fm?4#h%#PT~Cs#JfIoq}I+@NIV@R`SVig=}XoQ%gTUV+S)GZga^4>z=K
z7+=X4nwQh|Cq<r>(e9^svcsH=`;^&MT1tTfvnq=7xR4sTNrD&xK2O`1Y>lO{u~|Ao
zKRb~E#$`bG!`Xe{H}MV^$5Ba0_z>qQT()X8+?VD>qlt<ADi5P7yR=jQn{d|uP5)y0
z-G@Xj`Fzgdnr0<d*$6^|y_)FynFjUUp(>THM?J24_3~rp<CP~ans*9chyJ6w{Z`*T
zgUF$9cfO@Lt)dlUIg1^Rdf<^;cU&<iU-8QgSHq<v*>>{xWv<vko}}6e*O(5CX&*!#
zrca2=)UhStnE=h5K$TNg4dC-^pG_Qt;U@!yyz8y6dd>eX@lL?lzE~FP3mzZOdx@-5
zWhoU%tUydB{Io1ci?Q@1&$^V^ZSg^yi?-wTn6AeC2@TfA5#=4^<meS600&Coe)kP^
zihHG;NhK@J{v){afQ`aSUIRCgZBXoS?(?V?O#SEgw7b|4An5|lLcwp8t9esljf{DG
z`4bP__0{)ZzG=*3huv;>5uN|_w4BO*xy$#l_s^l7;hqxJ<_l|}@*#BTvAR3=YWWzY
zxw7JwmY90?X~BN?A6Vn_IeVMnTvD*;`95Bm(g+i%+&pQbw`5GiJ(<m-i{kHLr*~d<
zFc6}|K|6wWh-V8OF#qR`MRrK0m~a8Cnd%cH_xbIX`?_0_={^Vz!5DEDdPqOa{$kPb
zW8P=V!p8Xjezw1HjNdVVP<T${e)_PIkGNcREDEqWZcCl7CiiOGr;JZdP6WEQJkQ9-
z+PAq_kDeGdaD2aU-wRT0@h%e}SRP5Ka42XFd3CO<l#M1m!Y0-b>Bt3weSkoCL((Ht
zZBj|;AS0$ofwJa&_0W%wnT1ZSqniJHsJyP2QCMmk2l6fe`N?T9`IAxii^4q&G#apq
zhwf+;Hy+lM`O4cp&AfXV{Vzcj2%F<Mm4`5#&SiXLd0*<cB$6L-d-k=zFsj@2CsB-m
zMzZZt9-}zH@_7G`y|)gFvTOT=C8R^?5KxegQBoRFO1h;-DQT&pV?abjLW!Zfa{%d9
z0U2WGlu#N5ke2=~RPN_~?)&!H?>_c^_x`@)`2KPnu36Vw=Q@ACGuB$?daFa8n7)ni
z_3D!3%g~-z4(((*xaw~)I5(LRRE&}*Yp+n`Ri=ML3Hk(n+Oq<ntBnu+x<Ed>_6l!v
z_pK9q1EV=xUNP<#M~y`*BX(x~+xpvAJ9qj^ULLYuG3o2l3*~^X-VZ4r+`v_0d~tIG
zaeq4Zn$oLJQTN$4+b*Tid)jA7>(1h<c5%OH#9&N2`{oNfHlBs@UIMlb0F_oRVLqf*
zR0KPOxvQfm?2dVZ9f;y%Dp_v8AGhMUZ8YrH;*zp_7x!&bjbw2OtUyYvzhDZ>RZx3q
z1e7=g<IALKVG3H*--@(M^JD_kj1JZL6hVIY5kRPN?oRyeA1R=;E7(*xtJIJ~%Z#lE
zck@+LAA%P%p9Vc*2w0d3rIqlmP7rk~=As|k(jTQ0g}7kTxkLaJ<|p&xKe2@WXj0}1
zP;$$TW>n^BVT=&X{G+YmRTizzL%pmYL;B$Zo#%=F_P^_*m`_XlG6Qf?xmb`}H>MiA
z+{$U<j3V_4X<l8LUHCvO=~7a^^~#{uC657g4QqL_u4<%leWb*MP^K;44j7982~eZZ
zL7_;yr7Ru`av8(EmHIcs{5f{*D^Ns#vx*#eanW)SqYm%=6Gm!u@_^>HP}(0wiDW}*
z=7GCdW^qx3w1U;=N6F7CXivesgK!=P{SY0T8ixwnKs}?qp9yUGOE`6PbucSP`6t2t
zA#F8ubZuy}AW%$>|E;(n=35*o9iV{Gn*GXPUf70#T}8CejYi;$8d8-zEkngd6L8)i
zTk{c2KwKH^kd7ej<%e?;dvOBy7X4j>f5<hL1DMAV%H^S(NN=!iwBT)F0;m#Bp-Iz}
zk;i<r{g?MaM4VXHPM`E=zTEfoSxO6z$30bQ9`iZZISpV;Mp^M0*0`pakRfa0$hQCd
zDgsO~s8NRmTp6Q|SAPoenFP3l`a6(IpMAM{-Ot!K`TSP%515iPW9Qppfc96;+aEH>
zC7I_#uOn%xN(XfHkB<LsP*9ElDQ!l0o!f>zrXl$<CCl>*@v0f)nB9a6_r)yV>`51(
z{ai)c1OmlzO#?sp;8|hFEgGsVqyORK-&*{)w{c)Yw^+^(13Qer>re)c0~Z-o_1HvK
zKIwaJ6EU2xUGVTa?>wcd*6`-n1K6VLkK?J)tg@T+oCRU#M)vi_^BzXSWx6`P^DouK
zHhvgi+7ew4V^{?$I@-kJF&VP%hw*J&C2#1F2g@<FHi3YfJ9=<XcesYu$NAv}Mv?kK
zf5XC`<pE8+%H{yn$d@`lK{s!?|1WWYJEZ`G0~LWCs6*?7HNIZPAVewYe@}?4$`bew
z$^l3O%8d<0Fv6S8PQ8R2zkW;<fLC2JQ`Y#`Jje&1pDGw=R>3ykt+NEEqeQt#_MbW9
zmugED11cw!t2rT1tjju*nLs|x4;R9hP_(^Z8QaqHZM1@fDDKJzXOepLWPbQ=%H@B(
z@*gKEcMNClWwX6T3noMh<0zQH_bInQr3-K-La5-4Kpm(4+AnDUCH@VpzMS2*W)X?R
zSYJAMEC3e{gAp#q+12Nd;*24T^)?YWfXBnbbe)^_KfV);FY9mDQ;h|2X{xTzj|pq;
zK_cBMtcSK_+RTBYeKlYAXGiAAL36Yce9^SYx*shKpMQ6axZx}5Yh=7B-S~x1v*rtM
z4bP8<ssH;Q?#!_Q%k{xGg04J?(iJTz#|a`C3%Y>FKQR3HY-w>~_Fo+(S8rj1b}x7U
zg%k~=nu~F$3x6)){Y0rdQ1GYNy{jsgL>a>?#EOG-@ISHun-J!8`05Byu0;$;oZDYF
zX#1D4L`XN|pvcJfn8Se5{*L!i<NvXxTn`*~^t!h`-~=p4_o|zY>Cf4M#)kmE7N}ZN
zEUdXY$}DYm^*@8gI{X)HtXGa(7g&eaX!l;cHn-!Pfg%l>p4HMqjnJj`Mlb&I&nk$9
zqeTI2uo_@FUnm4_R}ca4Ads4&M}C1s=uPx2MmS~h*vzoq@sw2I!g>4Mw4?bya)N(Q
z|6gar3<``$&)-b_Un5EFIyasT@|b&7vS64VlUD2NTki+p#@u}En&h<YW<QL<pUL6h
z$qX=Pe}MY_Kyz_(Wo7zB<Iwyc2Zexd@sxE>d4bdO2d1{nfAawTt+(2AXgDY=SL?dA
z-4(6eTnk_Y^h@RYKU(An{7DkUylc9}0sM$atjg2!zmD>+ZxEn`iUGSeIAWW_tifkM
zF@s|B&Y1JNcq|0UEZk%Y8h$FT6rmKDllxfGzCs&K;Q*+cknMx^VhW^#kud9idCdIz
ztqJ!y=iGYMU(lxh|DpW@aQ~l%c1V%1eHU#!P!;ZWS@6I3;~z9iN?9tE1GtVCD6wG(
z+(O;`%w2iusgB$)s-Ou!r;0{=GpuHQIG?Gh>V?Ms8(J>DgvJL*nFBWTPYysFa7=62
zPUrkSlAn}Nqj7Z<f5i6bs-W-L87c`kmQKz!7l|r}q3$T7FxVp4Vl<fg#*)yzKpof;
zD|LP5&TX`#H?U`O4L9d}4(Kt_W4-m{w?5&2yoq%zPIB)O_6tHmF<V=>UOnjT8g)lZ
z;m~;18`+GwzIx|Fl~7Ma)uR1m)y5XAXt{W0MKk9#e;v5GMpfY0I`I+~ipBw)kBx>S
zbqW27|0T@pm;Uq-3T$0&?!h3@yZH7+sFXSuiq5)O;x(FRrEKo~yDcPNi$eQ*!CsjE
z*aFa4sxfY+c{4o&0}Kj96N9f-%&I&5csADnU4|OI*<2ic8h{P)dB<1r>@V;C)s<!?
z%>F|xLB$uyL5^#IKc5-oK_eKgn5bkPA;5ri;9UK4xBonXeAb9gf4k|N9!utS>FpqK
zcsQLXOG86L|HueyXlSUprGD;u4VjXnF;Bx+v1A>(Sso!3mKw9xA}{?dry;(cS*xMj
z>rS&~ZxwYu*XA&Tbj4@?^jLJhSD?gq*7t!SY|t+|)Z%bE{zzp0z`;VN%=OUOgOy~R
z#@JYJa~ST;#}^{}i;UG;F>T|^GinrS4H9y!uO@?%AH4Qpg`rK1w1kVg9=k{;Ye}$<
z*kvGG+dE5lUBAt$U`tMhvqmF3Tqkm!CO4gIedvy>2U=A3XBFn1ohAkI4c+Rm8JJ*8
zuDg62IC0lFNHnYp<rtv@wGW>+N%ue9q@6(&w%~L#5mVvhNpWR%{gh0u5&9s(-?inZ
z*gsh`^I%8A(I!}sUY?#B+s!<3KI|C3o`w<tv$pe09^D-*m;U9SXZsT7%5#5e_(Vdy
zna+eGG0I6R*#Qg-<Td}Lp&a4Oudz^Uu%}{h&XOCEy*L;U60JA5f130^eT1$sUZH}!
z0ykHsF@M}Hx;uES+|Ret`w=lpx(S-3oNSWay4?+1`<&UoHI&!C&D_eG@cDt|n-br#
zl^8AkMiFh##?0?#kp|0LTSFj&$+r(Wb#opdVPq+`PMk?Hjfdt11~;tuXy6%`kD9|6
zHAR>AsJB&Ou9PHUWwHDXPCxs__Ew6S>xOYITHU97Cz=bukxN~|fL!L6&dT@4hGbO|
zc;5X7JOT}*dSgD=Z03<sjJW~|qF<E$`BXma9G7eNn0UoRj2kLsx9iV=i%%N&JFOeQ
zyagU{+13#LiCU@Ft857WZj!I$BVnbSWL!6<#@Pr7E9L&8tp5VpO0<+d7V4W2d@UhC
zTwMGYns&#@25EmzwHoF58)Wz>brr<mA+v3IzmY~RGkP?*d4YDcVlX2Y|EDy-7U#sE
zC+%&KTj4~80V#TSbN4qrk=n%U@}KPcd~>p>TQ4Zq%|(J!m?9p8f+kg>)u~OE2jDSh
zGlMdvzt;W0F)21V$94P3fwz?T8vpsLRIN$cofgTf^oI9w338B6nk)O1Bnng(6F^#H
zug1D{_up-`V&Y*yl&)XJ_>;c<=_8bb1w4@0h$bp#lBhN0L=FD|)Xx7tp#ImfUn%1O
z-X=uddOtq7VQqt0HF`GAEY<xRUAY9c#X>o$;nKqGwi3^gC~|?se%32i&0%HP2?#3q
zfJ3#@lism$wot-@>H*k2TEuT=>=I`GbubnRVkGUY^O59h)(V*brMRFVOJHCiub3ED
z5EVRQ?!dDx2?<Lv6FZ+O*i9xAh`#IVI>RPog|^vYac&O<-YxSS?e^-<-PP#7KlH^t
z6K#T5hCKjH6gZK*uP0p(nTU!14W6!>pl1QET!=j>E1$>r0#hCmqHMR0dAXAv8f_O@
zqGL)_qLV@F5@T~ZA|c+pojO(jMXJD4Na*vYV+PoXEN)yusA={82rYCzhFt+=DfI;X
z;@0av=TNF54*m3P$^CFV<8C?tOV(GEUw#KiJaj5}L3*4T$uAi5VZP!8mp^mwes}ho
zyv1LfiJs^e=loS&Qh9LT&)kjp`J#WBajkQ*Cm$e-&cVS#A#6zb1OA!trD||9=b_>f
z5<D0f80R|RjZycxNX6r47oj^iWaiBlxSKmPg^A3j%{;7jh~G72PVxz8==SX!goK2`
z(I9>kW@<S8kT~V*cXRXooc{w`{!b=9uU2R&!G`opi%eAW)5F(TwzS=I^aU^G5<rwt
z{;<UYFa3kZj{!YniFg71&EUvM;Z{9i3oOg_cFFlN?v3sow-=HOv24#R^x)?I77P7f
z2mU|#5fygo^rPMwKd^EeJYqkg!S@?iZcee8V06v8IhmXnyy#`me()a4HbHD-dHK!h
z>1lm^gCaZ3`Nl6BArF(vL=#2C+pj!sR;lVkXnR`PWbiMP?x@!6q+R>X3#3wyG2aS;
z<+^3ULqj`eX29PFZ7rMES63t3+S;gZ-0#>{eI=R#lI(E*;fpEAHlp6GP`TCkHoSe^
ze(tfoHJj^HDF}0ft&VX1{h{}p@)(uouQV_*AXJam)@Pwcb#{usqrC_aRD)nt-MPzm
zY-|hg8+SEh`zIzi52)bE5J3;$xS+XCh`olD<7#qD{ke;f@WXdJcM(Di_TSC@@Fae8
z?uRGwqrc}Yl6_8F-8xm${TBRSIo$XDbTZkH{x!aZBs*mfpPqV%wX*VGk@bfskRny-
z06y|~`vL|CRhq%<0*Pw$0BSow)Nm-aVAOFW5N2J>!e2P}3MW)U46dl2Im-M|kv;NX
zla&I*0Xq|^j|G;jZM2AOw_92MDIEA=+;){xi)s{rlZP<!yMn~1(^FroC*Yp4Lc+U`
zegT%3wAtUtVrEF|8vW?dp#BoxpN!(4K0?jT`GVK*&BfoPAiyZfoo{I4f}68xf7<~o
z?i<P77<DBjsx-8;R%92dWF7k_7wE&o-1w`_{>qYn5x~D37$*;~s*=)NMWI{T0Y`_w
z5fBA<Dg!S6GWp`>k7;w{?O^8oVSfse9qCoim5P=%(y;97xdEzKq5jij1zM9Kyx<Kr
zjJ-z$qc!8%E8(WV#f&VOzjKk<;-r|}!4vo0RaL`hphiy4jK5}BZ4+Q;U2oE?oa?}R
zjB4^;c!z&sYmOi(fZ2|Yj<&b<ipd~LXurfN^#zlWaFnkG0;u+#C2O&a@(OhtTAF-y
zR~KOC;6JO?W5$SgXrk6Up0>71SlHO+?(PrqHk-rVF3!IAzXG-Bt^Y0q+*8TYa#jl&
zkBa||ZlzM8TQsjfv46Whyzp2BbeISbfD9$#<%RS!RZ`h7_V&`QQq$9^8hMb($6TTx
z?bwnL`Hg7vq32FAV3A9(vX^i!=7=AfPB03@KNv+>Oh$I0E6`A42zLd$>)#9rFFMzY
zOP4@S3it17{c_W9V~!7Ec<Eer2k-CiE8fF!j%j^&WAWE;@`C^rTU+Ik?R>-#y}U&B
zyO0QvDIx|Nh^BQP%1P1vYB&yv7kEgQthul-AOe*%BKQ8_oC|#Y(4OnYfQ0Nzt`MUd
z8zro4ZC_Uf$o~q+rBc&@Hi~l>a=FX_CD!puCH#$8`mX~aS%M4hX5~T|1yTs!`6}W4
zFWn@+2Sh=w{4(=^^H#Stjv4yTI(uEs9zA;IhouLC4<LC^W011->(O~uf3&*eu7d3Q
z1f$vmT5K2W@)!Mwe#Fswi&odvWI{tjLsGyQa6y^!DS+Xcn0gXIgM*@?dGjG-cG)Hz
z>m|vOtzDEXR^Urs**m`(7e|s*KNc#tW(k<_@VAT3VgFa~6B-`g+1%_;Kup|OYTBMz
zRVB2+{IR@ayYroU<3U+c!hbQU)v!8P(ON!u`!!P8KCAqbZj-Gm=S@3V3`o><z$%(q
zvrbZcxlN2>109d&u?zf7cQ3TYf>G@inioq=O;`}&sNka>A-MC^*<S)o6+HLsriv9`
z5CaaT-Ht!CH57Y~;OQOAKUs-CeT0hK00M`ft`v_=Xo!(?ec=NQs*BkuB?Us*P;3Zl
zP3;YO9e@*qGp64F{Sx}Fa;h@nXgR{RnaB%V$NUlCQaypP#yZA|;B!TRid+Y@h=`_U
zl;bPZh)cNVLdpLE3#qj$@|_r>3248}kqtJ_AF`Ms2gtk$9TzuDO4oq&bc6`Y;6f^Y
z(dt2V=!0A50h!2}x7N@2dF3sD<yzalV-8?MzKNeM|49yi`Uur$0YtLL0*E5i0X*U)
zswwmf&{<aiBh1dN4)p{;*(`gP`8V&5ijx92s~pe*j0CIcV<-A)vmzJ)*PZ=+$xixh
zu&yV58pDN<FR1XBVBfN;DPaBd0iMV=_D5aVS1IdrL$oJrFbI^%?|pF%;7esIJRV|b
z&#<hO*{z54qfzL88pfYKLh+b?XOBYv7x4tS6WP>cL^%3^3hu)tK-rE_w-6tN_nWvd
z6^yPx3(p|q7zTp4aF+||<p%;@hOHR9Aj46yum1u98fMNszmd>&HF3;-UU2viR7$k~
z9??At<qNL)Bnuc6LHf}#wP9l#`q7G!>_MF0I1NE0w=f`b&XF!vf|Q=t3I{+A2$zBN
z7ft8K{OVO@1HMLmfimE&N}~U^L(_l}@fR=X!a~8Z_kMDO?&06`zc-cm^yN3+K6Qo}
zBH)^Hs9nkIyy<qAkO{M(bYpE!G>FB*9o&vda{DhTaiQC}kI@CYm%YHjiy@Ku`hpk`
z6O$BvK|w5;Ly_-}Ik1%V_DICyT&<l}&APjAWh0UA=jEB4qs`1M;Tn5d?P5{mu%ePc
z1~-xK%f4cH-YUftveBzyFw@o5nf4?i-&zJ$j+DsNO7;2L+4<2QO$Po?H)*%oOf&%Y
zqk;ZRYsNxIu3pUZ#V9WP$JA^bGd?gZBj6E6rE>EA&W`-3zY!4~+aHO|t#hZ>cR^c|
zk^X;Rc4E=$8Wdnx9A;@pU*46ybumqv09&Z0f!hJ|HNh%<?Kf($D|h`CIOmuUP{Gxb
zX?4>d{mnH29Ou=lHRs`L7!Z~FW*43QD_M1u=HA4B`1v*Q^YagiIdlASNpdaEe_pfe
zJV=KDA(FcG8&^?^4&XZfKc#Bi!6SF&GG6_hf2hL05)F5**!-_AEL7sCTG}vmZfZTA
zO(ne$(qD89A%C44WAFR-Y61cRz>#jN%x3#Jfl|y@ei{R2exogs8EBH2>4V=%C5!YA
zWWltxfwU^8J>D<6*CP1otLZ8oG3U$cCU}BZ{r-KO<NCd3*>*vw;=dZ-mHm@tVGxBV
zd3t$eH8rJZGp!bFtDF3sFi~{E!<m#PPwv=fCdS5o61jx)17hXE?X5Bx-9jesp6!Iq
z{m3pb24epu$EY@h8IU3rN|-Jgy;G&KKn|SR#Z>Oj)O96Pr~gB1x|EvRBpR_HK)zw2
zJVY|S+kf)sf(KOXJn`^g@Ba;n2X6Y&M>Hr^rrx39sF6CH0TsqS8wDs0(=A?KrN;d^
z3y-fgYyD5TZW&<<B#HvpL{1D@KG3xxl||9mIOk(Pt#5*Ra2Hap)Vr99#tGR4<c}70
zvLxZ0E6Xcpu-lbsf1|=e=^lJq_?pD<zyQ-D49JbLPx!yAj@0hCsT8-<hA0E1D(Zv>
zq_6U0{kyLcG0?#R<z(l3KgD*fJ$WN&q+gVgmG{afXyu0*f-#@Bx6P+8JY*F*GYgde
z<nPm(Zhz1pT3`n8;={77eV}|5ZZDC3;(L~+zP3fmJJBW}l5J&$r<5U}J))o2)whXT
z;|2~s9ynQ@gNG?}gZf@8GaW7pbS@x*OT?68y6j7$S2T!mq^@5^hp1XP^fP~)_>Kj^
z`_6umn{xvwW5g(O(F6!P56T%?0hB;btgBB~m}uuCi5>dl-mAYUrZ`ygP@D@ai0Me&
zqYyJ*Vm$ABer)rbQq!$g>|qrl@G95(d`R6qr;t(pbbuByildkz|K4?_=GP+yRco+~
zVqd+}FE8yFtv@wBUN-bN3sRg5&EbjYS&%*SqI!546>gJJH0G;zdWwzVG%{{PO=(-l
zzPwdiBI!}-J*S-0*_yK>AKBoja1If*n`m`K?|$-Eh^?5)1<q~&?uPd0lX<l<-(g7E
zGk0M%O)TJP3{2bpZ#k=Oo$XEK`pFYsH@~ywFI8#4kQ;B2jXpv3d)nKZT?J4`Hr^F~
zsT06L?s4yzL?e$~KJXKxI)^YO2u3HZn^~k%DR4$f2J49EqWumIunrD063jZGZ(h@u
z6!Sja-SXHf<y4I($F~d#^}&h0+icrkN6ZinI(BLBUS9;gd`+Lg54NH0UXL&+7EzvZ
z>|wO(D`hXzp|*QYaV1v=9YU<&g}Z6m9;V>n00N>a8&>I`>zn>kXut$fUMiR>1Dr?)
zZWW&!m!OYv0cQvX`Ygoex7*wHdUG7NKkJ&Lmv+<f;i~G8!aajeZ+<#aL-Js2Tt`0e
zYQyB^s%P#suzr{M@*t_tm^4E)wHB)q2a_AcZ=glkW;^a~ltz+<r)x%*2yIJH2UW!^
z&ovL3t8f^u{RoV(@mo1znf>C8-CFxA2A}uQeOeX*OO|0aIy_O8EcFW^0i6n0{G0jF
z_vGb2V0itJ1i15X?6m_})rR7Sz5t39&#V@O=5^Hjf?(B5JgpMQ(G;)kWkCSJdt(wU
zyJ4LztXs}+hc9h?seBfMZ$0<Di`>WQ1Fh<SL(ot-npZpYk#MSM(9<VF^RlqyI|;ZO
zx5)MYd8FgM&TFv<pR(LrKFL-O72tVSRfYJPOjI#_P1U*m!I*`SwBcnDVw7I3(;Huw
z*h|oSR`AGxX2KqVwrTU-JCwNCq<@iVT*)y4&tt-aRh4#K%O|U#8eMsuWnaAMo`ll~
z%@}<3DY$+(Jf?G~*6@4qpuuUsrl+1@{i_Mep&M#ez#Qr7holn!v@xlbxaUDhTBOW^
zoh-1>DL6j@ENWCfyN33ITP;|?&!ZT`nv71D!iLJ^nu6A(<=hIqca1pbxA-WAyM^Gc
zeGd*;jV$(&8$P$opJlc5%5{jwbjKL@?Yq|Q6whBj_ng_1{X<puR~z;Un%osmk0+30
z`Y~~^;GL|r{UQ-P39mK1{%lb9M`|bT8)0aoD&wCXMEaiq2ZUjSXVd#7kp>Jyl-oN;
zBRqu~8A~Qc<<dRIyfEgEbWvngdj>wMr%X=U`#POM$pah469)W;HGwuN381F4?18z&
z*787dfA=buR@w9?ID=HscLbxev$MRnZUL)~p}Va4zt*xo$^3LoYv99ej$&||47$9_
z03PYl_tshr7z{y8qt&`Z*QjCkyKP=Ze&OEV`X%B<#`J}zuf2<QbvWtJKNt-k!H&%B
zky#aj^~5xt^xE9Vrk}UE@&Jewtz-PDNB4-s5hK0_Z@p2)XhxkHO_Fw16b((s!$GVa
zIVHl1;|5;^z^dFgP9q5i8lk|#q`*Y?O6}u4E<cx9`kK{c*gQ~Avi0z+tcK2MZ7gI?
z*^(<l&yL%5`>`>z+p|C>Q>PJ4gE?Z~J8Zn^Ae!9$;~IR+4%Xd-5!g<5<U;QXS+pUs
z3&1}4!Ty9*{ol_w7wD=o+|~T$J|qH4A=fLGP5+qmaAWj-n`YOPy}hkkxY^0o(bNe~
z9Kc^ni_T-hyncle?dW!sT<EB(?G(t%Ra8zK9cg?=E@tW3MyUwzxC$RQoMFFG5Myup
zWX~q6CAq}$t{f2{Nf>4#_TC8n6dOa`_Q7kk33(+NxPuO`l1)Eq;tekPjG~6y;ckMa
zs~-@0?zQTnV_FCL8<{e#bkr%f^QqcTvF_zX&2QJ59pFy#8ng?!@)zcfVpRx^=clNB
zUjYdrFjn{FVZ|bSdHZQ@OZ&wodNuNry2cF76yy_9Y*<hlW?)x@cpC8wZyhHtev?s*
zi3J-kz)t~^s*+5ar(Q;BwTmlwRvn#Lx{2)r`KF)e+d6K-3aedJ6^)BM+dRaIMB0q(
z6P27gUY#eq=kqQ9{FsL8h{v4iAFC$xqk{3JCYm+UjR#I0FW0Ks#cyO}$jvYHU8NqV
z>atTaJL4LyYVLYEIXHN;o143TcsLu&Zg(%%ibi5uq%8^5TfHMm1y{6u0k~eNjX1Ho
z;7AD&YNFuKy+ZZMaP;PQa%N@kQd91mMN>&%b*4T4&S%p=E_T|q8FFV7NRcyNiQQ@j
zroc~y=Z5{(PWF-SZu4o^MC5JobozcJ((E?aZ$4Xb$R7Z&;u8&(A~keKkwAfCbLDJH
zwFUaD>*b0exbN`Ve4w0})<jzyIElNv>?y=x#-m#ii+oTm7adB&2yR;UCPsbpcGS@B
zHgDcL-=_ikKfF)lA6J1UDCVI?A=Ay3Exl=Jfle+iTA@VkP21$p$(25=f?C?zUSF*%
z&)~ad7Q8*!?+a7;I7a;FSL6t+jP8x=w)r<qS=8mIBy2SgM=Qr_CB1)tuD_UQv?=34
zaFZ`kJ1kVlpnp;n_TI?~^ke`wsOEb-Md1!EK=TsBLNV&U)dY-(2F_?hZ{k(dWelOp
z{2Xn+2anoo0L$TsSw;;HiPRd^y9lc?7obM53v(ACNGch+t3&g!+o$=ZPwLdv!d!A;
zua!*W^f<)Ik3$I~xb84jt_cYVC>nSLyFegVlGnJ4V$kVF52`FI$ODxaN>zAkgp{Rj
zm8O<pn$`322^X5g<zy#07$*pBN34Ce3V8hDp@&YThsLwca)m1c?K;<9C{jjJcjaR;
zn<I7$Qf06rdGdRrs`8hIGMBRpxbet`2P7q@-DP3R@5?$Wyt6y2D}qn?=(xIf2!IMZ
z>&X!gV?hh$yuevFZ!=LnijkNg1TtEZZ5mHVfNQDvl!01_G66}<SKb8;<G87NM)WG~
zFp|CFm0QakCf825Vil@;Sb@L#-F1?~$W00-PqXn7HH!f~B@Pk}rrbF#vCN%0dp2_p
z-BM8~&s{~c6+3by%k(npu#^+o<22&V?sJE2lBRWoD&E3pW*B?8Lk@!&H7lD;i=rJ}
zfYN!#n`)d%m8ua+*R@rU(HSMJdeF`cO5W8R;=?If|1t$DN};ZrIk?l&L;n&&+zCm!
zUD;Buc!pWO`+0_DU`KjGAUVF|MZB2Mb9h;HvZ{W}p`3M1oMs|@$rUxJ_nThbnO!((
z#bNehSC=wtsVHd`sX}s(v>I2gjhL4OJQz<#9ED$to^W0M96p()_@ZF*4o-fK5601g
zO;6%6R%H@jUVTB}GFhg1<=D6YlkMTO;dFEIq=b*2xh^8sHA3wc3mQW=AmFqC{BhSc
zQ!H-P(9|Z(6gj+|*kaSWL0V51X1m44NJu>i>l$z6mG3eg<pvq)K0=4sYxb&iNCq38
zRW{D-lMj5JRNV>N!aE6(Nnz*nhUD&5dR;TV6S=<H(sgF+`9)oKW2P(C?0Lm4NSvpc
zMhKm%*&OEk#8fLE=E&HZY1(Txa&D3??Y1`ii(M5Aeko4#R(<1&yhA-ZCZRW*b+Q^6
zAHq6x+y+>eK43}l#fF`eRxR$R>*r!l&6uMBc3^S4_K0x&SLWsw^Q{cD3FX%Z?`=8&
z@5@{1IT3wt)(5#=A?*cL?F?T~epZdlS%pHRuOh-1JTc+B=diP+PEALfB;8+lAR=0q
z)sVsc9eT$#6*%JpUL1KbY=@$Cw0MWR;I+el6goBR{W%b*0i$m8V88iOY`r)5K7>Jp
z<YUpZuo#xOfvNQA4m%jnq=QYV)sv%|6eV~^X~qh<aAvLTsPn@;qHq-z2*13WO=M#4
zBn$qIN!?0+dRvPIzoYqB^!(iDvj^Ynq6G6q(hpIhVk{HHVM{{KTygz|`X25<DiNQ<
zbcS-Y$9MI)<_+wJB{KBGlxMJXYugfB6z&hSx|DBqT>@$!65mNbFxD}uT&L|(Vp+Kh
z9Gk@!-+iXnpKoX@EMv1+9r%(zGDTT-%*p?VC>)_Ut`aQU9>C98Qdglz(PbTZFs?$E
zuN#{~2NIxSFf~8>d^0DjV!5GI*w9K_=Qcup?JNV&VW5$z%=PT*kfse-YIi0vEM}&?
zXt!YO(5TV2RzInz*woXwSAt_eMX0Tqczn0q`N{pxJ6zaduP@t;W|XNK^wJsL%T5TI
z@hx&v%JGSJm@K~M^nK}JZ<~=Fx6@%ju@u9=2NxtO6?UD)doCrkfxdta;l>n?FE`{z
zBx70;y!Hn%mWvknkVa}XLwGgyH>1g%e%JODzlX5scMeSU1A@5K*0sv~JCMo_MkC|0
zdm94WktYvj=IIIzO?k6LD;-*fruT{pC+NNJN*pvFDt#moFm&tZI_ON`!RO%@MJAEb
z+hBg(NiDTkx9A=(j*R7#{%&vuZ53OTn2vBgsUeBGMe*)l(jB2GcKL^)W|dwA;?D1V
z;I3O!7AG4ojD<v_#ttWg16DuTdu-Q^n_H>YNfaam9=*Ny2nnj|K^bp<?YY;oYgL9C
zGgqwc*{~*j4d(H?ZLoS~0oIHTJc<x!bt9tdvvgZ%w0I7!8LvzdS%5n}3uRXd){y>Y
zINWGz;5J?E6$=`1ANy{Z62^oQc2#R9k8?&D4=}Q>I9n(4!WVP=>?IEMCc|^R>qe@N
z{B49n60rA#7$$F~9G-%gXh#`GnmkLrkGU-|XQ?>blT42&ls0C!Sa#@(I63w1E#hQ%
zQUY8@*uG)Mf!b=F+=@=%1&m%97ZrAA6RZ34Rp%#zhlUGTtu==z{ajnmsBX^Wg17<2
zQ=5v?(q~>2>-`etrt*Ft9?JFi=G!?*yVafA5Q?=I`4obN9>6!Bh0+G>^;+w7qx5V|
z)hJOl_Gb_)+A()zTWicf8=auB<H<;mVnC?XW2E9hOED8|xWl;Do$8#}uwm9U&*Yrh
z9GP(wC1P4^*r7&_tNJRXeW9H>HEXa%8Gdl_o8H^vxz@S#%FLnA9AAcKh>FTm-=&>f
z4X<#s{cg#sSAX%eDyPV-st6b7RMMriMY_D(Q~+*YD3T96_NtJy`)o6E?4Er#CXClz
zd?ZuOhkWzR;UhdzdQTa2#B+ewKs-L`m13{VIcwQO8nXvMTw80`sSMqHFfS5mzlZ)J
zbau^cJ-gw<q0*iFeLL}cHKV*ASf89#tmYZDCqa-l{m8OMh~$@S?bp@DFi~6fBQ{<!
zQ#N>nU#5Ndk`lP5)ssSiVnIbB^vezgK0TB_R&UFqQ~vktp32oB7!_o9n`!nH=Jmc4
zJ6Fo!BUm~yDXyrmuOE2Q^)abov$*NO?%8btV}@%YsRjviy{Wuh4#>~RO7*g*G$FvA
zE96+bp!;-IWXH6xbqGo{YICMaq0{4#>7*7PhA7+{ZZ0s*p)ud^_ADsWYchZkhawu^
zI11+n+OL}I-D@nUxf`?N@@bj1JK6>*s-V)_@g$PY5-vzL+)*EFJj&kmu<jmE0?c&g
zs&FbPm!Or|yz3IGM=S~TkiSGxMbx9X9&+F<?4jbNPU<E&w*N6(d@BP==cHums?>NQ
zYT9R8vzN%+@^R!vkA2~U_|0iQr*d1*gH%)>?twU>jVfNH+$G&KhGs}rBF%uuz%kvh
z{+*mRF)Mi-UoKlxq8Q^oku6NwI_j3H2<PlaVY;p2ZB;zu>$g-(DP7S|TUzLN;T|cr
zIv*D8hnQpi`$+V6-K$odS6-ISxeUz?9d7K9)@HNhIB~qO#2Of4sZSY$NYM{f>NlA;
zq10wiZ<s%kb)Owi^RXy4PQhP*d_tHvIJLgITqe(&&i8Dai;Xd1H}b^#^+D<0n~MBE
z$1+Iy&Y9DKbj`y~K8%PU++B-t0icUEw8XY!tTt8Fv@3Jbka7hc>*nD-n-O30=PBBp
z6o_3)5U{fzo#pId620a^leeSpQ#!rR<EE`&g}2%UD4zJfageN$@!dS$Yo>CW$=%Jt
zjVq9hKU#1@AR0Z+gTEngKUsa$H`E@+mSV#&y^n9Ot;Ey&vO80DGzs(;-+V*&vI>y;
zH$BI?sZfVvnY$iOPIx>#XQxu9SlW{o_PFhjzRZMGSe;CJFEHwrKCR^4bPhs(s*t?;
zt`jzNOU&9|V05~iqUWtpiD4ngaQXFC=67)s+|v`wN<<_Gg}1Ic9PUb_!|;gIQ+e4e
zaBr4$sMmFK*;CodYy4ThGHRh<G8XTshi}5{BrH(Vwahs*kP&};wJ-E&?o=6DYsoQh
zF*j|ldl5v`Z}B+$F}jg1PaLO%I^=*V26gQ{CvGfsqaA4nODURUc4%5M(@Q&<eyuai
z({8X3y*On3YBwuq$9Vf>Xd<A|h5+RY>8=UcozEQO%>?01X6?rZ?+U{YX8KY=1|}6Y
z3pWL@>b-_|g$*_oEv~`6X2*^0ZFS2DA}Z<bmC1w#2JcGPJrAXCSYPb6irhCYQ{s^@
z_q{s;LrPQ==e_y*CDdzf(X(W-T70|ka|v>N#yAfTvb^hn4AR_*bFn`Y^T;rsfFQY9
zMw#m%ixI~&V|KiS`;w++^NE#D!s=&HGSqKqs=N($F2~jF8gGAoE?%bBPIk|tr4QVt
zlyZ=Cm>566USjMUP=7jj0DUAmu&HFd4z4+!Ee+jDmk6Jv{^AuJdPB!Slx@GEdctsy
z4_#nZm=)3)-RTu**^p$p_z-8qQiE6ay>?FW^~KebY)v3(Ku+Fu-6w<wF1bBG7)Lv4
zCTk!qc1(!Eo#%T(kTS+-Bed-P)xUTr^Q3E8LUE$g`01UyDPMGp4R2ENtP#ErP@ktr
zv=@3@)K1}ZI@#bELSQ=%5_R>f3M(xPLdE49>3dd$cMWVV@D?4>?$)^Hhn=A2Tl<r>
zFWb-_;aZG(DV4yET)JimhUEm-#7iU;+Aoy{Zss%)>}5ccwq7qxq|PFZM0(0#lt4kB
zqmQ=px2o>7r?B?Km<(-trx>5@X&7ZwMi)EvQoc*-`u5#_%3*lP-J>c@({1Vd#?!Vf
z%B8++^Pb^05%Wh?#uMA~ybX<%OCiL3y906W%54tJ*YMp|T<hdatiR2h)yWz*RhYh#
zEm<b=7@%fFD2OMuI3&g)Iw4(cWymIPsm6?^B>ln?YxvPO4WGL^D}1VhFyrwDD}j5w
z&HE#^(Rb~1M;^2IwJJEzy*1x9d%5DUmdF`vJzah@QF#J&9Otj#dHOO;*rd+8WiFOA
z<E;3D-24|puX&gVWGZdHfX=gsuAg#xO@x18IHz`(z5=XMkXW++zG|hq+nAz9E?jqI
z>`Fb?ZtAJ*_k$UgvAmNu<1mR_ACY4(`_jG01J7{b?I`*+p>6ZhuwCBqMvcd%>}{Xc
zLlTM@R=1a_hpdiT)O{$H6weT!*Bj|XMw1KnhY$39r_=SeJl8}mmtC_W)in2aA9pT_
z0ZDZ_bcpNGbTdyd%ZCpiw%+4|mR928JcRhbAcIP-*eY(hgQ!2YRf%A(jQiV7jU4Ug
zslWFHTf-IJuj`d(-Gg0DVZSCWLPFKJRwRjUv6C$BRs629vd2Y#i0%W1s9WUNj873w
z=hXg&YDczFFdyp@@fNA4=v*T$2$zMcvC1Oh^T8(z59x4*UWT1^OSMV<fCsp~vqJJE
zDjG)Kk74!O{hD)#0m@vo`mDJQtb`2eVTA)$p$%Ax*!Og?tmj#~)|Ut@Zmoe++cJ45
zs!~b0w_O4%R0%g7B;4K@mPCGASOY4{o5dfL7^%8RlqyJk6E_;yWSySmYQ-MpTLBdt
zu_p|Pez6I+ee0Dn%AvY^Ulsgl8$VX7eMc{wv`9BUZBN$)eG6TnnDT3=U8J0<!z3u>
zv}N0Adc5PMebU1kvyMV3m~Ya%hV8uWEhMph#M+8Sm8N(Wy2QXv--}IkBMK^7pD#1w
zhe^tO{%-BPDzeTG2}$->o~^DCJ7_VrMi5+6x3}=v*}WCpCc9dZTA{wco-o~Xti5F+
zz!M8|w20e|n_ykZeMP{cb+$EC;dC@G@y>e$hw<#`THIE`IAxc=aOU&U2`wb{uq_NN
zsc<5@7Td$JclSey9!rY%YFtSyoX^L>&$j&dMAbNoaxkQiD&h9mDwBr$Q#r3U)QIc+
zyxq4mTdwC;=Wp!vzqClA@=8+aophQ!*!tKux%ILR8^b}$i<%KfC+|vm5MniulKagn
z&#Gp4#1gLf>W78h6JaBwBZ`LSAGmk}SkF{mA)}(%5p-1@2xQYNSB*5`gDk4cwr>Y?
zl-&0zeOVF;{q%d^(IWM)Gk6>!nmZlCRZ^!4+7J#Wwt<D}#IbaDPqc*xKsb?I=+6%O
zU(*~TweBGeDu|n8LMgY0niExTA&Y>#cjo5|**<xeuqx24C-};VP`#(G7LJlaRAVH)
zI+rgSt@=9dUK9H0HH5z+nKH`gcbsH7zf-e<mbK5#H$3^YufO?dK<s-_imK}y8f**t
zll`TJ(^^p-hK;)Ml%{XV&Sr0D1lF}omq$`2(+UKa8kh*ox8j8}3S^2!ywZ+)$a4@G
zptU#1lP5)D*2vQti8YZ5`!&5ZL)CuELk7bdLtmMaQu<v44?n8^2>Ib~t80HS_w;*+
zmsyf{T~0FYjmt%1_~#{zkvZi<@=v=jE6#1{az_LmmOWszWSJ)LTQfFArV&pcB+^ue
zSvci?u1XS7wXZu3nCYpzYl7+ZWRmBgiP$azP2hRlms9c^Y6^-vf#KqOAbk-z`|5Y9
zJcikz^gPzb5?>-DbR~8}+vht2z1~+H&mO6~C};|q@M+@^f7!}xIl(h1c=(|}1?IN;
zuwCav+;UVN^T%#~I!{64G%OTovYO*zt3^fQcXLpv_i8EkR$>?{dMAwBR^;+sMYuFN
z9qpX9{%JywxO%b45bMcy-K#CF6tcuO#azj~n^UTTaMy4e9PA-?-`??ds>>YTJ(sLJ
zjS`EHfg?84l`uA4o@Apk`bSBK1_xv==*&Ctc#L?6Z`yIhaiPk0)mtwM`IU0FHFCnG
z-Vxc%GiHq(uo}_oy_3IXCd!@=Qfn|Lu*el<B7SD3R}9jMIn;=z)+An`T1gfFbv8*<
z8w)OV@=n+hm}52N3G#}w1g@<JRBKHvnKcAYJlqfh8Rvs&glx)eRAC+qMGe=!7dBEI
zq&bwtEbf@cQj6A;n#<gAbvdgLnL6wmvcv>!KQzK4gtUp!MVdfXnLK@z!UhWctbI;r
zYq#fk1+Kt&!zS}U!-j|Ra%N}AUkFq=)ed}DF(?r-IXf$NJPtL^-W+mHsGcnHZTaoE
zAqFQ354-9Q2z{IeXjKS}mCB}dC5@^K?1ye|V+q_W8G-qRRb@nq;r8zI`Zc+BC~Qgc
zm=xRTjpAQ=?y0rN<(mg0@L9*kI25&hX#CDHdeS$ic%y^2hzoyw3KDLR-8Nf*G`YFa
z0JfUxGkI9j@nY*N9(F`}7|^53;Hxm)@J@GlS@HUj5(B>G<!39!Vs=AptL-Bu-Ws>|
z4AVh(YE;q8RpvlIl9Lb<MWI=U4Jq9%pYFSN?_SxX4_LQVZ4&~G;8G@EP2K@#J4)4r
ziGMD-t1+LSlyDfq72tcP<qXH2_z+CIe0`}JJB}NBY9b?fu(nXU818pAl_zwj6yss%
z28kESK5C6sl%8j3_fM(|Ev)GFKM9*ii88V_#0%I(Guk^e@!M58sA{k5m=x(KC`nUQ
zNna{`N?AI+JMNc#S+p3driv$pHAQPydAfc*{U}-)QmBK3PB2V!nRptqf{#~+5nD%v
z%cMFA`_p-(BYwvVe7vM?o??bUR<*&5gd$c>i>B7ijjJ`U(Y&@s6O4zLHk!=xj@yWb
zUlwO2D-dF?FtFZXn;!V!WF6$W==zCa<>TbhP{g>v@?{}*5hqhpZ`J(ZZ#f$t#_K{D
zl`5&Rh4_OR^VuVN%cUp4+Oi3v`eq#208>}1<<|%OBBvesk@IaMAeiy43(g><ed}}l
z6He>Z(?<t3NtceQScjfm;}Hd?`vt6J_NHDxA>~nHwVh6H>oO5_cv-L95Q4v%OIToi
zXsf9jmT)%he|)5wAL7bLxaZmzZsHp64Ibhh^Z54S7`ag#<7M(Wp^vjrE<&`Ef&5s=
zLHwRXF39pbJ!_Bui5qeK6|N69r1lV_!a<lY1#&HlAs?m%Z>-BBi861$Xa2C()BW?)
zbrQV{WASh+FD&Y7a-BBh&MjN3nBxq&aw~qiAN->F-mkA&NF5sNR@W&LE>Sy>u-<H1
z6BEFj-s*{l%4M2t9N#o^8mn1#V~AO8`W8`tg%^A_ZnZs^u_2}lu}L`2X!PG?K-MQN
zOX?UEGz|@1HofvbLSdKyan!JK`Fmbf5e~r@q6E?I>98fAx#uO4H}(lvD$GDmm(8t;
z*~cP9H8&$LS5-BaD@k>p@}@B76*Kf7ePnfc`#Gkmu=19<R&3#8wXD1c)v*&QbU06x
zzQjG~Nk$SO`4`{|Qw1P_&v|V~$y(FUng?b#<L~&qob|Enf~^k1Yqd;fO3jn&Ts5W-
z^@{h9IC(uyl^QLRs<~)K21upJiLNVFZJmp|hx$eIUeMc}+Mo%63{s1f`gGNGPnUAB
zGSDm<$<AzQ99iwOzbD^munqEURf!=L*OwHKiNX8hSvScNS;2<0pulKTapqAKdK?5n
zv|am7vU{>G?-$UHK`&+EXWGv?tj@^lQVC#d<JsY)<$MG+T)FbJPAODp`e25+_Ejg(
zfr+)4hl+SgV1#x`1Wz%4XAbDn@tAo|E}EqtJkh?`>+1~x`;&vUaD-jpVPTxkmyX<J
z3&drEi4+1A){yZ=QTnNrF3-1~<EP1C4pUxJU%T)?k;`L(z-eaU!srCQvM=sueXlK=
zPbme!lBu$@;m1#<_Ts;EC7bFPSGOn=uWIVO-@05rr&_WEt6F)$db;&dugp@JrXOkk
zdGkBV!Heab0t{vR?BHcmqDBddk<C@h>0Z#r@WzV?ssfroAG`%cNiPM7v=f*7*;RaK
zir+ronCD#V=L(Bpu`J`|7G5r3(K%mr*?q3EAf{7Wq^-<66MvVXVliMhZDTd0XnzUs
zwwqWMay6AuI;y-f(xS|p=Si}rXWd-Nv~y)*jeGG_4<9FE^#?vXyZQ=?RkVX<p_CXx
zQDT)bb4w9c??UBW_X-;5nMS*FC5_B&EZay+&}W~o#g_A{f)<W<+p&d;^jVtrYkl>y
zLFklB6VIuL^C)-fy6zOu)?Z%|KXoxGT9|gImTG7d%{F>lykt9p)iC1sRaCWOk)aAE
z-2EUCm!Uz))w>gKEWNYAPt$MlxWdR`*#3*4Q`|`r$M<&FerNDBT<oF1CzVcer9$pt
zhbGg;e2|Opnw2?4KkJ=o%!9JyMHt@UE;;L`6hoivM)(kY{fjrM2aUcEoAT)N^<G%i
z*^V5@?7%KAm`6Xw{4jkecvq&?r0LsS^OcxYkD1lGz$QBRaQ{hrgk7`XVj587r4zWB
zD-bDJS+Tq*Fz7DsFeOH(Kktm85>Cl{vEf|}kP7joOJ`@Hvd4HAH}Rp2u)$<Q!xB+0
zZ+l6tak!&;@8-*2_2GDhE{J*ee&&mL@rkqhS1;E+l)@y(=SUAACaYr$2<Z31kLbmC
zM&fyC{;W~d*^M=FtlNI{nZH+b7pf~sbU)c((R<%J=aU}hCLc;Xp1EE1Y<^KQh}-k3
zFa(8jN_Ve~;OhPR_Z3`RaF2Q3+E|3si!-vST%)|_EpwBXL*v2y*6MzQ092YDO7en$
zn?vT+98ZX}ag=vAy@3iMnW^6u!7J=OQ@zP!Ou)_&I{V<$R>KpMoUTT6z2~wNFsHQE
zOsXE_^{Hag92u45+sWmx;;o*MM^5b`I%jo<B6-!*EDZS47xr|SgWQhaK%`{P5*hI;
zEolyVH^5cxCmPJ}GrjLSvcK9M@Jorlm(wHP<~X;!Grex#Tt72KNceT`OQ((E<2W{U
z9WB;5VOE)|i=E69^I;{!H^eou<O}rUa5~-}K^X2pi-YlxbqizfGclA#(3*HtZ4z6?
z9f1AbH4-6Z+n;r4t`&50cNfSznrsDF#%ze)RQ0R6yO3#yL$5I@LB?Brg_g&FIEjB1
z*W;1rjn<nyjQ8I^P2L@uE?Xp<YpMPE5c<v#O){=3r*1GNS(#Ho@6*6^^G!j^Ymshe
zqZ|&P&-S10v7*-$)ZY)1C)*+mIR3Wlr7njZc0<yyJh|)m){Rdf>NWcva)o}5($^fX
zfGVv_krO_56lG7U?@Abz8Vrf2n0Jef=ViakQ<rT!tJP~9^OnLabG+P`vi0SSUca32
z!sWS>@~O})O%3m`%C0?r^9~{p>gDY9dnqjr+C0Og=KlLpQVeWCjn0Q9<!{3vD=R`A
zKHI9d`6TAiJ+!&5<{yq5@04;4rwFLZ=g3Gac;J}VisgwJw<L^Vn3rxWv3o4T$l~5T
zmD}602wypEtv8UintSw>u!GI*+N-zQbTPTcst2|!r>0I5-w(>-P9o7pii>p~cvcqi
zXj0lmGgQ)(bOhLbr-_op^5Lj;zs(-TWv*@?g?7Zaoi-8r0Nd~S1Z8n9^U1_bgdVou
zt#~3m%hQ-6%i<WTgnh-=dMsg&Tfc>`J*yFY8??XV8$L}?xcKRlT4Iv6<5&ij^tW*2
zcFos?O|~#%{=A#Skm8`i+e7$F@I7IRYj+clI-^2G7wcE^2p?|fAtzO7E!Vvs<xa$Y
z`#Sbq&c;?}vt>4Cu5(^=UUFdMMXI<NzRZo0Z;$(H9<U$i7V-9%>02qVg;pp%ahxML
z^FUZv-sdS~NQ)Q~>a6N4N-5$!L@P%+Nr&Y(q_4G``aVh!eFgFL-{xC<`f8Hqj-cc5
zZ26-bB$PT@p-9ouv+u8EAYa_~NOlI)M+RY@n|;D<k~U?B-X9t>w!Wgs;%e+cQ-L3M
zj>d%4C9@)?of7LdFPTI?7rqg)QQ0$T@mQI&9bq&1Hp}9Ggo@(T!l7@NfMn4$(}`tq
zG}mjt!%2nt2cK3aBh5eVua;IeKg{TSQJO6_ESP4iij#BebX^=;e@oK#naD={%k39$
zukzA6FKBm=-;&r)Uwn$rOTXKScVOFj_pXYT$JnbkWt*OJU#M2wjpbtT6`zy4wGPbU
zJ+5tI50muv5!;ePmC^Kb3md0@xL1!@ygB~ynJ_!Ar*XWE{e6=DTJNaOp}T#G4;_kB
z+LyA-0|jI~2DJspo`$v_k1yaM@;!r|#JSsvTRmi_sB*qaWMYdOd}AZ;p|kkjw#B`@
zL;CIXwy7)Z*m3vM*UG}i8r@a|*k&J^1n(kObq8bCLpc?wCJrxCsHE9$bgAAmn4F2>
zTRhHZHXnEs>GTappL7J5i~opZg`L1BgKh5V62WjvA=93N7oD1H=~Z4*XU{qs^ONL(
znx0e7U3(AD9*eOy2P|KlCp8JKYPSWI9^cB&-B2{t+;zFxNtME7$8v3v3mKtUph)|A
z@=fdqAKb)X-Hom&iIkG4BCj!>gU2A~LHVZ}*A|s;C5yYc(pjmM-u!r9_>kedkMJ&%
zzh<$WY~Pbe&q{|~KQtP4GiuAAvFq~c>C2LlPnnYN<Y7BXoA0b8xi?ODif%CXVmon`
z@hD$?ZN+3!7$Fq%n7eDQ*<XyvK~qLVKry1j&pvt|wt<j1WpmUwLuiH^$9~j2%jXVM
zbCBWcbDz|lflE3#KR=#UrDYD94e7yW0Mk&(cY^J-W1P`fLQ=TmiZ_H`3ml#-9Ie}{
zE7;o_lkc`Wwhb60mGHe3e-fU;J$Xa*EM}!xR5k~^DXvSheV6@{FxgAK1JYPnCH=jW
z+40GE%Fp2-avoKe=I;}5k_974z7K9#_tenu;wDn!KIOt`j~Ah_=?}p3JK2Sy#w!%}
z)+U14j~%=<WgkRyXlLGkNkXL`Fe`U==-xN+ySSLlSE8hT{FoS*>>vky)NuBiq_1?D
z%3bzmVxD6nVtm-QZ%+sBt>dSf%7r5~N!!1%!5n06GBTpixN-Fy9DT)6zk5k^<dvCS
z4oaA7Qc{n{YcJtarSB_!G;6xvQQ&$b=IBr@M1a$s9FipeF$=alKz>ncajfQNj^3ja
z?>Oc&KFrM*L}n`baCka9zz&CA$AyKmd_d#v9(^ZSr$PIm7I8_RR8da8Ge0Ynf~-X8
zkc`Z#E^XAH;Rwz(7Q-$|gfTb-JqTc2T;{2mLwIg#WY}sIzhd^4_pWhC*h#6FQd?ho
z$3~eQe0DFjFc;4poK(%0r&swb3QJz$vKLm=0_`4s4B;g{H*_WOyXGa_?+Qpbk;b|v
zZ!h6cagZ2bcB*m&=w4}l{`CYF8WF5xeVYis1fhE$F?;e-EgSA#ky^+q<$@ZtQb3Ab
zt})|C6-|0|-$qKDCFWHGYe^pG+9jHHX{cF&tPSeg;sg;Lj6I~PGx$=;ei<Qri|TsL
z+c!6y*b#aq(yz5Jd+CG1?{g$;_S#cC(tL06PB@_3vJk-~=4aY)$Kx8J_50G(Cds9T
z4t2b$pZL_Gl(GGlg9?a*1d7?N%AW;BFOQKuX;rr@RYR9U$6-sqO8!w4Bj`4CfGG#b
z7Br3$jaOvyH(w_UjS}z|$GWsfK?vTa=*N*UV$R?h2TH?Zx2~&S%E-eG>+tr^vHI)|
zCRCyV3i6P$yY^+rvqn9aE{x5zxE7t#M+`|%sSz|uGbC4h-nlmF{!Jbx;?#B*S5Xfo
znasR#S3%f$m;Pf%vU5F$fu4H`EuV@$ja1E&8nvHKnG2@BT^Hpr!I7{1rP21Rk?+&y
zAz1KDcZ{7`jZ4o)=~b@WdS0K8=3fvTO0fI&u@>#)BgvOuv}uJ?IC?1Q1ogE$UXGdK
z&5=S54m|VDV=r1flhBxtSuU;kO_;qWxjn{gK=i4ejv^w&BZ!NqJ)2a3hd2HIVe74<
zqI#o$;WIOIOLvHL3CIxAihxK6(lv;bl(h5=2q@hlDUH<74MR&K2n-F<(%o?R{_cC<
zb^o~k&6#y(t+Sppdp~<WpV)Q4GBn7!1|Ld=^flr{R4%w5)DNmnbA*71Y#z<dh87%s
z4(oM_L<vDgY?H#`hu!`n<=XY7o4qzyZ7$u{_-UZ_H8bRSjJ0=W&S_-82hkEuO|R@I
z(L(Bn;Yu2>L0x}e<%V&6$@2H9vI{x+js?ulxW66bjBj9E-x`SP2mc~S=!|B#9+a%M
zaQLu{{)-n)=eDa11Q~*@m-RX>9}}97?;D;CBtke4Ax(&MjKS-i^3td)?kN6`s=pGl
zx1T^#iP&!%l^0B#wS*cTy;SUVNigsDEs%Y<U|sZ<D78=HoZe%SH8XXzZfZyRu?_Wx
zk%jr~W2Pu$6D==eP3?YjwO@Q1yfURKNi3#i32%%YjKbvUwt{cemwVX1XoR0nyxnW0
zw*4qaQ~~&lTQXuR#RLp_4^$+g39}9klz&BDj)K-UHLJqIw-*4H)^k-;tEY*66?O%7
zV{{m9v|^RhnSf>$?aE;DpTJi^2hs}@IOh7-1{N*}A~{vk5`TDkBAza+VohdX)QdlI
z6aMJ%-LOX~2XB%nf?IK*<W$d;xrGz{O$5nb$i&3@&4Lxl^<~zzzG5T1L`8;Ig(cPX
z*Lu8L9I*w)(rs&^wjjrZjxEcn2ERXAvfjq>tjG80U~2oxEeRCcpyx~MX74F-*fzL9
z6m*O!+Vb=RjEw%clKmEElDt=Y!N@*B=;gK7UNF`eHO%V}pFr;~d$IDdJpjdzX#Bmc
zKAhmDp=gAy+uoCo^D9H^oaLk~p^0eV)u3b3j$79=5-;-?8%lO~ETcQNOZ2T~`;R<b
z*J)EmRP#z4br`{GhM;2y?ZGt$+oXh#zla@O2$s=RoV!^!xTBXOaJTLHjw2)eoN5Bh
zr?V0j)vtN#WYTh=kD1hOkaCF4Nwd+d={I)2rlTY`3H8l>xd;R6Biz7q9p3>lsiX&!
zOu7w+IUe(M`xsMEvlqEsVV|o<6e~Rc2ER)1!CEWbpYh+xHM(18s_dL0eyZc^X3J`$
zU)<~$ar$UvBYNUj3Xeq^A*!{-Pn=7i`}Klpqt?)|1meB5*6+>hIgI?X<7Br%x3Ru)
zJ$&<<UmdiUBJulYi4&FO^KOr8S$p5rroLS)i*pn;Pg`kz-yiDYLFhyG2h}{0tEpTR
z$=c*c*}5@yC{_1)R?g>|KE^em*UPw{^+>eZlZ|E+`pBMN(^nU5vVF+WzP+WV<#V+G
z)%QR6GV{6c@k+~Epw!U3^PN!7*w0NvuafD&v#JlXEc$_ji#B>p7W$J&pIXm6YERrn
z7q`RG&NCW<|8^~n3pUhpX%=Wlt3a*qq3K&YSnm9nD#6bF>cN7Q{l||vdWQeG@BiD~
zFxWl`sxYeqXN-XP>$iS_D(p#>3{qaylxU7Wf4`Z=K5StR14nb~8a{h~&@=#YcX3|h
zwJ|9b->so`!Zxxg4PGjxK_hEhFakqJ`1daF*Z`dW^b1NPoqQohKzEG*!ch#gHt*eL
zpU6@8p@v&2rhU(?zcBsWzBmcdF+Ba=yeHK7XgT0n3VwHf<=x+~w`e@z&o#7^vr}S1
zC)C6dPM^RZ3Dl-gmYR5V{@d@N#6?o_m!5<+=Rp#D2RuIXpBH=d4T^YM2)Dm3zR>9D
zeXf&q$MBIG*==Z2NFoUA5CTY5e7oDnjBjhza4R$pIFxMlsC(sNel<=Yt{A~7ZFxIM
z=j--~m(}pB$A3uqi%nkd1ql@fG9}jbW(_h5TgCan8%O>uiPU`4fV;uv3)I87@v^ct
zf6;2tTP!yMvVGOvhHLoTLo}|5qYfqwWsyHO4oJZNCda7XO5Iw_Gy!lDR0dX72j0vp
z4X%*t_-?v!4AcTHUe$-{!+#nX-7jJ&<y}l$i9&KXV0{=Q*2J<`5Dl0V&cJ~+=lCuH
zSnG{$=g|RDZBX1lpT{>4x8v2hH&`ddM(p3epLVAq97w;#`0NVbiOmB1{>?NN_EuUi
z#?R`X&Xohm0)Ykg%(Cgg=9GYd;6v^|*NPAWad5?(#QyhZZ^6U_uj4<@y^*8RHr~jO
zDUdN`DYj{`2zK5SUMVlnbbKgQh(X1qpQlu#b)zg3DD=#*{q!F+#$x<n;m*(Srvo5^
z`VWXN+*$Kd;S&Eafi)w$z0sT1^ad`Wo1{N>bPQMTB4U#E9XJ#!^x+fUg28{SwZiHc
zqvg<&)r47(n}i2`&W6<qU6Vy+oi6np9%<ACQ@;k;d6RIK{+`UU>i=R2?W=6*wGh7g
zVMY^{hVFTqAj-KU5QIrDN^GPpqehCX_&P|sWbSk-<7mOTb1n}E`3F|yP=S00@)Mp`
z-!>V(OMGs~@pDL?7kitRER{rBzgji;9^jF~GLDJ%P_--cWMsCZ8#FEPzzUEQHanCy
z^ZU*_8+17y@5I}4&@}Vp7wjp$bzZ9%>L|qR^FbHa<>C>;(TsrM2T{HT8ofHb=dK|0
zE|RBFpUl<jJLiTxMrPw#m~L7XEFW!-B1Tqj_uaAKL;wfg%Flwc4n?u^yJ`n@cfhc~
z;$^huAzQ9U$<uO20|nab(oGg^#c?`^R)ZRqR>K5-6)`J=DH#;3m0{HAJ3E}KtBB#>
z7iV#2zsnCS|2BB+;vuLq@X=zqrv=SKjb9numA`MJUKM}mTU-7rfJ^}Nhz6;Qi<1(2
z0EH&#&q8SDG}&@mBL<4JG|#`|ySzPnWj{a%nw+22>Aylw!)@XcK2X3nIokBN@PA$9
zK>oD|J<M~S;FfdfO4FcB#oRf4ndQL&E%C*oX}#%%nzvZhI)M}Bw-e$*+QIK==m-UJ
zqcO8Sjop+^^C8_pVw~m-W94e%4clWci@wza)eu%*^dcQh2vGKuq3K7AqB|^7#Rq={
zO(e%lTaD&Xbwhda&u9Oo$~eYSZRA@nzWlZ`2Nw6R*VO0lZtU)hy(0hj8QMz5o|EAR
zPPT_9TPm3CG-SWx@+|dcwbQ^Ia7(i^i8Pz9CIbXSzb^R;FMSccq{Gf1Zn(j=vrPFS
zxvyYohC>b?a=QfCQC`q6pUsa8GN#ht_KN)moByM>!z*>rKQBu1esFQk5HE!~&xyDN
zy*JhP{Wh5HV^Si=A%nJH&xm`t<T`+iu_j+vj<kx;`MxU1Sl64cEX47yGkBmM_S3mE
z=uu|soAdW8OADPGs{%MR4AAwO4t9WJoyq%A)ccm}{yN*WQwvI-Pp{+|um13H%s<7d
z>(YC}+{btZslW5=luoJrka{z?uTtxaMgJ3a*H~aLvK1z-E-)LBnsM*E^Vsmu#idOp
znXR%&o1pTqY<^&S78H7JZJ_;nWftWV$0Jy~EP}2Vwm19UTtn|0>S@MmE=#j-YpS}p
zYh-BurBcTBCEgw3>!pq)pqC5FzjXlLwrE<N_37TH&PXj7jhkos-nQ*s*Yi(H@STMg
zL$wUaqA`Bc*5*poqz%Q%@62Xw14_i3a^^M4_>}uGqJCaZ;d{2Fd$M--|GpqSxc}y2
zZ>1pH<~%6@@_Ie$HO~Vdnh3vRKGb0}UGY>|kz5!Ul(aT*tZL4B#{Y=*QOw2DT0gf-
zgC5yl{Ai;9N^}E$thOT1v>m_Z98HO?tzB%>XKSFL3>o?{=vw)1m9DUB`JAAACdRZh
z(bOiz%x|gi>`T-QSC&aDd8idfUX4^zb+R!Su=r)gWAVJSu!vfF_o!rUx-ekQMS-D4
zctOiA*+h`DYZBD##}s;?Zex~Z8sPdE&{W$_A%G*(bl31H0)&5L?K<X>h>jQG+E%wG
zX(yEdr$wIzEagfTgIM{V;V}EPS=ODHNmTnyhMFYeZdC0%6!FfMt|YVJy1k~MBmcWC
z$Kp_dvZU2$b}gdi$!QdcZbT+Z5PEU{)80XB9Zw8oeP+Zo=r=8VS&b$6EZ2g<bJNuA
zP%ud==0ilkOF)<m_({DdR-td{=@+YOwq7po4R++$+{crpFxP^Pn^8t#Ldv#s-!>ig
za*PSy{%r7(DOuuboSCYMinBjh3mtJ24u;$3k$B|EO*5-A{D3vLXa_ulpR({x;x0bT
zOyhCA7I9h8B+5L}pzME)9V8ZSJjC6L`dE?iPez9Z=}jMq*}GJ?J|Nlz3g3k5Mm3jw
zCxDak64)o}ParlHvl*oD4Cdx@RV`$*cVX@Khx4TgZcY77P8_qQwM<yYS^Y&Lp2k$-
z6p3se>d69OQH`J-S*@?<SV0%;si$~0fwS3!i$HWXcKlLamZYRTcWR1^Eli&?Y*BPK
zjsu2UOZ!@>1(O-0dnb@kxYHZ@R_Ep9hL2Q1Zt}t`;28}0;$jbPuUP<)rXYAkkYa7)
zp~x;i3uA9dH=6k=32;UPpr1DOCz0UE5rg5kzfL2_RSW8pq9+1gALJRH513T{u7Tww
zH*RI^%$T1saIc8h3qEb52L$)BwSJ`|pXE^al=fFIgui<>O&?c!`&P;Waq|Q7tPY84
zJ$1nk66^OCDJf&>SdYk0P~pZiFs6SU^ygNzQ}q|6!{(TC$HmFqL`|LmutkED+^h(!
zi;|sOLCUS>nT{V+PXh-F8Civlx8)L{HZKbtrH+iX<OU@0=v>BZ6owAd+z>CDSpB2G
zN+pE08^W8jmOm$ZC&YX6N|9jytKu3VR!p4;3HGrDFkKJhPVAE*!Sc@#wP)~oCQ7sp
zp3^Bf{u=LO<+`i{5_YQ@8acIc6X7#X@vGwx#F=%C5ZTWr6TiF%#XX`P`*vW~I`~$@
z2&O-Q1p%W?-}?e*WE2{ctpL*IFe6o=Myqk5j0?{x+&-s4JlZWM@H<9m|83H{#MUgV
zG6UK{XmksQc%1&1-G99kTSuc)z~c0$gSMaixgxEKm2)`EzqI|Yq=<S?f%5z7c`to1
z&&i5}dSQXVrzxQ$Q)OpyYM(zhI|sa`eKoYA)-B@RIg)jsL*#Jn)INPnRta`5<EjIm
z;8*i9RFWZ8rUAReAWRFm(Z3GbTXsAC`W$1oZ{3ze+w4-Tv5(T7r^7@xLfQ)YHy6jl
z8wO+5n8G6hkr`!ur%_x2x6KdlA8LH%>h1mGce$HHy`Rix8n&yNCY=@fJgrBEeZ<M9
z@9c#T?&T$Qs{Eh1e%A|tw_VZ1Ky|N464#|I<M(dEr~7qt1@!tsq1I%|lX5NS2OO`%
zRJY1;Bacxwc`|r8R?z1KK4c1ACDSPFjWmFcCzjKvfXPYTp7o5iI$*BU5&QSOk~<46
zuqurqZrGB=)|+fY|7bjEDoJGF@s&zqwfHg|6DMyZW95EN?bCT<y3IZI=(a=xZY$*U
zNg8I{*sAhr_$=4oW$@$~baT~nH9wcB-F-X+klCKl)Q_quXZM|%)F3c@GcXu=>O_&g
z=<<R5GnvcSyM6@<g6vm=Pwb@AUA0(?BHTVXT$){9L(DIc0=g?rQ%27Z@|2?ZmG(Xw
ztMV7y>(y87oS{mpBF}On7jo!=8w~kY82_4+q0~K00#pIE((Y;wvKxkY6eb^Hyg+@R
zh<8(Hlqoy==2}|74%14Y4uYn?J`tnZyor^T<l|En5*D7IHYBj!@9KK_Eh1v*jN6-W
zr_JrdwcGXnO=!XRgrfTX#5w+7>yg0Qy5Ydv##;}+?(waEIR)qct}&geH`?zEH`@K?
z@1D2)-EuwsJ??#a8z44w47)EqhP64|r<)$SHePLRwOskh(eHX9?<YNxrgitj<)0;t
zeNRqIeA}2K&m__Q+mh(=S^o`P$@DU5FR$`eQxe7hc@V~(sol0?o(#35UP*nJDZMnZ
z^!>s94Y_r%vhnfxVsLDhp5TQSxAwzeMsb<x8<nP~-T4Vb4>rO}3g8k)FnS7tTCmEi
z?}})R08x7U%!|Kk=!;=G8<qK`3D3Fz-rV($`MW^y*k(RE#VX6Fhpxa%t-M=5Wr8)2
zoZIbTz%x&)Wi@NgLD3gZW4N*_#Zq+g%9KAS@6<*N%Zn-zKd<yhCx>Q$VhipjYPfCl
zCkxev8dbNQ3(*ACo^(6^BKqr4X{4201wl|N5yin^R^-GQRPTB37;Tck5h*9#7}Gr%
zSvT@0;a3Qo(CrN3eYz}n=?uLqJGT&p;((m61l~iRTLM7$*g+)8uI4XBU9Z;%qNdpt
zcyDFzm1IWq_5OsxmY<!Uv1?-62rg!T9T`Fb97aRJR_Qb||An(2-W6_x4tn+R6Qf##
zBMJ-NnwG{mE_1xwC?N0maV((4!{Im;<DfPLA3ylY{XRh4it{2}IBTNcDC&8K*XKw6
zc{p*@35l!hPx9CjqeTtBVj)sEUd!sC_fR2QgP8PA^nQXf2D0C>Fp=f{ws<j1yI2!>
zVyB}WCa0)-Yx^j^O<m2DC2Y$-9b{nxDNgD0s7Z|V(`=)|{cxX(^)xxFkpjh!zqX!=
zL=k-@whI@mYdLJ1#fF*^i;Dz{tXhu>uy<aCJYRMxCZ0UxA`o*Nn&&O=sQee5=Wl|y
zcDFK5$1PzzzgCqbbZ$!h_zm!!;&MfRbrkGP0NSad;s<^X{N36FBdtmj-vQ``2*-@E
zzEg>R2PR}Fc4WNa{d)3?JW@r(i!9d#!tEw<&YF$R{!c{lW2rm*Pm@8f*1IqUCe^Y&
zIAE}O@SS*G@ki=RmMN*9OZBt?0r;p|16U3*A%e|P#!fby2md;`mr!cNr^zHgR>IJc
z6bwA~R<c#bTC&G3M`N+SKc32waR{h+4y{6V{)lXBujxG2opPfe5yL726b`8M`u#CI
zF#k!BNZlI6QLMJXeiULN+h==(`TT}G+B$wZov5dNz#aR;f?>YyYflQ$b8vfo)+yIo
zx{%XJmP7Bje4)IylO*8mwV+V>^>sUzLZPJ9ZqD^?Ae*U|cAe~Lhvd=etFVM>B`{R~
zRx-D=Ys_Fo%Q7uaTlezr4F|qm^>iyTpDEpx5)GKX8k!AlU1pp;05KDzu0G-1z>>hX
z&BAH%RZm@Q1;Ns-R?dv$_V~ZOV4WSQErx{_Np;e%f`c|>F{KhG$9Ia!@n&2GxCNc1
zWM}RQZP`;(eWop~37^w$x13MkJ#&+awSy$a7L)~>I$*0@r?GL<>(X;|abei?dorc|
z<uAp=v({_efa+E{#>rnF((tUAmj#OCJeuXnxNob)btcw(B>(n418mWRZH{An<x@dP
zM|fz6nfse&P)gu@sgZuq#E4-?>vhlRyAnu`SylUo0Q6u8%}jd@3SWv&#eVeC+g4bj
zRX`Dpo_lv3i-{tN$Am7bE7J!(bjgj2!uEXVkT=syEmE@McORo(N!!m^7!)8UF9=<x
z@vA2LNu@I7r1?;sN2HU_?AkJ*vj?8uOK62mG4&2?cPO3{!=u}s{uztaJ@;=&YJnW1
z%?lyv?OvDX&aDg8`(MNFyh?p)i0p}Fn~Cbog^$Rn7k%{WDybh_wuwc+i@0rO!yYwq
zbKa0}g(>r)ouT_@<4cy!k6@mkRqQjK_DUNyWg93rP|@XVR?S&=)^CcWY472`8q{OU
zES<G_2JTU0-~Xsjv)HJd%<5uAdT<r`FcJ&{F9idgx#jm%O3GL*LoAYzuRHgBUO%4C
zsauQ2uh86lrsweVOwYjwJoK(ynbm}x&Smpnf{RPpcXoBdJJ1IE|KKbC1GFqh!xe#4
zrnU_C(c2Zk&M)k6n@~vnc#1~cZznlzbx)ahK6GgX%FFxaeYQDCNDO}l2`&@by?duy
zs}U9dMU@qEQb1JQnE!co#~0-nmqNAPm9DlxX1lZj)gdI<)itUe&l_)R!qr9I0QX?)
z+A<1jtmLEGPkUSCLe5kyg;84<=zsywkO(KtbWh{O6`WR&!`63rT-;Y*4V&R#TS4b*
z>RmZZ@<!dx&zJk!ZfrzLX$7dmmEydIn#uUXn1HZ@5PaRqqc&bbqw=lIzDkyzJBYp+
zYw1qyN@+4m4r(k1xg_v;qyv7suT0Z7MM6)F$%K|6L*-Y-Bw!m%m>}K#+KE;85^V*1
z28Q$+=ycRIenG-rFz2pU*RjTaI7lAg)m8SnElMtH9MZ`igyfgTlf%0u-UU3qchY-f
ztS!U%EOt2FNTTfYGihj(a@J%2qw{ChDd9I7eQ3f;ngtP>ljDr-zVAdMTr<WRwbJ<b
zI|2augIv#~2C1}q=Keda#%%lXP+k(m?vlBk`{&cCcWYE8TvJ(t*_nT)U<{f>yHkvt
z{%+`11tCuc@(xSNtw3)AU}q0WX)a!Z$RY>r6lB2-SHZ|1o7k}HLs>8yU~nEkG9;@`
z7sj{I;3f9;M(|eZ!FqF|!gqDlB__Yg6r4=grrhzZ_$9rAQ<eeZg!|E)w-s9mzc63I
z_>ps<i1BLi<20sep#khPONr^*O)YjOxHqmDgIgt?v&#`1e)|_y8yYMvuw{ZX@@d#8
zO!Z~?O(M6U&=TvMT`GCrPmvqTdn=erRmSo6y3amu3VK;~8fTEe7yKX+$yN37iTxoA
z<5SA&B0?mu3FgylDm_$QH02-q7s*Tv7eo%UMcVxbkLGFlIqUoKc>$I))5Ok9$xT=M
zD<z&vN7>)$5%TneiGcf}3IQ-H9W@$+QjsYMX&X0o=$7f369<snuyG~95=bfHDEVNP
z@^XJ?$f40aJV6v5drPASi;JCSOfH`i4E6``(L=~VZLorc2VCa236i`e-^8<=IYmUO
zR&Go_D#URb_aLW~_naDZ*+}1-k=eVRgQ17#^-~}*6FBolqK4tbFe&I4n9r@_^~(k_
zIAKC!E7nZdh+!PRl#B_YVd|D#)P(oa!@GpFW1i|(%sf?#68(VJE}J@{d>q&#{S@#H
z3*_W>@NFZQ%D;FgwypB%32_6nmEr~L>u4%DjzuE!(Yr4zz{^E!YVr3D22h}h^=7tB
zCO7@1SBiN9<#dEs>%ZIXO32;Fgrs$(Xp~VR=Tv*hQ_)S&pH|+|;tMV<=iJbZhRG=F
zz=<W%YrjE(x`5>s`Ac?Y<o)o?I~uuv{w~cy;uS}ZGXK7NX=)*w-Eqb4E&cDDCmb%%
zMYq?0)m_GF1qaONi$$KeT=2T8x^w<OsRBT#c;lE*!d?|Xw5dMg{w995g60n$P?odJ
z+J343(DMv*s12+Oms!c6t~S22Ye!R6v22DMeSVTNUhvRC!W~cN*Bb+E2iB1&zjb=w
z2Rk6AX8nMMjgQDzT#0OxR8>(m>sNB6dCz)OjDxOd3Rp#HqZh7x^$PdL_NX2-M6XH$
zO@S5(U#)9Cij#^M)3R76m8WNhSI;^A;(gK;Ca0rzu+>ozd@HubRX!zF>$Eb`9N*$h
zS6Z&m$cJH<mh5gWRFb~pqn5E1qvq2w$M<QXs@r~x2yx=5^Og}r_G(LMMEb+@zOI?^
z@#aaw5q2Bj5J{ltHpotCZz7<2IU{rG@HA%oaD>t!!P|d<SE2oDu5NW{#gCH&!&GHL
zpX2|~Z>l&9HH-9EDR1|IAAuq9AcBqkBPS<P(;t5^4c|l-Ty8;Te}=7JCh{4^gC6l!
z;66A^G&`^Le{9W>bitoIY3?R;m2}+4J86;Kxb?}SvTpdSvhR~1IJ$qT(<8fx@47Ie
z>xD=Rt2-IirGea<I9d?=zixdM1wfgO;$y<-P(98BVNa&&WF}w4cbWoaBeJ+LnA^4n
zu;cAUrI`1mw4{$8y?raD$+C4`AlPkW{UTak^8BaStP<#Uz6Obs2AEM4g)nf98~(-5
zCl=q!`2s`JWaM_e=dlN}N|2I4<4#%e&6N&m@mQmly2#^}cTbUykwHgJlJR0?N%EJk
z?ybojfnSDWGI@Xuw$lF8`Vb~oVLSdHB`0d*+?HKrW%-&^COu#PJo}1e0V?q_U*?a#
zAZFuLiWa!mR$qg$1#`Q)pwOq{%ZH24J~bY~HX1BFInD`{eR3?3Nl94sVmQ?vClE!T
zeF{aKH=(gk5Zqxu&xl{jTmAIjfmp_Y;Ukw!PR7_ZoxC^Zvk*tXNbEf~Ca&%xV4KF_
z`0?=?fae)oto8i|K&Di6$6PXB_XsY;<Uok``@W;GfJXMgqz&FEqDoHq`jA)StK+9E
z5`o7r%vcVY^psw{(5Eew#dK1`J8a7%<0W3KZ>V5hBEwqSAML|mdts2j8H^$Rrk)6Z
zlGnRmR*dlGVnU&Wyg$OJM-yk|&CU>&sC)`$I*)}Hw3{i|sz%HzI5??42(J1uPFmF5
z{`ypA{27^!8uaIe_RT>TED;fSslSV3sv#5kGM83#3*fAh^Co<YoQXC`Ug^bM1Kl#U
zvdv>GGFuZak0U-Y&99x@2MKb!@Q#med8d70D`~eOSq6B}nbBUUz~4X9W+?Nzw{@D>
z91p{jC0dTyI`a5=bB3r6wEZw!)3Yd2eI=l4CE%LR{d=taQXP-0YgmD$2}`C|A2bAd
zn8+)oZUjYp>ou_feOt~k9vJUw!7U9g<&CKU%pZbxXp&n!dCLHt!QhqEsQdxUCQK-3
zhstd2s+|<xKJ%J#siXY7I`PS9x^rY*@ZYb(AUf*1eGa(QOf1E*iYR~{F-{;ut*<Cn
zv^JmTbRSu8m#k#X>4HS>#((}K=wipbB0uGzhOP2@GHfFzbp9Pv>q#xnDA5LxF8HWK
zeh{A|K`{dOJ7juVoG79g^X$+8jyjP9odd(&YOKDPXivnLK3*NWE-XFQx@%~g_xH|^
zXft{!2zE6wVYV=QigHcBJj=`?;TXq<_7(w`r}l3ZUadjkr9Lno{GDeWQgYB;p7>~w
z3c>OxdKNca*xNQ(@U-X32nSCj;iSkuku8S>a^aX<%<&HAzDv~q%jsjDII!IGHcV@B
z;9z8~(tsw)Z6}_6<BKa$t!|dHXf@zLmnxFX55+oZ&+n!xR~;TtTRJJbX<fHQ5F*d_
zpnCge!d95@97-qeA)@Zrp{LSXi5(qOyrMgq&0n2zRNr%XV(j3|)i3?ws6TrH$-954
z_v@bc-TC#27x4%O=%xSe)d#ybd}~VPdt6#2u>|m#zzW3vRJxwOkN|;__seUHtwqf0
z8$JB3fgaWS`=iTolCeK*;(W+dtgrG$8=-U#uR8m$kG0vaF&wk9G8Y%#%}nry>|^fd
zEzi9eIO(k@yE7w06HsfY*Yqp&p(b&ib!T5jIcSIf8ukx^%}atQ(T=<d{zx#o@Y>A&
zCpQ)81W<c5lbE&`d(3_9j<>#;!sCv(?Q9X}0B<i0*9`R2p?OzE*mYZ%2_}vG^Y@ar
zWqQO0BZ$el=*X@*VxJo*Y1@4~BEOL8b@~YO>c<<0<imbb&lUvz2~NAkEg9*C(bIj4
zs8ngEcbg-J*@rKy&wV}RouM)oY*<`X|JPnGhfB_nn<nn^O6Y7a#`pSc`MCGujs;LT
zXr!0jffF-hzVkU7X$GU|t8PK*#Eg=hhII}>(v%Jh%$+$%%W{L7pAURlrO%(mlApE%
z?cJ{Y?ThE@u||jmqW{Uzm|WGz+C8#uqMX8Y7(NN5*WYerjz!diq)hHq`fr0f`0-7T
zH~8P)WnO+S=wRG8Vq2p?oyA{t6Tpriq|1FPz|RVmg&XJhtV{2e?^nRz?mIeJDSQci
zkw$h171S^0pwDF!{p|91hiSl>{v~KqKf+9EZ6ILgDmz`HWA5zowSE;{MnO#nMrg7b
zlBB7~#-eDyU6%%g-j}Vq6_PbE@E)BB{lQBPCN?$Ea<Ekk6<d@ukREUZT>&>T8cf4R
zFS=tdCF7Y)iL6<YWGg=CE*y(v`K1jRRjOXia(2zi3tnFarpL?GO?QW5J71~<4Kt+*
zt@7vA5r?L0<#(_?dL3zo#M;Nh7g4Wv83v#y(35V$waKS&syh(6+VU3|)8%I~0Y12H
zZAu@Y6+D{Q<%`eM>OGRnZ(<pb$>dGgnw>i-%abL(;3WS6nVr%A@Ol|cAqh)c$ko>U
zs=T9d%0De{Pc=EXUtR_D+)-+LOFVe=$ufOEI3YTYc!yR)vjwA=&cN@+!tcxVcSXm;
zN)Xx`_d5%jMywk+=z>mE6a-J#`sN6r!vI)dfW|;5a8@%+iQbe4ki-&TbS_C0=BMJg
z!$e)jb8`4CC14kZ>D~%KYgiB0W8$^timfi5Lpz-{BAmO6T`Ux^po_`&!OvBC5Q%LZ
zJ^1Rrv#A7VzrsaVFmj&!@7$AhQ<IjKD&%|)!KR01Sp22A=c!SaibG5J04$;)qmzjt
zYNTN}hS7hvT}L`D2_+c&XR24Qwtz{R1skvyn0J@WFPp~S77OPp5g6uLZ<a^CT*Eq6
z5wL4+sB~hp(Je0N>X3MoBp`74Yv0NUIC}=5>B*@_Vc2!WSU}THVK^bldG|gpZ2!{;
z0s|RR6C;t8m$y@b=)c58f}a6)2E)tJR|y<=PT+-;doKa#%RKmuamaVzMP^4&&F}%w
zNB@w>m91ZbZd`MgKB1Crw}!OT?~rO1PH%h5Jyc^;WoZV))DZ74ZG0PXVX2hpdL7c^
zKC0XAJlrP!jRIk9jLcW_ndQ+UGalR~Zvb!ZF<U}pfF0vqZl90vxdWaP!1HkKtoe0H
z>&}`+R736(mH%+QXtQ?~*!3Jh`8mqkVm4kJ^G|Nu*-_SShfTyPGc8+MgZ0~pWSeWL
z?)!JjqXK5U{>oRYO!8rZ>$vZ2v=p#lar*i)ypK=4m1(3vC+DmcwTOm>81^x>GJ00i
zb4OgPJ$%SJO-eNW*oQ6;;NJClFWOz#w&Yx5GCmZUacV{Mem3o?jH(G{RR?J3HK5Oy
z<riG591TL_Y(1l_isHl8n8LO7`c!Qzm{UZV;P*-R+22Mi_CPG2(K^(}bt55%J$9PI
zmhKMv6+P@xaT%8vkfLu9V&NYAh}*MgV}r^pCNICg*W&esiGESR1?OzgF1^JLSDpQ^
zq`NI%MAIkN+LEdhZ`C9cjZ+fU5<zTNw^+K!qYhv!4iko!xcoVvAmskFqC1II+URkv
zqIRi!^g@DlwD;j}{!3|I^U)D6l)biek7*@f+&itzNiwb0uE3|>x@mCw<s@egwAhSK
zl5}Mz;q>n*TR|^FdBu=jFZ%$3I)p0Ej1|kG>(fp6;H`%t@*C<j=0&mDz>}ft*~RiW
zs|`mFL&udXJ{Z!A2JS-L^VspUb4m}=+Y*T0RLqZ6KGs}g{ulK_^55kvyO~BZ%6Z(p
z9D30py>Z+OY&MN#Hk7oY|J1ddZ;KVNv2<2?&g`KE(PSqEQs4u_V{dyojjx%Ko|>1f
z(pXq;umaE=YH1o0djBj{lKA;VKK{}0;-Fx;;NV1LVFZxkyk`FI1OIDW43M2+jpmPb
z%(252hfrQa@;oulw~r0phKz>{b<_Tba!&D5?bFGz_=Tm*)4$H#$F2jz)5m2j9g4=h
z9REHrVI=*FcYgnw{rsEa;G2@|a%5ISUS$tsVe&oj`q;{Uns+OI@pU800k17Z09NNG
zM812+xsxoNK$CK?ERGvmY2~LD^ta>PXW8}>YSd&C@N>MaQfnzBHJc<8Gl)E%NuoNJ
z^k#XO^i>#-@9z*+3dU-8f6%ky#rD@P_R#{^-NV{L-0ifF-~J37n~b9v#M7gqCI6dW
zuH4*Ak2(RB-e4nBjISC~iM`tB77X!Jo4Kg`66C+N?zNuDFZQY-EKD$ve?fX+$LdjK
z9W3}2sDznTj%CwP>QJgXOP2QN@O?@nZTyU-``;|Py*HQ(Yc=$x$>~Sff&=^y6Kq(9
zOYtFwx`?*k?^#cXb=N$+qy?r5S8u6h2a1tp4+`0Z*A>Cit8_v2bf~k}ERl`NJ&`-)
zf5CW%zu+2QtLBO{h=_c*i8a19A#fB}h?vj}$G}rVi$3=Uc7)+F<T+iC*wtcqD&1?t
z-16$jlOfLH1#{2CFST{X#u6z;;7%WQw6WmnsS|OBvcqkK5>e!T1rYM|_)%%ZnC0ZO
zxJa%~1pVNjvRAXuPWp<nM%0u?;#&Z-F||834mXBM?nhS4ymTRMUo3K{xaT`GWCQ#b
z5%M=W*i7gQt}gWec^_8JE;B@F!kzW1OK`R3Yyaz4#AiH;cup56zn*q*s^=BylIo&B
z)ks2md)6VI>>!?J7TS`5QOwW4Zx*R|71wBZ$*V}XWE?1$aEGQ)20YS5tDQ<NaZ{J~
zAPN;>^7N-Az|I4bXXaU6yf`JHijcXumXm2Bd3P%DY3=FSnI(Ye=ggX3^o(LnG(-4G
zb9RmY>A^Nmj0BJz^V9u~mZW8I>z(Pp+6qdeRVh6lM<CU_8XWPJd#)mO`?x;EEuEGR
z$9Q^IIm&v`5ztP(ugYmLmdG<}dqr|)XlL@7VFxr&pAt@pJH-tuyv<hE!VH(-mD%k}
zJ70_>|BH4NLb_?wV7-F)^CI=BiCnm-v6;FW?c0g$BtN7dJf76^KuP!#_=r5e)Ll2a
zpy)|$vNnry(kvYs_525hwWdpt3ls;f36^cJ>_~ooLmPETs4m314x)>)I{o;VZ?xcM
z{^b3W*jvYk$UTnO!JezW<}_;N1@ONTosJTl>`c7-zHHd9B?Fs_i?(ko8u$QhH-T#W
zs#n`mqJa+;)ny5MrrkdAmSl`6j}g|jJV!*FFj<pxoG*?lUACNE?1;}TozL0II7AB`
z*9H9UMk~qxJ99@~FcGq2IA|PrG=FfnZjjmAAonzH*AbyA<PBh3O<+7-?Xkp-bpET?
z#z<Ggbn`xeckbKatPrE>?)*y$SfI8^53R#dYUStDa14=q>0`q>j4=mH{9}$oU;R4y
zl9ls0@vkBm)*4_;ZUk9x)PS^}36q2?&O#4T5PL=0>8U8`Cq7+iwNNUL$xL!<_K{_a
zG|!`th}~^l8=DrZD3%w}oo?#iNZGOqr1-1tHpG;+x4XK!nBlv{8i9xZf8Gipt(Rqs
z<^a5jOWqd#X{LtA{83923-;ejKQhWn56YaaqAZO#@@1gqHp?o*B?Qe2Q07q?;(#BL
z8Rsk(AF>HSE8O|sSZ=%eA)O_%XJMT*S!D6u;<U9!4e7BJX+R6%RN4VT*ESc7dmw+?
zcf->V3XAgldXt`^MfMO?{9v5i3Gzf?VwArbm=}w@exSBLcsOg=^Zng=5s_u{ibk0)
zgtm!lLvHy8#e$MQY#YNfSH)hhOyZiz{q5ITUZiq_I3F=3dT{vB1;Tqj>KP4wu~mzC
z+Y?|%CAOa@NRfefa_ily_}i~Z0gf$l@2op|3uP$L<d4{D!IYuK3#NbbSA-o08UNss
zm$F^*C@#6kAoiN3tf~RmDRV-NF13a#%o%*H2U3Q)o&jy*BcjtT(JY_AdzY+++i>Yt
zGtX>V#&|gq?pK}ib>F%=73&2M&I^PJSDc(9(m}*_c?4a7UnG=jq^o`1E?Nlj3*UbW
zRo46(1)ATbyb>!2Ag~|w$2-hMqRO^6A*uHMEWi#2C|i%KYZfkl(+L>Bh=VcD-ah~;
zz~ImOOHEuNr;8HSpbK$5nATo8028Qt0F4+TMEH=$c(5vM?@E;4<(|7{*VrCIC~fmQ
zy(jQ?IMLJk6LDaB6oNJq5wKud4|vyygxk|Q&Lu{)Y8_`py(gu<aHj#;6=R+)3cY;$
zRaYAYpN}0=Wohj!JrU*VvZ=Ae4n>!<dm&No*(3sylBx*p{ZU6V)gnQQU(G7=C6Yw8
zO`isxN&@HB0kzj@eAH1Z7mw_u4Qo%cB*n;3^;pQ5y0fh3cbk(v$d9fI@V<&y`j~<8
zH^~h&kv3$g3?iT@rVW4|$phDnaDNqC_>w4u4=IBMpC^K)wjB#(TG1pB+i3_B7Zj5g
zV?upkQH`~&m97?ZvX}}icebJ7-&5DW)1}nbq9U>2g$0hL3w^c1Ox3ODo!rh<dFmov
zUP}~q)~s-QM^kR&cwZ@ttGmbgW}amP8QzsHs`cf9QyJv3JTJ&4xyQJY5AWpX6lR6j
zS{BEj3e<ykVv%|ofPr&kg%kK=daPvTR3*|O{h2=xZ#={t@!*L{;DclL(zL9bU5<fy
zczfCW(%24qF<|7eZ74Z*I+W@nI<P5Tw0FulxnBn@eKL;<cd?j1iY^F|s9<0H<OimE
zQ5<=j-Xu2X7Cxkqx%m4ff2meG&(d~u+De|VHL(`e!#d5nRg#%ALJ)r*=rmOpcR`5d
zHOI+*jOKl0kvV77>96_SVzVkcLQKl%%AqT}`>DdK28}^kQO-9yCA6wMN-|GLw0S7D
z=0e`tG~YYrk|_F0&GDKrZZ34y6ij(@-_2X-4!)F<K0f_;N&W7QV@O`{hfz{Mf;G<M
z@*(=q>eYk)SF!X?1|;q3O7K_0qKm+7?ZmBA$mMm+fUUW3PzB30zxl9K*tpOtt!9+1
z>0J01|EfJ}?zQ&^_6QgAz27N<QS0^DekqsZmw%J*hi0UND9wi$8|TefVF26p4kKZ%
z)5O?DwI!4k|38k}jndLZ#Y&~^1`$x&8{QgP@$8x8o4+eON^f|ft40_*5%gFOb0^pj
z)DONhLm_UZzkmPkKR+E#f1W_$PJ+_O#=XP+-xb+^DQ=001tAGSQwax&+1?JOj5}XF
z<9pTcj5Ya0l$_uI5s~iapW3QR9fe;MJxcN|G&C{#@4IeeBNiOkhDoL!3kI?YjFPs7
zAlo4D`Rbp-Y^ineHW8=S>$gWtfIS_rYxWX6@mm(SEzhEco|5j}r@av8-R}L#f%vti
zRca9;35^uUM~l4OH4$#%BBiWK+1?dLEVn3*3;PH~3sN~3VcLWjEm?3nT&B1@S*?^8
zF#+UCwxyh1%DC3k2Y<6kLNph)#9QaN3U~hYD2CPc2?O22gUwn+S`fTbr>V)f5A?fB
zknwCEXE55cfW+#1{DId+``%Tbv-n`+l+c&t*5;~)YH$s61&)g9f#B)2Ex_sF(@G2b
z+L(LyW%-hK%%{pog3p#gY%-x{;ggyQV=-yM7OH%{zjX-aw2BSieJEiT+yUedC0=}{
z`3-PSql$f+4RCD%po=IUumE_(A+Y!W$#j}|O0Gh#Q0|5>mq1L=4x?%@+Vsbf3KN<Y
zCtJ@K)~nd}L8EQVW2+&Ni*Sp`%hnHycQ}^ilY4(Ddx_d3WJ0IDQK;lz6$hae@{W_m
zD+h^SrVxBP4FYUh)SfYgG*T6qT@Y}7r+3idnE)tK5;1X@1*tNhsN3D?5xoAFQaols
zmyA_OXop;)9c7o>)Be*O#DL<*VnWX{_`tmLNjAQpdje*0E*f9!9+*37K&s2dON|`q
zD^|d!B?V`yERcAIs;n%GMZ8JnOB>7PxxW!d)J<BbI3iXQvav@pE+3NQk_x$?&UES)
z&iw7U<l9`#R&pWTR+{oMT4#>sz)AD#IW|1q0)kX#c`)Gf?|EID6Ol!v3qbjh`ji-D
z-(|pXU#^}@jgq%F`ossGW$nW1X|d`!u<;HGUUu#*-~?f>@pYR4)Kltmn`-(1?*Pa;
zJ@_UTkZpR31<PR$6U>wnvq&lVk?G99X;^j+%gYUDv&E4f5y6=9_DKXCg?D4sAL^MK
z$I(Zv(eF_lE%jlaHP}@zTjE|}XTG^#GQf&f)kU(lOf+GJSGbhfG)#Mw_e9poV($J+
z%JRJTDEK#4V}@xHPWV*KR>!nUrFdnVT_DlIm8MH=0AN12m3|QOIx9rcy$k=^daDgU
zds>oUY)#`vy&*Y^A`c|L<17+Kz_OrueK7+9{w}Pu8_ThLoMOurN%rGn3R1?Eu``WR
zpI_zg8Ogxfs?S$`_l@vXa@d~>uv4=un-ND%F(^7ELc)<n6rc;ADszaf6J7NQUF6w%
z=x6_JiyUDZiCUy+7r(DGbzv3k^)#8fO%-+XAA({hTTC5>F0q{1m$B67ee$DP-@i9W
zdGZ^LksJ4RzmqBQ)c<`T$}O%P+U=g7X^!DnwRvMPK;|eE6&}T!`P=@FHdB4C?Ty24
zn}dp20O|P9kpY0AOy_nx$ph!<@dM{cP5EkX-gyat#&5%+jHX4|@qC7qOCq{5ldsBI
zjr9~0ncFzKY|L8Bdx<bjT<m}Oo0*x#Mn`8Ac8WU86#{muEE&nv)u>jaf%?<B|4UwY
zA%J)&#^fyB6~5p3{(F~b@aFbbRJxv|P!>Ok&HW95FGDIGE3u7<by6VUz9-Vt{T#Hj
z2MC_S7p5qH13zM(xgKq2?mFX#o|@#@6+}lXYEX}l$INU6PlG(XEj#f9tbJy~&CH9q
zf)j9-nOr!ezX_Vyak^HNNx%^Yi{PnE(fsy`+E`~>dBerf_#`3DNs5lMZCtId$}yU<
zrY)nq&;9>+0c4YiUg0x28b5xSFNJ~Uw5gY#F<$;jLbF}Ot<Rn2u5E%Jcuv00h4mAq
ze?T#Vfk(_?K}#W@OX5%4-)jo}8yJhn+}uoH-HKiNI7o~SIpi$Sf?odq(_7;^@eP5W
z)c9NVE2p^&jCZN&vZur6GZ5RoVi_$9)JcuQ;<SV&6mg)hewxFy^qX|#PWpxjQhJp=
zn+l*JtEE>D(@tygb$CZ^s^8}4@Wsf-B64P>c7OJf0O++NT-2}w!dFM4bLPX&P(&U*
z)Da!H3`qstgNs>$WMjl91BvZAD|e#9G2ks|Fs$w&R#h+o6noGGNh&7P=3lRACjHX~
z4B@{}1lhqggdK)ciR~QdH*xTHC}zeY<zF#9*TuIZ@PGIPr3G%@cBC}Ylo1Et&JGVh
zxPu4OSjaq}xZbCgB`VE~58aQWFcfBk0K-IJjOiY(9Qw|ukwhOeIh{TUEPZMkB~PS8
zFTx;i%k3!9RtcEyi;3E4zA(s}&yRK}nRutSu##Ca#b1D65P0Qyfn^;kxLNGwO3Tzk
zeV~txcbEeNv->saWL!gla9_qF<>W+h=8#;cs5&>)g$b_t!VReON$gVKi>2uM@kUpg
zqcwxRY6v=1eojV|DoqQzC;s^Sz1G5eY<xQfmPn`Gr0L=B#!qZ3P_=p~Il!&KyTpQW
z#AzcokBhCuU0NmvTIg^}^?}<p5+v^f;zcq6{}GQJacs|Zl~{;$XbRjgEKXOjE^aVO
z_boNLt?*w*-<aD|O&;4-*dKl-cA_!Hu5RZXY!?78{hMg+$0UX=;<M*8*Gq^IY>!VZ
zscodo)D1W};F-bg%DN=hhD?Uouz0c0_WD0Z4H@t-wf~#ITsow^5i*6G^Si;!_Ny&L
zNYC#?@mF58sV5Dnf0+MW!k(3y@yWc>)#jb@`%ch>m-^$WrZoIb$h881Zh8YeoDB{w
z)!j`45cJ$hBSorQNcJ0i6gX>kDQ_)jl?BOV1b)R6S$c*#v;d&O!)StYS^8#@?_I%t
zReH~%#hU$$^KFYtv)|-zmh%st{m*+27~_Ry?pm0w?9R@IuA}<2Ov@9PG=)~<TSzAy
zuNBri{rIeA_<tBwMVzjtA3T?$cZuTwuFna##%`tbU@nRX5wZ1dTDeT`ex|b7*e(b5
zOVFRyArv_&2D0WRpch{O{48L)&ZGRp&V0!7uZziLYVF21`rpE$l1MbZn(`v%Wkx9<
za^BXlmnanYU}^bY%`M7y;-_|8EE}PF6CoWG1GE#C$AUTV=zoiw|0CbPqey>|!F>g_
zzFt4B(?aWeUMS^P_E`FWr%rlFA1%|#t`0oMdy&QrPxt2gq`wPBmg<CWfMG9UuCaZd
z$l(3#pg<ZpWBLFse%kh1_`q*97PX*W9k{bP!f*eY@Ot%2QJ0g{w#j`He}npoYAFn9
zOXm7UMbZamgjM|Yo2ZCpYXaXQl|<#t+%oQeZNT$*>PntkM~)u=NXbgc>y#k-15S0B
z7-akJG%%hjTeO}_kn2MLctZd?AOjB->?DNhMwA|{LsBgkz-;3XjeY?*8#$gGIe1DY
z_-M^}^!szJ3_`#=j3D_{zg8+{8aL_vg$8!dEU$m*NYR&YNf64!(_5<upj9vhhsiOa
zF^~ly)a2cy2>d_h+6m(J^#cP7f?~vG`!Ed$p;G{)?t{uk1`h}lW6M~_frWrjL}|o!
z>VUi91<Q%37#O`tXD&E=VyA|0H`#dKP6#V#<eP|oU=IpXHDE%$`9*z6>a*5Juj3dW
z%BfeTZui6d<`2wRM~3I*E|{g(QW2CA8Fj3gdCn0hnHbhYfqj86*3##TAec9NGJOy(
zl5Fni8J@}4ukczW+%1<<4fQJfy@Ggt%1YE}ioEl%Dt|wCb`SEPMZiGXzHhEJq}s?L
zu{wkyl!2x)Ci1@-z)m2V`d~)Xb~Pdoxl;{!CjPQW0R8V}U()%B8}Z?%=Z!}3C9O?0
zRKV;3Viy<X@I`3WVztz3+Q<68%i)JQ+T`a0_=x9(504<4dN?FA-hO`NdYU*jx93@f
z@)M1cMlyix^io^LMuZ=9H(i4XO&`cnzwJ+AYzZ6xl};ILT(Qr7t7)l!oXnB<nI82N
zJE|qS_M{>AA13&}I*mi?OPe&3E1vaW{90lYwRq{0O277s5i9&u_YW^0ndE~&`YEyG
z(HiZgYr-r8zCKOxJ&!U7t-PipR;C}()pF0Kdc~lXG!O*^u7@(P;Ho+_#$gLuM?ShB
zv|efp4&I^hdQZa*7@2}!mSTG4nG@d2j(B57ueri9$pZPV8^5gkCxLU%rv{_v-h3P6
zxfEwn8Ix*1IQqOOTym;7<;$8AVb1vSToMG;bEAlNc!PlyHPc#-1CUmZM_91!iiHHj
zG`;KUvRk9SX6EwulekCqV5H4jS*GyVnx2f~uEUnN%{LtA<Olq+!n1RUmN@l)*pd-4
z^-<zi*R85~o~O=!L&DQX6gLlXicrbILDQ*&>KAytm(RZ=e82gXsdOm1_;iO_&kUzx
z;Y${SyjZ-eMDgtsGZ(2^F?!fOsg8cHUjESX2jRNCHjtE5UoWQ6az}-JijNQ9wWY!L
zmgL(G5TkV0zxdu35S`!`QtrvUbLi@c_g4>Xv21&XC$DqG(7IlP>>Xd@`f6^L6jXl6
zgQLIYaq*o){ctJtXuoQ*JE-96<hBXOZd){bIB)#-)MvF3$^cu%D24>G-leJF`yae!
z2exFkbsc9l;fP2o=F5lNGl$vt)8*~gaj;?exLLXa6WX%|bw@^tCWfF)Y%r0BWmC(k
zwT{={o%^f5jEd*3$<KhmXxr70QyS-h2vWq)j5eG}&2Kr!mUBXWh9j4N|1~KwUYhO4
zqy*i@iMX*CGyVjOxY713l12X8t<<qtQ_&zXnLSs##YkEdBN5=8iJLtNz9xsqW&Wtn
z8F{7-)N<e*%9R2z)9*M}iuWi9ODPKA>=wCcEQ_ZICbsmHZY-@Zddo3<*ctJ*^(GL2
z@*+jnxXLKHukWD<x6FTV+%+BDXw^Jtw$-Lox~qq%>)Vs|(R>9$)n1``EVv#fN>`0N
zN>p8u0-=FDC6>FGY!a(I5@wllX{6bNFc_#!m%o8%e8uhn&+-I)mi2)##VxFCG;wuJ
zV_747RUXs@&pW~1`T!qt`bp7@#}VRIJ&R-`AGo@{wG+SjEdOIz1PobwpSI;-3-3`o
z_xYySEwyFn-SJpRfV{{j!&3^;@DUqoVmBi8*af;5SBKJYihXuhU*7Xfs^^my`Mnu`
z10MOG-JjjBMKqMB%8fLC2FxQ8+3zAZc)jJYlZfrER3>P!sHGnApfUbRo-ff`^1q)(
z1XC!{^f@%Z`i)nV8A4Ux0HXhH6Ed4c5<(hzF4*b;dFTX)$p2!?2HtXSzF#y-bogb&
z(IXy2*=0xKqi`0y+&X+Q={@P<I}+Rd8bEFYW=N1GI;bJw-Rj$CO8$B<>;U<+eWmG#
znr39VwVf<TY&EB9H=d)wmb`cX`(e%Kfir78`{1oB=hw`{Q#denIPuPJmvN)Xj5q5>
zzOZW*7B&4L04_xfoasD=4?#O?JF|kk_;*8Y8G$p5i^<kJn>Bo5l;sPvOA!6TYO{{s
z)_jW3b-v93c05zpN@Z&M9+KN_xq-!MgyDB(-bt6UXSgsK0Gju|ID5;esNV4JcZdPW
zQMwTk>6UH~5J>?6rG`crx|^Yq5-F8bN~DHv7`j_wkj@bp7(#N0Gr#|N&N|QY>a2Cv
z+An7DYOl@S_kCU8>+}8I_myg^bLif>6dV39ZBJlXhyXx*?zbQO-CG<)(Q^m-l>F3V
z2^_eq$+R~*&ACQa4eYs;D3(FBLBonzZ|)xdJ_$tNX46}}ap}1^eQbNq$p1586Zd+F
zgNVcK>A=Jetf!G6l%QQKC{k8f;<w{|g+9!rovev}IonX46Og`*43@ASAhz*JElHBV
zfn3AMm7d<eOBK7AKsu2@#TMe;I&UGhG3GRgy`W@L=6|{6BC*IT&wZ)1Ew^0meYyP>
zF_xX$ZbbXp%dB^QcP6MdPUj9k<+(*F%l^RypE4WKfSj-Ay!PAvNu}hSB%j<)m4AD#
zN%{_Kpw;be^)g((TJ-!NLnW5{XTTcYfX;QhfN51lx_~)B=KRFroIv7?6!Mcff*fpm
z)<2<Z*YzsF`MUX;`*_&fw=tSja#=Hqace%iGsWhh9Rl#;e|`?i+W%k14P#ch4&zn9
zxI#J5`KmPNeE)p)idwg{JSA_X98KFdokxgq%O(8(=Qjx$O9P(wAL1?YJk~`n*_F|j
zC91qCK+sL8p7x$RAyNV?HhoyaZX&Kn&|IriXeAEHS}z>tg<<%s;H9*O6Y${IM`yst
zR$T7B>I#mghO#_Lc+8xGjAOofSe=f2<>G!V_6S=$byJg9mPD`xz+ba6dQM*syy}y$
zBh{VnEjHzhg(rG=#LXC>d6Q`opBZg(_+X20C$~S5xlFgf6Y7-+%5Y{PRhMnz>63Gx
zvtp%feE=)zh=8-S;RF8I`?*qp^4PbA4@6X1$`(}Nxemm~y`J+}Po&HBqoMJ1TH6;S
zC;=ZwXl~sh_BF|$btsjxSX;0)5s{xMi%B37Y4aWAOZv&^JnTR5s+(+I?lk8u4-Uhy
zJ>Z-j$NQSX#Elq!=b<7F$cBx}xvb;0g~(sQ3ADFAOJ&<7^b`7G&lHHFL14!{0Lz%r
zBL{5lg!O^mw0`BCAR3t>JWoTo!}^rjRtEk-nI8fe<@X77%^oN4XDZsQ;RPud1cPl7
zTMm*|?dskW74Mzq1cPZ)bpWs|(*<^+yF{|jAe5mx2}G0KShe}{-IP<v^xW-?!vKTH
zqi`jalv`Bd^^z&HuD{OLNfsOUDtwTK^_7RsOX%Cc1icb#-cHjAih$pCfHh+vfP!WU
zn)y{!w^`(Q`M{&jK^h88)t3h%sWU{`uinjZaf&JJGkiEBxB27wKB(6H*4rSidY}M2
zK@E5s!v$cZL68G{$w?qIL<Ph+sHfqVzQi7lP@*ok)O*Z(%HUcGxLWX-)UqPmT@HJ+
zq+=M)Lpfd6Dw?%p%Z68@+S)wpudnCCw*PWL3U%t<M8*0;@KUYVW@!FZ<OMHLRu_R#
z&?9Kbm?~SS(W@SxKH7qZ7pu6~iq5@BIMJpW@4x<8s;O<nkw&ZuZ*=1XI-cZtwE7d5
zc#4*4qmTrFi;;sweTa$vp|HKS6MI$gu#0u=&K|-v`;CnsmH|T18K+HM_}eChbv!fA
z``V(WZ`EC`!r!dAizmH?|Ie76hB1P}-;q|0{fsH)%z(suKHkn#gvun3hpCjyhGu6-
z<8;;*|3<eB-w_uP18lqph|}{}eC~Mt40mdqfQEW%uBysh?RA{GRgxWR>~P=lUt{NO
zN{F}~WBBF65U^hro4FF(O0)h-$jdb48-h!?h3y&C1s7rbyaM022TA_fk=Uaggc7P>
zD8YVWtN`(Tm9F)A?=V&)1pNn=`-&Gfc>UvNo7-<Mt>&S(kim7z6AJNqoo~*MlXHhm
z&BqKx(N6DOtL$`#gC6((NfR|5lB~5sU=>NMgZp1g2b;qRRd|xmQ5xdc^F>^6*1(?%
zuMW?d8vD20-~N0(d%F{{-DEoyc=+Li|Ho8f`+*kmDzj&=_Kf2)9?IP}Z!nIZH~F*`
z4HnrwN!{_o&+nJml`&5%xQCg|PD*tw2D^wu%lz$E#vNL&TIiRQGEz!pC`8kwm)C^3
zpvqiON=}`jDrOm+K4sc)f(c(n1{$7YY91ciNTWaQ%YTwGUGDT1YE60f{rReD9?|3i
zX?-q!TMMy@$%71$BGO+EUGp}myTMi=vBxZ;jaMw0k^6j;#}hon#}q8$Dtmrd_Sq?7
zvQ#(UNKvfY=}yLTD~_bu=n5Y6zQuw#b05Eo$eM9gQ0)~qQ6j*$eeZX<Lqg#yy)421
z?#F=lv*V-y8ZGK2sl^_EH)Uu@V3C9uD)&U?><_r{N%*h31uT|oWfnC2X{Q=qy2~!|
zCkPavXkQw=z2M%B-L{{)Jb3UC1PN`Qj9jf#zi)aq7di4acIJrxS~lH6&h+UJedSAG
zx06VFbgID#Zm?`v<IDkG8y^*moIn|0<f2j5hnjWA>M>#5F36a{+2Tp;`}H9BgO<M2
zIr+<)<KHs}o`drR=S38&r2I-s7yZ4tiU$tCRJ;3~OkXN>Y230>^&YSYjq2m>{;Gbb
zZxJQ-QsjsB!JL;C?Q2E0Jnq8B`T0a(6#+hlTS|~)26p;h)aE<S)<SSMnZo=+?)}?!
z$qqNs_k<r~(G7(5L37^?TfC&H4!2%Agjc*X{*d{_ugPOR)nhQX>h1s=bOXcVTloae
z^ixW`<Ku&M0iyjgxgB=AP1>A9E@$?Rpv=BHnf&J$238ls!P#Bc&OaeB#d#D=ip2tA
zqeIRGWq!CYm2b}(#RF2c24cvTGub^{wC1H}I^I*iVV=8yne$<d%)e-P?H(bB4HZ?~
zh#ntqLMw}H`&zhz7Q|=&b{!u8YybXW#MY1nLNr=H=Ka3>G?C<2Y9(Q#&8r;LYhQR3
zJ%sRR&Rvu?E4J5Kvqp&qza|Ub^YFI3j$zJ+hW43LZIF{8;<nEOE5EON$$rusrs2aI
za@==N@G#5+Y|6dI0B(QgrL<yeYC%f-P{pd_D=`=$cLHpbtiNHv1z^6pl4an=JBzzP
ze+S_M#@NEmG08dUVhHckh1%R&T!1(-de(pGX}=F=ZzqVu2SdzJWM@rZMh4i|hw!Z1
zgWPPoWS~8MG!ZrTj3$E@wFYsK(S>LZEHPmj0YY759ePJveU3zuegpdQ4H$bdNhFMl
z3#~WtpR6uq*Wn-};JBDCVZVphDcDHfYhPRCCSJN^%Y?+;RcZ1-=NedkKIRnit@>og
zA_ugCc7&!Z#fs-ZeWLRgUwIxL&YLmd{ykdx<5{9*6oFZDwMhlSpDoE^L5c6dVC1(%
z$m|F%X0lS`#u<?D)q?H|cIrAi|BI{{grY2P2~CiPAzjAB8@HDU8-I``+~YKNP1mkC
z<G)RTf>76gy>xu~tE6P?@<|B`hiI_GUR`afNNbP^FtODM`sP40dYd98#6p*{s-9$*
zXc^Tms_;hb=A(7J<ql=RDSYmYok?u_>p0e^py*PZxUe{%RFbhMvuf-pxa|jTDSre$
zOg*$oXN4mOk2eV?ppe7ANsYQ#`@5*~$4Ao`OI$=&=UHsgcej=qHPg@Dnku_X1^VQz
zF6c6;)nedzDZ=csXMW$dspSw$Z+e-471I8jpZC8{j6KCp6lYSnPkPBOdv;8s$*2yn
z@s5~|>{q1XYkHr;tR_tHua8P=rtR4z<VB*WGO}T>9hZ`lJS%XL9)#YhMzUUQ^Uycp
zGd}$1M{9fXAQ-$Fq+-LjzVIHG@$=TsE$7yI0AkfN7@QSLdTo`r=cf^o-hKITG#zbR
za@ny{#+3sKU4<F9VvRA7cYUHTTo^gxdd-fa1P9{Nd)e#Jlc6AQEgqst<+DczI!^e0
z?k`9eLBWNhki-)bXSc~e3DVAej5oiz{^${`Q-JfEko)U3-l`f#$H#M34nDVy{RTwu
z?qHT>o$GG0CF@gVo<$Dv$Pvn_mW?tmz@qGv4tI!ck_ZkYlOH3hHaKthj9>D8nwniH
zOh1NlY)5oX5w0LD{ca?apZE3a*N=OKItT0-7$sZ{-ygljMa0L)OKrTJ-D(kYI}e&$
zg(f>_%XCfR<zAd-ScHxaf4#fnu_;(}<~In<CzM7bc{^@0)5LE7hs~XTS8^i{`X5~P
zil5W@8d;tQ4Xtj_I`_hwC)fi=hWc^A`O+D{$6xpCHhH=E_+;n%Q)nLJpj5sPZV7VL
z(KQ<g5zruHE%?~5wOx>QZAuZm_U}6rFsnNZ?uy$g6V}!J{#bD}Mz{BCV5yOowM#z!
z!}YGLH?cFkoYCy7N(IMn0QHoLZvOa4QsJqWW|@XMdEqnd(!L0TN3Zq1Jy^M=Rveb`
z)QgK9LV|h-#p^@vY-3l=F)6v3O*?L>8mBW`AKwn>!$jPG&PYz_-|0n{bfEt_W<Qos
zrFCz?z28ZVm?KS*kKi#}SuTr$#yamHT%fl^RDyT>?OnM~dp=Ea3yoax`n}4Zw8C+m
z-upF35Xs8o2!i#5X+AH#*g9j_ZkG`yq1wR4%zRntaG?Z4%_#v|jyeT$)q;IqhEuxG
zH?yPIiopphLiJ*s#&mlXPQm)OVA?dId2CDei4tzV=vNDOWu)a`g%dWO(C+F+dUFYe
zfI$(!bBcNcW?Uy|lD0@wk|6foo62U8hb{cLB?*qG`_3NipFtO3rzi7|mIQ#dJc*X~
z^+;4tfEP@IsD89P;Dhxd)x!&7#mXM1%-}Mj|NUdo51O^yf3{^uqzL^DBfbyZXB#g1
zEtfHkJz>?4RY*~3g%wV^46gt2*1`$8mwEe7kn32kW7Hq!e(&=Z>`?en!C61f_z@O@
zZ32obat*x$hFgu`GG>>c7^ASx-<wQg5=gj15yK8UU_yiK&=#uwW(`v}KiXT@)I1ye
zv8@*ub0*KP_zUNP0x2neWxyLeuM2<?0_dc_g%q$WU^Pyx_-%%{WH7l&atylw6g3#c
zg^>c<%+G2|fQ{()C(ByKSMV6!jrBSd4dRWhAvWD$gz!9v3V?X;I`wejP_HJcU7~;*
z>+IG)5~<F^RkFhB6hJ;0*`C&b3)bymjGbQ84~{=s3iiBLPp}h{p#Q8iDZ%VyqcW=?
zX^R6nvxUl$%ENdwi+oDMaQ~3xkAd#oii0(sExqX0E(1(GnYjA0rS;46d;G5svG;R{
z(k1|mAU`%8Ch-hmL%phjNs>5=?NMpi{N;?SogUkTim*(=&6)dai45_EP1#09%hmph
zyqJ$v(UM*2TIF822<_8&tb5?#>CU(LE(?A^CnMtYO8xkU0cAmWEk9k`KP?>3I&7sW
zG>#Omx;`J!&SYf_aX5dODGvVp-~|`%T2TRapVTu{uYLS<ozW<={Cr5GF89Q=l=shg
zZ?3auw8=d^{s|AL!I!{vw1Jf$d5BwhAgX&wV0t!WM?uTDyGI@nqJGqZr9=&hP?25x
z%QAje<`7>)^M{GK3-d1RDcuIzvXY+RK5~VjpLKWG@&a?hyjhrL49c?Km;292wwIw7
zCVxjBDr7pE5q^q~KXmjgiV0$vngcmMUJU(py)!kX2Q(QR7~s+SLvFV3T|t9bdalTO
z(AT$w1Cg&@EuKB`{wJ1KiagzH{U5>aL`6Jrh7VPHBo>x8lEpe7GHzg%lEqqkoNoU=
zlHaLe=S}mWa~^R}`_0Fs(&r>?e8J=Sr6*VaMf)6ZrzYUQOTtA(L9JZQ&3SzB%)#q`
znh+v~iJTaq`0WzihuDn|%1oJ({Q~5GY?A4hCUpqfqRye6e5%-e|7-XJ5jZ~s9b5lt
z;0Ux=UHnH$Z>NLcx}7_raa75~4GWC<R9adbI%PaQaLydHf{shF!N6Z1^zWT^xu)X5
z>J8?PIElOi>#bc#RfEn`xT!f-?~ZzHc0Rex<bSX;CvP9$#$#l}x~(knDpltXN773H
z#5bM7D#>-TPYfga`^R0%H7qY9Dl39(cC<2lK2pNf@(CYi8v|rWvW-Jc8ceUlNQ;xt
zu?g2~<75~0X(F(8vAWNNai`)3X@><YgvQ@JKq+sO7~EFo4Bz$6)I8oF&XdA}pReID
zdQM0Xdys+A(Hvw-Pq>6MjBf_UwW+|albn_EF+I2#m2JGm3a@`qqIqd7D$&*3b3lki
z60i&u)=7F~`93Tmtc&`8LH>;dRMOp?tqC~YUAAu^jq$(2@el<?!FI`fn<Y3#vUu>?
z8sqs)#_oS$!kxqiEB7Mw)&)8MUse-vN=qAaNER1I#6`NmSQ~!8d_;x~%I*Cpf>Gah
z@xq_A-z5d4^JbVs(&CZ>8so|*O*iSiVr5t=KY0fhg>@1ee)qF9VwdT!*M_Lq>WrFQ
zHXafR={MIzT&-(fT04X7Kgx$BojX$rih-}NK@g3_-HnP(q{tcH%4(@qqyQj@AsAdP
z!mjQJi(Gt@7Rp~PjfVhuSVnDc6Bd!H-RnUC=P+@7!qA}uFWG&2++Z*xKzzsAml&e;
zB)S=Enb1o#oxMXm@tcUCt^tlYnevzo{A+vhGb@=-ujVyh(L)IUB-{qV)4W3f!@e6k
z8#dr$3G2-$2Y?KA9O2)E3!}w_H@=R$aK-_&@m9`{U!Mq9x2PJ(!)hiNT(C;>$8)H*
zlVuq0VzM6=_(!11$xq}XeBPiXkfXtuqdves#vpB%-36XU7y}cJ`>_H(7%i5?Ys;b9
zAP7HT1-v{K1V)akfYsh*BqwPwsZ0|+ff0GWtQ!$w%6@S$_A+KGVQA8Mc6&z<X7$LX
zhWIa7Y(2J<HLtDOfL}^OZ)y_$)jTDjWO=`9U!#w15_WjgY0Um!zc|tP)rIJLV{h^N
zSvtm2?7LD70W@YSx3igX<q#dkJkLxz$xHA#i3Cf&66<1pRUo(uZFz5-GN624+l6Du
ztbk1`oFoIHkaXEH;SCP@lQHlb6|?Q3eWKIu0MEX}4D4(g9SeeB<-GidfAvW1zO5c-
z(L(T`8PoF0so&H*IBc~h)%0nzeNHz!bhg3Ab43m#;ASZk-6k%k6gu~iC9<<(!r-K|
zK-nXGAq&bI#&kuiOnBya_LfkvgtK0w5K?R^z$PEmYeG%Gk$-5AVt)M+{$Kj!TXS`|
z=Q5>x<!P7zpizh2pY;Fbv#{tqi7xKF2_XB8&(@F?_Tdm~0H@ZX)U9z;u_2_`_|#4B
zHLrg@z5TSb?(0@o*q$mu9j~0vE6`8Loa7xCDe!*oR*fsw)#Lj^)*J~ykn_OnT{-4|
z>un&$mk;Z3?8YgL=Q3T7z8F{5*KvrIz3wt?Q@6P0P^C%poiwx<ux9Tm5HARI&3^Do
z>0VyMGb<cPck|hRYNJk(ja>5)Gb!=$M8QM9U&DKjQptuJ6Xsb05@RO3{Qxe<3`?JM
zw8tggApoqvATN~btzk$}9S&^7;{9+#`JG(l9QHnWY$Sa<|H&Sc$@ts!xm>G;%ekZ^
z4uXGydCEL|KKa+dEn|O@rNvdA@1Z+=i9FHU8uE#X*JW_l1&A+nq<)RN&o5j#Sd)dT
z?}A%ZgkwDeYdnwad{So+N5a*|j;G8vaZzUx8!_XR+E3W{{b5&Ur?=Jz8P4cJ0+<^X
zKE=v^l37fjX)PnLC>aQqfv-4r96`0QsmQic81M#|c>O6^jvNQ2f$jEC*x1H!Gmr_l
z{AyQ5AI+Mi?WAew{SHF$)!$xD-O=|3$B+u{$5i|_axaf1I3L~B<d=``*-r%Rh_1r|
zxqy2?y3hcr2m9@UO4N7xf^*Fhe?Eis*mNW)h3UGSj6S5U0NnTSK62#BLlvs*$C&&B
z`+<HPl1L4`xYa${=u3@&-Pj~76IO?Xh=5ox8Fv5bTf?FR01Q2rpdX`@5L*W6g(UQY
zJUPZ(daKLW=ezS=xHhN^*Gx`lrS;-^->8!ldzgSF@LgJ6dpdW9u?~Q8E9PbfDQp=f
zI~gh-4G^EF|CmMWSU>-ZSI`ajZ#-;ld|=!QvtRO1*Q<K+qwNZJUjp`Fcij-KO{r-v
zW7R3UDSgf$?EI8lG87aAJpXq_e^|q6ql+dekm!7(@e}G3YXesh#{a6E6SA?5O{{!j
zf^qO<e+W$Q3Gj?|q$hrd%*NS}YOCgc!t2MB9)_1_mT;EmJ4_KnmM6D1>EG8dTo8G_
z8bWk>N@3z@aU(W9M8Zo?=_XO3z`CEAh>n7r^yh(yJ!qoWCBtvKkB)E0=OU}(|LQjf
z2e8Z?HgrEY8E79DiP(2@MW}}SB1o9tG@j3xa4)6kAtQi3_?)FaVfJO2P<$_!YUAPy
z)y9>u`V8_$oup|*#?fum7d-aD(jl(buiyMKk1=k%=n?s^Whhsz99gCtsMF8lKCrUA
z<H_=6s0h2cK}ULGM<xB-!^`tp1-k#jpcE8w2|+B_e2@1CYqDR^A7M;6W9xC!DiTwb
zL5ZF5rDtVX{&UYOkX1@ET77|&frtEg9+PAELI2+71dXNDypIO>;sYS_1`8*y84o^1
zXbl<C6XbH=KdK-HfsB#uhnd>vNL1I!D>heM&P$p9(`ds-bYp`9+54~?7X7&X{H+1i
zbIE@;%nTdmY{3lmoA_P2y|D?zFDvka>tCoz8xUBUX!qnUay?G0CnX_3Qy{SH&b}1~
zY;q4fCWo$)JMn9;XD_(sZu5)%RkbDO%XhuA<&bBpoh8=kY^d?BFxTPcKMMKCuJ^u^
zUh$VR?B%6nP4pg>@TQ`J99gL6eqaBV|HM2=k+{QJP%w`8C7U$>fo)PuBMFX931zz7
zc*|;;i8t6(@S~sI<vh`y)b<&KZ;k^LJLk&Fo~G6=s(4~uCq(zr``ff8ZoJN9?QK*Q
z4g`%qc`mxEMI`{2l#bIO`=z{3BhOrUvVKej?|pdZsh<6!x>G{8AVd~ao~%(s_<VcB
z&+?6w8Vh>*r4!s~CEwR(TtZb*m8TMLN+Bsu0cXJyma}X7uX*jbbA!Oel2o-k*&ul=
zswhuDwPDIfwUN8-=Euj?mm3VmI72IMIB?;i6rm<rOtMyS1U*i8Bq)$Ju?J7|DMNF{
zd)^slEQ3)4#3n5%?7Lay^I<vke1q$u#msQi|0UU{>AaJm{TEp?AFkZ~b_E9_zI#Cn
zNmUPBD*CJ6pjDle@`?7)k;6_Y1*1#ys{ICU0haNlRH|>ySeCto>HS|4tpvci@u=TO
zS64XDx0&eZTbjDnO(l?FSz!2h1mvtd{2(6h)`@OTX`oRycckCvsljm5f2Us*J)E}@
z49{o7Qn78tgv1_Lw>R&@)9GpzQ%wp0Y6H>im_d&miWk5VS`F5!iYwSKuh2Oh_>M^!
z0K#+;TK}?hIcD5JYn8MKFR}f!_NzW^IN9J%JHE}~PR`mQlX}ff)VOuz8Xo4sphzd}
zd^!k)1rma>l|irM`MY4S7nRU|CZ8maL48R6hhedS*8qfQrfU|a)KuK@@i%h7$uk@C
zhmIfBU&q}O6CbzyXDnQr!Nn}gkP&<2i@2#-)8B=ULNPVpT+*VkJO99t@ih)S`9h0r
z!uMSK&jR~u&&cNt`#DPA{GqnxJq_CWCT!+F*!D<;uw^~zvyYjB_Q(`ET$XttIIL^B
zLq3uNooj~f-dkWL_V_&kmJ1my9o^2-MgYqAlZ9Qqm1TnJH9tL%EC-fe<%R^wh&O7-
zfKc2=!?~{SD7{YAo>agrJF4>P38FhK9Il=`sLJ-U(P{;)GJAok4~d5+j~d?;hh<4q
zFc%~-GRb&icVmU<W`S9g%j~c{+VEd=Zf^jN7~zW0eM4>_AgnOBmb9~M8uV(-GWL(1
zT^}|}=MkgC%~j(R+EY5P@L^Y%Ofa}G^mL{xXT`2#;0z-Id2z?Ob69h;gEN~>+{_ZK
zJyz%DOv-yl5$g7I1`Y1eSHsG6n)f>~>$*tSPngcZuM;9khJ*$3t?DX?O?Ia@<4RYl
z9f~2Q%6b36oY7CAUF-Ts@NamZI!RVXkH*K~oz#!$*|s+~s1--0UJ!=Fa2-l5`b@X)
z$kT=7T=)d%yqh#n*~45rCJg5D(C>#>*iOxl6*l?veGo&ho7lXF?w)K))vpTUCb=uj
zXt>MTS805pLxcETZfv~!=Q}~z|Ea92K81$yg^y><t-_L}z5hOz&q)MwR;Mb<w!R=G
zE#tRzw2$`kP~+g}-cRGKgN2LA|45+WC$Qu%R$@J5#Zqo=>PjT2CFeNj5dAp6Xb>vW
zh!vstRQqt&uyu+8;4=Jk>HISfg9!~)3~17f-5?V;Jm$MHkc)u=>~ln(j;~!%e_Zt0
z3es*gaNnq^HnYm2sK4n9lqeVr+?v6^(jELWC8xLElY?P#+^ut+d}KnCm+=$k$;C{>
zI<!YJ;KdM(9qWVvN4ZE;twl{}v>E;w@u*_d=Brm>pgTfS?RQ+U84W)IsIOWCTH0m3
zTtdno<h$pkhZ{bTejS%~9rT#bArJ7EDJDh85_k4HV>C7zNKPm4m9wOmQD@YBx299N
z@d({M|Gt#MihRs(1c@H};MbgNBEW8DHrM%OwdT)Pr1B4L$tr$F9_2QX)c(XF$!X_8
zr;1&|&lTCK04N3k%&R&q>_n$O1KUI-{~jH>G|j=fS3;!!7nx4^^sBpz`R;W6tK|8*
zjv2!92iH~&0E7$9PaPKUC48b%8~O5=FWdjdr>EyXhzVE3wi4LSLc2up)mb2yx5LuR
z%Rk%QJ#BOYB@;G^Y%E|v_JKQbMeL!&J#4GNeD^^>r><zj!qcqv-e(JIlkG3VL%wW6
zaj|6CS^_?#XC$5|En7)prQ8Ep7Okz#G)YE0b?<<d1+Yu~Y`>>6nsF*#&Jagth)c%$
z^y~w4qClhsUgYZTkLgGUu)oFq2nm_i6I~8rQjDa2Ti~(S13U;GIbCAXJW=Vm6q2jb
z$g1nU%fac_-oZ4!&t#(7eXX(Fvbp)$p7HPH-)DOr8wI}dQe=6~Kj3!<b<QZ}*a~Rw
zO0;DDg=HkkF)P!Ca<IHv9ILcO(krOU5zE~eI_qiX<k~3x;7WuT?4Fq>qQ03l#%bv6
zOIv;&x}$zU`n3aH{HYp|!&(T2fUCw0-df7;UX6Swr)B>reK6Q>lHN2W-o`NZET7ot
zn>X7XOLH#z0S~NNayVJ#i&ub=evG$3Us#kFY_iWQ&-K^rnM*{)lKiYotn$%aS9RUR
z{bhjLvM-uxi!~))BdpCSbj!lw+5mhnvBI7N!=2n_u?~w6=u-%sKZSLo-Bm}O$ROwJ
zG=-MFf0%U(PUFyYM@P7MEyg+0zr7BE)J73G=CWzLqep*FTzpYmsdH3NnH|G!;(zSk
zWDxg&{Tj;kLyQoY6l1MxZ_d)?qOmxF6TID(dK--UDrR(@;3nTPj4t!B#;T=Eco2Hm
zrwx^+5JqEk-r+MQe>Yy8Z-bVt{9R4ytr6?`<rIf&3T;`xAb)XhQ7h(3&oB@}_#JJ%
ze9=^(n{B{OI?n9@BOzm}AvjmAfNG09t5v#6$zO=teu#~7TdQF=?~+OFf!_&4jP8j}
zp(W5?;_9acJI|8pDy>j2G0$YrWwn9wk~%RF7jh1hwWGRu<ke4mzw>lT&`hBj5Oh34
z+o6-01lfqgTmQz6`f2&T?U<OeMybF+OxFBFt+S=vXi)ma{d0xq>Y-RL4rGbZ%a2WR
z2_MX`qDKQQPLF!V&smzMRu8fZhGQpwmul6d<OF?A8uhZamf7+T7`4d$K;*qCT!t{7
zQHh>^^oUCs7gC+8XXdul+}-H59D`J{hI~XLJvIkTPe0cf5o+4VbNN@7IIrG$-L2kI
zOYHe~w~G`{I#@Da7n5BY4?Qe_yS^WGB||BXl(oN5#Q(pq?;Jwj)Ci4a7Z*%LWhEKq
z9wqAz>MeOWi+wLR+T6fkbDZ3kGu>ISw%5DBri(dzyMfVh6V=_?`e<%S=*ge2+|XY|
z@Ch>_)RF--gpH3){E8puvEc$6yZn`%93^%TmoCuSMb(nLRH)gAD2DCzKX`*SCMQ^~
zigDg8x4M$1H}bqC%XAm=?*4jk-aBagCnazBBkZ}3{htV>(`s}7#9S#7(QPk}>ewJ_
z>XGv^LfSu>xy-)O6-+Aud!)L%)0P4Lt4kQtd3XuQ@z!lL$pAZ=7dZ*Lus({$)Bb|1
zaNh1CkROpfg$bE^-u^Hr)bmPKhGeAej<0zrTA&XL$PJ&Je4?LZA>pNK_Luju_v(VA
zO$+NR*?_DcBfIYEm-4>L7!rRbRo7*Buh9I@KReTET;HOW(iOCN-wQlHVtBVy!@;KL
zB?O`|Hty!w&Q8fopS{l2vR*#)4SG;z8Ft>s6Eu6AcEDWy^?jl6W#Wbdvx(cjU&CV7
zH?uA*?BLBd4Cj`shu=+ose_nl5u-Y35QGP270~4_pg1+6<Itci=i2_O7)%aJ!u#Iw
zQXxo!iBVS?B#_2_vvd)bbU174!~2JGaEmolydWm*oD@ic`e7E=g-49qmM<3KEWNq@
z6ikhv_QlQ%@-NyR4uWqbR<Tt&PV+;DgaF55d#)`DScXf7>)GDpeHWOd%spb*|JlwI
z7hX<7HZu-|&Rg*6DY0xsVyy&=@XCsNA3oZACfosUpx7WKAEnQ{R5McvU9JX-Kf_<X
zd^`^y=o7g(><e0ZWCAYrV#`YaipS4vB!3ulgKHn7TS~B$8B)!rr#oxniwmulHTgd8
zqGJC{rW_%vF_vb5cmD#R>W}WUpOD<FCAbC3pS*|OvYCL<#5~vsfSPs`jlWP1+ims=
zF4db3<i~19Q6}Jga-jIKo_+pz%`eqsk>^jBs=Kx)fS-S4<@k?8!O`6pd5}Tt8P@rn
z6du%I-@FW%nxP)f@`L0!MFqM%h|5TM>-vr4I27xgF1humH48&)o)^-B<B3(vXklMS
zRj-|Ql52Fk)*obl7!FyewQHzNOyOU4T2*6Rd(fP4ukN3|{&hUs+a{mu<z1KxHG<2G
zIxoQQdzas{XGe)uk^6^<w&Z~!4UAy6uz3*^OPPUYjftEYN3={(V+Z-|<elsa?AC2{
z-s=ANNTlsRkT(eJGBiu<hSjk3=b!&msXO-XN71^h*sKPxQDe5Phc>^8H`4axo*D~E
zb?;{*J;Tn`7qI`M%f;wM`7(Hwss-owuA2;8E||<zr-H{M$$0|pTI*NPGeGmk=i|8;
zqD@dwHU(yylry=!$?axSRwg*uCaiz>mT6C5Ec(*NolS=XW%Wqjv?nSE=&XA`8&ZwP
zH-aa%25|y<Ul%$ZB+#<#hcs<%2ry^;ankC!+(W};`>OEH2R5HAsOnuy*uoGFqRzcC
z=zQVfJ&d5u4A=e>`jWM`s!cmtT~0Ac@V*&Hs!;Uu=kQ}hE&q8wNT6o+c>JcjZf6aH
z`wAjEVm|A8<;aU$tQBq(DbwiIRB0;ymUiV^Fu^R$&fBXliBW>wCR)!r>+PDH770pi
z=^~m`A58nH@q_~B(L?M7oniZ!g`4fn<}m<~!uRbwN4quMn#{J=)%g|3W!Ki&jQG^;
z@7K(kG+IIyo|--J_-_e&1X17yHm-h1&HApL6N^)vk}tNXlQW&z-9_vM%fO3@-<gpy
z`yC_poz^=1B5=&r5=Os2-3lMNl=MG(Wg@5THsN&iRfON9=JyIv_0pGTCb$s5uQy{J
zv;qlYPcPJlwQj-%2#;jHnB~Qx8yueYiTyQ#EZE%NWJjNKEtx&5XSR3vX4cexV3PeO
z{Vp=`eH1tB%lOOY=3q3p!TN+p39>PhraAujq3sCnV}B9F%mWNL;#RQd8W_17a;QgY
zo15rkiVmJ961VdX>B62gL5x!Ez^-r${=quEYqZNBbzOQM^H=zDM@j%b+ajGlM1D`<
zTLPmm$@nynnY?QQHLL)CGrb%ZWp-?6U@0x5&zg(1{N(E9<X_Te1)X2iH0Kj&{vOr&
z3_(i=#3VsgBN)R=wfcI`H*Kw3ncPEN$H_4w1kgL>V(F!0vBwh%VD~)+lSA!ZfUmfS
zFn@_Q?S>(EHrl`E7E+Ci>D2=dT0w7ZOe9uj2BZCzp&`6xaVfw^@XO|hlW(9MB*Z8r
z=r#EMfk2;nH$(n~k_p&6VC;BML1;{T&-_#SM`9Fc<!)Zy8Ox{E{xzk2W)_F3*52xk
znaSo((<ps`Jf{Bq@{<lU<a+7Y>ib1d;>9u>F{J+9&6*4qg)39`a8DxRD9u@jXajFG
zWG3$qwDAD<a6uEq>O9{wj?`-Iu891d&Rg9jmnIhJ<;8917ojd6cI6s?Gl}15*crQ2
z_cXtLWc_P8(AS*|RYY>U0DI}>GLzwF@Cw>N^D3$j1D!PCfu)3#9D^}iA-~JvC3bua
z&kxJ7wV1j5SXQ2#E0iK=SDeplEyz(f(r3{JIbI`SbE1yN$Nz;hpLT^2n(}c=t``&6
zh5@Ixfq#p)A<>K|LNk5*E8}k<&%5<aRKBI3<ed7B%<T4(pQcnZfv+_1+=I*O<Lgzb
z>%*GMdbmz0Tpqg#m}MQgOC5c=-ncgRC2wB?mzaLQo9t=?-liv{Tu`0gArsm%$37v}
zp6(Q|`tiIvQV*j+T(v$#G9Y5PV6Rudn8nE~ScLna&n~*Gb*m!8es<|hhf1cNP$I`0
z-fl8fX1d08m+N)}k*-Zpnr6B%Kl5A0%n)GKDa&1e^IK}$9vW%o-wFaRhI16x<64$9
zy%wSJ(|)X$cKMOWD0ylUdG<Bx?LtnSno(W1vvxMvi}<`ZZf_A5(!yow7k7TF7LfZE
z044qn4l@>t7yG^9wPKamBgx1^BD`=mhK7IrY&3hgq9*OopiCbuzY_n0{r-hk;{pC5
zF}kILcG5G{ZRuo4q|NO$x=+r?ky4f+<f_37#k$^6N_qB~TWI;HRcmVP47KWxyQxcC
zQB!SoyLmPo?_~(i|1|@#jBL0NJ{o`%VGkG>j@dp-2u@0zBwlrWHhrCg?CV|aCB9ld
z(y}`s62vP!zMYC^Se#E~iVZiL&x;?<xSV51XdRcRmZxuH7Lt4cO>)43Iy7p~@x?2)
zrIf50V#ccy78A)($!F@JOwP+*<LK1O#?7bL`yNYp+%zA0lr}djTTeL3-wyLFT~3Ur
z0txDdPFhji0b`SD1oOkvlVXA1VbI-sB1+Rs{(<eE`n8X-yLpYJLQtYP=Ih(J558K(
z()-X0>ZX-jsKTL?*63Y74EKJaZXYGkUMTMMohrG5=6x{JB(KiW@L;52*vWp*gFa1w
z3(J6z%-Vv-@*lCI6NGcz)YO!tL5cS1QatA&!AfeR0*uP6q=+pESLT@f^zbf}B-^=N
zUmuL*3F><6UTd;z_lbApw4cZ3*Ol7vQ-gP}M=dJ;iDZcWfBg+FLrxI8k#p<UywUkq
zz8`)N$swMBK+zqmj2IMr`*;I5B><)kI{KK)K|YfWLPowaw6>jej*Jk+8q;vYM2)#+
z(A=~e^>F#3d<(1JdTJ{H6MQlZnTmzqp6~pbWLi`6ia3MgERtqLr1ZYoY|4A*?A&yf
zK0n6dA@4t9i{T>386>Sf9*7gsjP_t5tqP0XSG^CgTU7!PCqk!rYs5Dgg+9uW%n=vD
zCh6;mNGK|}<DuwW7v@G_WEKgyCU9|fpJHyi3i5LEsY%|F$FzqybL8!=a!>Rn12+AG
z@J67o<1!tOcvS-Y5Or(f=Pka35VzLrq=E1VpWOlG(<3$lmy<($?|b-t_i?$%gN>H9
zdT_4Lmj3A-hku7P7G^a*(fyF=i}ug|OL*`(p8eqKtUY4dJsv?C{hpm*ILA6yt5I?I
z2KDpcJT*2`B_Nj$arUWz@5dys?pLR*+~?!T*<U*}Re7>pqrP7nRc#I`X(#iv0#t|T
zO;m2bUWq~k2)rEmHTDT?qm_8n&^_UXSuW9LDSRu|_vcjURIn2bV4qZ&QQ9)#Vvk4=
z&{=PgN`9;KNW4HhboU|wh13HFec>oH54tAB2=O5|-{E1Nalwo)=iu^;&fqZKB;0I)
zzUsTXv~%`+pg3pHQr-Y=HeS%KSzH)vm1Qp<?22469O_<W)k2B4utK)GyjAfJ9KJ2R
zx8dxFCP_0=O6-RAPr@U6hG(-WA`|e(vK6RJ1miEp{k8f&;|h{kK~JjTr|agcf$j6r
zl4;!`>0Knp#4;C|>IZK{{9ICK5O`*I*>OKUw(enc7X;n!OK`VU@j6K<c_-inx3WT3
zZT4oFDKv$K?FWM)HMEmzRn!VUJLi(~N0dAj>I^?zrvh<!a1bLs)HH9O>Z!eUgx<C}
z%s-oh`(CiEWErbKL$=FHg7!ZUEHZGq+y#9mId-OS1`~1ukIVyu+{qyI$0TdYH|1>3
zV3~HvdfW6J`W@^~)%|9zJbJv(tE*=pWCq~yrCfPkxK0#_7keo8JI1fal`BBp=49wT
z$SN`<`+~)srY;8pXdbi_^?IH;COMYeE!VGN1-;l{rEw+pVGEVugQY~isZyc!4MEg!
z_z#%j@oqXucJq7@zO62rgR6CHeSLPyR;TTMxDtG?jByLh^)UJ%xf1nUFe>|oB>Br0
zw}V<CNR>l?^ZRS7aSqczutpD=RrFuCoF$QGg$`fzfZ|b-&YqD00)5A_C0YvvD{0uR
z!Y`m+33~CI%3w7ZY}Hdp!k9qf?>Z6V@L4xUn#^a#R@+e&wb<NH2`V^1tFAh6{<-w}
zi(Zb`@R@|aQ)s{^%Wu;ownz!uv~id`zGfBJ)nAEu2gU$nMrTJ;@&Ls*-r2tlQ##*d
zl*-yK+O{EL&S4McsxLa5hePjXN>r|=W0e9fSz(uQIA|eOmoV~>P|*q3e<Ypv^VH|n
zDOG>BM+YL(($*W0b=9slHo{W2va$lO>Orynw)Nyp$^pObwBcJpw|bd83%ByCTzH_>
z_oWTy>-B&%SmTA6ZtCY<dF!s25iK2^gRLgn)!PsO>mH)+%R_(nX&TINnlpst53vfF
z&SdVC<2(#aK6<mYz9nWsb)!-AIvTl>Sd--Fchz=g(Ns5H1;rAN_2FoC2NA)z>!Dbo
zFAqZeb!LbXHi|}k7FQV%tX=MT$c5~p4dRNg%$MTL+f9OtO?ft9Oj3{d@k9r1Z$kMB
z{{?c@n8KckW3mmoj&tGnzMEH#l!x;YXh>66)gLH2N^Hv3E1p;oD40ZfEO#Nt_gj?y
zLa9WN`MT-8cFgr^--&DQh<Ph)uAnNV$J01pvZFt_ZwBvX#b@<3Q^kgDoH^I1Z?FP?
zf7HQ+TrI9>`tdRfmqOvaeq;#E$UjQ#nO!MC&vhCFOkGqT^We5uw*Kf_X&e`KS^AsA
z{cL~&^lo!#T0CGRy;k4)^p39)D4x%+t*t#b^MrQeb%RYYJMNRm|6Rqe6p100<KBg{
z8nlYJgwhp+tonq<#h))9+S8Kp5*ovm@r>rZ1uaZT_TN;5EMutuYW~m)eFe=mR~3j+
z?E}K@7WVERB8%f6AzwVvo9KK*oLyKT#1ae5v2s+0xQ%*@TN<Hjk!3d}xU;N6p2DC=
z9LvtNmoh|i^DV){r;+il1YHD8r56IxGWzdynsG>Wx?evc4Q?kQo?#!9K$vjVEHk7t
zRfJh>x?&C(NT1t$9L+vMt3?St4kk7G9(LYPcksKernAWnh7b&VlJJiP!7ZN0)QkEg
zW#b}H5!`$zIuBY-P|C^z*PMyI3-Pk=bbD5p1wvQ^_DNz;*~zsUbSSS4pq)F0X)M5M
ztzO{&$pRq9Yqx|yzF^n9#(r2H*y*+cbEMRu4LobVW*Rr#sKD8`q1*hDB`8>oFGq3>
z36i{jXCf6s71=hZd}R)jDEYE=Y}rYbNvZ%3>{f`TSZq=GmoGuu+Qqy(>G-oQp3En&
zHibBIT+ayWx%frt?6)w&&xC(McD9BTrZc(ef+1wkN5^?VbMo)c>v|jOaMEZah#`MA
zEWEDAig0CO#9+eip`v^VnHVYrQW1&`fAg@SLC~$;3Q-|KX*Y6y!Np*LMdiz#FJyt-
z)QJ38lH<}4fj(W$DkyUmTzv5Z6!I=+6&lZEA@dJ$Y>#XE1gYC-W4zl@*kYlD1`&J*
z#UrbZXW(ZaZ|y6>?k288<rabjb#6|m5dIeJr;S~R+9Hm0KLQu;$HXWv?1C!Qb@5;+
z49^IxBD6Ai(Or39*A*ySBVXR7$R|VaZZcPThJ3JKpNq>-8lHrA<~)6F)cxrsVCYvO
za^|<5VhFL3>w-zfRo#@^qX+C$`^roQpE4s!|Dva(FRAmCOV&RQQ^cBpPZ`?NSc-8O
zY2speV8=|?qn>SdAkCopX9Kq(IM0&LO5ygkr2A8MU4+2Pd+^4b>pkY#RXcr(-}Y-g
zf98aEZG)pb#+Xj7O4X*dTv%{W+bz9ot3-{$MxKTiRnyY`k!;6}Ve_LwktaDX2e*1~
zEd*B>5XM$ys1R&LLns%0RV@qjt)fA=vnAgRsb8X^AytAD(YV<svtQ9TT&U`1wB6T7
zr62nI6hV0^Hp1M0C52=tZp)_sCLBO}0wpRWBsLfBi2eB&e_N&P|3r1I_|}idI@=b!
zVpOI^#D&mM3F5uQ!`S%UE@w_}NYN+v*P%0c*Oi&lFgMaq{8qZQWwO&gu!cvRS_(~D
zM>@pk@u}917i@x0OOtBZ*M7EU5Sz++(Vtt@P6cjZIJufEmYO9zbT1hO{QnsXc6&ZW
z$4Gqs;^-CZofRCO=q-1<JZQ{<jZ&j!iI##W<300v9$<7`Snwd`de(rhHusJ`q;t7~
z@;;<*Uzlhk&EF{a#7F(8A!O)%7|q>k-z&7Zgw@OjsEq1jbulqcT8choQncn*FTNJT
zkM6*;de@z;52eD#o1h=;?OYi9IQdIe|EgfBxxv_Pg8ACuuuG1ji;4S-@0A08dh;l9
zXlp&+8dZ<4NpuEutRf7RlI7(r2hQ$}rEfOHNCn6)7+2}Lx?jxB?j$x<!8%j@Uf+kE
ze;->7^2}KNU>D@OwsJsu1d}Gb-INjdf~?$LLYm>rP=D6LTm|JdndU=J5W^`KKy2RD
zoi7(GHI_FJe<AvrQqg1q`dX|$TVW<cl6C41{$TB9hrY1mjw<VNy-3lBB#SzbC36B)
zHQ<$nT9N@Fn<2faj`#Ob=%`uyq|-mO3Jb7R!w+szxYz2RZO=}D`w(EhSa*2mRrztn
z@rrKYo-B~I5Y_FsOcA{Axmz54_;#~O^~=}w%D?#&xi+-c5jJ?8{XahIKZJI_4381Q
zW**n@XN&i8m0W!JEo-0HB;d<W3ib~pKZncI8&V-ED=YnP)A4?HukbwWdJq@>-_@u7
zxDvNbVfN6o9)8Y#*iC4mt4k~2_^mJIQ>tYt_2t?HUv-VVc#{g7-XqS0$~)(;xfdLI
z=N2b*NU)qXhqcIB-!UCuNb;eV)XQl3vzPVXq;F13<iPpMs9QRP)dbK*@>;BPRdqxJ
z^4eIt29)nbM?iD3pMFo|tg=F|y{O`B^*SxRH~yuR+NJF)TbwND$6_*(hD}Q+_y?eB
z)D=vmb?tAw30_-2Bf#%{w96Hn$Ev6^YGjp9i!|pcyBJ&UTfKMuFUfN}0FL+s>rRjI
z{b;ro=Pd4mYPQ$Bd7f@NM(>|Y_EKM84QflhSorv1AJhFjKQ?<@uv>YTjR(69_Ua{2
zYSTSmiVI?9&s$krbRSX@h~Iwh!FGb~+ZBggycP;8fY9+>Pn$NEY)F<5wx+cZ@D!Hy
z-OcqTpg5})8w+Y#_=_KH)E<!9h`#&)*WJL=_`;P>`)RQZMt%*y`Eo8`qAm86@MoH^
zi0fw{VSQIbcdcbNZX5esdkV}Dy=X@RqD76F`sZsP`H#EHU9s~dbz_7`BMHi&?1f*%
zy~nY~H)0e6=}_1fq3#06Vj;&`IUy-~QenveT*q{$IH7Cx@Q9@Ax=jvP98IeiW4tbZ
zd;I29_)=p&Eo8LM&8mtmXtYoh6?^n?YQmSoIQ;jRc6F{DTDnrK5;uq=T|aB(n%1;i
zd;^z&BrIgO##frp-E7glRK#VMV;z;v($F>gJt5jIoC&K#MN(0|uDx9(-yRAIJ5Rg~
zrJf5w!rKI;R=9gp!<dsj^wRIs4sm|X;@luoS{<V(kk<Ymns>i%7Q-iKS~t!#!|^a(
zA+D7j?0=PS#%S7Zs7$CzJlO7hv<NWy4npzKlg|jaL><rkBr8(Oo57`mM*+T={H#T?
z`tKz(;aQ5aXW}N{uq(3ISb_MpwxyuPs@|=l$=hqywv;N10Gp?EcI{3#RlT<(7|t^t
zc#Xh{M041=&8S}_ZF4u8^4pKgr+n4A(HR~pF&F7|@H;NEepjElz?Tls)4z1JZakIo
zrQm~6mO0$|4exZEzpjAZrt3nrsrPgRWTxB2V8=^D@>qi^(<7-b`>S^sc9`E3%|<(3
ze+Cowf6X|hqkSh@HfPgriO~%yE`;B82cMQrwue@AP|y8=u<t6%$a)56JA)M(e;R*_
z9=fWAx||zp{lawV9SWXvM<(O#XJbcPuVLPHK_kzmT91<lGZ&NUYyWYd1n_J2o%e*&
zERrrhWUHxKxVN=T56Llvqu2HRdUmXj9_cuB_w_@zUlzaaV<kLFZa+PyK9}Mv|Kvo%
zwPM?DAZv{Hp-#}4KoC6Zv2UfulxKDCCvkU1w4^E{=Xd?G?W$)Ja`N}iT;QTbyrtLe
zC+C_--pCNLpl@+p<GbVl-0#W0gCtKvh{4@p*)P=Uai@KkESuPq+N!C~t4lKwbPVsm
zS<@D+*Vn$I|IER1opE<ZZ~VHzn!+_vlSv5N^iN#@7-1q{qi(=g=g=^V#_`Uhw&-&{
zK0gfpQlDi=FH;cbA1==A7c6?L_km~<UtYdd17gUp-cs!IXmFEZStW{ZO^{vvh>wmg
zKX6;`IC;()K=6#?^zA#)+sj-<;&_cWe>K-TUDZvAFnKcMnw;iPzZF$Zor`wMXj}Tg
zxJgo-z3b`C(F4p;V9k5hKSGfpVAr?H_U*bDm%Mq{m*cmy<X;e7n!vj#!l;z)$lt#m
z8tnf?q_*lPyeNnwntE*C8$;IUfSib;V7$N?1PyA_ffCk*g`&#%t#<1ulbT#}h*91)
z)8^*pjsC~h&TeiO*GUWSz@?gkCVzfx0%uda6r2_kGAwcuBi_@NC%#59@)_q~{zp&Q
zu%*&!GqFpQUIn*oETz0Vw=SmTECm1gAK8UqbbD2|9ByCa9g*IKy9gR0<MWRMhUlc<
zC-cEX=Y`R!9CUy{Q9fAH?8Y#V^ykgm@+qn%WAVV)eY54${yVQNDpsLe0ylU?&%{^Z
zkqHH-@8-cHi8=g%x4WWEK3HjmB&*EN=4exDq6aA*$K*Zy_(K-?_P5_S_Pt?SQ5wF}
zWuh)LeEv>#bda5*6b$y@5D{#t_LOvN#lN5XoS78&9{-TO14@@%ay}Sisex5RMw(c~
z=T_o<?kw=4%~J+PEjNx<nG~h^q9ylSpHx(=dXBxwCgFwZ%GWY=UJI><e|0|D6?ETX
zT#q|GJLhPQD)!OnOe6hF?zKoD5!Myxlo5jbat2dYGu;4y2%n0^_3R5vZ*f1^A0%%&
ztr57~hCP#Qv4a?81<qscU<~5$CkL8gFOkZ!H51@QRES3D?J&;od@vcRfW+e+cr|k%
z+8@DyU^45x<<|pGX8XQekP4@<$I8o5kIq9XpmGe9YW@lsJWeYJR6Qb_^{w^&X;Xva
z&yY#XCi$hr-gU|Rc%z@qWATa+ZiRVxD6)wd480xb{p1Xe)%X;mG3<p$1y%^Ljt-%A
z{u$_OvNKcCjS+*XVfufU21R0oXC^!KK1lr_D&m8o1KExf?cN{Bmi~8VligcCRm|dS
zV0UU<u&(>(REXddQAOw)))TaiW}`vGd13O6;kb+=q7oi>2l8no$1i%~yMi5#7f$8c
zInlkXh1Ey7ip9W;=U-ekSGWvm*28!RR-%27)!nV%&9c)^1@Tg(!WviH<ejk#FZN6%
z*I3e(Hv0Ga&|G&m`Rk7k+idZd<a$f}de}fg4wq(<<C;9p{k7BitQQ0MhSIa6BN8z0
zk?LKanEmbA>NM>&FJ`mqC=ujowH#?k)j!H<SrSM)sPi9Uq%SCIck?8Q<k&zitL4k(
zwW_t;Z1m;enneixatCLxuwH#=Zyj7M_qaI6u7EQ^0ZxHInlxSd?H8bIzE(|Uag#MN
zy5TZE<JEDlADx<PUF$2;EAfnv)uPRsyV9fGXb$@U%a(7&(#lIV3IRj{=K=Y)`oXoz
zW!upr=X*BK5QMPASFOeJwcq|P=H3G;s;x;I)+T2pi)6_lNy#}WL2{PVBvHvAIW{?n
zfMkh+fJka`h9*l^GPDGVB289u_znub_q%s|XXgKB*80}E3(sv1r}x>rcGXi=Pt}&#
zUr$v%6kts9SuQ0G^I@|1FtTYrkMP`~-srfcoNAuf;KA3xB@y#d`arw@c1CgsTH-f)
z{;clC;9hH?I-X2xd-UuH648XyZZ+)j@e0rN{RZf#;;u)Zy_<;V48MfA_LmD1K5$nU
zlGs3+xl@JZ;Cuam$n6GIGnKbR_9=4sg6tai<nS~+nd(NNS(RCnG}r3hFq2BD4z=JX
zC&=bl!QM*zD*u`}XF{PH+2Ct9weeGxk7lHBi(EUwL@sLPkz%}(s@>ekie9eCz0_>R
zHdYZ0NdwSSxWrS!x&7Pm6L4@?qs`h_oUj9x9fjZ!qms);yJ9`XseD(LLK7y?VCw^&
z26_a4G%vJm2rHHr?q)LH>CCR5tRQM+;MF>G$C+fZhg7?&=A{LjMt)aBgo<um&H$bM
zVQUY`c&W7o>kl`FZ4AZg5j}G@W8Fzd!LbsMu%VPR;Jo8XB~x1_z8ZU*T%TZpODDZO
z;r=Vnz&-JUrm*~RTMo71i*xb>{b~)p-WHi=x)i@;bn)d+$##(Vk;>gX_)RCKiF-9x
zi%BG16sQCOA3vg6#cCy0o_`<e-&;uJgU>!e>EO*t78j>6kgacE#(T*tt>?Xpczi7W
z;AM5T<qPgQ-G`yy=yDTZ+>P9H_o!@K@hwHIvOE1YY(ja`yEafsY43W}SJv;sxp>$?
zr>pcSsE*$)hA8+J=UwYhB*r)$h+`w(wwD1N-Op=aeI1mvH!0r3I?0oxm!x+sCqlq2
zJM&YNdnqg9@O9nIO_&RF8Adu%cw8HCVpsGSdyQ=)dieKh3jWFER`N&)-Z})iN6s5*
z+|Uanp%sr~=9%jPwO0Vu)Z?TWvw#ex=cGO-GQu~DP4Y||Tja?;B3h=x4ny0QCHNEh
zVbe$FQpp=Q3Jkrckn%@}1<S$UTPI)dKz*+jlt!K2z@)AtF>mXVfMYmyO-0RkOTD{b
z6KtN&#JY@Q)s%ej_m|$bUxUA_h){W*{%lC51IBECv^A9>YE|$Q(*I4W_NBiL#8d3K
zhdN|kR_W9v&3^8Cudji%xCcnvhN4}$Y>TdTHlOcR-0kElCc@i~3Gt9(j8Da#)_d++
z=2XIra8dS7^v~kR4$SKMNGUct&6CF|xNl5yEg+<cH$G)I??o3eIgO7wAQGynY!lOY
z2uESqs3Zasl8KFjlJdL8Q6&{<iLF_2v1%tzg?-{?%|b*{)(~{eB`RH~KJqeU*oE|U
zbrk5Tkp3{O&9m;^ZcAoz3e=fdrvdhQ%*~lv-R0ZwU^AhXFjAlTj~nxM{V<Wl9&UMg
z9B5ch#i)XJ1;5}m*fn7&PWT?KX9Y5_&swQM(*wHIs~eX-ydat2xWoE-Dh7|KyYa>T
zFj5k>E8cAUfFfGCv%BX$&N=<dG%lNp#;3Sr$z6}Kyj-J#hUj#QWfrB+%iQo98d^BA
z<An@E?{3k|BsTk=v&h9_BHF)LpQ*P8=@Kly$`G?!8_U{kiS}YV`+mv<FW8=WFWb2I
z>M>nRlW(o{^^BNaKJXpOhcs^XH$SiQM8|s02-GIu>_z_add}!X>5kIg{lYLWk|U!w
z1B1=^CSM>Bt<->L>*Xe!;kpo}acjL8FMXK>dTjLRoC=Oxq{w;gRw3=HaFJ4zq|>d_
zyw<6-SX|X>Os=f=5Abrkh&OdMC#<aICYNZLlg<`o-B?USJ(fD0i@i9M%I559iq7WL
z7S^$KcL&mj0?({&b3m3BCD<4ZMs%Y*fYa`Ru(ppK@alH5sE;7SA7=yVZCKaSyM$Mp
zKW(~28@Yc7-u5p%EN)$_Qc33-H4=4ArM%9elw`oA%Ggl7+T_h5B5=CBq-4+1G^zcC
zYTFi6v7iZtTTIh0rG!}U5BL{+z56)o$C2mlU@cGQxbTb)yya66(btlecKx<?0#2L9
zt+(RcOPdbiOa1qz?b4lIl!c+8>@{u*3=D`^ODTKrfw?De6V3Eu4qe5<oDI2>LwM$9
z(UemjXSlb;)NE!Mb0*6^rpl;2cCRNSBokfYtj**QbLpY4-<T)lnvEa4BgOV`?zEA5
zBP<bi=RgA>F)8Vidfq-RPi|WIsA5g2Z?_4AfG`b~sUt9oaRfSf2X=`wyoKj8$KR6b
z*p+uBY(meCLiY(9Ox(QP(#_>;8(i*de(vJ$5jRo)j&b+3_0&O(uA(TnLt!*K%WHr1
zc6q;jbi>fismfd48N6eolVP5yn@>cf@4qB}I{dhAD}mvKMzFx}^128GVb`rWjKXLW
zAP)97O4m>F!=hu6LeLH)4HlyBNAJL~DYk?mHtGBtxBY8OGJM~;ZyGjKs4;>9P!j6z
zY~WQWdcs5j^lSqaJ5(dTZS!FHi2PQ-_0sG&L^)RLZHzCmwQ_TY26oYQry@mkS`4@9
z(rl`z0&)xjQpyl*r+p4}LVfAfSniKrmnzetI@j)Pnd=HXg!J0lf#XmHc(!i$JJC|9
zf-}M)ll9`xFGo$HZN*UobT*&tPO#5>Uaq~KBjVM>)x1siJcH+14J8WS9Wt*keyrm;
z1J1CZr=rcadEPe^CPSvlCzyM6u|G7QkKA;*Q@Ht_W-!=rU0&>Xond6{5yGkHVXy}?
zM{Z>2RJQ%9AgOkJ?EM95RFf<!`!|K2++>w5D`NUlM|#06bvaM;;qX16`;3z9qm;;&
z%x^M<z8(}rUkM+KdW<>u3oLt6Sn54k&00tEE!L~MxwIdWZ-09VleB)a?)`UT_^ch#
zO@$@CsME8zy@q4k%oQCAxx9d>SfoK{uvMGTwL>&1qo*l8i)(PVK-Q<JC!V`jI8P{&
z)vRwdHU}@X9TU}zS|qo==D$;$@MUJ&&xPD&CZkKjr+ns-{du*YovKUXh7t4-zEhNc
z-HPXZtHg<#Zip+W2<nL4w_DngQaa+jNp52<fMw(5F^q0!&(lROoqab$Sp@tpbRjT(
z#pI@36cSvO8tqACz-~XZ{mrd?`AFxnDUr{eEmJn^K2aTancr}sD?LJ!rC|bD&ssTf
zHaAe&uE)CG+sw{S;^Lzm<mae`dHc?JKD)urcPP@nWRxiuHW;bq=IY6@<xR>p;K+=p
zli8@T-3s&S%2cApRg`44a>DVf)#lzYMr9Z$c-we09f!~4jyhEyd*l`c4*_8xBR_6+
z_M7jiFI@NN8-!D~adV_Zb#u>l^4e48YArqm8{U`AWV=h;zWex|C*kB#y%^#%RgFR}
zeBRsOeH-b^lo@j~F4}OfEcAy<*V1_-=AI=T-s$kPz&R2YHi_wX{3KzIJ;N5c#d?=G
zUFY$ejNw4BZUGUi4-sQBrzosm32%(Xt!KwWo{CQOv#4S&39W{7%!J#$C;cS8y4W6Q
z4i;kSqcDrqEf=0>GDtnEdYDK^QbV=(ZT-pJuf`5fqj3(ZF=EjsAAc~VAF?ZcP~UFe
zo50;>)0u?&oh_J&i70wv4jY4XUKK1}Q=<b3-ftS8ZKD_2vl4!XE+cHB|D`=7a;rk3
z1Py^VOz@n2%7X2v^?{Q07UTPeu+v_j2ORtAj2_nesWj!!ZZT3w+^7YeCuBzr{P5pj
znEv3xMbI<J@{Vv!jqM0MF3sCXx#B}!@C2WWi77qY;8r@d1FABa>AAVc2IoF|ACq_}
zb{?ka9^%`26=|%Y9xAFB5OYgAShSr72l`IBx?hhCIS_g7?L8)btlyPsN3PQ}No2w#
z!2V!1j!;>UC3b(GRb=fegTLdh=gU|n{cYQYqaWV|c#&dXb<L&0h|6}L3Y$2V;#9xE
ztGz*3F+N}WAq=%4<G?2?!pnmrOXOsLa#qzc#H4lop@^&H7tfc31z8?@A>FxVbgG!h
z9#D!eFW2GL3&D~1_s6jtn4ScQ+MXP!<8yP~JiulN1_>LJ&bt$<lZO2aGXjyuIBZA*
z(FD-iEBPowEBZ#|N&B7U75Bo92zCcn^O>XJ73&CNNaXcEwBmHTE@?=Al+(#34w++Q
zE2ZjZRhB7;qywN9KE9(!3%#Cx!=QvL`?eYmdm!<vGF83)>t*YzXKFDsVGtf%4z)00
z5T+^ay?Z9n<F^iQ73m|)P6V(tz!|7)8OBl&CC)So^ehq0o{e@tax`-nNvcOWyB-7I
zD7AEgTIz(_2Yq2d&Ju}{#_i-2c*F0LF_Fz@UR>|SOBF)060UYrFd>Q9QH3a@v<&Pa
zku4&#f9P%jQskk6_8)j<SUgj&o4T!oNm2c(T0Gw4@Hl|O1D>$s!EuU|1twmwMi@{z
z+TU)*&{LgjKQASQXXN#^<B->RpYze0&a=eVSlJ_c9TmJb*aiE97{zj+snDsk6Wwiw
z<<CByw`fV+m!<n2PdmFeq0g5)p5oH9niEd86Hw;lh+H%U=aM(0K2xJ@Ur0@A<z>%T
zedC``zVG>Y{f>+bB{>M0U|)6im`OulaoUANOw$}?NG?rrMt){}aU;LI`>c4|km@Mb
zao4avOdz$d4DB=&NwrlCo3(>*A1@C1tKyt9@0_cYzlIc1g1^d`PdEi@`8n9%#Qy=&
z>ftc%+2f;QaAOl?&lIJqk<%+&t6opn9lZcCQsT2z3Ar44=_+TZC;D3)gDH{ws0H>m
zERu}0Ds<D3GmGTOee;YcZ<Q*_lID<wN>#_vw(vy4VDeAr=98%ET}bh~@7FYdsM$3X
zmolHHIY(!QUpL(`t(C<de$J!ZDl9|t7F8~r==H?3?QlPbp35(Z5&LwHt7*<t0;5Pf
ztxr1;n+UHH^4mn(i#R0i^i0{5-4hKNMw0m^>Mh43Nck21l5ih5OSD3w>Va!trIKp`
z{A|8<jbQtLAMRPOAHHQQnVhY_FH+sER&p<2T%Kg&Apg6ock0T)i^ux2k-VF(o8o>h
zpI3k6>k6l4^0ZpD@$!++KCe|S?$)y1N6aCAwn;*7?bbqUGORFr4Bk}6&mMZMA1pOu
zT)d>}bt5$3+MQB<yD3C1;fkF{{w-nK2BmZ!3)NYGBUXH4Fj*>?3kxx(&E7n^kD?|Q
z6PI`ywGWlJjxGl=y!Ii07YpcyQ7niH{up#{Xx~wdlpvPxcRlF%4NT1B=n3}i@3_nM
zG%@uQaXYx40W@JI__)j@j&spjs#Mkl5Hx0_Jucf|E-v@4dqYmNKC@CaGiyI4BuM97
zy<Y(wW=ub#TRknGy}4>;9Q|LpvaX0f6b8tpPnd^Gmi=;7ApPwuAn`zE`ELQ}AUEfG
z5<suaS+)o2B#!c$@#99!vHs}TB%qCS$9=M;z1g)B?=2dpf_H{^2}1FyLAU3o`Q5J5
zSM~8blx&H{)VLCC=C{gg)(N@~l7E+ex3dd+Lo1+sBC0u)Q9LbRSUm2x6}se<V2Cbe
zb2?~JcC(N6ISv9Jg|GMP1bs@#IH!vUfi=O4`Q<N(ZEN0TwH~L!?3``;re($R3DT>0
z9H1Eu9AAQ2r|$jh;sjgGo@2y2f-6&{`-Kkt@#8+88F!7!sSJ`BiqNq`uN^-zAoOG#
zYoBSfN*o;-oebP2ENt>Y9$LC(xrL%&h>eBdLUB=4jNoShZ>hz#|4M}eKPL&ILBjg#
z(p>386Ul68X|AkLPM{=~5ukwQriyR1XlRg3Bx|rXlbz_+@82wj;GpoK;u<F5xu!Lr
zjMN&8332nSD7V)v!CtONeJ+P}dV~P{@Ve)X-!xgF`i;Cd%!nlABd*lxMegQdrZ~!8
z<w{?0{Xrm#OP?vT`%mWV^&4_GzCuFqSix*uZP|72p$WY#v~));+*};>>X;-G>fi^O
z`3m<hiw6{v{D>wP3<Nt!lrRx}ymKtjQtchtH&)=bpo0;$n3Ci~z4t?*k|-`oSPs>w
zE_b}{k@U=mjLi*}bdXP6N7>V5hSoA>5l!TLdQ2a5C7*uiiz<G$@%8HiT>)0$eQ@ji
z^W)Gw7?lz9%<^pMWBlcuslG-AQli8}6Wt8gF&T2kU>rnGmwuBAWF;^WL+_(ry}Az!
zZ1f8|RP=%v7Tn9n2@5(vqJC|+o7);BKvXNuO>YodQo8?nUqPFreZib)L+T_mvh%h7
zb6Ou9gV1sNPtqut6Z4Z;EUB0&j)ef!yE+(eswaN#*I;CsRDrp<7x?7aGajHNPNSgQ
zwaY&WO#^1A^fpz)O*w$MFnuQ&Ya0uSx;F*`3<MBdT*!+IMsfMV9$6k+!&N#|i!Y92
z1wQ97$3Y0>yayWRI6+}&t*2P)F$@DiHF|{Rhd{H01(}ix*{kvmnM^KZF3XDS@~$RD
zECd>fgTtImQ#UWPMD!LwoK=JvCiSRt{2JtvTxQ(QM4t2*_$>uIp@;sGiD7~~#0nM6
zzLU1}b-9spY;3YV*G9Khi0=SnWL(qqv*nXVc$WH}=_q^#p~fK6d8<lP7iw!DE#2Hg
zK0#@i?^Ht0-RI*xEvOwKDZ9faGk37GyXG>P2*i7VB-vA&$i6N~$f=kLWE!Y%vVa(d
z)~Qlmt|34K!6$_vZTUZ6Sztpr;c=|~c!8X=r?vY<>V%yY3Ve|PBEI;iU`QMihU!Do
zY}thXaIiXrHKsYT?Z4VpOI?ZRP~cR)-%=5w)%k80CxLC?^qmE#wub``X0Bl&&;2wz
zsDZ=AgFBw3LlLoRT<fOs+~2$cmUcGtw7_@Og(Tksxl8uh*)o6uUfu8@!FtBb4rl|m
zhmVDDA5Kv>>bRSyZN$x*$=kHK%ImU3%h4V(X_z5G%29v#g5_cI-CF!}bc4`~iLSev
z=!=&FUaOCzJqPwOqH&1$Cn=rT7lfn+&bwuXtBw^hK0R#%6d#uD?YXS2u9hw@FBcky
zhUAd;#U4*~V%75%n@#p;e%qd(Crq%-FaMz31TE|MuLa-|dP{wXCsqkwfrKfOcLbH|
zf%Cuf%3C=zUf~wfBN{j-E4l{GyUVL?KZM|r&P&JBj>DSxH6ey(uHN4$3M6l{MW0RW
z(r<mPOWW%>CT$5i&@O-W2+TBgQC?#;7zgqMyZRTC{G!=kFzx3F<#fRz_4woD6CEYg
z@;|9PbX+wU*fJ2&=vT+Y6BJy)uaMs4TV2X5cnkl7BBT+6qaD5^<+U~9nEt8<Bz#xE
zWrh^c8XfkuO-3B;$JZ2riJjHygYIQw<?&&c*S~=k7x`mtgPuZxdXR=Zn9&ow<1rW&
z7U}$E2uPn~qS&S4*F68+rY(vzIWQml(!n6N>hW2{LdYE~M2mj?D}$ekD{}%_An<NO
zT@U%Q_gJh7O&sl5tZ#V;uU;xNRl*3<S`Yld2#+CoE);0<g{Ox$F{#r{2UMDDXIu8A
zf>K0}HUjowyu0gez;8#L+w);fE6JmKGoJ5Y8}bKOA)`OOsFI6+Uu~_w;8ABBPP6S*
z2SH&}3A#vdnqKF;{)(t>knk~TNz%Wzf@&5r5MqbxA{}EjE3raXwclsMiv`Sbz2{RW
z`Sh&7iBJ0F*?-kHv`A%=U?Ms}Y|vAZiGd(DE338WA8rD=Q{jt5zg`0Z0W6AVCUWnA
z2VmuOYU`(j)`NKAMaL34-$0K@sHcU-I{8L&U*c3!Yl2H=>c$h=q$neZ-W*U!25yig
zjOS5tcfThvYG^cbIEOG-*=%0d$JCqJbDq1+6l2B^NH+c%?8I~rdom2wp<L(aFE7Ci
znifT~nM`P`^4Z-Oxo5D~oAO{BjtGm?rOStsm<6#`@7%Z?q7*)J0$>|J^*K2mtZdxS
zx4)zZ_^Dcj{1(Ome^2Swa(_fS2k6?I{^&4QLe_?O%=57v^Zmo)Oy?HN6sukAC73g`
z;93yTht4lwzvW8RLm`JcUlthE9>@pSQ>?W}Frf((|1Nd|E>Oj<@AXciS%7C_y~Tg*
zqxv?!NB%YX4klPktZ5DYKHG)j*@i>tFGLCW4<dCV8dR_P7R^Mg3`P|g&lvMB_4`{7
z|GJ<t1!;oEwOWGQY)nL7KfytSvZ;4px-l>sS{jr+n*4XBp)4pa-0#>_{$f~uebd!X
z<~yKtQ2hrha4)F7Vk+?M)r1LP$U!KXrX|y>h$aFWKnzX)vYEdq8NYwZNL(s=T}Jv+
zb2(&L4T9fm#2^RoHd5|R4LKw8ZLUMB1~v%TU&@5PlJHi_(RzQEj6pqz;nFnx{+r(y
z0WT#!|IAqev`>{pBYdy_0X2Ml8^zboA5?FMiJ(a+gD0Z6Olsw0UYV!gU;X<+@TVzI
z&IF3uR)>ZiuBH*vB1yxEQpFBMd$5EnqagGJ=xQm%b^>`@bftU#wpu^`7Znd;$^xyg
zuebS>B|A1VV_R|b?Af!jgand;!a@p>E&BYcNtPHu-cbjZMA&j5U<afJ#_M?&`|<an
z{(gbaa-9_#qNAfHt>pZ6^|8W30KlQIHe5a|j_J~)9qhzsAHC2q3*vtT8ds8VC4z61
z{<7gkMMcLxf3^lVIXmQn6)9N?oIjU~3mE|RqR6a*vvhz)HFm61zJzIimH%JYJs}cF
zD13$<U5zK;BoKU4N`zV3es)x1%gtWA*p0)mS}Cza*f`WiHF&Nmmex(AhFV7?u^67)
zJD$eb&UbNdH@*nf=P0&x#c`YArL_T4(_3-0HJj$>RCpU*#Dh6&RjWJBD_EsCJ@}tp
zd61*f_7}oehMeKg59@Y=CQvQVH-vBccLf|fbU3Zn+0=PWA4*vV@F4G}VYe)_oYu_O
zKCg`j!BwQ2kMYUW^cjxV#AGV2z}??xng%2tB>`v=e4-OfYut1zpu$yE=L%w|dVzq_
zux4BZZD7R`tw-g!uEgbUQ+)jfGa0aAYwzcz8nmszq}X*lSJh;+4*@huYjBtx&Vk~>
zC?)i3x_>LHzg#S-z%;<5?l^$;Nc0ufz_?TtFhGI8V?OTT17o4~mqGjIf?R$A{&0ht
z0ma42$_im+hgXZK4Y_<mX8VQHW=gECFc18zP)V!z$7f^;AcyC`Hoko;>xdxL_w=mq
z2Gt8Vml#|<iR?m(?WW)+>uJ*+(K_xV(Evn62`no7^AlwNy<?a~e7p4WmduA3i#f0v
zY!sIiq6v#BdV-Gku-v9x^$oY4x&$qOKhE%nhnc#Vh>Z<f{fZ|s01-ha#`IESN$A}G
zPyt}2%bp=U#k!4!sCe>R>W|_ZjQy{@CE3ePI*-y()OR&Db97*AfL-YxF)s9f@h%93
z?@Z%z8~tO+fzQdNke|JQ*|I}Ha68EK@UU5bfB)Fj6l+P0bUn-a@4tji+D`(-=md?8
z)gwL*=*xq>e(fM?-wM*d>ekqu^B~El;X4b_=EJHoQZ^$Qb8`sg4LOECOKU*VwGhDo
zl8LkJ`I>N&i>mCWr9C|goV}UYE#6Ez|4;2fl&XwipM6@ctSZ)mT6AxYp_f<M@Gs8f
zzt#{Pkp3(`9~3^c!Iamp`$}LOf!ICLSL>L$bz#*-b|qu80VS=ZC9nENNePX-Mm*uH
zWlw8Vzza1CC=C2<a@B!nD8Km_MRT168X!~g=e;N$<(CKJ*dYhvhEyKw$_&3iD_I@z
z_f&Pxvv%FJFWegqLic;+n0`O}?^^)OjL(chI{!=I0&*t(rX~NyrAVX$+eW$x<1HGH
zVfk+T{K_Z(`y(WfgE1kddSF*)=gIkwh@!^EX`nr)=A)DzmP;f1*qsH+zrGIL8?fDG
za{^REqB}}~+dsA3<tNc7z?+x*y=DG=N6f|EX}pPv;A4T_E(lbA_`!+RNaHCn@k!2K
zbkpzKFE6+>F^Ci))4x2CS15a`Hg2rdKUWtP%>ZPO08lNR-S`3I`DCPA;cBdZe+Uie
z^ak@7`Gh0wjB(M|)GO=82@J8f+7``#hXsoELKWjTF8TWfER!i(9@76KzkGW!$x5Ud
z#YIIWikF{XABd4?X={_DJ}oY@B!K?^C@E_Mah?U74an9ZpLQKr!XDc(-YOZM?Ch1v
zBwHzuZ`9qUGo?duk)ery@khzVHU&{EQy3Xq;P8&-s~>blXrSFbi;;&=UYNXgjQ=r{
zH_d0*wBbctC3(~TSUGrZd(OvT+n~-ig~Qmj2_ODbJ4pXq?Z8p(t_QZlu@-OLITii2
zWtS8s7=C*;|5Sq)8`eME7}DReo12@P1q3g$8veCmX(WLfjmp3ZWOGOwNu=`M7w@+U
zfJM_vptu-n#q*BsJpvxF^$4jRKQ2DxN*t&l7e493R>1rBAifzQqYmzp)Os5J`|tmD
z0h^-(LP=kw8a@GhEI7eiz-Kn1CK<GYq^*%pn5?4gF8Wk$w-!gP=~nT?do<J>1m@^}
zV=s&6EVOo@?A6Xyk0K?_fIe4Dldakbiof$8IMYjxg~li@b%B5?d64V%(&^@|M=d^Q
z2Bwz$KezAYCp0RcoH;2G9Tb;z*-?w%`+z|Lzy<{G0@aBnNhWe0-F>BarS1L!HZdQ7
zX@DBE7<G<zG%1zw%h!j^U2Hbb;Tc@3O$1d#PQs!dfEI+=9)|xJ09z;lUs_gdM0|zt
z6kcAqtiUVOdb}{80lZ!S$!l&U`u|a+Lt&~6s_ZCLzBty!wUZ+w=6CMg(bm%w)GUx%
zy7cZLw7}{kXw~`;iPeFEV25F?-!J|53sq%g1xZK|FSIvbHlO|Sw;#nXATFR7#77(j
z*-%_+N8a|({?V^EB+LYPx;e-5j%bE3O09jh77<(~2bSE>8mN0H3-RBfWY_i(yJ32+
z_cNx0$n?4?a}m?oC1%;n_{n+f#O+Cx905$QsBz3+Z^PEy!)!RfI}#>cCHPLgsRQYs
zdB|s!S01^-X5D?#r@d#|5rGaw#%@3~dH()lzttyKD29E~-FL9PDAF#i3ZO~H|9iH&
zzZax$1wOu!u(^4oQcJI2mf0A@n*3)8_z9Uc|6|Df6xgm6t3-O3h<P0YTKab-`4v6k
z)_tp3w2?diR3g7WVMe_`&;Wquj>OuA^u*b+*A*1>l5iSXudlBsL;xf(4@tr7m!uEH
z|7$_I@PE=@6Me-2=M^|fh7P5Q+T`#o0c(4wNI(AJA9r0D0~bE9)w!3!GbY<7Bonnj
z9a4f3m&%6aEl$MLf04pAFO*R9Fo%khL(F6~K`6p{35!jiqw7cuVCu>N-mrPkiokmD
z+LBa*7po@2rMLc1un7^thy!9+2E-HYBwC68%5Gi*tfgWmfr1t?;5Iz#m*e;&EMlgB
zK!AXgOpljgC?36vA=}k{0^FS6C}Lw?krTs|fd<8;&@H~^j{<;=2|x|J@Mo{0FQs0U
z@27to<Tg3t+l&$esy7x9`m^1a_V+Ug#S43;O?%~7m|wUkJX&fV{=mN;&U9ga)TNmp
zk;5&npvOfJbeMc1sMvXl>GE|tZeQRJj`j#0lb}982&IO!jsCbbfkXfhbf*?^`^d`5
z%7dA4Rf^6?1_Xv<4b;*(wQpt*@GCGT4~1W_q!aT~|Fh4?-yw*TYF$3g*2*Pjp1U;z
zl*q?C`&xZvk%?gv+yj<uo~84$?J}5%PzlzXe_SlGc4UD*!CzrUjhwA0(mdSzC%0ok
z!WaH?7~%i2|NP$p6Sli$FaX5)1TG!xV<P^u{REhLN}`EKF%*}oaa{WJOYp#GWn<Id
zGv!Hh`0dfvltAVQ<O+Z`6iRtDfD>Llgc<L75o3D%<tR^ehnf27w*O$Kv|dL0FZBna
z2(nHz@ykX{at1zMMd1t=GU7TF*w0RA)Lrf9W=?V`?ICk+YH{rNucZXePAf32?<+8c
z0#%C_uT3(sq+k7S)TOy1nC2qx)z{L{HV^c;DkaLjfT?_#*0~qzWYAzf$ne&9>95qo
z1^M+SHUDSW{wIU~CvWzrAYOo!h@W)*3y|_FYw-_llE__BQq=QIaCe8LQi1Mjm;W|L
zzc29V@vBf>xRfc5f#gCLZPd_Z_!o#~U;|nOtIb_?nGOpfsb@p*=#MIr4p5O;2nhPC
z5<qJ&{*Kg3Acv~}2w;9(vReUg00~GYDmt`(>-v9Qko6!xD<~)|&8^EsZ)Jz->+2U~
zNtq>}6ro!%UcLbeBcPQa5WHLj-~%N+Q!M)<29yv%UX&7{Dh8C|6#z>G9Ih`0&xrsB
zPW$dZzy9|p%z#Y*a^Jrqc$zuCQEMFc1t3Y;bJ6v*biHY&_DSvn-6H)PDFkA6fVbb*
zQ{-B_tsoCcD?lVaQYszlrPqmD`MGeBujF$pvMxl>g4Oar`wW%U(+CsMuL`aP9+hv4
zdJ1T4G;iYf*Z!#5xWJ$sFr*){m7B{5oSWxYpmF6%NpuqeG3mvEcWxjBh@r?{kKWZZ
z{R4I?1WB;4T&zZU*&YY6%d6k+IE^a^u=A(ME<i}r^<%(x_|a>Hg2H!XFR~OL|C>n(
zMQ<;HDf5#|kB(YIkaAW5>btk&r)j59>%z*VUh%mAVT}SeQHD8w3cPOde{GxgHfLx<
z`sG%UV?gjIsFyDg{PtSY{<HT2vA)PMrP;m^Omfw2fE50(*>$UehkN5o8hNZ?e>A3x
z(tenTFQZyxsv2p7OYX*Z0n9Oz{7vd=9wAwT7e4w^yO2s(7g9qv41CV(df={Jo&3LM
z<$sIn#RUvZ10Zx4AUlPIq|wcMr30eVe`Tlg&>ksU0?Cn9kUrnUl4c%j#J?F_CJD(z
zBm)>SJvH@=n3%l0FAvu_^74LwX93`Vgx9S8sQ|+Y^xWdlHlN{srY;Ep>5WWrB<C*&
zD*?GM`!pW|xNiaCY*r-yXW0o&xCrVnm_nw9hD_t*<3Ew8=Kl)?5nZ65_d*JaWWc`#
zV)R$z`@bgTr>96107Q0*0VFl`&Hr%;FBB_WKl$=G65aT~zA%5E^&OC+`tRRl2!4OT
z-=8D#20&vddo2NIMfv&N<`BcLQ#w9ZP{<9K6Vmj<%x*PIgrtTRf#x5#m+5U)1x&=)
zn0CRtccDdv3c6Q!W*%JT)eRqjcLDDX=(*GM{S}Qi=m&4ScK!gk@Wp=&&TI?zDzAD0
z{CwRwAxI5?$2+rjooUlK8fjMkhbqJ5AnOlwo0<Wqyve!X)t~373K=yE*zlbceZuTU
zuCCR09#wppg2RjQ^15E9q?EhRQvGk>>9nnnAmKD(pToF+gQ&0H_!<S1AeuPMNm)$=
zfXH>HB1l?T(!k}Wl5qb^uDL<vk%V|oG>VJ3{d0w@Mf!)f3`YM;Xjw58u!#Pc`B$-J
z31m8^Uy-qxacOx*fh=HK9jHkD@nJIHCxq{UK=q|COzUQpDks{Re+HH>qKKKVcJE}d
zLw6bS*8Yr{z0ipOMg8z<^DH6h{L9l22m~GzLi6eR1F(!uJfegW+vd+P%}@en6qG6;
zd7c%i(s3CTeRUBPeKneH_{IW~M#<E1nfmw-keMUnMW}}u;Kn?|K{ys$#w%UEzCsW%
zh8%T52Idfe6r{^SApggC!F>2TO+s2T6rXmfiHKr=C&&gGf71i_6yJF15C1qBG(Mo5
zw=i29o67L;@B^$Oir;sJ-~W4ZGYg~W1(tMXQ|=J!=uLKP>!m-#DFy`^TfkQ={aRei
zq2J(ET3KDK&p`qt8L{Rg{|Dfv@7uXeCBXbELMYqSU5){5ThFQm`91W%U(i^B9N3`4
z7;zWXQyxMbR~CRv;8%uYZ|>b^h@3o=2+--xGDtkS!nx>=CLk0xM|)vSer;c7OMng`
zt^MjZ`U<?vzHr4YxdNbPNhgmr=8w->RQQ*zTgyQT=LA}bei`n#RGnN;E^<Q&D0`Vj
zo2kguYFJp-Cwaf|>OU{IDlkT*Ak(^Fp!XuE{+CWEs;ty9x*g)8p}G1?jDs&+<PZsl
z6@3{!@QR}H_@A}aAwGIJOw4Tyj0j)+#kB+C$R73=$wfn)|1XKi|BfT|-^iOL>a##k
z?|rYwg8<Q*vgQqG8R5U-KnH?%nMlkpqMf2YACqnSv?pf|{U+Xg2nND8!YOO0<Xt7O
z<vDi^jo3rt2_z-s@!xnNj`6~vmK!NZu>xTQ8D2T`pI`(+hM_5;;fR0kyr1O!+!@}4
zWH_-m0ez2IkY8B#?HgtRb?_a_eKTla50Nd=FOkfmg{|1*Q5A+++E=e{B09y!CfduZ
zF{5T~RE}x~P>1D2Zncly6(HYIkiFdUj`T1c`wJzpFm3Q56R^<o`tDa2diBCWKmV4s
zNDo-(#rO1VkSCQq*wjSC@v1%Gz^DL!qpkbLTf&u|A6Z2^aN@bW7byL9g1d<3Z*vur
zBbCl=o=Wpk)G+KQGk?4o-#i2C7&l68Eil-;W6ZnC0d54Zv}|6t76uo%xK3=mw7qq`
zLta`bKC-Z=LjHCL{6ps>tHP2>jO-Bj#EnYSMweD0E+e)0x(jiSV}K}#1Ww{j4}W@!
zHNIJDBfK~HvhmpFbOsk*Q$6e{rb5Qu_YFF*JQ7pMSJ%`Tyf7*JCV!@re`F#QRMM!I
z5T7aagH|<Cp1?Zs&`Wq5X>G2BY7Oz&*JxJN8+N}&Bj_mKRP@Vaiyc!*F?3zhw>b9h
zjSG5;DH9Ozj%YyT%P22{y=;6R54qfZamW{om_@Fjf&Ozbk76-rcdAg3e6bR2Ds{G^
z6o#G$Fu%CP*W};kva}Xef!g<ObW72<VpqJYC1F;yd@7+B{uJ%0X&kmj*9&NM8n<wR
zsR8|kaX^*v3mCRRtF!4&Xxs~FE~ym1Gi~cTGxT_DEcv6H(@oy??Y?t2&z(0PRK@!q
z^N&rOtY)tpGO1K+JnDR;6NRcsX#ZUYSs~(@9ek$R>S;IQ(=A<D@h59dxSr-;9(%ek
z@f{zQQC2;DAZ$te<xo-To#Z~kLOd)!Ost3(yL|Y;=p<Yj=t`F%6bsA%4FO2gNWAoU
zglMgO?2qRzLAl8r@UR#`2e+X`M$c(ALM+>OGFI!GRBwllf^g~)VX`UKvjbIccgUlH
z>z<_uR!$oEE+VUx8V1Yi#K>g6Mc<0=>$ZqQNtYv8mTIW1taKexlUIAi^?Qe@UBuW~
zUs)u*1ei1>>Vh;ij>lrhhlFhFogoRgPVP`WLicf~k(5lw*;x2EG?<R2m}o+N!jK^V
z`nKBXX$<xCeU|3#6A|m~1-32s-VWEm!CMw^)Ow9V^Dz>#cS`1$=l^_hD-izlQ3P`w
z?sesz!^68)>tAy8*Hx@shwa>?OwuAlYxs?M%e%!q2dssi2He(ORjTlf5IsP~yzf^s
z$RQE}hZ0>omT_)l72QAgz$Puk<ayjK^r$hrzv-x}WADTJv?^jF&ISzW$p+R*QGdGQ
z;ga;H86uXpoee~i$UxW=8xxUnZj<9Zch(VxVRMExxpA_7w`Q>BN#bTIf5+KsrPG(3
zoKD|Lp9eC2g!r4j7UkYQ-Z1Qe(kX*;jIf=IzP<j2_(qsJW%-fci~=C1K-@64Zs=m8
zR2hVJ>6a)NUfKP4z#%~>k)7CdO_^gZU8E*jbehQuDz)HnL@^V<#-hSROacB-u&Kj`
zrJS|FN=EFax8*Yf(SgdAG^!BEvXa?d*LsT<p*6n(Nz&r<VQXVQ&@G>>&$KpZ7}sY!
zDODefmAOTnE_WzZkt%%Rws$ysEoipx9HIg>Ay;J!d%ZX{HtX&cG<~r6%nFW_Bfq&m
z&PL64m+YN`iGlU3&u9AM1HJO~4`Go>z61#pTc+!PjwJHGW~cG!$Jx;#&R*YZ%JkON
z5AwIyaYCZ1YZq=w+wDtz$Pob!WN-FtK3!>3T(2K2QqeUC5pm=`&fJtO6cw{Q?6Y}S
z>FcQ!Vd(rwk>;g~ER8RS7s~h+bj->Mg@1`*=#oG=$WaB)+Rsq1K%LAJZJs{@DubSs
zds7ZZ<)wjc;hnK;_1I$&O^m0B!tXu3r~f*f33Ny}w+2uoP7YTUe!GsJFU<WV*`hXv
z%Dhz2TGKdIH`_HjOzdsPk{dPgM6)00*a}G<oDr)D#(SDtO5Bu1fQj&IBc1@dVKSq*
zSR)n`lFYshwlJI|ZhCf886p)JjnSgsRmtF<YTC-Zy?&6N-B@`v-($S`m{DSBZ2pcX
z;)U?}<Enj$R>v;G&KR;s8oRVYO)biu9^)d<uMyYZI!%j*&rOjikUly#;P!p(jVKi4
z!?7}PI+eW;l_44^r9vK>(PuFAd7Jv$@@Ws8(2A*fbIV~bsG15V@1B8n&OLL=aP_V2
z9zoApAGfg6V>31zJdybCrO%ckDnPcaAv*Ap+%J3>wvbh&2Fk(Z=G|l^(uH+HJ~PK`
z?<0hs%cpIPV7mM9f{Y*}EwGc}vTrK?q?l(qJJjqw-Jqyj3FuaA94s~V5o1(wxgtQt
z#i}17hDf7B*|Y%S3ecSq2oSj$f$b!J<M&L1U(m&+mM^4f=@p8LF;##`6d|6ri4xRU
zymHF+d|2TY&UkWa(Evwd`Kkp)u<eU^D%f}zsxG~$E*CLShs(nbU6K+BqTiu5VA@=>
zrv*BtLGcdQq2kC`2w&jX&c%iW<>>tpF=jzKU-K|;lz^S|p>~V#qsGG@<%prXD<3f7
z+$^mL&J|9+iH7^-dFwigRr?M%jY6Cryr4GO6tZ~TpYh3>(4>3esn#nzMU^i>;<tcS
z^08JfkFdtxi(_nmtD!0Y(tZ;1_?h)v;vAFy44T@LB}P+mx}3?czRaWrVzA7fb`e^O
zeLf#<tdVj+M{IuE`+PogG)L#@!^B8-(26diGAM3c*tqcaDIeO{j&s|b<OgTn8=ExW
ziU`L!;Z=hqc2?yJxi8Ah7$bLbGBfD7xIk?mQ8zpHhjq@N?e9Mt%k8gg23aaGrLzft
zY=&=mrjJM_dvFs?WC*&xj0{M;7W5TkD`oGxE_fEGhd>d-vQ}npARF7v0A~FA$M`Jt
zKzy*uGl2%}dqi$HKvEMgADi+w6-(k5#V<snSK`Klkya2~+~eN8a6^(DM7y6tYvr8>
z=Lcb5s?O%MCdvz{ru2PwD7>ljP%%9XL(OokBX^G=vU_kAsGS~k#z}acAQ#I)4=hX`
zVA)$Akzme0eO@-*<J^CG9VyjVtiEH$d(VJH<kSH3AmxYj{sz#+p{2=Xwzn!=+m{2(
z^?e8E84)s=W;Og^+zi$Co7z{?gDPUK!p-fm(GCktz|=6v06^JTta)JL2>$GN!SSK`
z8w)C=BW}V6*ySle2e!u5ck*NMDp;7Zv`~tkmBDk{H>bAS+IFT8LQ{`ZuH*K}BA0EJ
zyM^RGzIuDUb?<b1ni3F37UFT!72_Ej$emy=b%^8Si|?0N<<xU+lji}xT|mR5G~Ji&
zyfPkcHOQEVTR`5P>9j8~$C(x=$7k63U{5&)1EQmS6vWcVSv0CxC@y9Y;t3xtYoFMQ
zXttGA!M+WNY;s)i75JBUO7IRAaZZEgZ7>J(^A$(9HIqzicn%sS*{0YuJRc|Bb^FD7
zvc6zFeW0UXnErIKU=^Tw-*dle|2FN&Fk98+n(r3&$zm`wlsof7FL4$VP$a|;G{xnY
z8iSq9$Lk{9LPD5Qi|z|(i1-N|U)aeE-};Z`oys^~^KECo90IKa8T`Z~<_-#(6~i~c
zq?Syd2Yh~1b50JX&pz?=CehgKjD+4g|Dr2QHRk1Yj}-uN*vg=WL1TdH?sJexytmFY
z-U_;f=BCoN(9OLC-D-4Ads03pD!DdBTH@oo(<gLH&dxM$mVaw3LPy-5)Mp#Py%%3i
zWtJU+*G5k(8XBP^-II*`ktf!dV3+-5RvE#hDp_Yc?o<CmMREOQ*BjVl{rk~a4tUaq
zzUr8Wkc5fh0@t1(x55e_!1I;~#_tS7GNuI_k2D>{2`}hgtIo7k7t|3)?EjajBW19?
z;g6eQ$UsbyL2Uk*VpMq&ALfQdJYjc0OEhsE%X%<KfIZ>u$FRJrV|{?54sf%sE9Fge
zwTTUA>VQ?x?l${;I;(RpseWuJ%E+23Y!m0b5&2-MkNd|v4V=EN?&jtSoptMK&eI*i
zdb?2$r@e7rvgb{RuNZrrcA{h7tln#ipI#pkyg`PHECi+R*))1&$1l?GxvIi>o(-*M
zE%%Y<un8g1=He3%=bLD9WLXx)8+Hf%Y<p=db2-R_;SG4$%mpe*JVV9aN(E`Y-|2sN
z_LPnFmZ-*m6)wyJ9jh@zGjwC0^fKoHB{U^gWxn7qkLxiAP12QX@%dAD>c%+I_;EPe
zEdg5o*)vsKy69A2Z8y?+fXmig0C3p~JMv)KqrYQ_aZCU%?=3gF%@WTJz1^Z)7WOd<
z=>1W&7fdZ~6<j&<*&4+T2XPB%@1r<NqIfC@NPJt6z#Uh^tua^=1<3@@sIwgEXu{EI
zq&?f=$`agq4TuoFFDcdOe^REcd#K=Q*zg{KdeWt1U7d2#FRVAg8$(i6hwzpa0<|Wk
z0Wp&7$Y#*+O>KY<;2F<s4wg(m<t(1H8U3hBE$SXe$^E7!Sk1FY^&aSXzt^Yzv&T_I
zfu63U_FL;+$`vIIOHUd&Lu#{uhA{GknZ#Q{-+wd#XNJJfr`DtzaIKy=Q7*3m9keEH
zoV->KeSNKH7AUb029EEk!$<R!Q3LWUp~=@mLcd)wWvhgiQ;!&d80_)R$Mj33flUkq
zPggXDiN_}0VaHv)0&68@Wjz3?TD00kkbxJSP7xxC&id4kPOK5gDLK-LNrcxL&tzPL
zqz2L~Sqvrlt45&>%Lxirbp`8R@Pe{#d<}fHxB;A1WOw2db_YHj6OH8TIC6k`e54;$
zrJ>&m?(dxmqn9Prya@n187V&nT-Ic=gOhmpmK=0B4V~%X#<R_eYM=;dzp{RAFOgG@
zp1*KNOvtwmI2@q%H8`5z&`|dY$g8PvdepXBw61R>6F4A%J05HMz{(#1l$$sMgqj^a
zf1ka@PtH(SIwlIAhz)(l>9%4vg$(#B!4?p-(naz-KRc~@zN)_V_3KurX)WqrU-DgU
z#sc|p>1y8hi@zKXX$TPD5*%@K)u~>etcS00F>nY!(Hb2co8ansX4Pwz%Nug~4Zsf}
zW84Q>KywW?Fhq%T-T+DXrpU#iijX^`^SJVF_3h*+sr8Jz$S0~{p!&8vfqD<+at7c`
zsqbPvbp5(vVK96K`)h#Z5vd<LTZmG$UYbF|z?2`atTyy&phT6eA~KM!aCkQXjH#3`
z)5mUo;B@aLH3u}24<TRVYp7#vGoJRzlv<=w785aHC3#au7FG@e3fJIN*&t715nXpt
z&%uSZoLxq^<zNaMaMlbMZAczyH&TNy{zT*c=p+Nv!*Z33yKKDfEcW>>fU8MI4}=VX
zV-nC~@sz?lrH}X4%j4N}JQhkRddY=d3YF4?v@j8>rb1UajwNZphhSeT1A;-o$sOfY
zReTF}r0>59<v9=)LVFb<{X5+t_^4xr{vZjVZ;cYRJH5|ap9d+1&H3(D7bi@h8H6$+
zr3o}iq40ec^eC$%lzy^Sk}Xq;XX?eOCLJ?EI8=AoM2g5oS>tleOxy`s=7X<r0!2cU
z!NMcaq##*34c%Y{uZQeAB^2XL$x>NAWIs1~XVo@%=+~AslDu1YH2k~?z`?V%PfJxU
z#w8$6sxhO+^{j<;`x;V?1W-LE^$61DGRR>YIGs5hs2cbz+J9Y^@6kS$LKCZ#vfVqw
zxef0Gc<FbtPW$uv1_Iut+mzBqO(PoO`CIx8?oo!oS?M;PavNDt_9lW&T-KBC>>05y
zz<vZo5F7Z?so`xm;zejRK^E8SgdZ{$^SzFIQJ5z^#bvP**uxm~WS*XAVtucuD*>pc
zdLMVK4Jf6ft0`?dYk0FR^?8)gs`uDMLHiNW=wcD%6w@G7yljL7Xbdj)WcWVb{^x3A
z*F9_IGTi>|*_Wr5Sko>Z8N3+G>Ffs!i7c<R-)qXy7=f%?oKdP^kA0oSiw&zhemrw)
zM^ro`BT!1>YB^r%d^&%pQl0cFZp=*bGvx}cvh=hp>3Cck$Z6h0iH55SZ)#-O|I&Ah
z?^TkG63X5aUGR89TE|<WiOFg%m-)u1p+!<={<A2x&9<9ed(tsw<?DVxAKMx%5<hHS
z`yAj1oxV;9lfLUB=gvlfNoQNx-Yu9l_<1NUXd5f$ovGp51VliamgYO$T%ugpDQ+#V
z#FP0-b&{0zv@1q%Z{Loma>*v&TO1)OmT$E5&9e!}V$+i~GHRoFZZ^$zYg~0rHK&S*
zB~caPadre8%w&Z2or-wV(Gn7=#}l@w%OudxoEB{b1UlgTKYDKHo;`Yd7juoBe}S56
zd1kxpG}gRD=*_6mc+!m1HFoHrMw>l#B)J&2HfQ6H1>NX}(aOa(omN<V%<QVDsk=M!
zXUei|K27VQ@%MwNjjT+T<qxHL?;qlaAcW?W%;8PSo1z1X%$~e~h1A*>W}li4AAZw%
zVDUj@@!s&;hnkq1o#nRUXGm4e<z2lQrjBWHRnsv~sv~)|j4&G>(?C{v-1N`389p>@
z9;{&=FPpL1+c7b%N${~kn^$$f3qm?kMp7J+<{qD&Lz2cNK(UBGY_BovE@6o6z|Hq9
z=}1zAXpZ4LDx65C@;Bzp+Z(_3j`{P$v!(WsaqZQaHVW45u^RFA1>jGWoKey#wfNq|
z3`Q$#)~*;SHk2G-|JaXB(&rd)4{>wFIIp?&={yC)bauL}FQU28Z`MI9<gmd^n<(Uv
zhcmG<v3V0OFCWo$H)&sVvbKrFg7T%MJ>}<TQAH&`60+nHCS09u#aZx<?+L^4GB+nz
ze=sD0cHheyMiv%yns<;7X1fZxJ9U3*jPQ4sDCj1IG9GG)`wk>nZ)kyshX=cir=M*K
zItPzGmfDq~HSj3*$&LoP1}Rm2xe3MbX0vVF3tO;mPHh(?@D|%U&FbN-@RAky;PRzM
zp$fZ&hX;37P57g->w@T#?}JE~XZ=l^k;d!>WYd#{Zhx1Cn?EWFa0XM}LMz=Kj7cYO
zKWQ^q5Yr8=kumi&Lc>DnQyWfo1qoQx5Ltm|@!x>p^?9oJW2C+PJj?pZB{Q+AS#pDe
z<2ln~lNKpZF}DY#@#7e^jhmlepAXJBU(;5zSPxJi!or{r#FyH{fA5z?#5QO8U|h>#
z9F$;)#fMlsOCV+sMS;IBc;}Y|VjYUH{hEd^C+8O}Ati;@d+zPJ+{EwRKt6*Nm{wpu
z)Aza7X*1b#^RwWr;B1R4M-yA*kDMPpFrQ(!xjZG{Anr8?a8fF8<f_<Md)>O=bItH=
z!w}oI1ZxR?n?bsL5MEgLl4ID9o!GPWXq+a!0~>d;O=ZeQ%oMxc)^18+k`yLmDeG9B
z)%)@~Qyj1aKRgVmu@0YKzq6-_$r~afB6Ks8kf0!W1%>eTt}e4C7V!;l%~>%UMc~#u
z*Y!bF$krH|+!&hZFb-sU7~bi5P<zKmyXI3y001o>aNiyaubl2dN+Fszu_W>9X8_A6
zbvjg%28^w&Z@6OkWlpBJ_FiX3Pku)f+VbG2qbQbv9ClB7Ck}Kkxz?oGdeVp;#o|E@
zy29-3YwXO^a?D^;$?^0AeA?9?rc14SKelK+`Ao}phG_yZ_;sizY|X&?yw?%ovl)h#
zT}^Fe%$`vcfjf#!tf<FXk<K*P5=W%s-^yJQy9f1TbF&~r$4GO&&b8=E9g^2oW1@Um
zvEQ{DRF-Rw%^f19JykKcUpV5kC3a2XifMRg(+6ynnWQnxw>FcwX;Z43E(yFKsGyUo
z{xm+8K{gSGvsUHUC8n3A&mQznXCPg=3=PxHu!MOe-{`s0(<GwXEB6X+7rioz8#?C4
z@jHGmM=UsKsY-eJ7%Ba{tjN=%Yvc5lzL(sd|B(?d@ycxnv70?x&p%{Z4GRt3V|m09
zc(CO&Q}&>$^;oet$imq+j=BGv468*9KR5mp-I!{XGOV7s&yy>knX<(no%0yu_?b=L
zVp1^so*e#NWq+UNB~^>3TmMUSy)>OdqK(3v=N4}}6yb3KAp}u#-_@SK&M0^3eQndB
zj!M41uG*9!85XSZ@{BS{P+TxKAvL0QEtS!^-(k53b!pJy#n5gJ@-BQRRrFQ7J{<}}
z$lGg}E2qn^A;XWbiCokU9%QE-Wp3JzW>!5{R8ggw=l&oZ-QJ`YeB3`%p}&_rGunCA
zKB2YtxwcE`L!eK=d)rhAQS!_p-wj=rSJz4EsoMtf4sRW)d6>pMmlVCty|VQ%vpYIU
z4r5t*AR>jUct4IRp3;p9jXR1<7X0u52>oFIzZ%bx<qD(oTbo9*A7GXEM7PZKLe@hp
zt}fq{<bq{CNZKtq1?w7~#ko}5JVFLo?Y!&vlMZ+|y^~W6BIIP}lk0x#HC#jOR<4rZ
zmTTF1UP<@bQpZu8tTAEMFVg~&3RWy7S3`Lr%Df-POAr&+{EAH4v!d#41;=(A3>iCn
zn~)y#IU_}pTQLu~Z@e0iG}9o@FdP=_Zxe8Jt@Q0WNY7Q9r|1*Unc#6y4o5Fw5lzWF
z&8%EHl8ibv)Yb2MWK45duMl)<HyASuC21G2HX_B*YX?)=26NBmd3!(A6qYh9z}(x#
zTk(8NgkTq?-{mLm8SqzUneu(OgL^Ho-P3O{rYsCr&w^n^KE{2AS6PVvVYImjXR$CS
z%?iDB_p||bQ^5>(e271tV6zq}bF7QaMgB&cz}7wWJAP4#_ww&Fl6Sr_iu2#k>#nP8
zuP?G{n<Dl}c-9{xY!k@j_Shw6QBur3DMazp19#V^?3yYCh^w3F!=#GVMxig%<AsGa
z$F3;KS~1Vj*A1Y~QHb`}ue{?~UaWo9?pD8L^VaJiQk>Mx%WS}CQqNI$kK#-VH))q{
zyw!khBq~aEb2DNpmw8f^1F_bkKyi0eHSve8XORTZ2T70_uDaso<;2~|y*5zSZgS@n
z*;=FK$=X=W%ePtL<>#oLjTSbf7H_X1+%xfwlh&XwF&<-Ei@ipOqM7qGQ9jM=_U^e&
zJ}Z<spAs_H*c?_i?{seA8D<?MqV|F{%C4@WU2gSd?d#6F#PX?Kt0Pd1U=~F_;yL}L
zXzLCyZ~SX~Q~?SUD%@miIQV%&#8%(23$tsA-<a5DI^i4#5yCoO<8D&U$aTOr7LlP(
zAF5)1^tdBMV*O$IwjkLioEvZGNm5y$90ReL2!>S=>Ot_gPld&=?F4;MvZ*zeN8emy
z7WFVFRW6jKbRj}?dEniTaI@TL{2(MFxse*O*6~0g7PZt7Hys6s$0x|zUm%rr=gGR2
zHHKUKwPQ;Qq=}Jkp#Ok^ynNW}h3u>>`a5@&=;-Jg#zw`S{$G08`$8|D&F;V+7sexn
z3#C%GX=>|<hvQ&w3#f+1x*hf@RH0REu5@=CllVR5e__l}d+<aD<Mgc<>}!VYI%8#~
zZZd}Ac1FTxwBD8&3k2=O5#7>wId3VASHO4>+SBF=-J2hIl{_R5<)c1CGmJpMb2%BW
ziRK@tQE)?5UP+~zSzw4XP^Xs}jj8c1C%WC$t&ghBFMsa0<F7#DJ3Oh1-1jJrbNq#|
z*%K;F_pPjlp>g}R2>Qs(z(YeT#S`|pbb0+u8uDg(9TMvqHZ}}D$?k>io23RLf*dE-
zPjNaP<B4Ri)$_*%t1hbPeX+|Tc(&+jwMeXk>Cg`IaglRnHUpb9oQHZ<^B9Zi{~zMM
zDk|#mYxg${T?0rrA|=vN(jcvrAPqx<bazQAsnU&rC^>+D^w8Y`L+21uLpL1$-#Xv=
z&bd2x=k~pswSKeqyPv(Ecr$qH^?CgIYlaUFz(@^6U^dmdP-!_udE%E$K(cQ22mR{)
zB7j}W;LqD1m0XEM8rOJ9znrYvhZ5zt=FBEuu^3bwrf_cwS28oVnlL+;EwJ#;yhkgG
zk-V(D=dzN{v>Y`t9m3H+46P1)nUv(rHQ0MrGAWuH+EAMFDQX!9P+V?kPCm8!qJ6L!
zx~Be$nEjqMr(0cpl+|y8(qi<~@oD{q1A0PVP2TTY9S*KG%!qT@F|JMNt6h)I8tu?C
zn`K#H|6joZ4Li-O11CdYJ9pJh8p#HbqP)ENuqurdIlHu8{~f-aisj>+H0F7*l9Rr$
zL8FjrlKn~xDTY;J+oypc$8Ibu1O+FOf*os0ZoQ>!G%UibOe}t<BVB!1#u{l<S(5^S
z1>0_sQcYJj{Ov>tSB|k<Eoe2UuUM0ZRO}3%Sm_*@6}>q_B<MfIPj8f=TCR0<Y$$Or
z`5ifq-_9Y1p@_Cu`4Dw6>8<ZyWa{=T38Bf0Ucw0$<$>;@=&V)e69r`=X=7`=W#${9
z_EipOL@TcJ?JS<qH-801Rk(LGOs`*7mg+Ik-#C@uCe@4v_yjA-dZfK*DJ?c18Iv1c
z{Du^72JDk5tQ}IieXUY1q})zG`aE?9yUUdR`&E)@CkZB?FqVUV77{&IY<Tw3{ww-f
z${!rwlRqf-gb#Lc8gvq|Yx-7fra+z0)5QIeYFGJ0yG+CA?XL{J&8ArU<hJ>`Q`j-A
zy!te1hm0q=a_jf%@1_;Ub1j~_`FeqJ;i2C;`S3*NU1QzPFq@R8<)Ce%XQ#sBG!KJh
zSH}Yj4zs^z8m@EVx55}+M;SlSq+L?Yov5!p$ZWG)xZ1vo+aPStEn7CPkZ+vVm0uYh
z1)R5>pMTA*5*iJc2{kZ0LJKs#uUil!tKJFUatYdfwv*5h`a)bsl&l^5Zs$<AFWj>!
z?t-BwYK6wDeYsKT!@d%Rllt&bDu!>-K;q=O*yrPpCO^i_5!I%biG#Mk@#Kh0DLJ3B
zRUm8lq!LUsIPvQDl8bnGd8gq&OLqw_%f_PCG*Z=Y%q9rjFHJ|3zcp(5s5fnv+eR2w
ztp8coRKY!R(1>;xzUDa|i{5IC=Qy@)vu0aVzr0Ux4Juwh@wJY_tabK#yft<=Y}0R3
z)kwk)<%6FLKL2<$T3S<sG#(<rJuPxz+{!oBLdulc_lXeMJ_RHw>|33A&+E<Z4Gm(i
zA|nY8Hk{&r)pV4q3YT4RJ7m7N8?!NJM=>Tk%M1ij_Z)Dd;9iPM_eA0_-v@mf@4<Rb
z@4n}!>v1;?2cGYE&I2h&UKgl2n%nm;Lc_Mf>%iGPM4OWP>MF2--M{jUUY)Vf;Q|lm
z<w!WFx@z;cr4X`1ashCEYrhR~YRjv#gu}j#BC^@JbmyfUakBVvV7u?V@b{{#`TIvl
zb^Y{(vW_YzCnv8a)&7HX5hO4c=;H+or{V_Dz|W6D5*VZ+^*B+@H;Z|6<v1lMQ(|}d
z-u60>0|T0Pr@H%*TIEWqq$x+!DqU&oOJ5>}{BM}viOw^meWW3x{@BvTA`d$oJEf!y
zTgn<H=TT2xeR8e8qQjnP$Y-}k)N?s8r{luUUo*v1BBrnXIERKspT8jo!GkvDJ*(++
zp1i7!e7ZTsWDLfF<x^&0NMB@(>Qdo4kVjv8hNiBG342E(#Q_H?&Nt(Sm_*k{JPrIH
z_@);-UvPn2XU0P9;`Pua4OHuucwx@&jP^)Gu_fu8^{_eD$n`u9^icaW=Xzl=I`nSi
z?N(p)Ld)^$n$gt<xi9%1<5V8m;tS#Qq~-beeyi2w8bOI_s-}8pb!D$4Nb5SoykGpx
zT{f))j~e2^izKdLMJ=a#c7zGDy7zw#Bn^|8a~Z93-scock6)d_r}<&o&Om04edWlo
ztZBb)4h^^Wd4nMqy`!`7IO}WRoR6v=nkg>e7z2XV2qkt~1ZMOtQoNL5ka}v`;j>vY
zBt9#n0UA)r#d&-W;#xShboGnMA#pH?s`2Nl`@ab^R^Uw;GV?A!&Ru8YJf!|FxxF%=
zCQ9Z?)#~}q%AK-yz=WzD!iE!I1i`17$7%zDaVLHCOaogg#FDDznx<3DC~C2>;np?r
z?Wol#PYiyuDYYwNlO2^$zYdG)uO$~|E~dA5FId*H-GWCws3sQirYc5)Gn{o!7%A^e
z1kQ(5GTmIn(Ri+Y_mv^_<%>uCNqXN?zHd40#JtLoCIsw9w#-l@yV1VuKb!~(-9MB)
zGrv^`RmMnwub-~}x)nf*5JU%ITEOQ|>iV^K^|yUb!_VzPccVfa#ihTrLc{;G+Jk+q
z6jENyfq;xHzErk$5sP+-t#(Fy`?x2#?({2^84$*mt6ZV}iT$FHi#CCn3)4TkSOB7i
z;Sb|BAhebHHnG2Do%ey%JiA~%K1kr~U)==pk_TQLX}^0-wUywE3_9<vk^~Ghz#Gx<
z6<s6Uo+FrqqaOz$F78bOxCx)M$o1WtrR-?E+&SItjT*!b=}HC4&X9DKyWJ>}+|It;
zIDP%ETewhxZ^)Eu>bAP;S2YKoAm~GZw?)Ir?T`4Y&Pun$Pj8fRUo5fazF4*cYJMnI
zH7>nqGtz%Qi=uLI0=^A=<0UFMGS?AgeHwCyj_=(XDU;Xf0@#Of05F@7UmGz~O>j~S
zMy<iFYy2|)Dr`<#i$B86e;IlW{;<G<B<c;TFYmssnz0r*ndV7x@pPme)e6;5e@eg!
z{39%`A7zR#nk($`5!5<u(&DC~1B7B}mi6(a5wr#P@~(8TC<SRke6Sn;CwVo~dR3=O
z7b?xC1?)-7l)$-1zzHvUOxvfkBF(*@`i2{K@(n#V@f7Q`OS5Fac{kUyrBT7uR<?9D
zN^ro-oJr@GFi&ICfA^huWur<IA$1-fzf215YgxK%U4O1H%{{I=Lx&c3H2J&qYEwme
zP81@2kH>ClJAD){7>pUbJ;#N+g#U_x4V2VjB1K*D+J6`*R&S2?uMCl4Qy9P;>W(4W
z;jfW<3ciE3A=UTuu4y%!cytlriI}ucN13H=Yl+90(3<l5GSk~(#YN8c-L*Aymd+Yf
zTt-I5bI*T&MXk(F>ko?kTL(y2cDV6^cWnlTI6#Khp@SVM>0AQE=vi&I`xNaXTV*oa
z!PyRrrDlAnC{NebU(&m-mGlUsVV@h3G5#ar+>v11x?pJj?+190XuxHnK}fde^sH<N
zwV&l~Hydb=T`PLV?SyH^nI<foN2$760v3q(H6b22hr4nXg`4#3-H2n$>sCfq9SgKX
z{4OSl39GKjIun%oPz9HgY3Sw8gZ*iG7wEgUmhw5}Thzzwl+ctzt$Nb6=+MBg^aySK
z7wA9F#)R#y3eMFO!E@h@nVol@Q$4CE*H-1`w$eFIHZojGkEU$Y_=!|z8owNqi#KM%
zpUg8ne`VrJ`gt(8)9>M|$XCNfx<NVu^dw>^cVVk$qV3FG%TNF!^jzxxHCk8D$wZ2q
z$vQfFXi08fDfGSl(}>H#<?@sn{=r^&Yq7lJ&T?@?Gq1Zqb2#Af#Vg%2KkBQBb+(?m
zo6YWZhRqaaFZTz1MH8C83jaDf%S%{=#At9IE>^gXC7VM-(5S1&2-f~E!Ao+p$%&J7
z47d^Jjwy_WF@#El5<WxPqv$Z63lNmEh#~1$fwfKl4FJoA$F+uXk5^Bf@j|9B0lxEr
zx(Nz;#67j1p32E}4SJw12q9%HN|yC=&T;EyVw1mM3dIVSXbxNggpsZ1sgC2xMsU+7
zQ$8N`l@HyBdVTX%ogGcjMcflzRWC_6%Z#dCWX6RHU+~GjA8i%H@X4r5sE>a!sW0t1
z2kr?wLrr~`s3oTa48(Z%U7Ru9P*>J;=qBT?!+^3z)Yg8{sNlJMN>r7PP3pB9Mba7v
z@AG5fs{YAUuG55HG=(tqJYe3bB{_(Vw?Hi^=3M)4)GX;~Q9ZXKw{D0*nyY@P`5Nup
zPSQvGEPT(ota8v`vMb5Z&SLs>E|M3W=Z!JXw`sFjxQ{Fg`Dm`*yx^PI#KHCzlns0L
zZl_obrfgYglfh5>QMlA8?1sdo)A_{!=&Y>Y1?pJ*mEzg=h~dC^L5bfSFR0&xJVfy|
z9B5=Ieuhg*uQy{{TtlD__HzpT+Cbe^S?t$^2V|h<<GhOo&!Lh9ZVq$Z??H$khlT}>
z0setvq9{edhKJr^Eo^3*DbF+C8K<FdenGEW{Cj7<x<XEgc?EB>!2M5P3~&*NJcc{h
zC*q95p5Xb>8&r;-)Gw~X_CWeA+I2fRkVW&+OkL^{_S=pNpvGSn9oD9oQ#q|8k4<H^
z5Td)CX*TG3=(HygV?n<u_~yFtxXt3|AZX*Gf}2g{4?;Qf<i*`wDHQH&Da@Z|D0SZR
zH3L~Tq1bx!2c2D<TWh-6;3;yc(}khtK8mPEIO2jSu1HLd9#vxS^hw|Px0!*$+8<a#
zG$C8wWU;-i)@cNOMZ58G4H}XLnEw{z-q);(S3J4kiz0OT_H5L2>3uiZk0r*$uAn*!
z`{dwG?6o^V=#Y#A-Nl%^E7#Cg;gj8h-NIdq&HZI@nS&5TM`23$uWGN-K6x&Fx*8J1
z^wQwya!uITT?xGd0ejd4&`wnfg%bBjk5KOlnwGFAh3Zx(kVD@AA*DS*lo@^7)vYoq
z&zZCZtd+LOoK5nl$a0$Q@w4F4qe6&zZm;N!-ZGwePXKBzxWp!7!OO{_*}^vt`D(9$
zwN73$x*+$(JfQ+a((rcy2;aHF!4y(PR^>)BT|n50rSt3_W%BA*RUGd2^}ls?Hrhy?
zgzQXiVQ<wFdx8=kWQsQ;;R;!PHNSxrHH>Th<y?ALv}L}zDcUO@i*K8$QpSMKcfFE%
z+zl%Kd%-m$=T$OC!U!0?fBzio{7})mjMa0uXoW9KqVzNCxson0U&2G^RYS-i`9f4w
zl<;tuTJVF?W2#_Y7?<;l4Fjy1Kx3Yn^@^MPf#Sg_L-mWPEl!Kw|2oF{|9Fh7lch&I
zJf^)RFm0aQ-xUoAd}9oL{kkwbBBGtypCQeb03&)elmKXAHHYS!F;IL#d5kR^Pperx
zc&UWMc%vb70XX+Z{)I1Nrsk{TIx4_doX^>|dV?1cin_Sc@_r4spCNxeQIuJ6h{4Zw
zbN>iB>@^g2!(Dujj;9CanYS15AiCY~`@>K4h+T$VHdmk;?XOhTbbMEW02%_)3HA@`
z1~(A^OS%9QONGe^KsbrRK+rZY&{UNR2ytG5%{3&2t->t7#Lr~_vAyztKc>DdJawR~
zv_<};5yQtA7V<&k0ZFHJKXx~*K(0AI;x!S%2xPGHSkX1wqoNspV9X{<z23^u>2B)#
zGHoC`2)NnJUnZipiCs1KIpx8Wuq9vrq(wET@~<mbBk^`CqwYDO*!o8Hko({6_vc;=
z@Y(uRnIl2*f#ZCQ+Mihj@}zS<gti^?A%3r_SvNl3zfo_eLb(iM)fQ;s%J6mR#YI&~
zn1A{5K#!P$=eTGug{1apo&sC$HXSG2Y}#r`*Tjh*m%eVdUkQoB!U&)zDc$HWneuyl
zkOM&YYB5$M<W#{o$SmF;Ol?wT<8ACc%|KTK4rvTF8GqkcJo6rxeDgHPN0}SGp5U6x
z7h{J1eL&D%PQlgI--i84tb>`e-0RP0Wi?`__JG7kVfGvJq6RH3!yJN2CrrQoUYVmM
zNM10dghk7<qBIP_&3V)WPu}B`4!kSm1P)JgIDuLM)K{b8=7u4xh_{YIu3P!ptJ_jN
zC+@A?wR%4G_mYu*7q5Myq-slx*TRqgRKBMEI~6T_^-5f^*n!LIoY<s~5HG^jnZ00s
z%@o{4U_Ja<rYVOYEaVX>=I&P!npkcxGAhuc{Tby2=?sGtmrCr3RG9q_L6pS#eAdC^
z)k5BMa>@+e$u%Rr>q+T?48iU`UNUc%Js8Sp(P0k+kiAdnZbmm$c$j(<*zVjM67#ZO
z#DUSAs{X6{iVxL__GoIfhOzeI9=9L#@tE`64yjTU7;)eO@@E!BqDyT_I2K#g4<oWK
zEah43N74fzI*O)}N>&n2g4mcgFF*Q^==`mnc5<?Gxkp}SI&c>fnJPh?KI&QXjTmk6
zcmYfAqWo9PNOM4AuG8Qa*MY3g;XSsdqwP?F^~_*Xg!cPRqWyPWB;0Ib7|^`$S#us@
zdd|3^?o4&yiM5T+wlm(0e{V(p#IBr;LMMG8$LY=(<NB;2t98m>+(4TR)_OBtD-s~L
zfUb^vV}6<P8v*mRisDeFnUY5CKe3+Mx%50d%VWf+Yqh5uNUHizk%Te4HgAnh*|9Vk
z9R19PI<gK>+!)(}QkAdciBJ<bqt$fH8C0?#*He_AT+mGMsZ1av%(xcYbc`E*MjU#k
zXi{hM2Fr<Ln)~iGc~(a#yxt7_>wBu&^X34Lr`LL$d9O38zJ6cg1)ioZ3k@~8o>iE5
zJ?LkkgNEUm+T_Kr=Z0tq^*8!kr*P};@m9;zI5aX@EU_$z*$Wq(R()#ok%A+_CRV$a
z;?+!~a9Qo&%en!K?!vX~2lVSZ9lfyat?gW+M4C|~vrl<zEm03Y?(2G^U;o_e@H?*m
za8lA;W+@q}GYURABqfs!w>L9paM0h9vUxUW>i;r{T)Ap{;<~VC0r!66RRCM8D`|+M
z{>|Sm1U<f72r;jRsy5#wrw`bu$5$PMU`V-utM&SsZ4&uFv2ft6-+;%u>knU5e;)az
zT~XC-1CP?v$^9<!qY-syUcIVH*-K@7Z{8i!B>HbdbT_<Gs}i#0xF;=nhq@38m%U`i
zLi+-AaYF`s$ps7Qv|)vIY+tIG#Lb%!2Qp(`n|(z@xG#F$tPLu`b+M^Eqy%Rg{PMN{
zWK7t)WxZ<bMr5yeR-VV%!L<>j_90f>T^X;dKuucb?oc@5Qs~B2E;Ilg<wwvqdRo|{
zj5TFvNMWWj798gpu9<pvnQ}va;({~dzdiX+0`7;Z+wO#%*#cqNtc3c<N!cYo&<D<w
z`znkvWbdAKerK6cPYw?c&l`E0%t7U)sSAiDT`bqyr;?JBKc{=K)w{iIYd2k{&+crO
z`hT&3SqdaDkOKs=UcVfa493c{l=$Ei^Vljlfy@w3*CYBEETNG8Z67hrIxNwgou~$%
zQ8Z~+uC>a8LXxq!A25in-|M-VCQ!N-*15mc{uoeiUHrzVRc`tNwOcmMz>?wIRC3hn
zxW>h^oo$aVeuLe?tmYsS+<(SM7N_TDwv%l^=zxPC^yXo)BF599c4Af{!iQo9z))m-
z91loG4delulrwET-u$3!Wu~BaZ0xGvZ?#^MO0OPwbtNC~3;AT|&CQ_a%{l^kJa+FW
z=u?PGe9HYbW>HZpb38<&4zxK_xDyc{ti;_5g1|(b7gp?o$~602zTb%;iZ$#f%<QFt
zK;K~gvlj9@R*cU<&qHe0YBNd(8LEHTwgq6joYkDMPW|B^U%C-zhc30|3v9>R8NI6|
z*%hNxz!X+9z}goeEow9tY1i{ieHowX?=PIZ>1mO5&157;Jl+&!GEGnBe=n8Sp^eKF
z*uoKctuZ8t#EKBX6DLG^(vxbU|CNZtA|HMA&iCx;nq#F+tdjssK2L0@Z~#LWE93`S
zvLjdM)1#UQ!$-=Pl6FTMav_d3*Fkcw{vJPlzO#U*ASmnQF9v2Djcs{3{qlKtY?WP-
zzk$M=-~VkJh4GmRmeA>_)ipY_SOr_+_dzI-eNlLBMQ<p2@HHBZ2}Up2c<g^QztCBp
zqA)M5*_8Qo*}q16g>52!Q=W%O8VSzq8r0DE<o%X8(WwE^sZJfY?Pgp9Q;(G|LkXJ^
z$x!PjyClT(;`^@85pFAL;{MjSi#<M0WOsj#Y&c8Sqd+@j4n~(3lmqE)mihDrximdv
z+4_-IPw70a^a5dvWw9BA&y!RCJ|FT#V&<Yt;0`b8*xP7Kd+u2;ppDWpGfghrI1wD@
zS<0zW&M{8o<kL1U(0oZX_SVAlPw_=?Zw?a`L6~Uw3YRGpX;`l5lmhpDBCd`0Xb4N(
zyH@=ip5@9Vj!gKNhSI3!^U~Q>4;Br!jL6rk`2zGvB;7ydR7}?iP@W_Irma>_HFrDh
zd{XC5qDUbAozJfTIU%COQN$ujqp{tkj7|EE36;Uj44SX>e_FHIA+j>BBeo&d=d~zF
z`FUrTi(q}>59)A>iVt?yopLgro5gT^L6`K%Ln7G3t+v)PA9?XugWrodUsv?7d?us9
zWohp|lCE9A!T#wHkCaOzcPu#sYO+`y`$Q9`*_%7!ivb=$83&TJgGW1>dC!9Z3yLMd
zZq!~G821Q^$F8XHJI%PhJTqQR%tky1jg|0BUdqrXz=M7}opr|3k$wD+@DoLf!(x)-
zi4~~3S$Z=tx-WX_RA1;58+IcT5x~5JRfYQMVPw;9akrBeHcOd0WYku^sKe_0$S3JX
z097JAiEQ)=uOGIjl<vLM<l0&+8|pTYtt880EP7HyVFC99JeP6kNgf?70x+RA`l_#$
z@U(b5VHrQ80#4INZ~4u0^DE%*l>l~<{y{nbj@Y`cPn4nh#gzYZdJ4Zgg>+5*qwFmw
z^?}}8;Q>a&v%F<|@K4b~MY&I}+xV7zduSDp5=(+u&7*^lRO#_ajP!I}?uvem9}m@1
z4cPx(zJ8*8OI4M`AlH)gCA5_Y=^08)JJU{4U&e|4DqXu)lWIh4&u}^HgFh|8@5dL6
zNXESMr$<>8;q9I7!5_dDVbn7IrX^>Po*l!Jx}=2AI9GMtqh2ELO(DqPx7oFyM)U)g
zLXX+zHQ{0A55S^$k%Tg2nD5<2sBD03y|yn9;@`}Wv=PA8j5Bu%VcMm8$-fc?=OgSR
z=M8GcwAjInMZTtN;r+F0>BZj~b8=4iRhiQ(M4tnKnWi_t8yfrZ=4ClorRk(e;Gy0R
zKAre8NbW@f02}OvD`0jU3c7ZN{Sm>8<W3b-Pl_SN%*T?pt?)G-6#v+8*n(k*G#m@4
zjA;K%fAP=cPcL|Ja^p}htQvCljj&+3<)!<#fQT~((DdeWO`+{4642KkVDmWBX(4$k
zVk)*;RBY|-&~U`DsI21fXw;XK6g+|NT)D=JCfBtpvYIJZ^Xy&s{*9{A5}Q(m4yM2M
zfV(rka8}E;v^T(+=Z6nq!G!FCVHaH~uS4E#`LpW|;~Xem!G))nm(Xc-<T~iADvnz4
zD*}OlE_QJD{*TfGK_W8DvO)&mjAEkDtJryS0cVvEvFp4_z@?hFd9BDCPu^0t+OK2O
zJc9;3F?um_rH({CAD#4E{y}tcXDOmko!8t<OONQ*Ks5d+HJH-;oVpZU+znB`=to8B
zWAk2Nj16%Yq~_6k-@XD+M`v+#NZPK20?%L=iTyWaTvy@80c0hVu2S++KlC%pbDw6d
z=uXaWpRt3VlqcJiyQwqX*X0uHDVhCu763iMVfba&_ZFK8YK~gK+B2WgCkV>911UjJ
z8=TWWKj(b8!MUQ$8?@tfWh)A2*<}V9Z=cRLD*1Pzd0_09b!SIxT^`q$BBzFqN!M`2
zL#x;Sk$*b2VUP2FS|KEXi5L|xddl9X_;#L=!SdsPdYEOOQNo2Z?RgMN`*`6(JlXF<
z!eN+s>5P9&Arz9N(zu`KPe-Y<fZzQux9NwnpSVSi7+LZNwcH{ukrxGgO8s%8Zzoe>
zW((c<P)HPvIfMzG_e%)fe=&Jk6f>qCc5^N4zL7+mVrD_1!rE->`JMlgPAeJ=|6!vG
zBWdCw>u1x5Hgty1gM9wfJJL0v2R1Q>)KgO`mU&+_x+j=FCDk?Y{-O5gs<XCTyvwU4
zc*3B6>Nk8S>fX=(#?p7AdipCi93t0YVNOFRyK{kieYi`BM>r6mb6A61&Dn9|5m(>#
z^ydEroX?}Qg9%LLu($v(wva1}iouYJ+~6Lb`{_`N+@sz$gLEGCuhl7|*=`#9CEPq`
zc7^7#y%93M{^e^}KfW!qO9`|xr*4W7@V%K8U#|a!FUm%@BD{qKiO8~!vIoKGPbL+p
z+WCF~dqg@!#|++%Vh#8!Uxk58{>>&Kh<o?b5H`bORZF>c&{e2O8Et2l`grbC&7@8d
z>_~jlEJS{8auYG}Ee+R=Rq})LNiFy0@8$2aQftOn*gY3&64pbQgur~C1ch+wv@#zF
zPKg7XhjFV8x){0BJhSP#@zu_AzXr}vH0g|rkBvR#NwDRQRz-Hx+hWxYr-tRzluUrR
z#=qE!U;K4CaXLQ$9W8l->srb_)?0xo;;uINx&i;l<eQEyWO_l#noR%gGmC`9*u~NG
z!M@$=RQGv-W#_~T;;8qYlGgC5R)X~n{;{2VD);dAj<uFT+A`DJTff>{rM&4kS)QX<
zgZ<xh>)yQ}UYdS+O|5QNan{%BGJly2<#Izq@BuIe^x>2J#XBsR#A4@}NEdev$fJe#
zF<%Klh?mKSJdq&szc?B+fUgTwEs%`V;*UUff#!$h4Nv~;6?QQ}2IA}4qQoOQ3PG@;
zx~#PQ?2`k{ek^RL6k(pv*Z9!#!-OB6y(g_Whl^6R<y&9w!p%48k4or@_^q$4({a7N
zxN%mk2lvs084NH|s)buf?5Yoh+FXS3-e)RyYPG~i#QLw(-Cd{}s9Y%IfqLh<-<b!x
znGs!pVp}EA3`aE8x~x?~iMIyts#=|M*Ukd-PHt+rwnE=pP_?op_K|u@0DHYW_-G+<
zPt{{Xjo(f{Z~4XgNY@}&tD(bn<>;L4M3ng($6`|s0&KrdCe}BcDLcXKTC&P3u7BQ&
zD_Rudq|YN)sq|>f;?~pPW<M`-X~Y1J3Y!sqrx*>3IC=y#S%gSW;Q^1eSM?k5`T9>r
z)A_A`BU4{(fxJIOB^2ZiEmkX<fc?A8tNpiyj?jF*R)!b*2L{+*s$?y1T9FNqnRdXb
zmn}<ZaX|{Z!X50_DfDrLcnGi()4rNe*&yE#SE={GH%(|ndMlmh_?ZIVJ`nP7u~~s_
zwHACyZR%M!svD1fIiyV0$}9z0+2xyJXju9b9h-z{m&V-L$yxlA=?`9%FXgq*OZ_sm
zH%XylbTv)peSYbwSOpYat<y8z-QCLTJ#lna+v^7v_4IgNlJ>Lt_jG%Wxh2Q4Tx|AN
z#U*j=t(CQJ|KAVEuJ*~`=h`dj_)vE~FKoUFGM~34G67ew<rU0Asq)E{V`Ge)EA@ih
z^NFM;0f!)Ac?9WNPGrcs$PFGxmpTA*U32qcya%f+327W(qY+(&4nyqxvX^eBGLGrJ
zQ6h~kiWv)wO`i(a<TAb~@sdc<^5+}92ZZdv!-bD~6eXtnvHIxHx1(+Xe3<9#Z<!7f
zBBz^2@`x_bpkeeQx(+&Z$CC}2R+qmWiYLIIqe3-a$J0#^EP1rHGsY7<-+eFl>k6)0
zn(rcQq){#%incc18fS&Lddt!gHq|<zEKMv-!Ur?O*beynj|tmqc`<u;5DNU}$}jxR
z-959u>28!__@Z;nxo!fix+wjOoMJg#uMYwb3t4)njnaweI*(N>8)(`Ne9Q7|zOw|}
z6|i>e(a>lCCfgm+b;ouiT3`?e{MOup7Kuq7VJ#1GFg<4GS{?JI=&uGR1JDBgQz#V{
zghZYx-W8c)5SqR6-X{=$Amk|oppm$!6~CISvbg%yU9RST<qo*_M8$w`N0*-8(9x^s
zF-F<^TL*M`oU+R(@tR5V)rd)FaaDHaiECU~dcV>`0J92oqWhQI)$jP~)UB6F-vEY#
z&i}3#kHvp>3LBEOq&=bEo|8TZ^CsVU8_1HPgIzh43Yb4o<RIGQSf88a?8IE=Tx88q
z{yp`)`&v`!uhZkp7LskDHD43Kq?C4p4ZXk$c{`2otULt-N2}5?&7;L?2od#RB7`1c
zE^7!Gb5C(m#Rt?1q)cFKRnwo$R<)_Uv<`gFM`)u%Fb_?B*Z*bx_}JSX*`kAiOu8-y
zVYaw<itUVxXtsSQOHIBRZnzb}m#v55i{ml|fNZrwB`JM*L)pNWxrbO0J`J*gdoBdh
zE`{PDj#|-D2*U{UW6jsn#W+S|;G1wtL?dY#L8P}c`wVnJI5t7UUgrFn#r;%zmf<2p
zV9W9;Srim|!sV}cz@IVB{=EX9nYdOGOoRR1&iRBluW}BolRoyMA8yqk2Tkh+Ck1L&
zFb)c5oq&R>d_Dhr40;%G(2#zkNrfQ>)MD6UKJ@_DQvhI-cjs`gM^B9;!76?C^+~BG
zZ_XG_bpy7&J9UTgLSv5EvlV&qBz7wI6sDK`VL;VP$id7PMv)!}JqR^w%Ja6jwip$K
z24l+vgyw#bY)pW1Mj)%1MHI1CR4LttZL0`t9+Mc~<JM25mR7{mIgPoi5`shuX~8Xd
zpRy8sI}Ameh4q0SWpf<f>Rt5-Lhm~TK1C=V7Nf-X=Y9$~m<IUg9P?Q-P2W<6e95ik
zv>aEE)p0&dd;x*jGP_bj+Lesz_d-^eGNo+77Fj5=yAlgP&|&~iNFE^~4wLHp-36D#
z1H4Fqy^7DoUCVW_)`+~pAgfJs--NTNNDQBYz0Ov^q3*yt>1d-+hu@233CAb~#TdR+
zZj!VxS)6W!4)N`wzBp_4T?B<7n1l6roF`E4M<ag@0%i91fu2}a8|yon5$>PRVCAH1
z21LASfn=oTGpjTMxfi@htw>Q1e!a2#`+dJHVXQ&9`Bu6K<q<oh-qKLqZa8B9BHolq
zN`F=b_VZ6w;8*<pNm|l#%>j+bz`ur1ci*sXWmL`#SW}YiOsPIyV#ePWz{W(wH(Ptx
zv*cu+TS|+QXL0LytIGk}ZuB0d7j@f!O)$G1#72Kk@{10p!?`opQQI5M|3M_>wU@TH
z*rbb#i<<<(Mv~#6q$DJaBRtCnj%#+E{!fV?I_skn#p}z=p+;Tu(}OeD7inoFx|K|i
zmdOf2o+qpr&{PCHLWT02;cT{rigaxzOjdkZg()6#^guJF<+S&e^--cV=11rJbUg~N
z3wXG`a?^(%ePlvS6KdC2GD+$+C5aRXpx&!7ZfWB?Xhj<~ul`(`l?ER5+@YkKUsD<E
zVhyj@&7qYwU2rFi=gqnxavSu$(8o#%2C<}ibxC80fut$@-%*f(=f8Ex>BhPCK8RTq
zh>N>tzSG2%9~p)vG6V*)jmg`4hPHOxk-rU#LTXZhF$U)G+>0WAd4#QhGFs%;H^DSE
z`z<U_h3jK=OihJR`OaW*_b-9#M~0|+C*d*H8-BmB84cfRJWr~tyH{N!krVM)i5)OH
zbj*-o!Vg%7($$J0hZY+m5)aiW>OoBkJrm5jY_Xvxd9Q<MO(yAN!rm5rlll~hx~Ubj
zy4)c6ykuzl4VPu+8|n`OHV-pfV2z~V5!ZU*^4)>?RYcX^P!fK|f^Y>8A+jclbwUMm
z?9P*>$NI2m^P~M+w=C0$;Y{x(-GZQw)6O)}%VV=f!|e~*rztCN{}?w*l8`A2O5rj{
zdqI$f^nJNv<M;Y=VK2tY;I-a3eyqQ3^U}z&=P|D15b4eW3PIhYoWM9oXM$HBk9#}7
zH*bAxl5PFGPk+Uu4*Z^I7jxwsoz_<@9dx?yWyBI+9ei};C#PS&ab;lDGOUaAo|oeJ
zz_gh8?D9^$?~<2pNr49qU-w8)pYrpY_+K0sov@wwrQxtmou2P<=_1pTz}<Ex<$T*(
zww%7J!P>OXHm{iL9!;ZVT;=vo%v~%RD&>$pY1=tu=|a}W&G!WEHS=DQ>pUXB$7*yz
z8#|#U@GF5!q{-Z#n4+hA8O3@~;eHzjvwZiGlQEA0Uz$dGWrp>zAx2Qmr(IZuFa650
zhZg#76fnDt4gHrp<H5^bwAkNK6c}KH&Uud~m&q19;D*N5SIpjZD9F)eTv58s-8=Z&
z4;tJyyR#T|{#%%%E3E%x`>Tm3?gGGGQ!iI4GC~c^uM^x?%#1&-01CGclawJ(hqi)!
z68OXGY>&^)?QE$(z#i^ph_)uI-GXYKxjX!t^SP7Luyt$?qB<HjNFTkdvhFR%D999(
z<_lnuB`3nKv|SVP=ekbf49TX{k-Jy>kI|ytlG1!;L0jW_i)8mTx%Oy$INw!eOvV5(
zMDUUwWp(<q(xT1J<DcJE8WUZglJwO|%S4-}u`g~uf5y}JdWKFJ3U8HqCyuV8$E&5b
zRXqH{(aqz}GG;<nf?`0u!jdsggRPvNmu#8T{Zv!c!g$S>0~M$Fq1_WA?sW`}<^-(h
z{YQ&Q&v5p?{x%}Xeg+F3@{xu}Cd{CaBSZ^o^M@u(Et~gV!Gg99?t`yj9&C-BTu;=o
zmHntr8P?>L7<rudS1j63;r9@8XlJP&%jvRp9?`>*wMO)f%vJgPz%}_tGV5o+9t8fz
zGPy>IbS=qI65o&>L+PeK@Gr0#=g%j#uvGdBSGVwR%YJZ65(8gQ?uS*s{?F9=<AIQ<
zBipfp`k!F12_(X0B)bkoy5?R7>~$Lz5kxcHocpxqhE>Jx(ci}_4p@td4_00V6w!7G
zc)rQ_`InLs%^0I#min6B>EG*FmvYo!?_J9$omNm&mZ8o{(*^B{ovh7^g`Vm^BR=#!
z7VFa#ukf=9T-Pckw2%yeh9S7QWo%xr7FkeOo4FGcBplD5%a)yix8E!DGF}>pRmFGp
zc5|-(oNS?6`qvwzrwMUzM*efR>{_UI&i(8kh+2_hk<*Hr0iF4|xmBa;v;Rl#7s3_#
zK=&Ud^%B8#RcB$~tMv(n4vYzq57a$fzkE)AV1Vaohcbgz5m9ket_%-7TkQuOx{$u*
zUh8+!Hz}kPho^9Qblf9}_qeX?VXVtbZySqddg^PPM?ryXA77pb9i-Sx8QTS*??TKL
zwp(`R`Xlm~tjm~!rG`8k4(e`G2GHTOZn&-#kUb$3X<!6B0ziU4hZ1xTmnEBs!=9{!
zI%?`Q*q>A;uCobD4>Ak5g|N#=bYNi@^{#BzBJY?Mk$rsZtaoqP)~G5+cv33!@NtDf
z4oi|`u6C7Jp5}#*kGOrX9aIt~UGsh!m-?6|WDGrCvEhn!m@<2$AVYn9IzOme1Wg;`
zk9lvHHY)M273bmulp0oQPX%r^B-x`5KT@2i$Fh|?LBCnR*g2jF;>yRK2$Rh{Z7%A7
z8Xn0sFsyV0Xn=P~RUmt2KmC7*K^z{dC<v1U;OPOS>Cht#+WQq`P3k>=@jy)HRMuqH
z={~>K@LA;U@(AlKu|VX{PsiY;rf9%=%(8C8)W=o5(AJXG4Zd=^zl26wWKXV<MqDy#
ztYYH6Hkf|{+4Fvbdj!rPq)L8Ix+a4O+X?L+!}kuTQ9F(y#Evsa(_(~F+uma2$ed1}
z=n839Y%)7-7y?~wS1hJp1zc||En4a5|0NiYoA@eV(ont=#yMsfh$Jm6j1kOS9XtC+
zHM+RHDM~-%tH<Sh<x_3(A**mbTVtHBw;?Q#{^rW>^h6N#*#HBwW(WNCPU(=2u!l;b
zs~Nm0qm_9bTi?Y0PZ{E4Rs6KocfD~uZ??jK{vGY;yZ48QZGl(`FC*umPh87bH`5>$
zA2{zo_ZUxg2i%mWov7e?71YUaKnYE+Y9;3c`eKVALZo|-Qxw>9BVGG}<;Bj?)mxru
z`hpMUX>)hkc<x5L2AfD@BDIskm$7D|=SxQJea_PTU?$OVuYQ%$GPzZL6^n?73QXa6
zveWXpwUXptpg^;3ZdcRm-o11Ev`qW~FwIh91?gH<9#JqOI_X;0$+BT0*XYC_)k-Rt
zGWAJxUsnx#yZ6F9TYlqn_ycEK<^IT>!x?b2Fvl3v32QV)@I}K8KYFBj>vZn)u^Q6(
zk?9MOz1<Zg_ri!h5gyaU@>J5?leurrMW`xO1rdnv#*`^EL#-NW{lUKIOFN!W=^dw5
zXfX120Hr(=80k!Z`<^Z(LgDY%!}jmvWJ@cdKeT|5mYyM9GfUbNf$NicVQZy?O;1<6
z8fPp67>m@6fCQ{}Xb}yvw;zAh;i`svMBI-4nCTKyPUWQ$Bp33#M%s}M?B@{t#L5_5
zZbq}FuJ0zsEgKNJMrEVm^ao+GS$Rj6yvzjY&BT$oBZr0LK5VhF*u@pt>YcTp+ddu7
zYx_tm43B(e7=u9Ej@#RMcwK@hG2s+Kr3ITb^j>e9N7aE?MFR=gxEU409rful8#J6)
z$ynq9`4MBAWo$Kp&G4HwD8zw&XY%#I&Z4$&y4Kb@H(hT_?#+jyfNFB-B0D8*JS&sM
zvO4k4XH;?XXv&r)SCp8h1r!@9DG@rm{FFwL0wQwM?Wmq$Wh^%vF=t*D4a^nKEwyj~
zUJD`jIJJabKMquL{O69p+Gn{>6<_}Z90u$m;xl-7c<rC<d@4__o;`cUj(YWC=*J|)
zI=FVQ`oHAO|K;n`kwGIvgN<tA9$ECp!wsrQaT#Cq!(e3*X)}C$RVnhUq~dK4tlbs#
z2+b-j&b=WtSVLJn0QY8n0h1FYdL;GcT1+uzef7;TVENJMLS<2(MQMKGd*@}@x9YJN
z0<mI=OaV#{>$eJS#>Lup7)W^yQY)Xn0r1Te_T(K=vmjMXmg<A5R%`sd&#DQy=<0kh
zIa#KI$4PK|>*UICqm^L=2NUD7+9F8@GHkw4^*Q_7N7VQ`73&6#%mUu2<Cao!xVG@#
zT_ZX*Ayfh%H(|?ZzaoSvWy|?{$NFGVgPT~h9=hun&qjcL@c}Jl@hRWecXK<r8od}A
zBvSYaVZ$i%DGr>xF<XF*A~fYyV(X~npTK3a2ImF-bXC6(0vv!pBm4VRE`~RiJUwvw
z>w}KRLaZeQ`@gV&Vo||2*I!HnLN_(<U8>_W5IoJhMs|dw=B*>LUqmX$0Sqq_m7||G
z6cV&YZH5uG#eQRy8#F#|--Zic4be^(V19;k=UKMa>snk*Dx)E^zuE(0fB9fL<HC0O
zIrs(NFTBAHKcu;(7giMW^dy8OK$GRZ3*V$)^m6EGm2oTHPW4I|a__zYawc_kVS2G>
z+3uX-zg)10^gfF^25GqtyV^BV5G3>7RI6<stL{^``A%!o6`1Y>y+6o$<+LeqBZegI
zGs-_abW#<H{g;M_ftMxXInPTJMLl~WeXcsV!l%>&pv?v_meC%nyU@71QEzuu%I~ey
zj7ROt?a)SREgDNa-+f$%1@KE5=CEEW%JB%)$}Oze$&KnLy#A4P#&KvEZPmXiOLh4L
zFNA=n0Rshp@4y2h+rIM30rujti+NA(@jgzoc@W?;_7FsfL|*0^uncVP@<03RF?9Wr
z(RmiB^ibZ5?KRaB)R{%SouvT|yd6;&Kn@D#Em#LzYXM#ok8q3DT)l(xvzs~o-u&bL
zQp1Aj^%B@yee4Jv-|o`bc!GjIS$;X5cfd7rLxL$kn#781teTkD(-24<9h#JtD}?yn
z*lT@tM_ZHYzWhUwH>>1xaXq-=Zr-2lAAW*Z&y&-AB<S2RLnNp}j>u5-)|WIwqst!X
zCyNEqo%d7{KRrc=PuuW#FBTyM<#-uWo=p{-&5e&1(W}c;@he;9oPyErJcfH@*2@M8
zCjNiGE#6;|m{@M+mG)~s{ea&`9bc9=vh589k;c@RR>QsbWq<?k=UC&k-JdtJlsv*7
z_^c6^Y8sEr_jngSsBW4pZO+I6;1JB-l>i++CYDVX7iH`9Sx(Phm!<4)4OA7TRJnGB
z^{I*~4Du>0N0_a~xYtwln9}|PV7a^VMin12E&Vie5yV&Yo@fYbJ>0UYf4vd?XW7zG
zgk$NWy~|yVNl=G`FuvnS%L8dT?GDC2gcdhQ8Ar3kq4R7#ip}-LVD!!59rA+tNZ_0S
z3vZr`>R0?uFZ<sM)A$&&Se}H|GeJ3@v_KJX(seL~;2jxn^l@r3{9<vBwmlWjP+f_>
zauy%JbSF5|Xz3gq8=rFh3%i?`%B)%JWAM5P&ll#Wq=3+R<Tw7T6~T4Wcaf((YM{&9
zq$Kp|rZGxu-9})aX6}ExGyd!P0x82WVQ*bstA0D1;*hKKPRZ`<)s(8fB+nH5kW}wd
z=LbE}3Mx}j>{T7kuvs!6IESB0r2O$I#yZ!@qn~#?lMFewo*O&(N}Q+j>QL!=c08}h
zLcBxbG|0Ro%Kc$l*-h~xQUjAkmTKc(Zmw@{*))jAvg2fRLOox5+T{J8B#EFvB91PG
zP^@!5Rs6l_OIR?*(LCA@qF%js^oZO_uYj?4p$Lr6JN2ueF?7aq9{5MCOu`KF9tuwB
zSarpZ9V@IFGD|qBH=C%m0@=>xjf|;Ww$8NA%rZ@vxNB|t?Jv*}58-Toz?a80B--NY
zk;ka^Ltz!BX4&688{z;$-FQZGE>v>dRK-oond++px+EL-GPk!n=u**{Rv;S;4$}p7
z3G2ShzE?v0NgU;a{EoT+!*eV^uy=W}ez?Fn^07{onbZzlOR>DNw=<5(&1W#OWoh5O
z0L5L{#D>1jQ|4TBX{HfRWm{O7f`WLhd|o+YzY46W0~|0#`NNVDwI$8tcOL=YNoL4D
ziT*9b`LrY&;*nGWCoL6eWR!&aM}CwI;OUT#O{*qk#A>Fj?RcVc_ZY2F^u0_yE(V)Y
z&pZ$>FQC<P&%A%*Qs^Bi0LKC7Mr5)yP2F=_dzRiCp#|7f+uskIJ#D`A3Rw&#U4Pk%
z2;rsbzDf8rDNJy&8jiWLQD7*>0VeMuIuAZC@sUtJ9}uU!YsT}N_TGLZ`&4zP57U?c
z>z_lAmzqv7LS%i&+>jCdHeNoVM~V<o2xFlD!$%&6&m+RsI}IPKZ?fWu_19p;MMpH>
z<iF6ktKH60R6K}Ab&Q(g=L&d8JrezKJY7YDE~6=gvT@t_9zPn=&7r+ucq^nQy*)N`
z*v|V2zN$Y{R#d0RH#&UA$5;f41AOMEiv(a|;+DnbuDRQ=?s=<R!ltlH<E8KM%?Z3h
zkyJ8_EIAp?gb;^jce)tk0C01>cYiWE44hT`Cz7u1si{d$mAu#5T`|d~WaOp~;0m+z
zQB5J6>AF9kY&1T8W?L?7h)*zKc6vb)cfZ1A{H2TUp4prlautv~nMtl7dU2tW+Qj1!
zio#;0H2TspY>)k&nsjZpN4&qARg3HQ&a0ts%pXa>G|QZcqXgXah`4qaftE+T0ULnW
zZu9m!I|v=lm)mTa`BCu=Lis>_h?eZK28vz8x)FT0Z`~?QVw%u2NH|De=IF83^qrf1
zPn5^UygYBOxE<`D6ocjU9N60hD5Vza?W6&r!$52k<TBT2sy|zWCw18Opj77OzcjpR
zW^a7FKRM>K*wy3TeJD3b*cIO=UbEUOUql(>sv8E5jdC?j1eMX4M0j6sHyF|#JE0rg
zgN(J1Aj&RZ0!`#r;GRQ!DiB9|-bu_VB{hmIZ##mw3Uw=6&TD=BT%lLm|21lts*Ggz
zAq)rljzX374cd&Al@;#q%`EALo~|w>KfhO}<;b(@_75?}@&6rq@E<dn%e88rEO4H#
zz%&vup8qH&j-{5Y{s-H`gCr!P49rFn{7>}VTy@s|h0@X4fDy&PrVwR<1|M$#fo^x^
z-LDKFx_%T_EV7J#Tyt(Zm+h|lq5S0-y_w2(d|!=<)E0bYoQ&hF>kXX00ub7aP+M&g
zj-s==mSgP4jaz*eoHgX`@%s^m_Q2l*UadxQMM>u+(s$U8%2zA3g1Ocm^D7_Re5n6<
z=F~CrM~hkrPVbJ{%nzp|54AXqJL|GjTRobp0{iFk*kq1VOuPdxDElcwj0sNOFn(tX
zk+2dpmr!41$4K!>K!@cDf%c@uNj{GEd`@a_eK;C5-1xFH`)7e8w3NE5Ps}@q)5i=$
z`^62EemnK>0UtQ$W{VI3?G=aZg$`#R|6sr#X;s1?d<R@v&FC<LIWl<bPv7s&SS2#*
zD)k}ZdZ9z;mrARCv=9&PS<Nn6;WaFVkgurd^RenA)yn-TfcCw)<Q%>;x!{;fIS49@
ziK0o{YALckdk025E{`X76<HPpNFRUnG_>Num{n19Yv%yh1-ek4T_pFR`EQ4wPL6n%
z`q1Zp4x1IPoFo(?xLWq30;}Kf$7ocv5}iyGw(lc93O)LX)X}j$e9whENJOF~TksSx
z<b5jt!^0_ze;fNi8*zV5|I?lawC|M%*Vz8B-(cCgo&ri@03)w9&g3JHv$BcvrfUGq
z1KH-akS7C1_$<Ll5F$3TrQZ%c;z9w7^3O0KQ^X|>AheeJ=GNVRCvzN7K@8AkX3<}A
zBgNl1Yx9ec<|R1wHo@@WyZIIeGZck5_asX&&KJ<960k53)Sb|s0xb5}2=#e`KIX3t
zo0^v^CB@=os-1n`ct9;VqsfU>5X7z*ng|Ee%nl5<i5I9yr$?a<uF<6T+3?)0uwr{p
z&+e$FQut=PlNte&v;Ej_#aqizFd}J`XFdkmj#FyIjE||)(T*M<Cmw!fu6Jm?9rc5+
zLjIYNZgA~OMW@cjsnv$K6FM^&<%EaW>#FTeA4!oPxu6-$Yr9_-DaRg?t){y9k~934
zq1K6Tkw0jWE~A}5vD3?S<|Y$Ij=qFkLV(7k`VgKcQ?Bol4T6TrHk-ave!ZRZO?fXR
z9}9*7c$J7_<GQM^TC@M+!j9rMpP)`JAi$sH&GKPTfPm+wz(4ja|B0jv-ME)y!c0LS
z`_Z9sTgrex@0t$QztCZSddd6JH=vNfneb!aJQsipcr_N+zO*{wqV#rpAYT=F5m@4%
zjeSK~W-?7=-Wc3iH(o2{p6gDsvtydQjt+Z5l=5#OMVY@uD)tjvbnt<nl)uzVlDuLU
zD)?^4Vv`Jw)1zc5e!PnBYjoO-rsm7|8FjIjo<H}rW*<6<J7Ql59FT13n>t$46BtoY
zPz-{)@QSw@a|%jzE4f`=u|J`W+DF=->f|)}g&(I$Id0skUj*bQeX*Y{ZhCQBo@?JE
z+i>Wdt@i(XD2@cH9p5lacZU?mPOqDQ94f4D^>eQ>d6Fgs1|>g6Z05^F_xsS|;{26B
zTk(ATIw=de8MgNt;_zZYiOFOzPCl>A5#Otov47|h0dXlwzCuHQiNQ@>B&DC13JQrT
zqp^fSdva?o22hpnRa3g>xzTd6Dw!(gR-AU6l4V-${mtX8qNn_R8UAN83Ol#Hd7kmp
z67`x%@{#KlKzjrVRx@uPe6Z#3Sal4fYb9Ki>#B8RgYZupP6yyS&L>+hBnUyU;)tn4
zkFIwX0iAUmRRc{;97#FasK4-&d<yy#U2Vxn)Ns`_b!=em!ro<%Mw%H**#S{oWNezX
z(UU5Cjs-0kIUHShrarW^X$y`BWEP)XzOqjUF;A+w3#cE@L`nK257z>Z*n=&E&H>_}
zh4cQAX`0^xV8JicnS!1J4*#rvHcMf{ECp;+ik)9{E%*SyIWb4r#iVA42>k5^JHO^s
zKi})bh)*|Ge-0RrT@SSHD6X2)_+Ss{>b}pbIWw?-w6(aDHeb8P(^dnB=a@q5L53YJ
zEP6TVrvt-2-W=#=BV1)K|1CYy&RqTRHOjHrAcwT=HtvBAj^*giPEnV#<&%Ow=kVYM
zsLCgOQ47}oRtfrg=A;TVtsRed&l)~F%cIAuCSe!D&vSb#^Q&xRnQd~V;T7a>ZhTb#
z+RCrHCDTV|k{V%J#!TBBx-dmApT%A4D*@F_D~d;s@A`c?^)VB^D>s&&=9k|-kK6Ku
z)zETR!ai>0chOk$V6n+ZBd$Bn!4xrYK176u2!41>Z}2vZRgHf{F2sOW)L#|4%)QLE
zb#?kBd3|25^i<<#G^<wuL<%pv({2bYa$LfFrA;I#3TQZ+;2mMTNMTMywCBt)T?ygi
zcOHdtKY-{^(F|r?AvTroXn{N-y<U079<E~5oYjCqu8T9Hp{rD+-j@0OJ((4<laVuq
zk&6e~QRUwL`;=c+^f$V{x=vcuV`N2b9#OqUx1j_KTu8KKkb4ySz-Lxe#*To9Qhonl
zyuD{oQ(+e^9BQN^y%&+*doO~5N)@EH&_#L=oq+VFfFRPl^dQotg)UuCAxH_mL+C98
zZhYUlbMKwGfA06=o5`;;XP!xN&a?M^)?RDvcy-ynNzE*jusg5Daae$MNCg^*pUB~8
zpCwP7cy2oS?m>J{riU_GNh1FvD^fc{-Q${5o;+q49AD-4Y*zgk^Z$_S+AN7yn;dpQ
z_MzkQ4b5K35auAl<G0fQo6=3<9SS{S*O``|MA27Ai2%@A&ZOlkxmlx6!Gc^^lWT=d
zekO0`Tgq79G%M}BWpWLzjmJxxui_?0(j4aLz406AZuAq1$zy^31%zM_rqF%~2*zl$
zwP#ae?kly^i~)>~-V;1%@tU+Bll*IfJ(TwIQXKSPyzP1YTI*8d1b+q%N@8@`*LHkp
zt?g5vbe2r_*7kKiMDjRm*8lxqZfB;R;CWs!MPbUfQgA(0kI&VQhd3RZ4RjH+e!E=s
z%aVQxGDb=b8m%n+1$}UN!|R1V1A$4Bm-rxgpYQk2mTzQy>y!M($y0FaUQkawz0ONM
z;yFG(Cd5>oWC7N3IVDDUO+uqpdox+NRUtQbL3?Iop&eo^b$*_UvwmKRhu>MNLcUvc
zp8Hydc8N9r_p;~zAD7+v*CC-?1F%ii9T-F`wuzNi)2rdwTT0+=ffLjEq*Cj#1AB>v
z`$_;@`$wfpB)+%Tyktpy_cd12Q9dveK#RnM)&n9e?UNxS$%*CUEoPu#q2a)Cxh@M@
z4}QqjiJ2)$*?SHwlTg&^=E)XIawHCDZKaA1et)jQ9`+x-DLE{Rjs;$7{HIb|hNmA3
zgh@ee|J5If{ZIW7N`l(f=ZSNWd_=cCSW7PBL(OB>s=FWWW=%DD{B~^x9;>nN79P6%
zVY8Bx?gvs+JveLjynNMf+rjA}iaCQ>9gu;aH?V|pk$}iB@Ru&Xq9=8PA|2K0mjJ^f
zSo!MdB64=*7t!*-a}rX@45{#F+I|;lU;k=cNwKzj&AiW8AP)RyJQ!&-P|8IG5KLza
zGUx55eAsvF22f^ChG+vAPos1gV5AQFhM~fpiZM?lSwX{*<IldbSjmmDB~nBV{2t3r
z)!%5ko1xvKoblmFWr1Jf^5RSJ_FKmSrR++o-UUTt!zK_uDsm}SakB5;zR)aYk|~#k
zhVliEchp^<efl}j%i+07a!_VnA<-hL?k5$@s6cQes{b$)5HThI4ivI5W}ClYdF{$8
z3%J*L4v5)?gnI4SZww4X-|ggHE!~BwK{_NjG-swMv&{Tv=+!N)Ve@s?iKVdW)JHTN
z_!k>>qENd&2aNVTh6`Ux5juPUlf^rq1VkwAL6!h?mV*w<wY&k*8iZaQW808LG8TBZ
zQ3B?>TP5z}xlBF=h-SaPbl3aqgzc$VzEkHney%WD_Xa(F%8)x9*f3ZGyhtx3SDvGT
zyI_^xAB=t+fevpQKk_m$US#d3gIlD2n7(1w3%#fO{d<E(-&HueloHism(uh=joFFo
zPoN%t_~*#gK|YS$J?zvKT5_&ZqKN%;(Qm8F{2UKqVUadM3a{eC9pxRComO*uWRGk@
z8U<i2qpwc?KL4?U1%Q*CdmQ*5s5?eTJGo_s3&Ri5vd35^1(2a2(~~wO?i%jD(s3Vk
z?{l#V#fvRkj!bg1DDC;-@lDOxP+GgrxBjW0>4c-A)Y+{_Sl}2#2X<S~TKfym3``{!
zy!n3O2@VvwzZs><WNVq(ZY<OnAoS(|liP;L=Wq{Q=X|E9%^d8bM~$GoV$6|E<*Pz;
zD;vJQU|Sb93!Le1Df#6lgc;9{FhFfp+INiLD5EQY&(>*8zQMwK3}rO;CB;H0r@>CP
zZj1HXFHxij<wZK^vG`@?L>i{#bhiiZl72jx=awfwz1f~B0o&tmU<#CXzpGQ<6Bm-q
zl8WB&`i$Bi&<{RbtCuWrVcBIk2miXRncYPcjX2fHx4)$$?ky%)pK8kg&N3s|`J2Fm
z@{3Z)l}K(ZDR1^GcihI>_}43bdozZBW^G3Ik4fnH|HUJ|MZSb_L>v%kSA3ax3_!sk
zXp^VR`mB2vzUu0|a@GT1yk>b@dC(+Os{?T#(bYHkb#;J!aSQ^0J=6bM3qIV4i$P8J
zc+z{UYcg)x24Ll?zAxP4ldjzJP8>ZKm?v$|9L6^`Pe0>g82KA~0F*fR*~8zWtW7Bp
z=Di*U#4A=-Ww;-Wn_#BKREi%Bpq<+Gyi2asd|4D~)2ZNMm~U}9z!35yM|x=10K=(G
zeRlf<8;(nJtNmKS9S<%+4d@>m3z-8PVJHlw(U6=v)<h=*#)!8H8T9ArCF?=`DSN~!
zIWtG@UxVui-<^)EONtS)z;W+<zZ2yTPA+)sz+TTKQELz$$A-gs<ae0ujml|x`U~F0
zI->wPfBi7~HkfG=*;&aIK_Xj?%MsZwmbMXHEypAGTV=Ojsg;=pPL^bDC4}$Kmybxt
zF&~a-KBp@4VRHx?8Ljw)!;#sJFn`?5<cV!M41gl&ew*EV(=tNR>&^m$a*@@x;Oi7~
zdkmE%!-+OZ^>*=PF*!Gj9`VgO*_PV};Y?(Peo~Pb356a{kk*L}dBniI2#S0{yu2XN
z^XzBNM@(;c;HAwcz(prOTIaJWlOF?-Y^SU;&}vTy(9U7Pw$xFFW9x@M>EkX>{8=vb
zGnqqqekq(T>uH+`X;tmQZfWXXXs4fI-Q~)((h7b=LLQ@$oOVN<S)sW<3C>dqTv)k?
zv*>oYY^wa}j>4G7x_HO0<h|q&GiXR^{{Bt0W#)#L;uw_q%+@luQDZ7oXt9;?_tlU+
zup?x}bx+HcZF5bcreJsQcN$}R*)TP;0X5hyV{0r(?ryb1jNZ%GIudv>h)wtNNy=N|
zsIR^=<Cdk&9V2{c8Ng;qd?`Xqq)`RHMUP;lMytW0t71W*zAqY`ldr^BUIc@EC&G@f
zH$s_ef*Y7F8pfZYHrzYxGHK2Fe!%1(+BJUV>3>1ng}-z^c@KRe)nnp9c2iwq{WaX%
z-2rG{FsLv#w-d1Msacm-0-r`0VnN5MDQ`~50r1oUdt%o}oxKGBJRShOPm%fF;-`B_
zcWPujP#~BCrkg}>U<^dw=*ydbUcCqk&G(*)AF4qq9P+a2`1P9|4e<ZvjjFC%frP%!
ze;QX@zYVFKCvZ%it#{@&_I>{0IT?s~U|HzZ=6oXAsQ7lUxD@p;`ze|=g_10PR*!|@
zr+SK)!1u>ubxBweir(x6=`_&n7s%b4V`!LAQ1z(KVn6`XUOX`X&WZ!8V<iSKy8Sx>
zGwsc|V1bk|HdT)w;Z6-d9`TdweX*$0#~LDC`R*7_j9;mYPluD^Kz*Rh$6@pP7z<8N
z<5d&ukMoaWwx%%F7m}L|Mv~A0XZZBDPP6qkpKIQG)k!?f-AWIgWhlG8#CRq!5^j8y
z-QewK?2qQS1BY%v#uXrsfzem}embTl$z44CiR`Jun-0Uj8hvt)XA~FVO$EE5oquNA
z=C@?O@n$bwf-eZJ>Y%@N*s%sXCPwG$B1~P9TQjF&P3Ns@L(6ho?36>5Ya#();(H5{
zgpiH3P7=ksG6lm=SfP6-d>N^d(;xdMKRe~s_O8(l?4cO^0y@b~MUGqVDlXd`Lf^6_
zajaQ#b91A#=_kH|YwLZhPOhtrMJ5ZXr0`K)JsI)!Un4Cq4Gi<^{qaw4E8p5dk&If|
zSA6_%KV`#>>$gt-ms?x>gK(jNA1Vk7cmWp;b{d8qpAp)+mAil1-Xs_j=^SW}Z;*F$
zCXqP_j1@lS6BP8)Tk!N^pr($ZgG+wElRZ{e&aE;mG5hkCJ4bcl@^p=E7~Ar21F9vA
zf!oz8wF!??oR3M-?~&6<On;`rAAo-4*hun#Yn^@XX}mS-%by&>91#<bqgxrtTsiQi
zZob_;Z!qeeM7(|jE+Z!+f30Ya179;zyS5Gc0_yOK23}x<1|ak5$p-Bpo1mSE*r93R
zxnO%nDz6*@$6vo9XwF{0JMyn%)mYHLax*C2pzm)VeO@ita1?16YYNTsrOD6}4sSQJ
z!ixy~jSV8hi9j<cKxHR!)80P9|Fix_UWL7@K_MC(@<~ClHT_#ute)X9j78qBNuiMq
zp$C1H0*H)>>^(!w+P9vs9&>Rk5d~lhLGtj;lNmy~vLP)fg&l~Dx&3*jse&jaIqhj4
z|NKjg-P{vAU3l}V+K@ey9VobvJQe!@Ih<r;?OoLK1HZp6EY!bQ%gkW;?DLjS)Qq@b
zbEX6^E(?Gx0`n|T6v@;PMB&c%d(T9=*-|UCG>XvtVe5tOMP8p;1%ZIksv$WY_+<N0
z>=&>Jet~oy;c=g7y{`=!zz(XUcwG4qu$IC^{~~1=>H}HC?z<khGV7sld>Hubr{hxQ
zBU7yJ2i^@oE$4N!-vl1DwA*(MZYX3~XW4d$-l{zVa4NBMd2H9}a(u9iW4;wnYEUva
z=4lI=JnaO9N}otqxOQX(MW%EGr~jQMMMgD?yf4(x6A&=ud_^FG2e>#rf-Q}m0}<g*
zMF8$&ZB5nRo>=@%cwA4l3K)u#)TIYV*@ASVfz45>78rRA7TG=#-*48ACgWgGfeO3T
z3l{jkY1wv#57a-9KI1zhQJ$-GRqAS>R6kQLqzOQB@aUL{a*BHqOUHHeYD3<hdi19y
zjom+C*dwk6?ugNxkTcoLPDiKg1*xtMpg%ES9~iX@(GDI%Yz(kUk9^{sAD!eFgf)+~
z%*bX^k}{0@=O`O4UgrSM=#18u(#W=g0D8Bx8*ZhPN%9$35#Nu_(0z39RLn7MeF_J)
z@BJ&MJ`*a88KtyUiE(KY0KkFT0D0^oBrpQ~0N@%xwTj{YDEc%edE-Gp7BR9u4V0BH
zfS7WiYg+fN+(L~qL$)a`SJ>}YmM;p)$u6&hMC;F5R|2{)z6K9(=(k=x7JFt%Mt<(~
zE|`!oBDt*<^AjXr;Ka%PM~=Xb-&}D#eoo&_4GbY9bNwp&#&qo}Ej*Vy`Do%7OA!km
zO+RI><~Br8|5Y4^JrG57V-ZS?!$`Fn%CEvsO159gZ^7H22WaM&D(N(d3ddR-cXIUT
zYe{f$8@+>h5KYjgwBe|Ve2@3B^4`b1HyaY~YAP1tQx`B&+i<zLVsD?8;jYdHEyV?=
z!!#z*bsF`CTD}h@wPNA`yY7~!55GBvV9fR3#3v_5i@$o@?@Ua(`K!B=I=t#6u-X7R
znfKpdm6_Ob^50<9f(c#={|#2IZD4-k{{*W*7rX!Ktv5as{1FG05#m^|av*tSU#gt#
zpXGTRwlmYCFOFa9)UMGzsiaLAy%Vnhi11IW+ZM0ARtEOY0$ytWjE|jN1&RsRe2#ZZ
zhVu4@Dk&)0=*Ja2z~GU$b%${f>Mf&J+qED*6&iN+(>_EU))%{1>b+mX$I8Yx%6)Qh
zfiQK%_(Jm#04nHZ*aVcSWHv-x8kq`jF#)6^u9AZb&!i2^BTP#qqGhPb!bmAGCpqc}
zOL_f3uU*ttI|l9WR3c?-1BkB99PoW~{VmNn5*&m)R1DEdSZ*@?g5~DY^Zjtj=EQYU
zqOH|1O(I&OwMD8j5O?{<bBwd4q)$ZxNloJ82sgbr#+l)p*818f>|Li%0gNpF8h;O2
z;D&gxJGH?(s7`)%ffG*|k2N5?rMZ%$9`SRamg0AgtBmp%R>*g<B`=?VNYYs?c04Eo
zFT(8@MoJ%%-24+|w-8yIkL9Wl*;NMyj(Vwyq%J9VO#MoZp{w#)QwLO{*fvi&>ByLi
zB<hHCH3yqYy;2Mw_I#$NeQz5ertp?aaQNd~lJoNo>yLQQOU;U`E!j@6RvY54tl_{X
z2{RHqnW1!1#YbbIFP|!l=kpd!eg_ufi*@5Khv7%uB><Wn|M0mP5TR|ylC;Gl3B!ZD
zMH@IocknW9-0oInIf-ux9<!c=h50dJluofbvIickf??5Gl^Ko*E4<X1XWOZWo<c2^
z_PvYfIF|>nB_Jmq%wJxkIMjWC5glj!9H7aWovj@YMI8sFu9j|VYqm+uM#I_CqGV!=
zP$Y@5%_CRh_}MSo{eWHG!Vbh8r5Y0Up0{6?luN(|4~ofQz+m9w&nVCc7oBK6A0TIp
z8d1B%88C)zDb*__bHlT1NQOQM=|>+-@jQ{UOL;!1pe4`1;MB@>VNlm_cpZ|h^{yxe
zfBiu(t{^_fMZg0iW|79Q`n`k66``qYF_)vtZke;R8*)Had>i1;9|0bjJRis-S<(lY
zG@Xy)>f5gRLvovUd-STq_kz+W7-O1Bg|Y*6Cowj-!(0*L0fLlCMd}<+Vbt`69Rmu?
z&-rDX*0Et+JWA3oseL#(fD0UqZi^HbHkdsBJ2B1OUaB`&Z6l#>G#;s@B)L}Om#-+w
z3b#CBctg%AWDL<}!;cB&&Vg?-V`yNfTdvUd!UCHDOjU1yaazGFhl$@9#=b8@zl}mU
z=QAwWAp^M&-kv#lIj#dkyt*{8?kITO8xjcQ=?}$%kNPBTe4WJP8yN#4#!1KFPyqCy
z@e-Y^`iG*sJ&c!|2Ue5pOyaIQ1wd*Lv8+R~{-H}7`|eO_9pe@R6!~HTbJoPU>;Z3|
zwwa#%u5lZI8x_38iK-m^`?%WA^!jd_iuIJC_;+~io?_?(&j$x+;vb0U^f_IL{fmD^
z7=-yc@N3JPR6C)Q*ZdQPA=A6VCyF)`^c&Pa29u2FuSaGL5VNg5ZRg;&q%YDH7MhDL
zJrPUH5pHf-=hDpKSM_qLlHS#db^q*k1q8I1`2#eA{|{j*#v1bFaySi4Ne8)yb|BT=
zr3`x47Gp5fW@4vLB;|AL$mDhB5SCk9Ec8f0zZmrF2c9`bOOLVNtr+6&_(#oeFCk|&
z5(VV2VoWNKA2qtII2gB+-;{$xTDx#{5qsaIh@bP;*9WxtVMLa;kM)}xR~o|9EnWuD
zUI%LN*0vNe$3d8xFo0ztxd!n+S|B#zaU9{RxOS;1wNH9L>Sd<yLxsm9_ybo#fVp|-
zA~BLZD4QxmC{Q>fwC-IT?#|)vXTFyF`5#XzHiF28lr{-4wrb-~*uxqh;>>@TO=Wc=
zjFzzjg^T#4D~LKVav3zP(2+?)D1wRKY%Nad1yPb!xOz~n;yhv3IZ`%pX1p0%L<-?Y
z%fN)hdH)&m9!9|%dP2B8L-Jk{dSP>m_m~|mkmcByaE7Q#P=zio$4jjiZ%$JP5dEpF
zf%z=2t%o>k-_Q*$Jl4AoZ{jg_Bo|o615nW8$n#LG(v2oT(3`hznDTrN6S4mi>&0o{
zkpL!e@;oc15`St@$~cs)P~4s_E|w8cb-u(mBd5X`3EH{0d0<e<`e;X%EmUz8_$rNe
zb$o}>p+9oL+lC*WQONbfo@v`tc=Y+X{N0faofmU&cP~yXr+7V;@6~l2R;UW;u|ve&
z4p`msnairINXApO2;(gYoeV3$U{L9-(J4W}3q#90FRN^b7n&g=Wk7;!O5G7edrhT9
zk)^_G{b4;*m8pi92&t$$U^Jsv{K&z)-U|#0PyJMip#}p88&@D``2q)WN<~p=9N5Kk
zc2bT6J%+3dW?8P5RBO)0uPa!N&gQ{2i9&Bav$gare}G9=kZ285vAPhLR8S<8;x9$3
zDO++`n<)BVeZq^z_yW8YoEfPD_}fzP^ao3(bz^$4_>6?}zEtY`cAJi)L_W>MosKT-
zIo;nX`N>G%2>LyG_$`t?8*FwzgIajIvh8DaB1(ktNf8S~JLf(C1!Gz$GBj)v<FZjJ
zV3r=UaGMlye1BYfu?}Okz@zCI_hff6c6Ic8qx5LvsZZV{fp=GPXxc40vV58tsVV7S
zGa2FCba>iw9XHAxFg`T;J`2#;m21lPnHYeAc0`-(Aq7Sx2{mjIRGGv=1U_Lbr=AeI
zj&<`D-BA~j8v~?>eHCBDDFa}NELwDj6tlI}0sHCQ+9n)r#K34-UCtk&cZ!N-^a{PE
zVGf)VZNK<FsNN=5&uii)Jy|SMGcixW9NyHF9dZQZf8f2cfOup!Nc%%paAzQgH{1go
zhpk=lO7$%YDG1!J)oXcUeg1Nak3E-RSnP4yv2=UV`S?!M92|qOK^Y;err{+7xypvd
zFxfjt{f+Gqg=(Qw<pQBT^T3&)*AlEqO!Ar{1T`LhSI7_0{yYB}<Oc?~-UM9CIR$qP
zHW3FS{udAhr?^67s6+$y7q6}yk{YUVsY1b~l}><p%{2>J#~D0++IJt~ikO+14cyuH
z%36P2WRQSRVG_67g}@6oy*MNRMYIeveFfX$T;tLEAM&n2l`AF&5dykr?lr5<EEV!S
zFSx?E9k4;9ywO)PonjbbOO61**oDVd{ZXbPH@NcE6w*VEF(Jp4{~?Zi9Mx*qp>`SZ
zM!z6d`*@*jek#0zsA#$18|7J5h|+hGePRHT3-O8Fl94mW&0O-ELn)BZ3v;v^J*<Aa
zF@bqK>TNNK_w(>g`{5GtBI2*5)eKKZQGG#p`-($te+$VcGui+pzJz7K3lT8L>H8@R
zAJdUT)=)rkl5j+-dHS0oDU&-Mj5J=7<H%=Ftp<aB9N#95pCG8^so*vCMDLe%-O}vA
z-Hy{NXXAH4-4wnz_{TEbLQ7oYJFogg9nN9y){`D=`pR~DnKsx<klg3&VX%*?3O@iB
zDV;#hwZ*bWu_<!Zp4-ORUWm3hRwfya*v9u&_lYcU53PB&R7&X2Nk~0lB2UirQW5QF
zrMm@YTDxo9`}^WPg{0$x@g_GT!qiThc;NYm?{u*9D?tEu!=NEdkUKojL?g(@hlp=e
z{_}vQdiv9y$nfNzI~!lKWNr2T=>=fwqIMQ-=rYPBJUP$p0V~O>c`#y-I^NrUU>%w>
zoVQg5Tw=>;!$pokk~XRiqK}lcBZn3+eo|2l7GHS#waNDN^y5;^*}F|DTlgWWxqE42
z2Ls;2?V4YXjnMgQQi>33rk^h-Jd8AokFAbqzg4g=<P|uMWF;ln%u348V9cCNM9>Q4
z7(VaY-Of`}$9&>~uRZUHf(6!p0r#<lE#`C>`Y7kczTzM!4Q~lCY;BnNjY|yr=On#)
zc6}cX>SaXzL%O|voiOePK89hu9hb&G)WMw9@AfDkDs8aSaT6aTLp9ue@y|4F?ZcXh
z3-IkI^E@r<&twAKL5#yN>Bw%E2cdxuZr2B@Wc#8%Ef~9!cC{2HwNi4{hgtv>4gmg8
z2Ov)%piA^{f_)HYd0z)Vca|O}m}>L<l6E8W6xKz_0`Ke8i)$2q?_hg1TkY^7EW4!k
z+Z1E4r8)g?CBtWagr%dc)g|3SGD!3{0;*MaMd%aur3!4S)xS8wy=G?YbcRo&e2HCk
zU7W)jwJbgf2I#7@&J@*p*03gVa4&>f!4Lv>Be02ppcl)_Pa@!ff!^Cc+41`Ig`Y<>
zhehG@jV;fKEUlN0i1r4gGVs4m8(<;(^y2>8ror_&)@GkaiqmPuxwqX?Gu5gf>m>a-
zG$U5d)Gzm@c_w7ggM&3Z7Au6SI{yFAm0;$l6lMO~{o)u4lNRfZ2Q{4G(vU};KT2S2
z<*$j<4Q}3ks7PH||4^dCuW4qK7L>dESx55iZ+)U!Z&<UEX>_wx!~x32>sSXvM+WRu
z&;n2zm`saH)~n)Hriqo%)x@TuWR}4(@cA#J^!L_h#l(BrLZP2$NfPv9S$PvwZ!u)+
zK-SZeSgEU?1V|@w!qHHdDaWWlLx;Ia(apXez4w!J-0i{htV|D6R4KVGM)H=?f|(*w
zy86p*)S_JR+s(m&oJ-(qvsdo(PySq=Zej<9TBX=O%$+@mIPWoJPIFi<Y}R86F6Z?f
zgeiSc%@u_bfP&rQo74wAJOCHJ3d!sFM4@w=9!3Z6yTjXAM7{IW^*fgpr5RY@&T)(M
z4<6J2_NV*2Su;*P;(%B)i;zMB8thh#FBclv;~|zL%k!ewV;@zSR68E{3<T5Jy`pn@
zso`y#CA-7XcGS9i@{p?fef`VUSC6QwIF^Tf3|b_Tgw;&fCX=7f-;yXmxrHV^WXqP%
zR#3x~hUwr*@h<@=LjXKh!ZV?*m@*^Dj;$gyj(;1;v|Ozi?Vz!~Nl^X7jX>91Fl4M=
znDnh6dgZtvK?wuMURRCoV7SW{z(NbO=YWnZcD84S!PcmAP69}$#Nx$=c*v>b<Z;Wy
zkRVBiuhO04*Uc#LIHtbuuImHir5=0s^?(a2I-4Z(G>&_VT;a&A&DIb3lcq8SMc;Q!
zy-k0C+i;I-4fY0~bJOv1MighSyqPH7deQc4+i5>n@JL{xF2aSAw#Lh&zy096`fs&m
z!XTnQ5Y8M8#_*3GF&~p=U#obWcih&tC||d7d1|tU-I8U4?Sgl*Ru<!W3xdHTeUelB
zaqPED9UII9R<)PDnjed3o8CVLw*u4Nstkv1g@=0)fs~`^;0=Yob#IDu>2Of~{V@N2
zqoJP(7zVUL96)MsNU}EsN@s~C_>v{t05$NbV<Wk<mU}rA)AWa(>AFDc87M9=WJE4|
z9UCOi(~pA-O|`==`1ZC!LE0_dmPa2$elrsJ-gzI%B}TU*{^QO1=c~nAzfJ$P17JkB
z$lxzt5|?t?9y1^Iq`RwaouPG119J8v9FstD>EM6tF`#-|W{ffeQ<*9@KA-tfC_N(}
zLi=Y6ISU-aingX&Ws>`i>X7p5?CAEU-L>77_gtM)(q6Z^t&aBJ+t6W~OgzQj8w|h&
z@{TTFZG48MlSa?BCPMNeMObD7<S^_aOt;=mAR_1|n~#J7ltU2=GLSMb+8nj!uGJXG
z3+>>b`?b_w+g=K3zVkikgD7yz*Lx8Bq-_7<*^=?`k&ryadE&~$75xgGwEo#T5=C~B
zD7f83ZvV_XyG0n?vH$<3yZGlb|IQ5iz7fz&;fhC-)_s56o{>Q%uh23nx91ukpMn>6
zPz-;?YF88%QY47LhlY-3{l{XHT?+1=m?cWR^b5nRWjeOL_F{QH0wg(xdd$4!Qg}$u
zXv7&ZP8Mr;{fC}+`q6U`5jB3c`9mC(0guLVXi@i9W>F{ydAQK~g!XlZ##!q9EN(2&
zWXYedY=$SY7q}fDAvZ&5wvp$tz>BSF5~NIKETKO{8ru#e4KYsUW`pQPG>rHIT<62r
zTl;!HCA|N;BFWk5kDHQBeP*XN_fd7_Vq^M1%yvZG`whMEE6|jr17*j^19xzmwd6v;
zzp))3$MT@eo_va!2wnZCKXvzuNwvYue#VM{^h+#7&{%V>@qhw!ZpPsnwZ1Z$Ir^4}
zM0$<8Tl7mZ@tcC;40B9FJ@85k<B;?LH6)lWn@I=f<v~_LYB#(AB|U|Lf6g(~J4+0I
z)$b?k()VJwBaHp*FABPL9zZYyAW6h2N$CJ`@M5QY&wNVII*-2@`B77;nXVf&vG<vU
zpDg0$RKLHg+_<TG_y~qM(*XAAF@#)g_NAxz&Dtk2&FD5D1}=%~#saispyU8ztY|QO
zuV_ciRgZ@N)>VZ=gd1{@K#a-Cbjx<{(=593mOdjE*?S;z#-}QvBglYbI-OpBD88vA
z|JlyAIlu8PCSqhmyu3T+C_$$me%d@XPv~*IRpE{q9niN%wR}TBa9Ky<QIjw7w4I!4
zCGW|@r3V`^qQZnTO)@Lr`NEb(21mnJ=&!HHOaV}BD7M?%sHUORzM#nQnZU7vayh`h
zu93`=c18gXW-g3*IN=FL>J#>S)ct-QFv9ZL+#O=v-9KG$?kxX3Hai{#_LHE8Wl<4?
zaS0^+2{&^}XsUtk1A~VHOGE{a_#)|oDD~tZ@86SCJV$oA`~SMQSJ+ybQw;W=Tx<+L
zwgn$9I<W=+UC0gy6L*@BJ)J<MSo%n3*y?QtO{zXt5<E~<*k@)}ou)52VvZ!-cL71r
z7WhQ_;)Sp!WbmMq*5e_}boB7E%&mh>S!k&r7NuI9GD<Ia-Z?h#C1-%#NajxHy6)th
zpwGM+sGD&QKYV9Dex{|7aG%%(bQy(>O0U@TAaSkF-By}eu|&od*zm?O>m_|ZbeP0K
z^$esR*z&wy3zAY_S@`2+CICg(t)gk*+Xy*tW$&Hny+dnIcVGWw4>19vxJxXZQj|o8
z6ccdtQVWUW>X9n=7bFcMxhzSPYWi)USaZD;xbUW(Vf2_7(vq0+L!YNPB=b0Uwc^wJ
z4ehfR$`xOJbh_SndOSj?gy8sci3|@dx2=wc=aS^_bR`j7RDF9^pfG&XOC9!GB~Y9X
zIz<+~4d17Mk42?Z-8~3E76iZX+PkEkC{S)LK%b4!FELEb<u_%ougk-*&+<;EgiE=~
z%NOi)-)VY>rF#ipXh9s>$Y|gQA`?*RJ8YDg0OLuUxSTtrlaU4<U_1A|T1^1jcuqSh
zfUqs-6C*v%y~KK;NQ`JJP&yU)wx*Wg7@wt8<0mAC#aEHTkmK4H<JTNWc_WUWyJu?X
zfCdsnpF{!s#LYLQ@)9gU!snzfBeC0c2=`a36<!En|95cev*&Ewc>cVt<daQxG9})l
znc9Mct`lEss`rAS*z&W)6le0h(D{1eu8tNSsB|yIEJ?=37uY5MR^z~Uu%O)epiFGR
zKpY1@Wapmg5Iyl(rvJthI-93!Ildzd)X0xYL~0T4XW;p7o!H!r7Af5B9ti)<t9NrY
z6%V4r?^gDj!3{<wy%LGrfekTDqqZ7H_p9bbe&sp(6ECaaT}&eUk}6CdG)|NWxPWXf
zb1E|D#Yd%r_cGGhPs114Ll646Q_eGBl}}61Wtzto){Adw(WFL7N`Y@!u1=hETb@9}
z2&>C4sNfFUPQ76{FXZYh1GjR&VdrGN+rdmu6*UDJ?EC3igs%qZ&a1t~zv^wVQl41y
zxv^pX`gcoZR-fmiG9F$t0y~Frzg`6r`h<lVUL(Be2gPnDB;3EPhiFUW6CDq&sCK{z
zT}c&`&i~?j?BISfDnj&;#3&(On^bs@?@BW%wW<+1eJAL2xOj^Ka=lZ)H!jVO7o=Vd
zw8BBf>VZ5`O(EJ+0h6ZANHq7ttSty5{Q1P=bM9<+khgug`}0oDJG@<@)h|@}1HW$y
z%vMVdb)hIc?{-=UJ>_?piZ$;2eqfmnoRes6$P;Ou-~uVmnEf^nco34Z6umW!ax5k}
z4YM&<i#mvU;U;bAm64Yu*)4DX;vGU6OS3oejx2$0*8PJ<2;u&Q9;h4p6n(ZXvMxzN
zxzLL5U&#wq4iUEu>-*DAr1$i0JF;(o+JNg@M`o+$Tyw5Hl;~xy?!rkV+OB7CCMZ%>
zh?US)DN(!@!kmcBh_@IW|CAMtVrq9$g11V=UU*7nRL$Bj@Arw&_nnu`wZ5~oQ%-d+
z=226khChY+o(i=SANJAKyIGb;UoL$VSBE>T5P+VBI!?;w-u5^Ix(_rK92S~K6xR|D
z2BTRK62s-G^buXLZR;2P_KL@**~lu}V9cXVeF8~F5=u|;mHiv7E6qemyE{camd!x1
z5k`o3Lwk#di1NH&>;<|cDhG@cNI8Nm?fEhSB&iOq+XEozk4C?{JTl3c<)@WMXjjo+
zwe)td>bdSi!rD3mxVZL<3p`iem@wRLD-H$ieGGUfipr_|gnNtedb?L}`FNm>NXqRv
zf}{`q-D{B02v>*_|06f5ig!WsMm%@>;3Hl)W?Z&6W|y0qylZsxd+e(Y>7>cG=t_!z
zDlWUDAn3_?((Sscf*QLva0M4w$+jpP-=%yk=ioG6xP2Ib;H7KZU+SEnsj+g957>Y6
z;`>*pG7@^J-!~U~KwVE}y302&*t6wb3+4o%*ZDt!J3HSa!3P^zlY7^hoz>H$-02t7
zyH#SF@`#I(?EjaGc^{dB(s0J_vJ#=V=&eRR=(W_ZyY`p6<`OSELZ(O5%tB-rJKAco
z+nbx0s;b1szD-Cce%aY5%GFXO+`oA8^`)6b+%N3;CTAY&yNfP*e?t7A9j=_?V%+c3
z=tCSL%-rqr3w-`hEb1Wq<-Xf{p#sMtmfoJ+FO^LNYXUW>K%GxS#}j5C4wU+jutjTY
zD8i0>Dil5&IW(Wj@U}qt{=pv3=*E|UITDwu`_6zabjG1o9u}P*lYA>a4FC$8SXIrX
zlLozC5pS7nYPW(dYg1FgOkGo=@p=T2G7MJNVq}@`UX`L36F&upZ+fh#2U?rO4^=(M
z-U`1vm!fmO>m4zQ!TN%{^GQK3w;76rdDQBffx-_uhk|ElYMuHXzfrU%*ccs?b=m&~
z<+7lK!i1#oqzuFE(JgZqz-~kOPhIq`N>P3f$(|{q3QTLMNS8`-hAJ{$n{^lEhO8jw
z7PPJSQq=J4W|$S>>`9SO|C$*NCgKF!@b1hUsIjj#!^$J~UT7_~(rI#o6r)~u$I5+{
zsF_hBORP>zVPXl)#+lhrZ>_l9Mc2dEh@|1(q`$pkI4=+AOHgx?6-kZKe$&x1v#`54
z%ET4b5a)fNerLmrN0~ZT>tlh<pg&SBT8JN#JipdoonJ8X=%&2W&p^IrCRy)d2;rLw
zCyvQ6jGlOCg6(qT(IdEfa9X&&8li-;J)57^VzjMTnN!3%mnf~H1!t}3k90b3c?+ai
zH@bD<B6;u0CPy`lu4feRN?cMP-uP%w-LBcEuDvAN9PVM>@u4g3X9kMb66&zjSP<y`
z*aeF3-z5kw&@=}bK*smq&-(H}SrcAC&|W-8`IfOSjCWR_aICeFE8G43(ZN5i7<Scn
z?q3P(8mNcGd^|Ruq_Z>b3R8`(DgHE5Qtu9RIVKea`XQ;jf8t_hSJS2wd7Z4ZpG67R
zH}K!TbD~1JDH*`6Ugfj)5u&uT@Hf(EZPHn?)ytPTqkOk5w}04#zPCR<C&GmHY~i_s
zyQT@zx!v|Gt%lJQ76-{(5cTn9dGXOK7<j#a)7r9VFLg&l6F8eMHBmQsI={y+279cj
zlyjYOk=gy^g=W3{^GTvsKV|>k=cs~H{ZQ?g0w=|1oZO369{pnc9EJ89s8EePIdlAc
zRl4+NQJ6{F2EVDHu)9Z@`opn!7l}vD&mi-ge^(_xF~R(E4}I>z1HSTGaYY|?<DZfh
z+I?>E&k4Qo2VZ)p<C>~;v`j@`If+aavUXkfVk3T8nKh0pMF#~fNw<9y9tl2Zi=Cf$
zOp1{%@m-AOT$P{n7Y8CW5bs((hj)JMX*}B8@LHlUFgVC)z)@)F-&ws08|F&G<f)Zs
zQZBO8@T>P~vfq27TZd>)7CaHyC=xD^d!clrFC9U?wOl}PUTT@0+-_5US<j2A<n7Uz
z-oGq956pHEma|fUxK~>6CZDWYtEOoN|AUh*<D4m_>LA0pxQ86IbzLl5FH6J%!||!%
zcgL%BRe6QPJ$Bs2M<GWtbzmXO7Lw6fOW}Sg9<?L*Fu4M&LJEJr=B!)aiQg<y1IY-v
zg1;E<BrtU?VQboN9vZTEuId^khMSs4+;iHq*1Fo}yC@}K9neWW|DR>m{y!|cA;nF!
zuWy1gPgcZ%swo}+1}pC+?7`Fzka3qXs!~{rQ4Z0+f#5veQnn%;oP~lN0lRvr-+}3F
zGWf80Mwy;V;qLB(XlnC)SH6rDhE@kQDjZE?TkXF4NZUB~SM!?H&tm-jW!P3Q!^6XY
z?c|fOcv6j9pA)1P<R`oC*9pPiu4t&T3pg+Ip^LLlqyyt<zM!5?o%js}zW0)fi&Obp
z7>?sBD!5jG`@8PbW&Vp2CnL7x8Hk&Jkz05xD9uvA4Ls;~BQ_Fqd-<qSgSMuTe_LI4
zSv90bLpKs-#s?L@BzO(7Cd0e9#FR6`TUHXrw5w&sFh}#{zw4J{)^^BJ?~E&xS?8qL
z&WC%vEb@Qx|HL$&&(d<6hqBUecNd8*7}H`F2w<mz(*f3RR&37Kb@@l`^m&GrPZGHg
zWD8KISb7q73MrCM_V-(=ObK&)htcQyKmSFhO%-3N9(lT0XQgi6#=KLY){-ZoYkt36
za=}(mxhHy#)&<rwYwX3AHfX3E2rdfUINoS|F@<c4c+y^?egR>e^77xA;*i_$pRa|4
zatJ^JZUmrfCOA26tyt-9cSpa)h$p(<wqG_~1dDV+16V}icb93Eo@&gY<Ie+#1ft63
zrSD`X)z#=PXJ(mKdG+3O*1c!M!i0EAm!2goyyXVUoB26U<XsFwlq*$DEf56z6qQn%
zQ-8-U%}aF8-`rve?phS_saupGqU-`cVlKDUo<v$n{Yjz!(`_l#L3^>PRGxScfZoBI
z3=K()Bt##Rn>Z@niMIA`*taS8z2>D|g~GasXyAuSDt|S4qOba*UxVxhmufvy!%6lh
z7nO#>&Y8H*Wy8L-NHZ(_PzqRjDj;pJRVBEWi597S(H%6Td2u-_)#ANYU-%|y{)5`S
z3-s}R4bgGdGT&8I_t;>mF3;MRFYm_AJ=yPgJxKON^?slj%#7koEG-(FdUUSQ0`oCt
z*k;&=J9fu+*LRdd@SIeZ<kaQuCHEhd57F11!nF<E%AI3Ep^I<iLr?U;zd<NJ>Yr%_
zwgF#YcN90B^S3UbsL{Zq!V|=mW46S%2?L5|W3DAcF5OLvwW7o1FPx7=Z=$yPrqO)A
z_ng;&4xwa)k9?S^dX&S@4BL;V%G?lT$4R#&OL-H@-gI2pr-Af&q9_Og?bLu~OkYi5
zRE8~i^_qC}_9}&Fu5;)e8h#RBcn$J-^ke>HT{YMHr0W)o>k=I{^EO29)y_vkq-;=Y
z=zXomvmoIUI=j_*;hwciC=DX=#+hLGjGYf^16f=`lehp;44FzdQ>X6-${>7!YuiDn
z$C-46)mL922*2>~pvutgrD!8az;lnMc+Po|0bNMzFe;0<*>^M2qv>Wis6XGZ8A%>)
z&^;mAXPSYP3y^#Y@!C4&SG9}_MozPjr*?-sp#9q!5Yxi$gHKIolb@7}57|&`GIGIR
z=>2H9RN8kM5_5kGcM-QXan_x%cG+5O-%>y-#VL7>735zK=|;HN)|)l-`b$d{zJM-$
zt~Ew3Nrm)30K4z6@ir3)8;1F;3$oL$UOQC%eyW1|-%l^WJ4}66JdkywRb<W&Jw9#^
z<e!1XVCM+CQ|A+T)j=$SX7v7;px)z15w76E@o8!o-Lnv4^F}wbu)%93XhJCEj60fK
zfSilC>Xa)UIAqOaStqP69b(;Sm%rHVtLYLm2JrJ7YA>JPTh>pGl6lqJ^Jj39TY4})
zIEU3hCcnwPBoD1oqq(#2Zk5UWWc2<NpO~-WuY?n2OV|gW0(Q>8V4RoKbninvdxB8C
zdBXG6Oi#6lW`KK9o%gM;pc8b+wx^R4v~$Cwkm&I4=HV`)XQk3I?<v^MX;o5Vs(Y+K
zB+XvXv2iICmD!@yzSZ0EA)0w+3lB0JTw+D`U!pj{^U}h3js`I^!Rn+6$d@Cb_z4p?
zYnjf1{e`zp0TV46G$mcr&R*mjU!D8*=RH2#Ht8)#<3Cp2fL($hoj{yhAqx`zwY{ZK
zCMTLUgAdJ25+y{(7@P!V)aMrbJ!qe*mAsoKVXP8LwHoof;hj41o6hsgN%${*5PHi5
zF#NpL;eI(j798~&<5>&R%qt~Qe=ERQi<Y*@TkP&rT*+d!O6n0rL@?K-zO%hF(=sb`
zc={3H=P8%BJjgM7Gc?L!s3tV@`=Ov*{oRJaMU5^*kxJ~IPN;t}V!*g6RIKRr?Q5mO
zrTtHwv;~tn0g{24$(ZLV9giKKFKJVVrihA|n>irLY$zET^-}Jp%{=&4d`#<vP4(>F
z?}V<Mu8wJ*@3Glv@_aud*s{H3>JG41t88;HmaH^NtxNOtb0p-f8XB`4tSj?)TE)$T
z9-beDMf9`a)`>Ucu5#F^!?=PBX{X%g`q^uIAmlrZV|xhw;p_18E8BL0OR#2zDwgO2
z^dWE50Y?SP+t|X%VvIn%``Hm(KJguV$Q?pGmss60H@Pf!y0S4}BYg9RgAHY#%HZw+
z8DPp%eN!*&_?gg+;Ah#c@@8E_<r^o1m_K&TQ{MH;KZA9v6qOq6(?dyV0*KmDpu)L|
z<prVGd%-NSOk*NgU(P3V6j=>wTj*ax&==%0o}D|x4j01TgGg}AK+0HnQd!U9DZtic
zOP1i4(!-qQ03G@>KH&b$R_6Rrha~h@`XYHK-%Xi@@_|Jhiql`VrF@G;_V8u1;QUB0
ziTr+vv@+S~({5UpM)P`^=~WAyoXInUQrmAOAKHbRuLf>{k)hW~$NoxO8_CaD4e}2)
zZf8xox`HHN%tv3x<qv>l9dGYV9?Y)L_dZL<p9>I@o4y(wm;A{%(m_nP577gKGr54e
z@4<Y~l1_yBM@V|`-Oau?YMn!K+o0pzX;}iBlDJ;cE;mo!@14_=%}%o@;N4}U3%jci
z@i!{R@MBa(=3h&_TcO3aS#_J$fk1F9`0}`<(*ZxBjI5l<tE>Gkj|N^hUYOLWr{Sa$
zkH*qskPmacrx|q~ie$1l3vGSzX+oC;JP?LjW$@79sSm~?5BEEMEKi06e0;b|m9p$A
z)j^+^JWS#?BcNd$TNG-6<KifEsc$})?gKNkBKXzT<Z5F4+CqBSa@ENs=f+7_pfT?a
z4*F<Jk$&WlaK8NTpq;mV#`10+r@LS`%Kz;}S8^rnj@KgYet*~@F7weZMr3-HHmi*_
zxX;FueFDYl*X1@^jJuP$>Ee^UzPi2`UaX8-R)*$Ac|;DGl%n_eQ=5mc99M#kpp)Jw
zzbJbWYTKbkm-_8MO^2F|^*7-vcdUcBJGy<??TUC(sGO}8%b2=rYsg-Ju_IT&Ap7}+
zkQRlE(=<6!X*0l%D;u1`BD>wWs~BG3lC$F#+tWAYiij9EnJ=oT9bL^!3>^*??u)+S
zaI=g{-JPGeyKkFiP4HiBtw{4gdL$AZn{PGg;74nD^tS&cLx?pk1~?~#a|rZyA7APF
z(+Ii-6+KFcj=V>6Gub-nG9)0OEN>oK*;o=CJ=OeUyc?Q|AP|PVysK|Cf`=ZFHWE_G
z5B2s<1EIL%^Qnrslm>||y;{Nhn&+4n9&HAEZ??9HdHxsT>z6kS9cQdEKk$j^w&{DQ
ze6CFgWvJmH<uS?6P;H!w`PMV7m+2FaszSVcPAXUp%S(q&vv?2DrHyfX)(g3}I~pH(
zcaCoz5`vX?1d2y!rMbgzUOBuBQV7qEL=_RbsuS~@v>1U}TFGOv89m&;!4CT*q2g`i
zWYa=l?My)});1j)G*3PRve!J5MhQX}bwyj{?d&_HKgoE#VI|Huhm`_J<Cc{}e(Vp%
z-bxYkC?uv{BNW?vS*-CLcz};AVRz;tVDLe-i3ZWk%Y)Fg(;F6Iuxo6ZW2}=i{N`lx
zV&oO}{SeH;B&7Q?<=U<0vus8PB32iA^UYj0Tp1>z&-!<tbu6Y=(z-sM+UmQOA{|}U
z*AQ$R7I7PbDSfHi=H{-EB|IW{i_M_I?2i&^(T$2vOuj1-*uXfJL1dVsOPO1tiz`0Y
zXTBubY#O+76p<Gnw2`;yDt!=*I%wW9f^#EH!SD4?=~@%fk1beq1`0zYspxNqAl4pi
zxZPz$rhL6J{CD)fe_1=$f7MvI`Dx8Hq2Ic3Qt8jI@T@VvPnE16b;$e;HnFUFt2oqm
zQeWFEDZm_%Z`3coRTD{cjE8xfF83U)+iu^Oy!FO@)~X<sSLzx_v~ah|hdSzn7MF;@
zh2`+4gdy{nWHyT@qUh>c!l?8Kx!Ch=qL=9vAm-Z#78ds(Xw<_J(Y{!Dco%rzVD3&4
zAJ4m}{kqa;dR&7aRBp}&3AY8?bK<ww4=@z_FiyYJ@84i3u$K3U%zKd4&y?hQS%6B`
zOnu=4XvlYG3Z(o-Nf$0ol_Txs#m=&&LR;JI@$S!W-R&s=&5$4@@`>|(r@;_>CzfRJ
zX$qkSJsw}s4{tImEX~4Os|c*0^CA&<GBc%@61;*0r2%?bv1FA^mpvl~YS8V&q%yCO
z=qm~o7N$HOaD`~^%)NXQcysnu`$F`ro?qJbB<222Ek}_6O)}e;FIrXKGGpga-|gkb
zUnVja5BW(A-MyEzFj)EyZak0Zjfg_KvTi_I!<GLcw?ufX$p$N>vK=e5A7gj0&RKh6
ztGxY+=c3c|aM=<)XKwv1=axkPE{*l;Snoj|>FA)eHoariE}uEo{eF~l*+_FRVsvj`
zy<@g!KGLbBF~&c>q^G$RBVMjJkN^p>f2Bq*vXrJ*B7KMTA$~~B>H|4GL1)e%e9>d+
zM*_FM42Q}NS@U~k{SArl&cHdOQMP%SSWgK;>ZW^d*=Sesps>wGd-|J9yRcJYt8XwF
zOO|?1uui@xHjBVfZf;V5u8_kT?{__(i?Xwq$6^a0#PPIQFu)xw?Hx}tG-~bXDMvf=
zO*w)h)@6E=;fM3#?!m=fcXJblskgOOp-8l^rZ8Y!x8rofTp(MMgDLjq+9&J2o`X|b
zL{>>N)2{m)N)*Z%snPpU!Kyt^P3O4aGw@W8zQ+5BK$inYEq_y$%V99t88)1WAaP39
zc6>>*oOJT$F8f7Sot(zNDQyu9<4lk6rj)UPAEFHB1)c07M}BmHGmO7=a2$uNH2MM$
z-RdmS;jJwmQI8tNJ4v`IQob~L6n1JQZDU24B4JPe<g7HUYm#-(a4CkR?ni&Us2Wz=
zrx4H<qoI7oLpksLw)i1eZDIZwDu)7bM#|kgQIF)sV?r*wWjH_UKi1kJ*5tz>yGNe8
z0b}PFd)~M5OKvjGNy>K5{j}W>mtW#6VhO;~Hj(`WqF9q&_T64s_ULU)0fzLcMFtDL
znvS>qDzb!wZ<FV`lrz^$#NPtvzLy)?^hxuT6XIsn>XYx+n7T;mk;m;ez9M;MDjOer
zka-XwrC^L%rR;(#EAnV5`h#(mN~OS*^)v=r;V=Oiyez+F29I6$)EI)LFBf&+Z;5Dq
zLSGH9E$4?rw`|G@*En=Oj48z;+FgSK`iigiCSfkP-8&XjQ*}Ch{i8@riGfpz%w!;P
zG6PS_GR_X0v%5);bP_xG<~eGCKGy##IF_;u8QFCDCp?tEEgJFd0ceK`-sBmN{HwNW
zJKrrZD3*j3LRfdrYMC%>7Q*fCuvSBK=2P%Vj(F*vdhrDR6qAA5r%Dk^?~QiRYlXP;
z&T!(cy<zUfS?HD`s~uK>a&X$^V!Av2ZJTN&!IA9rPw1$W`vS;Z+Af5fd1RE|^~cX5
zDG=KYM|0{U64&-NJ}4S#IhRmy2A;E@ylnbU4(V#M*A%YOdXbt{n{?fk=}Yr!V^liJ
zb?4c?yz=#<C}KA4U-r4)OcU(<BF3p`)hwpomT3rK&Qu970<lJ52hTUxVm1drU7QL|
zZDe2oBytUN%+gs7FqU*wX-C#i^O}Zf_GSgtLGoTfw82Rc!sB)#+ujV0<+?Ft2o+)z
zJ!f@?2^YGM5DFumMp5V`q2<vEM+bFK_W1zpD0xfwn-3S6&l-&>5e$p#dv3cArv1W6
zr5Ppi1>!|KNmBKz;^b3t&wRaGU&P>}3c8Cj1wR$PQ8u!*sbcE;;O?S0Cb)a=#3aH$
zLt17eqz@bDc?ZB7JHNMbbqT={Nm2&j6l?UQ9pRb;Cl!^EcdfnoFxFEtD|HBU%5rU-
zZALb`?Lda!wx-lQ?V7jIR-@;6jp>RAZT*liFU-|s>!XF1FGS@EER(hR%S;!34R~~?
zt!nwI2+oK@>q|R9fTaoIya(WKY}qackO<0(ZvUhxe5wS*xj{f&c8g3jB#Gk#SXgW3
z{z2<GWCdT?H&jE;dk9^Lc|xzh?qD!zAv~$kPSQ+z1laQfnjU9pNu`t(AAh&Js52M$
zGGtBk`N%L$?imyZXWT~=bh(`-*i7ukb1qL#_FDZlp47b1$Nui9>b}@QSBH7~+AAa{
z>En-UOvr~iET)7G+ELwy<TgzBHM^qQyQ>X69ra-0uXv)JjuLld1tNciAkW-XEYy{G
zL?#@Gzbz5-BcPJdc}8D(o+;8?kn+)k87(w}0Mu&rCJ7>fi!-QUcQUb}=2C0zK@2at
zWdeoKv+!H>pd0Zn&rU4q02_+^jOo(Klz#hAWs1WDglK4MDAplwAlW7ZVJ)mY*HxZr
z`+mO0!Gx76>SSrTp&s`QEl1&YfmrK|j(WGK^K({{nV)H&aenfqGEEkzevTT}7Iu&R
zNC6Zx-D#bO?O{C1xwKinc5Q_oJhxs+S7~R1l@Bd{w%mOl{vskFl6x7wj$M=U_O++n
z^M%#0gT&yVoA(SkSeIX7OPCER8<|6o2o<<1{|9w%0TpH2wGB&`bO}gzBMOp&C?X+U
z(v8w3T|+l0rP7U3L&p$934*l1&@Bo>Hwb(eD)>D2efzBUeZTKt>;2bav1YEBxz5`A
z*n1y)pXW8}elu*~Nqxs{l<ZqXgrjGK&4B_NIZwggPXr6uNL$M9+d92|Gc8S1>=Lf1
zR7+4Cd7448Ef3v^@#0#r?|l6ne5^fV|8`4$aLC)UeIV87<M|2Li+6<$h2PX7AaijX
zvOKZz?%_^T^d?=G2uf@{lObKdz-Qc<>U}CGg9wYX-c<kW@Bn*-ugOf@y2fF7g8P?y
zzA3!9C8sy#P_vr4_+o~1G!2@>W+g;&pj4LB0{6~IU{-fkla)$$3z!qfY*Y183B<hm
zcHCND_%Itt2aPrtKp*ZlKQMgrWJ5)j#Tq$p>Dd7{lzos89J1)%HI0<i#9X4KiS~e2
zlAD%(t`RmxRm;ovc|!zc{}Yg6z3Y!vtRs*bV!ja?i~}})bg;wZq$u57okpzf>_KRw
zuUU#sg|<1l>^PQ)d?#_g67VZd4BOY9%roSL7zz$H;uQCoqfWtYV~bGa84j-A3f`?A
z&1fQvaAE5^LE4@U@$j0fqvWUFATFPKw^+I6sdv2Cf}Er9>4qV-$Eo3(7TFLMR8B7S
z`xv3D_1)>MqmenE8Dpg>$b??~#z3YUeG`gj2JUhD10GN|o28hV8NI#2;QZ?uEJIKq
zGgs4Yk<X_}{a#Re{pjXksl7es`1;oCenFI2Rn+aXhxCit&)l&zQB%30JMG<Y$k^7L
zhaJ(jR<pa~UX%yTd&Hz#<<vk5=drzLj<Sg_#>x&_%`2gS@-ObJrxDt7QwA>y;50uB
znN}RU-EVz5Ijp@V?OOP@+CI;E&Jo1Z;7cjTbhLB?9`9fX+^?srWC72;G$Rr9aOLh-
zG{~|_TsS>yn)9NYU_ePM@R3y@a%nbJ(ME9waP)xzNEDj{=IPN#1v|ZmQ(`98mOUsn
z7SvqDm_4!iadRJdCeY2GI&=Njk67@mgjUB9Ezi~77U7u%*AX|u{v!%riIfo|>DY5o
zid6L3E!r>+&LfMwaE$(Gs@S{a<W|F7pWnh##q)={-s|zZf4UJtJK|ItQET4EM6QD*
zeDHZ!79S7c*I%c+I3^tK$*Z?yU_bE&yCS#eMrFQ27b4_zWuf~-g2iMy0pL3E6%o1H
z#k=IaXX)v5!wxz+rEERH+p(!g^uj#(;8#fEXSllD6TB4H3hEB}ZlG$U#A;1`a!)M#
zKIcoR`|=!nb;m2IIQAA^qM)mOdrg3KZdJFR8$AJ+UXUuyi5InCb5slbd><8S#CUVw
zJi5W6TB5X(p9j|Z9nGjeALm>kSAx-rFvhBNRuM_+HJM_>U2zN##KMO44)`wV7HFnT
z_RwWFedRG%TZ7Y}Pij!$x`tVAxd#7QP<XSGMEwU{^jmmHbbB#q#%wnzSRd@&nb6aj
z2z1w4LiO;PL^Xu2h+KElvH$c~*jetJguTSc8O=7v>=Eg_X7FB}3*|;px9etfOJCeu
zEc=8w>o;K_eevh^%&1-bLvKG&D^ge1q-iRMDv=fnjSN^!D-^0EFB}=(NS_Xxv2nD<
z*h<E-$3ay(q9$Q!MZLZ`<J(-PFNzh8AEj7Hx|!sU7`P+R9VI+a*M4p5pjYJcbSP-X
zE>tT^rK;G8w%2IguRvZa-1Brj(Hdjw`bl_iW+shS`saDChZ?6I_SDt}qBUn+ls?y)
z_HLq(#&xGByMso$<ZceUH!$^NEj;%=tTMLpA=)o;9fuE$Z;~qX2Ohqv;%YMlA}HBt
zzxhlOwWJ)I@$8XOB(1~g?~6t!A63}vG0-3lXG|%m<qwsD4SXALk8f6KxG%$^Tgvsf
zQ?b8Q!5xYoc{FTpHajy;j!=a+d)mTBbU&*jZ_D|H)i<?HbaHB%AB9345cKlG`6o!N
z%c#le&MjL+dzTm0r%T{m_6;PXs1h0FsHv=PJuO>sK`l$6mhv`R`vgov5F{@8?Vb}7
zakqr38m#b7*U1KxKHop)HnGHxm=%+!i~{9M%CQUw?|+*j=j*28yk499I3H{4Ilb|!
zcCx0<hdGE-9u_w?#gX?sE(e_nJX$Cp@$o+Ay90I|gS2IVxpCH3OKal=1nmgvNY1?g
zfsWa!)d^B+Llwi9tC=U|A!U6oZHb%3H&qRA0|*Wm@**k=(xEne6f$w1yzPdSDj6<F
zOl?zb81|l0=mF&QQZY(=XG&`AH#`*gDG#fgJ)}A?^yj4J-|U6P^2}L%c(#g^Q?#K>
zv$RO54DyP>vgZ*Z(!(6NUs^+d+M|p?^1#kkBkNT0xhxZ8Q$!1k%4A{PM?*sXHN<M0
z=48W>Y2q5oYhj_%FzdzIrKGjxXwFjR**B&s=h_Hkw-j?TVnSz>h(i=|geG>Jckfah
z7G~>w>&kE2%cx1~`ykxr(_o)@7a>dZ%^}xmqPeo=?N$?k_~%JYaw41&E>s=gzH4Px
z>e!4uul-!7f{JzE4#ALZu*=k|@->3_^q+fSb+d?uyJXKQOUE1BLYI3#=&(s_$t+;S
z6ux`%z{gXe(c4o@c%lFP{y<aDUfh^oYKR2se41a<zBj%o2haKN=Aqq?k6!squa@<M
zz*8s+3E6f;z<eF5wkoN_o3&<J{>&+>rG_?v=TBKr6y-jjeU>(c7j<Ofia(ImpU1t8
zx?daUM0-Z>N|^s>B8uSIQKP4&zt{4*n{!8j_W1tkH-`f$-C>056Ffc0lP7xgE0$Rs
zwI9FNw0*J@0Z!t8_q^s%iFA<+oM!-Aj&do4PSrp%Ot0I|rql9!E}~BoTHS0XWqDCd
z0&CaMx+=Az>mNqWhOV!WKejvL>ldmxXMHn^Nc6ZO5J>(U_LW=v`(Rsnww-NnKi}W<
z<J*TCY0@gy`r{5bH+?$d^Yu|FW!j-&!U<@H&fS=dT&GGvqO9OQDo`<KwV<u?1%>2e
zT9H^5m~$S#)jfcfeZFQyL^VAt`$m_iYI@qA&$u>s(!8{HN9R+W?IIfWSsDWAyjb+C
z#AF{R`AhT4sZ=yag;b)ozlVdS)usv2o7Z?)ZQp5UzTS&>UMi(uTf5;)i5gzU5(eg7
zbUYPXZV5cJu$=dm6!Jv2H1OeM1~ORW&E&AO+ErrhaFEt$uY6GZn1%sJ*NUr+5|;!Z
zApFmE7nLDYnBs@9ox(1iBdrCLeV7f8xRc0;ln>U3Rug5{GV@6D3Ik=3aZ9rX*K2rD
z!DGv4a4C`kD6(0P4Sar$G+Xsc8wpKt9`xR=OhgL53QdEV1B=x<F99DWom*#^NLlt;
z2C+y-R#OcfR6!bkX){gzsmsT!)1m<$80W#LF6Raw9OGXh3e>hkv#$=#SPZ@QMK~~;
zTd*FV7b<zdM>m9V^e7$mm&aJawgFgMp9e7?Q~C!=Jpi2=7vhWIkMJn1PxNn_)Ix1?
z5RmF!ppk2*c4I=07#u;FRmQop>Q*kWsBhY)8zL638&E8tI5c#Cos%SQ>Jor(2-DI0
zJM<==RiBxcPN8B6l|0G4Lpb-mbZVYzH5}t1=>vOSLshYpG&GVNBDrt+*v$kC(LQjK
zKDJi_t%w^L=N!3+C&V7Px+u<?Yy-WHJyR3K_Xa!?q-vXzQTF-nHW~B{G*qk}q)XQA
ztqp-u`H5im_wFik?CZ)JS|0^Vk&gF$f#%#!ls2GVF~MLL!s?VFk8@s*Y*xOj`TmQI
z+~ebKD&5vu%hRM7G;4E;V6T;mE$5|(mtUtRWbZhdrJ^{alFoZIpEQY(&I6Pd=h`R1
zD$#E9`060O)TIONNfYi;Pu1KE_?m-}^6c!XKjW8goB2*CL^h<EHt&49N1P{QUPjPs
zwR&SD^j6!465lH!<Jp?i(nR9+U(S-6Z8j5}I`in;{=C%R#33^7g3Xjw@5M{hZu{EA
z-(J5uj|E3Ix+i?v<t)PXHQKGQ*>`;3JlMiG<l?r$#WB54H9ZbRYT~{KEu9YZW={5r
z+RuF7dt<ua>_Mt!hT{Bsky6ZOF{DuuIJ^K6Qa43Cq<mmgFXp)h3aq*&V`ZC=RR96&
z=RLhH<)!1jh_^vu%E%<3zm3Hq=({1qJxg+kM+8d?1ym&sd-XY?2*LL+8G+72I)z1%
z{7ZrRCPvv#NuV#?T;h}|9*aDX8>=}Q_uVLSqF`wDbUS<Zv8?I|b-h1_Bm;q3!%<3K
z3Ao^VUzwev^Ba&!e5h)WNAsE!ZJB!NJWTZ326jQ!2`t<8E-ClI$R6J&$BQrh<Pgl$
zh3AA|4mF}GGQ!Uv8SrLrKkLA8wbicdBByZCc9RUmrV?pG#imMj^m!BDpf)#ASnPiw
zMX;D~{bz|M@V{LOaK+{e%KfA{u;BYjJHp9aZQM|Dm^P5N&5MW7&o<ig>p;(T0@e04
z>WH$~t-J$J#f*lKu(L8gxh+>@IFIAE8WL+JzO|-Y3ctt6CKmGybh^x(Fn&Olu8ieR
zbT+f#z9PWs4;gWoW(?po%}hURV#M|lx!ym$vlTl#7_BZC5huEtH*xFnJ`Iszwc(Q<
z4?(;qk~rLlKE0l-q-Jl^FbY1YeAugYG;DYJny!_EOB$yS>YhiN>m8%WmZ+eg%{+A3
zcw@ML=B$HJiBsIUY-nBM7Je7jMBCStV4NaQ80!@ySNx2#Obv#Q<USD?j149%ot|*g
zlu);)xnnd-4x6c4lZ2$MLhbS55pLeIbM-m-qQ<O6R@nDO?K6}Qyx|)^37I7sblmHr
zrfVR3rZD4rGQf{*GcM?>69H0IK{%ZK&ga~h%(s8HCMQs#{nG(HnzzB0$3Co4lN3DN
z`-Xq=7`ivjd@j&U^|T=?=aAB#R^Bq_yq#u>IyDG!YM5BX2-4$pe%MzX8hpqhmpj8l
zwX!L^Ol7tq=;oO;0ZQaWnO=3W^(N%_+~m!Ss$FYCJ@nz6S%KW-JlMUY^>m}GJMPh0
zZ@i&IC&>v#0;VhK!BDyCB2-A&+Ixtfg3XbK$K{TQst&TCLK^R+qgXxZiv#VrC3bkV
zX7uys;QMad5Er@Xg%8&^N9wGSG%M)Yz&_WPCaS5QCM&P=STEWOjCb0Wyk6%d#Y;q-
zkPf*yH(1`R=NCBE1h)Aqmd*f^ADxBs^G+pwM_58G*vV)U!!MKfa|w>cp-cuI2K)jA
zpXD0DXiibU;6?_i2CT?&^g(IBS=<3D(g`0a%*>5EHh@c%sL!Sl+Y_<yEFsfJ;J~FX
zZ(MkRZW^z~bB_oH{@$ItX{3K}BZ2*JehLe2S{ISykZ{B!FsCf_?cN9^rlD!?cz!q9
z^Z9iU-z?H-vQaC2UBDN|7YLK2K6rgMY*|rqyLb0*f}0waSv(eY2LunaiytdhnN;N1
zM+0R9p;YZnVjL|eOq`Edu}rtoGLf3xH=FR*o1<HshVml71B%%_ZO3!^%@6%Yzb9{5
z32lTl?<aNVEf+qd%@efS_)fxde7AJ1<{mEL&EmlBnP`I|8?B}fWci2Vk><^CptPrw
zZvx(ovTDtwW4K_#HX&CyX#Jf>E=~ks2nOt;MX>uaTTw@HlJx2*Kd`!{4gruJu4pi`
zu3@m?cG@Z%U7<2JWz1TiF0wsHezNd=Pm@CQ;maN)$4nZgcs)^5ylKIpaeGbuIt65n
zZl$mXww&R!E5`$*8uKYe!Be$}8(&^?EZ;XrJmX(0L1Y>1Yb`@HZ(cdhmTrx(Ofp~9
z_AA@eJ^qrH0F&Y&`}~!68l?)mq)&QAA!Xgi;3-x#s!M}E32coW%!-;cKLnObYjDl#
zc{nA~UDbWkIVCHvzMbyr>T*akUE1rXd~l;;;+`UJPrLhgp3kz_A%9!L2i^>vneZIT
zM!VjJeJPEybX_D(`w51wPP8T<xpEDgYP2J7`EaYQHuN(nB28cOS8atr_O+cn>~i|8
zH7E4xtwU_G%2Zh5q`ymk$-K55?P$`M9pC5lQ_+1lOwV^dv<iK108&urf|h~0jypGP
z8!R@+P=m~$P#QG_B&L#!^a+u}Mn6e-iQ@8~H=j&5K|8dIX$dJtNY9o7-YU)7v)0=`
z`H)@m?W^V2u#~-z-I(nP>80-m$mf*n>&+Dkii-Bey-O3qtd+Ui!YlU~s?kGyFaC(A
z5m>?DCB2wTr9`l`tS{Zd#4y_+l>Itilzkw)oX%}+;2-VrvL9CWKwPI3?EknoUezLD
zMtX1i$V(rM{hQYH<C}hC3sT|(uKXiiR0y)uI*3DA%B!q}1*bQ3dfS&?y)09j+i8yN
zfD(YojR~4xhS`04mp?CLsNmPZTeCUc-yZ-P_2MBHnyr;B0L|Blq4F^|vuz_6?g&}C
zHdH&)muup7M+B4FY>o`Pm^nc*cxu12iUSHCWQ#C#f)mbAG=W44W5@Ed@_0JI?%GIE
za}MN|ZtFp>Qs0{-M2!^gDWW?}X%(Xg;*2!pQS;we-w?8~ZUKE3=HWRIiSt4e*nEW(
zeE*#bV<CGM&SK4sz1f*n4Ss|9nbw5#YJt<&V1z>1XtJdu!_Hf=M%Y9tT<pu~QB+b8
zH<W_`wi=$v0FZfI^74z`Ya3_<1XW3~xRmHQ&r-~Dy7)NGSHDzYS^<StWF#<BD#1Ie
zqAmOP!>4?JO&fj22E|!x1@!fTEH0=Vc?#rWo$Ly6GlLe3_L`6G`Fdm~+HR&<?_d_9
zyzUWokJ}X{2)3crp8^xgV9KF`GL5!!b*o4SvA;H^sNDI=R*I|R7rios6{}=nCBl@Q
zjM_eOtwzek!(6oY!|~EVZ->^xS1g_UM>CH#HdLfOdw4|&&%ETE5;{Zk4^iM6vwGqb
z9R@A-OEA*8{jxY>p-{W2@a9e$B*?4B%mdp}dNAy{=u{Q|+TI5*vcy=!;}*^}be18_
z#EuprpsJwR$Wd*TyLw;NA1Z%+R}9f~y9wB+U<rQMg<7GTx?o&LN`$fZRXZ5kz`L=f
zf<uA6LA}?X6*r+$Qdo&)>z<+Z4nzT#1dh*ET~*mMkO!V(F#$6;;MBA%7)V>TK=HKe
zLQBgi)M%ZLJkW|0e>$@EtRg8*8?@JWP_aAZwBXUM)M`yWf*0lC5TW1el^=3sH{4@B
zVxc*(eN>GRW+ji9g&DBAjnftYx}_K<2VHD3-i#8$?~~atzs`<-uEzpROt+oC{V_1Z
zRqV7M!<a{`a4#7`Vk#?Na0_9c8|*NNjKY>yROq;U4`aiP*-+Qm$6IL-E|-idf|lB$
ze5X&mxvi@Gz^pORwqqyEVZfAs0xDk7GpD$#&<r^mFyO<%RJvzbUOC#^sTMNhpgf(p
zo@`s)oH<&6Zib1nAE{`dk7#OcTM@|U>MeE+<s7k)y~m}<I`U-eUVYzC({PLE9p;WL
zT~|b=@l=0X61K2!>jE=6hn^-L-l&9Z&qU>9XChyt%a!1Hnx|82X?Z@_lM;;#m^d%h
zR1S&tIr@|hF}=4eOaR2Op9h~fpZj5RZeUU#c}dMXtgUv>k<|G-9G#L&I^}430z(r!
zTVc}nx25BqKOk%n98h_pUyl~u<}sqV+iY&=$2k+IvOdzIK6mnk`CM;>Q)50SSnPIx
zabL1Pb`I|B5l~j<t<&m;J>JQFNb<#_hd}O2u1k86tC(#j6g?TH*ylnX=f)d1sxow-
zf-<8mtSR+r$J7()yXAoUuKl}Zgw<Nyaj0sIsn0?{GupS^3HClcIUR4R*q-yKkHZb`
z@POwG1!r;UhI!D9x;*ddT1IN(PU{HpUot31=nJ*}Y6(NzdXjk1bFH)2-MC34b%sZi
z*+XE3XI)y9BhL?gQO5TmL?>ifMxlxDlEVAS0!k)=t!{30gB$|~nP^a)PxiVK4@gk>
zQJb|MacAVzY_xV%Kc#+)6mD6y@!=XRk?o=HQiSbUI_^EKBSSF+OiDJ#x(BmTUi8N?
zl{a=ZQ4&X0X?9*yuU6i;MjeLhQn$ys+u=b!vaAz)|A_B~-7F2esl)p_U3};xOj#$x
zu1MH7oeTttFlBwaI<djY`pBl=9*KlM>+iOhHMmXuGPYFDoUss0Ch)BB+G?!#aRMov
zRSPArmF0S)_tFy;_~ge2Wo%$wPix^v26sg`3;oLcNIF!nE8g#^`x<<<#^Wk8OJhbG
zPfQux$WwX*pIt;}11%n2o&~TEa<Yb?AT{1ZH%pU9^mK9)@)qYDcQP%+gtWw|*TxYb
zOn0mL6}6W;4DX{(AV{L5#WtkFGI`DJFjWqO6*8v#c(0hUfPHXL_(|Q(#nX7a4I-z`
z=LqW$sHbr;j~Ky(U9dn``Lbxkv`<3w<mIgyU$X|^%5<(%9a?EcGMbF6$u%T8zN1LZ
zqKmae+|hNqF2brbpvRUWic*z#C~4@XU;`?1y~R;>baN)qmPDRw0b(a6$4eP};__Ix
zB%v?0<`CR5yUnQ@QVe~?Yt<L|I_5r#v*0nE_X8!I5kh7pm*4}3s)3Z(PT(6dr+luX
z^VL}K$HSuR^*ITecl~+Zk@nQ2gLn&}9@NCENn5GC#zDu^@i7H?2UfE#TxqmZoSWK#
zUGSyX55Z(JiP&Fb-J(lN=CYrSjMBbFZ;^h4NvC)n8RlL7)Sx^qY>;|cSdVsWNqfIo
zM&Xkd#=|225&gC%(u=~B@HACR9BFfWwZ;*G_CH(#y9%;`M~!F0Zc7o%B>4iic$8az
zTKmV&5-*qetkcE<1@JdJ{AC}vWhvpNiBswgIe(D4P`wgF7D&U-Ukgve?^dwvl+fZl
zdQ)!0`*Cs6h?uxbYlu5>UmL^qMSzS$HN^(&NT|l$`shI=vgic1v_|V<&cL$vz)(dT
z@6*BO&JLiWMxC>qM(pW{Lh6x7h>O#Cw$`;;F!9j)mnWRqcN;ve$zo(<zm{qEnC~h1
z2wFKc`;chA@mZC1>Prv%d8!)ohLwcccR4tk>UZFyhDR@(>Q6LXGn$prB-PK1@DfjX
zW=s!nGUz)4hxwIoue&|Hz7(SU=%9Cb4uO4?7_v-sGOO%<_dqA4F8~XmI?6#jhaBMN
zk$vWT{+16S)Dpaxw^OA<phd=ul2Kc@Clo=dx;eOZpLUA9>$H>1#y<LMS7|95R^|v~
zub>5^V8=Ckq&1H;peg5mK2`bhM2IEF<6V}WE^w$b?wi!4v(<4@5Cg$%11A=oVI;zY
z;3p`61{nYtcs^pOC&P&1d?x~Hj(?4=5Y%w1l^iD3%%?3_cptKZ=5Z%1j1duy(9S>a
zrrmO|gA`IcyN?@CQM+hUh>wj$Oe(*oZ7OyM-Qo``K6aFbF=iKCI9qW72zlAgVVO}x
ziC%}On$_NC`-$)<w^ilUcwPj#?US~5#%n{i$%@J&MU;kFHFW*N$QK33BT!)D2x^rh
z?K%e-BEi4J`eSpW;!S{V<BD>B756I+R%An#Acm>d*w3=6#Gp7cvE{fF*Odj|g+O*?
z5}F9Mmb~ys%pu<s#;SOMy>HM!*E0t-89K#d<ip*Jq|IOwn~I&;6%VIzMr&2C+ngcl
zEqpuG8*goW+}x0!sA{zc=U%U}Z8Uvr->{eet@693@I|xzPtf~O;A@WR9<qGN(cU%O
zI%rbZaEEV2-vZ;c;hADhA^xJQvC>Swl9?Oz7C}8X?t&+{nepQD=0nBrp#z+Hr6$tJ
z2DhO^s0X*7@B+n*dp1zYUo+fK4?GKP8>IF8!b=tZNHA+mCyxcUG9K{GmyTXe<ohob
zdd*KO2!-cwxA%NZ^HzQHimFdPch)-}F$upFlL^Hc&-gi2H;}DyUlJ`tw}k7VW5IjK
zIKX+)>!ZB)K6_h)(IV2#qW8;uojBJr@p;`eJuL*3WUjTa^7Hs@w>_yJAE+5!e_Q?G
z9{-N%?ZQOZygt7Am{^V7E8SbhcSaQA5ULi6eysW-Mn7cJvAcR~%;)V=%GqG3F}(Fq
zBe1gg-GPO|M@fyca&?=^8If9m8~y%P-KAX)dzxGC7`keO-eACT;i5-eBZlb*AWi@R
z!1O-229Nu-V%Osz1<~pG=)fjBO8yTwkz;mU@4GFC$p0AeN0!&l7bE7@EEb6VStZAd
znM-*AMpy~#y$e4o9Fimap8u$l^IKZ^YOfAZZ$`=n!7o`;c5WsY7PBR8SEjHnbqkL|
z?`eXjBKW}3)mgwECcB~n=U1Oi@xSgP?>B3#ZKP^hh<SRxDViEEd0>d|)V^!z{P9+n
z)g+X>fFsxiT{#m`$eMNWgP~i!+W}QlgGJKmCP;Gsu*<qAj8G3g)NR&qE<hogt~I`D
zGc$+`)OpiHKu2;L_KSqSjM;v(xS_kO<o|x{ee8#E3d2y*^twS0v0b9EZ^)CtCavPu
zI?rw!mXgGAFHK9AEoUwLJbH;zuO9m>_-%748z+rIKtvS<5K|F7%(Or~PuZ?38k}z`
zYZ_4VwLqAMO4Qz29<%@2DU5*VDZg9HQa#=k1|_Bp`w<-d5WX{v#cZfv@Yn=ojzlq|
z0;3a<nd2oO!v_>fU*~Gm=nix}ff7AH9;b!QeQr;39vf@&9<PF8PGHp!*v(bUz1aUi
zA2dc5j9Z2eo?KYOEAH#+)?rQx+X&nY*_2{wS}vTlapd2yQG3fnTPSrteEY-Li2}KK
zJaH;!vN`e3vUgA9+42?}FPk@ByX%3Q1~0v5q@;vvTml;~FyZ030E>KTSy-hJc=t`n
z>83YJ(Yu;ytb<tvj<IJx-e$E^SXeg2E^89pZNgMke2t#_bNnwJ>h}%BLUIB56#tH_
z1=Q#mg(V#&I@5Y%5X0g%wF0@-Fn^&6o`i2S1}pf%?AMB4A!CU6^u?4Upg2#j(!RtP
zQ5|@aefy1cc1S3m1uX}d`)tQ*VphYz7`opWoopZ#`3l9Ej!jyEh!}w_^G7N5T<B8O
z3lHN06h7Hz1Px6SYCiCI6q&kt)`|z&FnxE`9ay-f%Q$=w`*~ovEpkFuX@mPXS{_bi
zU%kEGp8sCD-;a)n3xKSjoWNJB`$}i}v=w4kM!aDjoNZ~-qHBdOU+^0m6zz3v<4`!S
zfjS~Y_u+5Xst86#Q%OF(Xq}{s9C|c5aUxG<X)b>gQ-S>5N2xi2uVKJsG-BAfC2&s(
z_Ss~aqoe`uIT$L>3a&-1g4pk*Kn8z$jx5B8YG7z_EQ<3x)x0hMCIBg;K@7d{A2SQ`
zECYcLPQDAaJs_PYHNb{*f&lg{5b@;efEOP|x>p2%D`@zzO7paywri`ONHxYzosg!H
z`6_3gjYpQaYoCE6+Eog(owW%jXRJD|7gfrdM^xOEziDKQ_Sse4k8`w={zrp$cnB+1
zB(~G0O+KEXqP129qR|)<XeiDN-_}Pbw)VZeZu!JCdY>Jh?Pm;qnAlZ1O+Q(uW$*Jh
zSP4LHQ}D%vb4!4{Jj6j$^Ab8Fuq_Qghl|WMh>jVmA^g;U)QwTb^H-vqg$_~KfK<tp
zSW6YyE6?1U=)wk`xsC;Y;f-a*(zAZTGjbsdCwRmUka_V~{B!nlb)W2a7Er8hC>n`X
zTR^OBV2Ncb-FRXPtQH0aWQT#HkyS2pGf?uUdghDSu@|!!1G=W`*Hl&7yzXJzaJmNY
zv}_y&2Lg%nWb@`CflK7MxbMmp(u8{D1)b8DGX;!AQiu@nLsG8khWLEzJ#wZP$+fko
z1-4#kbAU^^M4POMeyR<CCD+mj8%#iyK~l?-mwQOI<p8!=1c@^T5MrAq{-f<)KCnXL
z5z6iB+f@Olz#gJf!nORDUMiNm7I*SdBAp-7upvzIJDJvE`4?38=sm}6s_#v(d0ugZ
ztw{mFUo8u&tqmA-r<<IR_e<KsD$QN^7K~dZR70upvsB>g4TSyHCyH8Pl_uq<t4o#q
zq*S-Q*w~g8Y2F>}zgz9vxXBDCE4^ibm4sJ(B!#B&hJaxqrC_FBap-A81vjFuxWQ=@
z(@D!GxBb4_Rb@SESN{dQvolkbwH{af_wNSDVY-T(UCI~sAS~<}HDyEBoP^HIsHB=L
zf%>mqGhy+7L)ot7h!UfX34azdY8*L@_NBP0AyV=EmNgYfUDmu;pzYCKfqY88J|F|h
z7yvC#pzgUkcf7sCH(`Z6Ic_sHZXi2b)r2vP<2o?_aiuP7<*TX6!=}<Se{j5C^|AG2
zoJLc3W2|~+>aAdRI`sv9in}oSa^2`&N@zAH|CC@gZn^{5Ics4GYoy~@wXcasy{-L_
z`ru&+1b(8^jn0fce#2)e*hOngMJ^4|tIya0@O5GZrD#r^;mbW71KXRsW59VQf~Bj6
zO1Vz)TUU!PH?-<hAQfS8wHJ86kP0ky8C2erPg?v^aLE9{O_Wpn(nO<dAZ6SInkwTh
zUJlW4qjE)%;lmzzgp{vIhxhVYJoEMMvhlklnnoK3Q|9qbra^CJiNCqlGpejuZ<rtd
z)VF?VM0oI6yTfML^BIH*!1k*c?05<$Sm$RG0Ng7~;da!4&F7CIdyc>ae#e{jdKz?6
z@x~?p9MbM#I7p%5&B9K6(JdBhM9-#JL`cP|X3noT6vTj!u4do}{m6>thgi~a{1lr~
z7m53INED$$I^e&V+sC@zXG{m?vzh@KZLaX31%J$GbFEHg{HKv>d+6)VomV8^M;XS-
z%y?<eF0=|r$j+-9m{WxBCyD~K@MRww(k{vv!;@mgPA^IpQ=8=MH<(i-RgdPsEGmdy
z6gw_+t4yB!2vuqXkA|YzR0bE0x``%n)eLSMX*OpA2QG8;0Oj$CW)1IEf5P5B{O`v{
zOi?ljj=Y7oqu^ejUT;tJvk?;_IpK>&NI5i;JAO1GK!!3Db<qe)H}SIl;M0fk-13i_
zE3&%cc3qntUhE1KAPRHbVg<K$#`U^s-5Nv`{)E}&B=i2&g=PN3p9y_G8H#h2(~Bp+
zq#sZWUi?mJPy-HEn+ssAnNk^F$dkYr)l%3Z{&<(ghX6!Q6ECT2KYcm0u}A+)q<|It
zihe;FltI_g=jS7}RZy_3Yz<rIIsF+(FKLCJdjGHgzlwm^Siq}5dR*jj&JSzxV|};)
z?sY=)vll(!Q=E)6$Z-8KC4ZmOC|?dRC5H7ChGE_ZFib*`Yf!}EeV3_^ZuSf2(yI@D
zf!`(GJ*3eGB86U3>71XE4786I2mpc-%ZniaU>v!`B}U<Yn9LC0Xu$crfRuU}OM?^e
zGmXnhrC|b;s{^@;=v6!-gvdh4^k2vOZ?8eU_?tv9EhLB|-k#;<Jc2rbK1D^mKaq?B
z#y(stb9rLR-^YF*Xb3IztL9b9@`xFj@GoD?*xA{;UmqN&JQ)GI_<XPwdMbqcp4~$;
zI)-0SC`VQLuhVcfAL7Vm_#xV$6l;!gZVf=dO(XxlN#q}5hkw{_QbYY7it}ify2KyR
zqnZfj&A?d8jRk*Is7@?%+1r0<^=F5K9@5yLI0pwPnOx1qE9RJ3L~TGMXSgD&4l1BH
z<t^M^h-?{ul+KWA7>_|k&R3G8_!tqO4@^P2oop+0^(LUH7x4J<KmM8_@M2BbyNasH
z=ve8bajyD7!1SvhlXRYv1@HV&s?5XJgx~32e&O!|4ziI-2hNHSVg*BV8w#Mm77vJ9
z0a3?Xk$jcM0S?d3WXSzvU)mni0Pd2bqhlhs*o`<^S9r|-$2;BsI~(Y2a-jP|{oWLa
zd}>Q&O*Ri6qWxVuFNb`JZlY}9B8JIUdO23{4D`n!U6w(~*U`W%fCm}ME<gF3vMs*(
z7cE;RD|&TBQFe+0ihf<uiA_=*)2b5r&ntnT04m}G593`)J?2B;mxv&)4^{w8@Ifo~
z(;xdDqYr4<a$=tH?^b`-HrDN_0So(LuVC|!zn#U8LDa^+e<|(hF~|<e1|<_?CXj=5
z?(F2Q0xVz>rcnChbN)DZh6n(RFbeQOFX4$%9yqjV-<<_a0rdP^u2RsTn#Ml3jCY;&
zPnZu}chUd`;WGMGqxKgLrdy%|4j$l{j;g9>IL4zmXN0ng{c-*tQz5W{Jv8F9txYAj
z17aUu4VY01=u8y{ShkqbghTNOF$Z`n?Pmb;w^h98o%jKhA_|Ig3K47yf}g+p>o$rs
zh^A;bH=pf8LTzkpdcX3I``jx);m=!;9B_LSsX>EB*~Ro8u1j3<AP^%Os9~_<ucP~;
zEtm?t7Ay$lGA~Y&n8&-&nMKT0T&fP@zMR!6!=t*@Kp3hiI+ZY)-?D#)eY@kgI&00w
zb$8mT`(%C{@<Fplv-Ub$I(0S_2_+D{5NpMt+0s~A*L$uvHAE7NCxVt%i%E?rE^cdr
z{y(_wWEbLu!h;iQ^v|mCQg6fizeoxFn>F}%3}InKd_x47VsYX}mm)-?1b~OU>I(+h
z8&3_G3+^#Q{!wzAP%aiNzBSM_egwdqvo4iiH>sllEGtAi(j~nf5Q=LFnZM)tzpR>g
zD8WzT^w3hH{gMM*1cX}H?Zqqf(3%LhN}UkqOfH6iKlg$Wp%!mL=D17yvh|CcDnvg;
zimQ|ZOrNFv@Q+f*@)WTQ$~E?0t@)SQ(!Bf^{ZTLhF{qva`jaA-AR~P_KYueQm)`xW
z%l5-!#erkc_7n~70{D~6N2O%Z7wBq=%g(-#`ky4%g%Jo~U5eb|FYi+q1l-5!se#KA
zQ(_?vASirWvwZ~-e_d42B7Q$P>|+6;G{Y~+&AEuPl#8e26P{qg<q<#k{n0^5M@Fy6
zg94fnwfOPY%ob)S|4f;Sn!C}642=b4#Tnh%&mA`3v$Y%_U+`O%2`qoO6xhx#Y&R=j
zKJ&y$^WphXAy6xx6gvMx$fEv`%vmHjAGmFOR3gg)H{cM3NHYJqN80$aJmwo=4h2Pp
z&seK{>T8xvKrVh4hpq_xKRa~gzhjF2@ranjiDV9NC2AG5ArSQcP&dUoWL4fjiW?qC
z4%12r|4IcxHGptYl8zBP)&QIj#Zb0Ci-*&z9~P>+sQCRbA|zINW%&vr>F{J|06+Z`
z1HdC>h{&$^T`2T__N}2nS!bYY3LuuB16n@3_HX)s87FWV>u~~yyQpNah5k`z7axmr
zwK*c^rG;MN!QzFvOa{M*rv8Ws@%UUW^OI%2a<Ts9#TdHN2Sj1*x;m!c82#qL*yvT4
ze{|lMN@EpGVW$PmQKF*hdzRnx^e?x#xClJclNJ1zK3v!uAW7L)m+!q;i&p|SG&P-6
z`Kzj3tsfAK6km8+#e5|WzdU&HQM?#Lh-D3_WHEp&VHU4q{@6YqJJ7yydqLZk79d`L
z4-g^w8ii8HfOIp73yiMj=dY$$91UF&GzH0=v|wlFJ$qkUEBK1J!?UBCMNW=+htoqN
zhm<fi;-j+~*-nb>hn99H-XAJ`YwqJ~D-B@x+>sfF&Od*#th_9dQot^KY=cRN<36Cn
zm{yHF54VIU`Sj#wh2pJL#~^Had6=We#;joNr_U7rh>LaX2sn<vPF$KN_+MuyznEM&
zY^Uh5E={@bMFw>O1{u>|{c*Z|FWk#;fTFyV*EUHSRS;$FlRSXi=uo5lC{qn^1$!*v
zM)GMKwTyqZnIYn57Ym2Q)wU|36XL?32P9R_Dq^YnxGn!7@;~tY5tQ4=pY8q#%FlS)
z|4t$_pDLp8(+3d)M`6-@7!*ms+8Qj>i2n}}kVN`zOgQ+~7lSi^42Zk?_k2Xyj=+yy
z`^;Ge;}^U3>kVq>|DM{2qB2)A2xbIwIXB)63MxQWEo$+&Q!v&39Qn~mAv(~e(hHo*
zT}llN^M&slP~`FbF$rucD^*SAkLcO?DE&4GY;16t@Ap?aA6~esbtzA(JT}hczJ1fd
z09*%10Go>BZ<PHPb-mDJ&`ltAOZiJ5faq;WLWFrl@_-&F$13h(0RLX;t0n*WVll-R
zKnTir1ONs46B#-s0O;bdt2<!?+zt;c#iJ{Qz6u6@y@1M*z${R%5FwD^BHrU_!+I*I
zo0v!h)9ug`{38<}u8dw`4w{;rv@$g{T?<PB4qr5T-GTn|@J0Gp&apoaUv$QL``@{k
z65jxag8qK`qQ$sgoU2<ve0BD=?g!7j#5_M4PzrO5%$1w?cQ@#@I1@@0V5tkjNd6Zq
zyFMm-ut+=pN&x>hj!c4dM2LoD?Xa5q!_(0D--saIdy(HY9C!M{fa2U+AHEVi94}fZ
zg*+-Ism6lwuf*S93{5{RqCGpf95^@w<%zpUQ{jJ&^!vN&qZ~XShDF>^=V1l=-;mwW
zB!T5TEn$oQqog_oh?^6^42?fEe~+c&X0iP*k$&xN$v;TSSS6mdAN#7oI+UWl0L>x`
zzzYv=j$>+ftcFIslvZ!OlT6WnA_9>yOtg@MZ%B1q$va&P_l35Hn_%S9W5EduRf+%X
zRSP}7(00mCG6TQ037QR6r$=X0R2fDhfmO7t|0l!Ti~OFgtEjjvQ<z?H?CKW3gK`s)
zZbT*1o2yU$s;V?i$cqvpO(OAcFXva>8R)$E@=#HU3}~vTKtJ%$gkv-S5e2kYz(Mrt
zA|QAnoMd${eHs|2;`RdJB0{-VS9XOVe;o=a0RLGZBBZu<tgCBWKaA(XD`)+md*$5#
z{Yr%iM`Z;sluU8;-p7P<6<;&u{_Q{gUZWqvi$Oz!1xHr(d-$tgAe#&K%tm#LlJh+g
zjFK_;@Gn33{Z`Q(aEZ7Ygw#SWW2%oo^oD7IBZ31QtCMpSTXbP6LuQYeeFWJUe}nno
z<{<l$ChVj8|380z#7mS@_kOZQ?XSbUG$X+EuZ=*#$-1tRO=Uds6sg~DLRyjMf?yn7
zvnreNI3SgKZ2FlG8K4>YCgTe(#q-iJxsp1DP=MEwevkd}lFG<Y_(i2O8G-}&)I<$H
zoD}eNr3tn_b3*}y`F;SK!U1kY_;%&|h)-X{#CPLz5@P^frbJ$i?vHEul?4HaU*lfS
zUzs(hAfQ(%85;9V0B_yKuG#+Qh!*`~G8nPo=-EV)SN;MuAYHv^?F3IAV8RKCG>K&|
zKlvMw3yA|^P9hNJs7a!E=j8vEP~h}M!0}%g;q(SdjH>g^h<_yS{t5s;?pH<tRPv~c
z)UjM@N{smhf{{IAWxP1od^or3a&mv){u2FbFIbGbHpf>0EpBt62t7SL>gwtVw}$tG
zlL`$t&LT#g%AC+u0fvo=);iru=G}k7R1<o<YDD1)%?kZu;J~GB^1mC2fAA0g!QR-3
zse*b!3{hecg+og3L<7wA+r9V7zir)Lgh^cLvd-m`*tade?GJx9691(n)9}DTA>eto
z0f{Vn6`|$l9tv1+9`^5ptnXz0JdboSkr#ICpBz7-8oU`G)_mid@L3SZimASM&-3?n
z_;n?Q@HfLK-!vNid%e@?f5csg5c-htyEgxGiV?`tqdVO)15BGyA>YS8Csd9Y0dW1k
zt!i=A*RcziG)1AhQA<UQ>txkt3Q^Vdk2=Kz_>4tY(he}b0NgUH(RIs`z$DcT39d3k
zS6{h&VMHU2;#}fTgI5A1t#8e$y}D{20T605->=|{JarLA1@on}UyAb87pLNV^PnT-
zu&K&pHJe{OWDx+=2`j4Jl@~}Qxvt2(`ZL+51N1d2yl{~C67vB6UG>tZedBGs>caoD
zb^4U7<eSMiy!>Lgp{T^T__sJ^#y@)V2>6|kF7i25rO%pl{;x~Zgl`P^)Z<{nr<4s&
zUdAmp0(paSEp~T<`TK%QX|vs#Z+s$7da7*C5jOG#WEtQ6$$6*?6Q9I{UzBiN4fHY7
z&#a}qndui5w(;qqi&Uyunj34;cf9mc&!I-Mq)MmQBAs&V7l|^7FJLA5ktwcU#aD;R
zMa15){zzT_gWUkxJh(8+D#L8~zsvxG-%rkvIdk2JRDuLnO5baVs;X&gJ83>DR^RD;
z+(}T5X#w6U$qd#NFb(lKZ~3S`12NazQ8l_+<iEVgxu{n8nTl094phUq8g*UrFMiUk
zf%u{CTQ=b2B82yfM99ycH)K68ncD0x*w5-9@d&e2uwwypr;4-{HvzrP%zupW2W|(0
z;TLz0QAiegMPpuUE*vZvs*eXEM7IY|V*<2@V7S1h*?a|%?sO}cA8@8HpaG9XSqT4G
zWmxzS-*AAFl9C{vN6i-}G;84gFQO2&S;41wDrSPjpQzQsOD6E7_wVCs8rUYHIG3?Y
z5B@VEi{le?0z+;wdi*<wNk<E?$1hb?H9}ai;Kk{w9;rG=9#TWzC({qY;T^zs>_lkb
zp^GUfMOf$b5B9F^*1yvEA9}<ke|fkdzl*5aAU~Y@B04VURmOySIs5B`kZQC`YyI=2
zKNk3z5F^x+CAg%BWP$;B$x%sk;|?W)xk-^*`ej%eXc%ByX&P8!!!@_q?f>~+080;W
zc~-;k4Sq>x%wymJbD7pF+BhQE$Nb=Y<!mC@32@pptxgE{E&us(ELc~dt7=%-pMVVS
zqR2QVB<f0(=Px30kEvd!i3S&+WQP~h<A9kS`TQBiw86_>ZuXs&ardi|4KXqCNS2hA
z{f#9lj3{~(XFqmk|7V<BZPf0=-b9*6Kn9u2_gmZ_95QhRpwm(<*^jT(%Z=`$QBfio
zHe)_?8sbu7JyzHLHVIeUIz>=G&3RMEX-0jxpv%H?CB1Dulp6B-&5-o*(}WVhdmebN
zks*K{P>07IjJ3t;K&yANl7;nZG`~$70R1#>0O-o=r98g$(pXF{d`a11L=Od(SnG*K
zVoZ1xd)iTwUW5D7$UJB<ODI2J!dwep7ICBvAqk8ccr4t5%*|;4hx^+r)!`k?xE+}c
zjlFepV3f}7O?Y>^sn2C%ErR3tN3rdd_xbAuP~Ip&(*J{|lr^5b?QPWcJ6C!8!t5r#
zN7-NhFjeK)vCj`Y%7vFAs?F;ydtUt(d*zLi_n0KGN-vSS^C|D*vcFg50&m^aN||y@
z0^{k2ox9-I$Up7pHdJ}^-J9_7<-Mcr_7t}O{?rWEV^FBkw@qYS!8lqaME5-W$Nq9s
z1uCOaeV(xuk0D`kHrFTQEo*{y(5gG|(uR{+w`(0%-En!vbuI1^_mVrMsS*{cI86;w
zJrK-27W6D9t7m9%<oP|^G!EW_$vto6ByT^6(t(-~(%40o!KIiyphi+d_pDT888<zN
zM+i2C3i8xiGE*F=LzHf4+(i>_H0<+^j_i(_NL^xed(X}qO$vF^M9iX|hx^n!=rlhM
zwL-tleu;Y(v>%QOtDAntt$8nl%-*OtS3WWCHQ}>|%Sw@rerq&)o@`$RA3Og|O+ZF;
zitQUdfO5X;Vz`V5h&VxINN6~BgqM(kI#2uaQua6Jp6Xau7F^18I0r+YCi1s<&xI1o
zu_4Z5#X%{oIoF8z<3-s;O)K=Jp_y)(Dj7)XlP(#V^(|}}ZCu>k3MpY)A3Qr85nnj4
zW5Ih7ID(C$)gWdbS)5O+TSp2fY-Fq8ReX?r&7m_h1q4m(oM$t7OZ@d!T|qd7rCY7X
z<A(loK~zmsiGvPOBz^~LI4qsdL8n<#UL-I8o|ZzS+P~JNcv#H=brBA0z*G;}Ok77P
zvkarRV7K->pH|nOnP7_gYD*Y2GM?IirWKd)G7vKqhQu59V5^RdLr<%6SRFWRy2RQ7
zjE${ml-E6#S2p=bBFfO~y%p3~i$XX%no=l#WI;mGq7iZ9n*)!9-{5h3ytdWwF=Z9m
zi~W@^s1t(~VpG)o6ErO6xYYd<sJMl|V~;JZRhuj}vaHkc?E9kcEDj>QXhP_|Q;#Y*
zVm1EU^6o~NWD<q--9_VfWO>ErxT|POXPw7-{CEx#lv=jIIK`k$>K92;D7<B54lFsq
zLq+_CwXGKtbgK~eDonIv*Zg@@jmlG^3)h|-59mF@`DVEgykF|wz@D+$Vd5(?qT-EC
zYRhJo7x@gS`d%pvhb+hg;l;f&Tkd;7_3pFkSGsiR;sKgJPY3PmO}}8`1d2T;Wba-)
z)Zp~q|3!Y;o2;hk`4Cf*u!~LjdgwALHn|BQlN`9KtBd62_a;lh&=er0T^ddSJU794
z%MXw$zozbG`-`-8H~asfSwOg~s3R$btDYJgXN{8W`wj-JK0PQd-jDJc;Dv7!N38@7
zV)Cgy51$=mQ@=0sfN_DBMSLT<yvvMx!m7{u8F*=xBIfQmCXu%_Vv@oNJ5*1sa73O;
z3@ZPc2Np#&OiD5is<HNl?fbe7wwwf6k-&yUGa{_YPL_*G^GL11O~sRjs5y1av107Z
zDo+%IT}-Hj(MTrEDFr7{?`I_AN{*qj6mR%nFV~-Vc2GQSxXS+u+fN^r$^%m|8LnZv
zjPt(KG3tg{2enY=2kiVW!}4gnfKh1@g?NUGZqBoL7>ucFKcoV?+YEex=-iC0lM%)^
zq86?1<q|QgctZVj!^p`(wm8WU!>6yla-X9{1mbKiL&7pJ#PrpA=zwF`9`l9M3e-Y)
zsUPIrTY=A?>r=TG<OeC=d&r+`unFz9Tsun~fqBK4AMA?;ks)VTyIv#uZXR7+-5T_H
z3?*hsQLYQd%;-Gf-n|u~+Z^@blz)u^ciefhBQgDy@~*RQiimwk<bxCqqZV=VRS~Iy
z-P?ah8yj&NJ|{-f`Kk2xY7g-{g+X2?`(uO(B8FmTThXHa_gD1Zf|FO>mQ&0Vb)Cr)
zOp88xq{=8FKTJOa;+(aD2gN>9pJ4&Fd%%7HklAB*O+$?%Bc~EwfGMbNeCHErUn#mj
zj(YN~iF5+X(s`V#P`05O>21#OdrXSq>gp7c;wCTnqCI~_Bik#ZE)~1`V@D0aB&Iq@
zCq8{iChXPxuxf)0QI{T^33M#0x^A1hG;<DeW;G67c#z{bd&QU$gIpa*Q`Zec?%Dcl
z?8~<IYC6kna<89cc0xbrcLWjdzB4`^6ON{|&WofDIU#bf44^Y12r|{LuunW)Xdia&
z`~IQ|J)zynl!~!GM^HB`+TU;gEg?>+WAx20wJ)N!C*<s(y{kVwFF&C4RU;ldL8EPW
zfJqw{z&5B<mrs)p1K)+$0cB#v1(5Y(vC%XA(Xy2~ej*QDb|c9!-}K#SiIR5=*#r_t
zJ~Q=qxeuo59rS5#L2oiv_aqDM^wghBH_RkE&RU-yE>>=@CQNCnC{prnb-N7_Gn|8?
zI2TtC-dL1^N~-))7WxOVyr&b4j%He{dA+<c;4DjC18;b;-LBU!PqfOOHVz4O(r|Q!
z_y(Vp(zNw{yzZdY8}^0?+;=#t`ZOGc#84BRF#48xfS+E`{L))*YX+>Bm_}Pqv4tOP
zR?QplV<Vy>`1{_6(iq_Jpx;2g{_}0RH~#F74MV2<hq<SpLw<aUj)L%NIF{+jE8E7S
zM_XKl5SI<T+W9bcleQ3isV6cciafVWft!J5(QhD5_$k@j)7m7Th>PD8e>`VKK%JPr
ztu@4Vrb+a`VAT3^B0a+a{Y+uPyQDB(sy34X7I*S@oM(33mb;xJCkYOvx8@v{efb?$
zSt;={Sp^Wj(t!_#acHCJNRwmOTzoyf36Q_JtYYeNK+FaNn)Vvaa^0+m2_T)0ySrl@
z-TB4u2HEbhW!uePx9znX8oCcZM?*fAi>YWlhPQ+mh<lDn2ydp}sn^Y%#CCA9Q|E0_
zEeMz|7NY&$!^INo;|D)?Hh!t<|M+10EEi+HPjszx!H?Ja{^xryoJq-$ZDA}fHuHF&
z#1rr9wQ8RSK3`p$G8LrTL_cu-2FXx7F5Gv8gJErD$xJjy)Hep@ZEn1Nw;dd8gw;Km
zih|C0E+BGb?$`FnYVU2SB1(V%ee(23$BEfeX?-CUFKy;76=V6B^=}xJ=~?gTrD58`
zlHcbPzpmFQimzZFkSPepRj-$qw2ep$CRAeOQHz45{Gob1=m<po;>MeAZo~x<AwL%5
zsqw-hd*S(>u38O@rek)C{+|~3V!Z{3kPofMo;w&&V+Q0F-=BRg>8oTr<G#xI=B{lj
zbAZ4M>5N?~HZSntbYl|IA$vedXm}8VpjERfs;V@@!!(wgb$~_TSv-FtVy1B(lN`!V
z+s0W=f5_$2$t`hjnfr_F%fv4m`ML-qma;K&^mq<J#D;0C9x;y(6W18ac+fKHU;2|{
z1U9TxhKI)l_kQJ|pPQqQcUicWiOv-hxqlN@N~4AbO@dIi;3#O85iaswr$vx{A+5yA
zBr5NQJLKreE@HjB7#~W5R-XCwj%<hXOP_<!g<l-TLdhJyNn5+2PQJNq5pnd&XG({~
zwhu;=R>Yc5`|Mpn_MM201bpsOL)jgc;zesZLGaQ}5hMo)wB0Aj9?UwKR-4prL=a{Z
zZ;K0;`$g7<bog~OZ13JJF!{c%Lux3u$zm03d~K5Zc@hiRrp;2$b^}^WaGgZ1H5X!r
zCGG2{Xwa0=cYzTU15fK(2@}GxZ?<H@r*%|mH$IIqERb3y%B)lw;Y>Y6C-!1%sp$y5
z+e(f&+Kvrf47X;agMcEdQmPbv%`xFVd#_T_Z%y4ytphDb7AW`3Soa_GC$cYMx#|_z
zA&yh|MeN_!YU{Lv#YP;pskDdiY~KE2x6jWZlkTPNp*ZW(Fl0K}$;GB$t*$ntwJ^Q#
zdD1T*Mc8FSY5M5BCA^}6e#CLa%j2MMbn+yb-4OkJijPtLRVG<x8mR@=1QV!86x<*K
z#u>GT?wqz|KPC%9p~~C+6jR(ukrrT2_tN_ASV5!Uzd**nA&590@?nKVABZwLtk&uW
zun-{vVb}I&6@`xpoZ6c!{d);){QT<pFj4ni9u~46xRZ8)9N68>{%9tn#S&vpGfQJ5
zg->rw%UI0SW*k4<U=e#R0#3o@*+J=>)@12>qR%W=JhAjXWMe)T7LJl3c#UDv3w>@1
zgF%fDEWN(*T&<YA&5mnZ*(cO<jQXrhlHV6^7oF@*Oum@jn3Ov-DSqvc(EBLS59;mS
zr4@iTgC6~y)PC&xvhiHT(+JC3={G0Z9(D38R!|LP<v9797aCPfx=W{0UTZ#WAEew~
zfj=p;c8QOJ-iuD3?B~bWZhO5xm}8oJGu;cvglpUPd?%QXH&*Xk7blT?)>j#<n6fM(
zzC`<0VgHb*?Utarg2|(I`e~(lByg5^x-z(y-D$><aAzfYITezqe4xp3RvKr*ITxt9
zTv8ViZV21&KB<mJSbJ-XAXWa>UZU1(b9Y)tlgqYujA4VjM@^vFc7mRnzT${F*z6<E
zB7@muJLZ<}6X)OUi|T9aAdeYEQ*<i{e3e{gPdr@T;08!p-;P;Xr&UJ5A)tup9<Nt^
z<7<EHZkk$C9s&WLN-NHR>bf@WJxE|uMQHhzwh6`L*guUkzC=)d9aS(`+wX?x+96SG
zz5=R)H#2Re1U9Sux)UzRS>Z?4mp)H=6Ve_R?qqAv-@@e<W_|DhyZ2R9=de|6cKDrf
z&W>f%*^{yFMaq$VIqUHruRq^9I8yMEQS=?s-dMdgv={QW@q>d%U!RVq$!ig3Z(OqJ
zH~1({&vh!$C-U>}(fAx1fC{IzI3PSPv9_KrdF~S&*avr-y&|VzVtyel>AK}S$aagM
zI)Dj%=;4z`zpUB6iROi^!y`xZoV-H?%P=&P`9E~MWmufuwxylm1h?Q`K!Qtf3r^wg
z?yf-!4G9{8ThIy#5`w$CySqzKc;PNx?>=XL{dJ$+|LcN3YdvesG444hIsELE`Iu(%
zyQbQL19gUmhlv0IajON&<c;?S#SA<=(;Zg-I>7;SYD@^i#B1pRmBAkd+X81bjEs;Z
zPn?CN9R58E;FWJCBmP?|_9%3q0~qP+FiR}QGaka*`8Z@Qq`GQ##v&uNN@DYF-Jxb0
z{FI0?$G@j(twyW7GAH3aF0uhu9gUjU_@~Pd!SmA_M{6e0pI0JM4IAz8e~@4QcuTy`
z|A{|3sUrv_7yN>#YO8GK?-z2I_DLai7{z)h>XhB4;jo|yOO<wBg($MJ_w-L-;2cUi
zO5DB$%VDIOI<@hNvFxQCQF4dNn<^)Ik+hfV**AAVqLG1#n>{E1V;zs0@8mNoW3bsb
zu^XFw7!KILy<^s8dRMx_ThHoxUo^Mg=MakQ`{viD0kV+}2Uv7~hk}x!6#sTTcHsou
zCHg;){m-S4AzFQhEJE|6#=Rx;TU)UbNGH!@AZ`Y@)6!4*j7m3H4=Iom6?W+_N6Fz%
za$_dxuN!hnd38cpge04>U`ub=ls`a;M!C*|82uLE9o>ElCDrt}<;c3x8<-!)T{?p6
zp1&M5N7{`u+a%4;QmH*fYl%A(&{?6^`DTlAA`Ct`|1JSD?iN+{Q!?JFA^L?=(yLmj
zV@X*n20C<+{BvJo#kR2ci0DcYzLHj>h(_WQ6TQvh@ILR@b<c70wbn0T4Aw+HIId%<
zvA=SyRd(FgD*6qk8&Bg(n5eG=@m}z>F_s(Iln;NM_pdDJdDdY4pq@Zv2Y&kcv5Rcm
zoBge=(Q0AJC&!d`l|R5nmz`T#QvdHGi1}Jflng<bL?Pxy=mHzI93gY>OWonKqMR@<
zLPz&Y0<&<N#FN$$waTp}7}~D3UOVl5{nfAJ2@OMvunCtQkIJEckipLIPqgnjwKKYQ
zBvyZTSI|Uf8M~jPvu>L057(IY%Vmz@HDZU}i&fU#L65#gCDI>&X#Fj=<2fFqyz}w_
z0WGzE7Wx0B`lyf&q|KlqM#Tb4Xh#mX3(dwzeC;${-E;sHb5`sv#YjQ#D`J+x{OymJ
zMVYtm%EqP@-d!Dtu!LqQ3wog}6YvTTKHBTVZ1J+81aTz;2H|f;rd_Pm*jsx;V*BUK
zw8#0L3ak}7-*|#E_EG}XhVUfsM@{(##S6-O@HAMlHMsL-YFfP1aRkwGI-g}jEwv4_
zjafn)9PsAvt$PNmi7%6}2f3IDYl97qL@0qrivu`M9mK;{DJg%D+ppuy(><QvXMg-M
zDL&u+q1SdlCH-)xD4jgNq9pv4l5`Mk8Ko{IGrmxe$|}QD?`Osn<m$LL#bfaS!+I3k
z_x1pU9yjAJak9=`^A$jMQ|*`=<?26hT*mep_3F|I6o@xq65d>3mK9DPa96E+nWA&5
zQ-D;1>g@WJ4a~Mki$f2tk4UzaVb%EIZS5-fI@FKG74(*&@nsS&*iV*u%i@m#Wad`3
zxDP%!Ff^Py9efmW#2A9Bi8*bW*v8LWa^KTZ-z|W!->Obj*CR`6qQcvJGlJR~Rda5=
zYppCxX3qidR3vWq$scM{4ONccxdS?DJ*W-C7?-kYbJZBTz6N)#l=9|@3>D_}4jz%;
z1L>RuHA0cDtMxF?yErnf%MDI{+Wz)Gp@@w&`}~{JXT70@AMI%7ClX(w7_rv`VleUB
zF;>ajYyHD>>$B_CDs1U%7gp&Fkt+5`wdznDH=3&Q&rE@Fm%gmA^6}f6mJA`ZeP{16
zK8Rz>@yB5Q&)VT%(y|I~L-@6n8GjoWAjCjJgDxyA%mn6cSnI%$kdQEBhF>BhBNHjs
zq*&*?%!mmc@god(w^{7zk(Q=cPTbqG^AbxI2N;u*8DOlB!+!~MaQN#NVWOi`;FL!U
z2b{~6m;{FoSDSs(E-ybB&%#oL7W=ohEVRYN#Z41(XMGD*6VsmL<>ljRYCfonFUT5)
z-tIsk@5Mg0lQXfi8P;hXL>k-M7ZDN?zEiTt5RC@qc=z=70%c{xms}?i5{M*F9taLL
zZ`DzkceuiO#Flp!KDe$KT`9*EiW^^F?C|{2OzP|5WMq`x{~h!O+IzCvOjJfe9!mtp
z09s#%1?~#C2idV69qYwKQ9ejf63%UsjDp2Y3sk{n*H?Egt4)K8w%}5DA>8~@(W>^~
zQ0n?m85Z&zHADz+K`tAWFDOJC)*T(ZQgSErTSuVNvO>SwZP5^70J5S!w>g!JL6ei$
z36Kcq#gN+D{uq>ay(HIHdDh|0rvmU>&#E;!0k@%B*(h6mktpIi9|&vB6P6TAD*41r
zE>X3H9tY{ojJA%u?(3m=8oyh$CX@`W|M=|mxb}wuP^W;q`$Mq8v;t(-|3xp*!Ah>N
zLe~-J#!xZj7zrH?jHH|o)EDn;_oCI8qMOzPy4P#n^)E8XEACw})z~sg{uY0{)3|e@
z?7C=njzxS+%)@d1o|1u}r$y6hkMOZG3(MoiD!z}57}u+iZXKpMaFaDl@p>w7<JU1g
zso>`PD7X2T&El7=KC;NSXp_$i17jZVMN*!^T<3&J48vh<u3K-?8(-&DB4;wm`lPZv
z*ec{PjMUZUtU<GU$#SOSM%VNR#3*QtXY#{VKsLHN9I_1{V}XIU?(|lVGrS%z-<7GH
zeC5c(WH3~}i>iZCP`en3O(3d|4YF(cm=@q2CS*ikg=QoWS}x4LlVAm%kK)4N(){{9
zpz&-tMPme+GV4JLef;n}|8<gyb0Px%=HFx4^9avf-jQ^i4QEr#IFDdzOYiYPy*uW9
zPoti*a?j?WirDvhIGsFWVGuz+j5@+WqtLgepWU7ex!U<(Jt(lAp5p(B?GbQuS58Z|
zYkW0{gJR79WTq1a$jJObCNP1FhWy{9`M<@?3figt>`sSEz;uZ8=Fws`<N?-xgPR{v
zA2`fb|86S*z!J3$8BQDit6khOpmZIML?5-DoJ=UVpO=)Z%Q~pYWR)-bwTHV~$}nHE
zP=(uOgaqQ=>U(CPUu_dpVpy*NB5^#=0^IZxo3egI#i`S|TC&wo{44Bz6pxJYcHsW%
zkZ4iC)w-A-<3Xt_Bb39Fx!!f%WU0>CkV#TZSlCeP@l*;2UGb~m4SZIliEnO|^~3dX
z7k5nrQpL4hp=iz5mq>dXn}SQ&x~UL+mS?1;W51pFm77xJa9<2LQDgB#-WgQy)5(f{
ztoXC<{!(o*U1F&p!f}%P^X0Uvb$2+{Uy@^heZX$IE%TR*AM6G-@_Uq{L2m4ofvY%6
zAMu-2!D8HI9j|0;Q0Y^3%T3sLY{#^E#N2<nVG805w(dQji9Gc6V%%rBq?z|N;&%R*
z$H9qI=Z~!w@`wGs`>Hg^laUl}gg(YPw)@TBKLhVdm=W<J{N$40MarVx6!}ftcv_>R
z)#3YBlKUr`RXf+3!t!6vF}(|-^O-wmf@;i)MV>6{cJ??&Cit;8OF0fRX+;ft9{>bE
zfwUuce=|dLC*Mf$B7^?(*Ou8_-*yk7G+LuGrs%q1Q7~Kol=#hG8Ij6Q5&h@nY1as^
zDYxq_F?pXT&;8<_9Eg%`{D!>3`k|VZ4+`sZ=o=rhIUZ>?o6&YwOg9)>C^koL8bEFA
z3CfTciS<{wzw!2pA8HCkPvU(?#yHYYH}fA?{|4y#dc8#&44^v4ZvO;?#Y-6yWO5L@
zCN+y0zaH0#8I>T#Du)tRxY1H=Q^_xY`*GQy?Q!$ETA*3NBoB`9nU6TXn-wwvJ4iTr
z9Ug3$c{@Ieb;D^t_#bdRauC*?ft?6n$;U+`3@pjz=WmKVcPgo?thHUzF<3|H0q~a)
z+lk2s3l{1{%11XdRab5t#AuR}%2xi^q-<V-g8cME-WB<RemC}Y)i7}Xk!?&HF#_2z
z+KQ;<`Q~r$d+ABh?wPW~1rMFg?=(i(!;LjE8?O&%*xyfa30}unE*dxL>OfKkaiZ(G
zPBhXn;_mU+<Ql$&QX^#)AQ#?mRtNY&2}vi5!fNdCSiZ-0T#<yl71=wmas2-~VE?{m
zm<S<w6R<bD5(EvOw%djZH$NirJ^Xnf|HBl~Z1l4Z`=++r`C)THE^fZO?Ovhn>Dmwy
zrIdYX*7|h4oTAGa<(yc`O6(LM|1OR~P(wxZN~_;I8>o_)6$C%IB5u>+nks@aD&kL{
z?shX>34sb8GfQH_6PkB_i^~$>R;+r$kGVsw0?|kyVKwz7pf2s~gTo?ig0e;Oagr|<
zVfgYqN&H0MW*BjcNX5A5jgjEpPSUJJh@j{0yEywfEXGUUwv8_sTE@qCV`{5LYNvNl
zu<%s3Sdu_N{dt9Xw+Pj6n^XOg6znj~#19DSPf{vi`aq14<+BE=@`S3~oey)&c6~N`
z+7Ot~3>t{1ohsHX|GYM7q$Dwwm2;laQMW|s1U7-jiQWInAnzP9C`av<OmjYh)MR5-
zIv?o?)Rxm$OgRhG7FDKAKiNkf=UPl6nR26+@&^UKDTN~XnO(Fw2TGjaAQwU&XT`Ii
zXV#jHg^~)D8N>H;CO7)4PDFw+MDn*9C|0#?td<rQ42(i^%`lY*--cmIMkmY=N(%XB
z75&#=92X8RDz|d`9>@r!gBn1eFJ*t2970(dWZjf#(rvmdK@{?5*)peu-Y+C%tMu9~
zloKi{xoHJBk_pGRn6%g$@R7G4K=)Y$gy|<eD8}K9F=<`Jo;}s#^aU-#D_Bs+=y!KS
zIjzTdUlza<JWr+eju4n8-xtoZ#`jsvm5oXbF0Wh&j3%Mw-p3b5)SARCA-IR*@G{u=
zmO{fIB(%-UOt`Tq*w7whYCdpDQG$fi*Lh49^bA$5%J0Z_B|{d@?-`I9AP{ZQRFF40
zH@K^Emf?PkK_2iQJKb}c!|A!I5=K$<ldLHa@^UX{Lq9sJfO23;VNa>gMkoh<-e3JK
zG`P4rHQU-JD6QP_b^AH=WhMq~!$Vz)@Dx3iM{YOHls(PohZH&O(d-kscUG9%G*0($
zj>&n*fxp*KU<cQa8oTI~+kFM4y$pQ%YJAA=Occx4vUZ`qG2!8>^n93K$HxKU7$UnO
zJ3fdZQ&MA=l-$oB<sn8wx7m^!>rrX`oJe-yH)yY#`H549GqRhl<xHZKOzIWe7PY=}
zp`^!*e~lQ~HJj8FE-5ma6qS)Rmyl~8vmau~eB!nE&v(vC?sJ|oe6higzo+tV#n6}0
zot{oE>yOe_Q64JmMuXOZj7%XlCaqWV@SWfXsX5i~yZ4ta0T{0#-s+_)u!(fH89|*q
zU(BcilumTpZANT^?TVi`nj4!?^ExDFx%ZLQcF8_eW#{dc=vVJN-K@_$Br|<zZ6wvn
zT$^`VU&6-FAeKs9_uY;eGEud@!3LV(_f8imYf|k1a%-`d{yh9Tti*Es<C40`)f#gb
zx6?w9Yx`lDp#T8(0*?zqDIiaxJ|{+yS>+?UuWc%;WSWk`HD7lwM)}uEM7pEU2XBE%
zj3BpM>{X`HZZW<V6&wdYrlcS9dV0%MC)FaAFFQ%9S_|H(wq>4vg*>bR#BWcZrN8T?
zzPw{CbgnqCGzA?Gv6Lroc1Juofvp>UeCZMKvI{gQ)CQmM)EhL6gNA#Fd)C-nRE!Od
z`k%)hSPF*^OQ$`W0s@_m`4-~5GOsVLx9^z?uMSGvy|1pq7nYnWs2$0VpvV1J+A4>$
zhQY@Nhgg-h%Vy$pd<)~tafiiObHmQrA{B8>6blWisk+IVibP0thQ3Ky7#RKE3eW=5
zXE-?l6*oFC?M>Com~8D>sr1_h-{c!=4gEtgjXJJV=F&R_0gVzDJ>hg=-=*_x?(1xs
zM<SR5#YK^n!FQ#WGI2>Kxm$o1qCSw0l^V30=E{-E?W^So{DzfQvg11_A9=Z-UnEkI
zj!14^^kR9{5$<zxhbH>0MTzG0%s4RiK0)FoV^i6TNp>j9Q6)LH$bCtqoGsYzh;@-Z
zE~ruIxqj;S?1=sBHa#1{xPx3iyh4|LyT5hp@sgL4SnPpMzk)f?dvcsPOkn#d!DT?y
znd$W`EQ7Is5G<z8>Y9+Zih8&b1I8$y52lg!UIxgjogPR|AYQvmNu;mb4RXU@8Mpi-
zrZY>2p@xvD-(HC>eTD3?^A_LF7^vAL^eE#GngTCv0bN3-(Y<fFoAa2<&(%ayi^6RL
z@UB6+@b1{FlYsQ*-c?kco~%YkGvT|7$T6=sETh-9*3iK)aF-`oF#Wy2A}_6E!5Ys<
zQ%<MX=wm4{((Uh@i;!sGoMl;AkBD)TW&Qa?&d^zbZez~vNn>=+_4!HIs+Dxw2i#3R
zjk#!$Y|(#GlL8L#u!ls{#M9Fwzv>qS3@ek4$;B}$I!-fmJ^{WVfNu%?h`y3qE^~Aw
z`*DCi&KIP|6Tlm#m)HTIcz5XyuV-rMH1AY08&lJ9mQP4Ws{sB$vuScesKPGDRI+^U
z<p~zd>{QdH!kwPSCXVP74Mttcbv73~Ns?#K9}Jk($yEk`fE}6<j8gg$O{9JqfwZqx
zG*W5%_MJyKSDHi*w#5z;KT&?LQ~Nk>zAAR{@dYjcx!t#<Ii@ljOBPYC&ZtGhNWHUZ
z>)zLd2=<ZIEwfW#_<iBQ6zukXx!sQY_BaPwz_I9LLw5;ZDq#2Hs%|$}2#!5&aNh>N
zXI~s<u>Y<HG&8B@uxb`4srw0_P$;q^(X`b7!IsNe)D%29g)vUxE^dQJrsbDz>Svnc
zrP}F2Q9}z(%RWr~2De<>5#I#QP}&~*q_=gw@~^L-C#N$pum#Hg+xwxdSoJhpZ)<Hx
zKM5gV(^1(p@)htf{Q4bxJ1rC<JG%974NGE9&Y$OVT`~A|+Dk-0^gAf|z*Fvg&esf;
zy7gP@!_KN;<_^rXya=5>My(avZFnEWB^qBY)OFae?-s+$aQnOGieTDH<;CpJ-J+Ui
zs()5N{^`#^KZQED&WcYU{@f8%rGn&5pr2W7h$4=Tuq-*Dh>vzu{LzeFwVES$bp$W)
zA|l#4A6s$sOT+2rY@pS<mI3WxR?1JHhF=6H*i|pzg0{Ja&V)pZ+D)^Wa02TM<|dQK
z=DQ}&7Ixm<U3O7TH9KDzAn`?QNrrO1bjx8t41IGQo<@)Z=v(Ss$uq(p+c$GA=DZtK
zCl{u%jluJEwDYI*1G-}N&LP9UQ^M4DV}|_)^;Zp;%U$Z<m=rPO`(F=Bt(Pa(r_%N}
zVAIEH;GKpazlNRiY#@u$6o-8Z*Y#$VY~b&Higf$DEygOzahi`7_tYHsR*|nug{8Nm
z-8i6O#oGY2u|?HUTE8EGss`Hm1=VRiJdBFP*hS2Pd;Q|hW6*}mss7W)=W7m_fc})l
zXq;LMP|Bsa!j~QuLy=@F@GRCMz1lu)#K;st#4OS@oQM>98+@Hr#!ZIzN<jSk{}-12
zZ>b1R+CJ`~X~9Y*BiASvjg59_XEZdX0ol^VKZ%iZ`De{|Z^>xzVGLVzE7%lRb5-{5
za1j8xu=bZAuz3{_JJHt~+s4n@6AWNnUyawYGWH`(s;EhfZW(JC9#I`NnhFWcgE*GA
zGP3B<^yc7f3iH+t#;>KPr++)m52};HPcm;LARrLsy(`NWOx~w00cx4--0hBL7OTyT
zlq;VnnKS{cfHO$~HbF;s7trEoyRZ_us!V`4oMW$&F}<j@7CR`?EF2mae>tjV<vN<m
zkM0V4I6{2*<7{^89xhVvfpWwVWtzAO*xBb*#n0N7%!<PDN+4tk8Q}(03i6F4jNTjU
zaRol#LH)z<=jTfd6K7mouTtM=t7xoTxJ(;I9`ncv0KK-l>jYKc|2bU%cf8cF-Y;SD
zh?tn4le%GcJKI4&4Im!nvr5&U9{5vN_(8@zcwtxHHXCwQGZ(mHWJcw=WZUzvtQrYW
z0otD6Cuh-uJ0^7pMg%e@FvJjLk8pl7eA8zu@h%zNF3aYr*2EZ7dh<WsYFl`CauLLJ
zv=vV=ZHFR-a$J#<(1Mb%{)+-K1ak8%m{5M}V-U%OX>$-E%G@X4vh!+#Z>JsON~5VS
zk2MmYN#;?Sz(X?0ETL)%vxm!$HlbG`F*rR{$1)!G%NWK~4R7g=#t3-wc+uRxJ%{dt
z%|)<lRO2RiY0BCM+2ft_F`Kc*h`g|0K?g1UhoflxkEs>_n{!0T!>eG{1bstX$#g{2
zD2XgcZh91VgEC#iUB=3ycczy4<C#|*2@*VfJh!UMRMNGc3@IExX0&{$4=?1BQhgxy
z(-}vcd?+tEwC=m_%8=Cy%gvTBh#ON%6l1g07CZdl`>kFnGVy!QJ79PrD}-ilqG1Vb
zCXw0TP?d)rO)s70M#+G7S^fh7`Jcg8QGyh?Yt4myw~z2ueCFadQik#VcS%HzWg;gh
zsXUcvTQP8V@g^0$)`^ShfcP~#=SLa@X$<jqAZvfCwQ$KhZ%}MNn(;g923vdkaNOJN
zabBdlzHX=ggcy93N85M}5`T*hh3?YSJ5$NC&Q}Yqm*;&=1_B+UpeGaHb*ZY{oHx@4
z12!}j-BL^yPpy`9nM(`9i^rwomP~r_N=A(y`66dM1hp;Eq)W}N>upozrs6UyQ^RxY
zn0GC9lXNNKnbHi)d1@T!y5WCq)~rqcv6&toFUDE8R}5n3fhLIpg_M_7jOxWwQWs<|
z2?Djde;A9z%;5_H#~pA69q|h<C$46R>y+`+cen;!{R$56ogxCwU0{69OMfof1aJ<1
z-kkcZ2Ns@(FQXeI?p|WgAvJL%rd+qfYrk)gwI+pZOH?k4IN=UZgx3}uigRVg#h)fw
zj4%$dDFoEm8+Y(|W$#-873m)L3&A=_AD~}OI1W9{rNCl=Gg~3Cl)`fv!ZR}XnT?~Q
zU$U&4u9Iy7x7^@W-XCP-YwzR!o}8SGk;yR`0>UxjaPIbtF=0`^<Eu-q<>|UZ&-Sai
zBITG?BdM*0g-SOEaAyfhNO%`3O^C5&?9JLF(5yNx&-_-Vv|a>i>Wj$E!qR*2i&>Qx
zf1%dH+`o)UP*|E~DHIn+z^v!b&lUx$A2kYMCN;*8cx|HVlQby%5j*i^?*Xsx6yJ;`
zdZQe1NDS_l4g&ek5s9zaJOe6UFwHHX0<SM%tTsd{l4Gewam?2_=dkm;nZ0?Nv2RiF
zct$4YRltd!%Ik3PImfxc_%clQWAo%9;^OnR_OkLz*bf61abF(X1&n&4;nq_WlZNV<
zW~wM*YsCuhMYJ9l>Nxx8u%EV-!8DiGrKd;O-FF|i#}Ax>h|US)z_q|N!-`SfINy-2
zA+(p0DFaLP{N6AiNBn#?K7F3H$M-zO{=aIFYz>PP^api5HGhRd2#)B|OGKl;xX6;{
zW_mFv)QcIuq-#Rw%~`(?3ep$8`=&k*LYIW#Wg!*PgwdVFVZY%b%B0dlW60=wckHW(
zj~qQia_?>y;THN+s_{UKF)hqVMxG@!&WB;dn}77IM}%B_7fG;QmxVy`u_7)~-R^(Z
zJrtv8r<8Bj@lKfEN$;Izvsb+tx85)XkmiO1&G?<kgzlF0IJfD4qtmYG$#z?#0u{oW
zgJ7*UEiGZ78~{_R)l)+2T+T=)ziQYB<K17NVIlzl3~M@M9~X3_JgLU)_Ku?#H^Ixf
z7aC8>qY(A0!K4(@g2yWcKiOJ+Ncpm;CNem!L4FVupH;7^TF3dHIgL(fYi%RFPFEWR
zg&#!IA|3|UOO7+d0-FM!uIB%iG5!SZCHXSF?G%|RiZEFKTm_=vA%znZvZ?fiZ!VIi
zX6|snxM~<!SQOctkGN^?j^b9VBX3cud_RWq&3jj{ONDv5a2D2M2bYyE*R!ggr^6ee
zrZO!(k89`OTx1y%Q^Sk;AJK0XQ2lB=odoy=`t1<=|IH$Wex|KJ%q7?%b8XPp_K)Yu
zlU@xV-4bACHn6bV`Vy3mqQ3f$ogC+=qq~C|4f)in<)t{~d=P&}hcN;8TH;}dVHI*@
zLW&=fQiR0-w<NEpPtk(S`})ZI&E|8%a5-er&eC~ldlqTq|9F-ikaLFKeu+WmzQ00#
z_g#*Hzfj**?-dkQzEA%ols9wU;ez7f-`9lDh?Co<3rU-AXWN5w2<=b)kI}CrMo>kV
zsxXkY`!VW<mObm`1w|gBaZeiYZ%8=QBW}CdL<Y9IGJp|(3KGNIw(0%q63uDClW^;D
zOrY@TO(9SBN6O~|HC`G<yXST{4lqvYJWKVFc8+7VZXpK$$l30d9&=qe4$#e)yNxD6
zc9qXJ>aml^r!VGsf7jzu@F3c=RU-9TL1SZSU|r}iGpois$(dJI*P$_+!E3p{I877X
z!18eRh6SZA$<v<>rcr}_5u}IG{f_H@lVJZAH?gEh%y?A%;vBkR<DzI=5jY@#Lk9f^
z&h7|Di&J^z(D4kfs}3=Zhn0n68@oM@wp&Sv;9Mxjx^B`QrLcDi{g>oGz@U6N#D2N}
zgqMF=!=MhNTTbJ&`Y;Gn6@Su#%*<KKJLGsT*?$9EL5)j>57-8k@%sT4z_~;b8&i3c
zq$BG;?uA_yyX}wX(E+EOs3x#IQuc>$EM;hWf{$=ti{;#K8z^kot*vXE-x8j^(BZaQ
z-u~S5rd#=YjT&*cx2F>~56{*9XT!uz>Y&@)2@FGm1<ZD3CBN&VeVI!`qDlXIcq&75
z9H`6+)B}4Ay>~EweUT-M=U3JK^8CO|*vPIbjDkZ|($Y~Y*V<;Y?!ns75%NS+Q1exA
zM78@-0M>GG)H_=FuLA2-@+VwxyXdEA>1;DJ;02%`rEIOrplvJOlZo%(ie;HB$lb?j
zV3*QXd{sS_^fh}4+)}tpM&fH#e(#Snph#0OzMoC2PD66nyMp^=9h)8J=>2I&PGx6w
zM_Ds!T?b-wj}O|JBZ9Bl?_mtSkPa$wYWAOdqD~i)Y4ufY8-0-xsI$~pKMN!Iv?l{q
zSSr)HaXP3!nt^9>koc(Z*(4LMnL;VreYW1{N2_g_)T!ZwY>QhXJnP+v`e}Wwa}<#)
za(*yX^}<`&ln|*fzA_FLR927>D2wnpVhtq#H&?2qc3B3Xirk0Y7&=o6|IC~l{W_95
zz)3pCtRm~BkZ|$Qrm!d6pQySn|E62QppDz}w}@C^E1@NOaob@u1yfj9bMSr3nt^-O
z$>HH&ubEmS#R$}EPvV=z>5pcH%^O8i1K!u0ZEYfjS`g22Ni>1ykM(cFm&-hR1|AGb
zJzCcCyaf-?`VqcsF4f0lhDI=%Q~iAxsa~3qceNI%5sHj!%D(<S36C+Lw@l%Ed1xW;
z+j?JV-Vkv+ehkoHb3y;`d%Z_DhyJ(UdSYb+1pG}!y+&jIQ(er7HY52N$qf4(ZF(@w
zBsw6{b&gd@V|wai5Cf3JW0-2cOAZK+ILp*$r1#Gu!=ln`K#jy209y7NDC-4b@b#i?
z6&$$60gZ5O0Z2*xmebtURMWX-!d7<MQoWCt-Y4?Gzm4EGO$?ht@M6|Ff>8)qg-!NB
zzVAEfHo7R4RzmYDPl2S4eB*mc)WOStpxngt^nFb%DQi!OS5X1S`by;Tb3$i&a|-E)
zciVB|I-3Jt3r=k><~fe|XV4t~GIp|0&DO(jLEAC>p@y++{oisIqR^$P9e=HsMdsWA
zPWR77pOfUNxve7Kgs9-J7EPRRv_EQuz4BRIHsVchf2dn_(X{$RV@%HHq=qDLVV}gY
zDT$1K+VgqyIObo~Psm5KnWzpiKdDI6FGDxAPq+9qA)}cbVVm$>Y9XW4MN}4w#)4JP
z@!z7b#m6p?gRS(}Ip^?y2gC>6s2U+7_(u?jJk9(l^8mKj;ZOgh?b#18j-gnn!`-5B
zJIp~8<_uY-<sD;k?vKborZJO?lS6YS4-qAlUX^&<*h<9l%T6q}+;AM>dyK%9`lILs
z!V1VLq=1iBFIai4J(R!wYGAHk1@>;02j2l?#)3djwGfOF(%N?r5+kHPIGIRA994_j
z-$)<Td<Am)#a4lMY~cJ$8*|jivU%8M-uUCrP|m)GI&4CUQxh8WZWXJGiy9+L2bwZ;
z>D6qtG-t7>@I^nu6#69Fh}_LgCcDY?R$cn*K-+n!mhWeu<jY$$>-OT0$L$#%L6}I_
zF20{xjMWHmehISh>fq-p#)UYN1X^}hT3nR5U&fI33*+_dfmIwiRwaHoR682Z`i>oS
zc7@Cbt<u3bRv3wBKF_kCeaCrF-1H?N_TzH&`_Z}4Fy7edg-j?~c*UEuEjI4i5_LfU
z&3T0&n6HBtkb1?W6=yT^rHaCL!j+XZ4Lnb5cGk_VZ))`Zj;KPqRX_r0{KIf5E@0>P
z62oBF50=1X0MNrJ6wraKU@ZShd-Cx=yaAX9{WJ)(G~wqPAmhl1O;ntw0mn8)d5JOe
zaBLh+7+PF_+3lk}+EE{G?1-jPA|De|{{?m$Rpq>CaO>s4>~Liz;C*y-wDupn-OqDK
z++HyxTwNTS3cpg<=I2f)FlR?kCoZ|?dMXK|KDaBDd(6@6EBs2b3P@u&mYnLnCW(E_
z_E}TUX$SmXluK-c3-9`KqigGFY*tBHePOuY8s|yY8n%BtB4)3#tZRHIxYe~Xc>x{2
z8G2BPvrinz#jPJvRzdK;*d3)p$h_;ZC+J_G4pA0KABMjKfOC$fJ-QURm7<&4lWDjs
ztVQB>t*FS7C*80eD-1xsY53TcwbPTlxo^*MyuZ>|ZREF~FNweM-)!`M>5=UXGW4TW
z2APJt#4RB^cp#7W8e8kb5?vpn@aaT6>MvA2PqW~f0ZB7@pVKFL;lUUtTXa!Ua)sM%
z6}$ya2Y)kg9L0iaw&xdsbQmAUpy%CHM;jT{g1x&~v)s7~$gCY3$LUSyRuVSaTv$C$
z>cYGEA<LXC22+bxo7MWX*?9|zro@su3{SI{Ci^tHinQHi5SbG-C1X7%3}E{SsNOi_
zC0eRV{jiL}OdJx1F}az?d&t<6T}6E4&LHAYyb+eIX@IP{!JXdjT+h0!PFEZ~w5jE<
zCSTPw;G|1gi!;9DA5oK%8Q~od_%7&t#n5v|8Oj=b+Fz@oXCP$;PNZU~k2zv(HDnNY
zb47`U;otn+H)vAEMZq&QP#tM=I+8VN73HWjb)gU5{OmrZSu$VpaF7A^Ao-2Mam--*
z$r{@y_;-Fyu{f)G3~D1YtKC3%%MbLa`$D3|DuO&OzQ`kdKf`Ipx^m=)QBIC2E|~v@
zr+?rJ-&D2NubvH0f}$yEHwie^UMp11`#%NDKP(`t*;)kh9?j63^Yb%-6N9#q8=g@^
zfFc>1Qz6{y{LBC*!3tURmHjrRb?FeepP$xpPLZV7vYX*3$%tAMN7Z~OR#3>Xju-#k
zX23o%rY)3e1KvWTwwW&o(~l5L-LF*sGFwDQL^NpV+RDJ9Q*x@cX=}SsVa~kbRbe!m
zqmW9u)K{^GsW=v-CZAKq0e}Ps@-05+0h2=4HZn*G=0BWSk$7!wNy~Tfz<7ue?QooS
z8Q^kfIAy*jX=T2N(l0S7lIeAcXvO|!&{;1rYR&$vuBkRsQ_wuz-oX4CbiZF1(L!Rj
zKb7y`!kHz|pWbzPUccs7Y+&_s+&;6^a7)$ow5v<fTD<?K|ILX9Bt=(fEXHNl_Fwdg
zCC0z#6GJrjAc@0FLDz<R4CV!`jpHUV{?%V}Wfe$)F{T~~2}_90n~}q2-g8$?aP>m3
zTUBUa^5y10Q**xiX-!%0%jYI%H{rYKNR@A#yDfHTxi{>`>Qw)H(!Cff%x{-g@JS!b
zN!<p0#_g|}xHNX*E)LF6VyO)GIo3`GW9Q$f53JHAzBg+2({yhY$WTajzuJ)O`HLGe
zx4W>MEllC}tO|tQw;km14Q)5j_U*aVc23M?Ml);wI0Ne*C6DtRr#j%Snq`VA*MvU6
zEQzn;Cz)!bBOSziec3_Y@+ZNq+`7-)_W4d6vVWL#<QfSye?iLBjjgQGv{F=;H8)9?
zi)752&x(-0UbRpmQ2ROtXa_oG<Dwmj>aohW-Vq6V5Azp$K%g-)RoLcTrIxVDapiFS
zi2AFI8WPg)jhy(=`cvD#aoyK$vQB4sh5s{okNYZwu?i<t%{(Y)XJ@xzXtVP{z?X!P
zkpbvej2VY&`-(wshCyss*4(^C+!m{i98212e~jwH!@yO_38(%z;#VAPZo2M|o^nQD
zl=xt0ae13TA}5X%-jZbbCfrM0<B1KB6N9r4kV)tH-gEeuQL8bB-(}^`THytyaO+pO
zII7Yvm+IIgu|%4{n^m8)*w%q*%nOx&;;yM8O{Q|6E0s%ibS_%bmo<M_)L6Ij!<fJ}
zrc4>nH|E~IangtN>p?7+PE4tf{cg+rtDlm#1GZ*>=Z|nwM9aJ#NTF#r4t~C5N|6)R
zJ5VoH_S_v&Hy;JQnx^tANK#^YJ<|I|1gc0CbZee_Q}?zGIR5zj^l+$ITQ}!jN(?78
zO5JvR53e)$UCYjYPYa9Trgc5eW4f1%9xkGE|CyFrpQl4`Y|cG;s9r7mU71$BYzXd;
zp~NV>o@J8tUR7~+ID#BK{T<(y>ivc8;)i+_6A^x|X(vh@-?3Xxx_+bVN|BX%QRcM}
zx4?lvQN}@#INuM{u<*41mf2)FS@+vzc8xjU{jpAiI&nKLsiaZqz!c<>*~3&|1@fg9
zdiu|qN38XEd3|R-$k+J1${<g6a_xfBxX|qNr6t(Q6iy&JC`HGIfq@<gIpd6<I7>LR
z-E;2nQ%k|PTg}!J1U!@C7$MoO5FVCZ309bIG(BoG-hE~GCZbcBylT_#E5YA*+rt;@
zLeO?kdLT&5{;-2X&7K(Lh~aRe`s((a%HHFJ)R;>Jvv#$jYdx)u@9yi5?E{B(`&%2`
zC&nuBwYqh-bz0ci(!1&l+@b!}1bL?jYsSHC;;Y0t5)M;Ik6<p5Cf8?Q^gfKW{j5(l
z^i@TVlz2!d0{LI#cf@qUtj@?i>xT-X>lGQ?bfiMXdvdxe_Cq5MT^q-8x~y4OpuP@!
zda>Fh8gTzC1Tw(YZ_d^G^>_175v2hdG1D#n-1^r&Yz`HD`nd5`b_7*N`|xUBf<f(=
z*&;-wtAV+*DRFuep?aX!-6knJ7)Pkh@O1XCpzZ&(LjDn#ZMc6vM_9jZ!HbUeDT0c(
z71PqrA#8pE9)*U5Jp=}CE3PNF#f~R>BV;~Zw>00!2$|P+29=Cff$OQ~bp=axU7u=;
z%Y|1{Q&Wd#iZn&df9c0rcB;+KoA^6_)R;`e5|Km5G%u_z<N~Ue#@&}*$)k+Jy%R|3
z*yyOX*KJTXl8}_Pt+jRKw=}&&Lk=adnDIYIlj=8>RLh2Aa+d~)y};x09PNleQ@DEQ
zBYl%CnVCMGu~SsexpeuKF=SkAu_v*IwL8$%cQdT^Ez)gu6f|A&Aq165?XDg)65NxB
z#D@<0*4p-@N7O|fzP>Bqi0}jE^@BKTZDy(3C~bDRs)`$QRN3Eiz%{D2$G;v>{ieTc
zSa3D5gE4ka?MemxNRA354T>wiQSS(8_=heT6FOuar=#{?Kki>rc>q1Q|ABl0wO$pf
znz}}o{wq1}{6CWOHC)4HtZPf>uTL#c4Q)gAt$v5)#(}N0b>8Gh{qW8`_+JQ2nf7Ze
zO+vk%Oh(N7nU;E@BYi2+atTe-G%^xtd-bbC;BEyPQ~#^V==QfZ`axfON35)#brqIe
zLr=%=+jo1it+5D$eT|3bJ1giI##Z=-+Z=wO@z0Pti775*@hZuhSJP-YGY?kNWgejs
zjJ*(z^w7teqO*j=f%pl#dG;a1Ou-5qPqedI!61=}a&#p#C|%9jwkdlOj?T{O%nV%N
zG%v*zDEr~zAdVBPe^ZPxXCy0domg?_ib0Pk#%;yH0<k;k<4J<RQG`)PPkC;rGSTPH
ztoidb*6-*)YZiyT18*OW^BJ_F-jvM;4W**mx_E<HnlZpgTlYB^&BKHGqj)qwZ7^%t
zm(;&Cdfy3OK;}n#?`x{|yH4033DwOCWaoAdKU{dN<F-T-2U$E7wfC->;}>})X9V3=
zG>#rbY}*S^o326O9@L48MoR#h;xOhCXX04x#hfSZJ+P(~6AZ&3rqoPYnrA<Of}ZJ1
z3sq3S-^)RHX_rHr+ZL|w7QFulV5@~h@<sr)(IcHlPkbgROF>!j>y4KuI*#1Z2a(T>
zE4VF&+x~J_PT{D|C-cmOV~u5k?9d>)n<?nrSJ{d1n`yIm5D^%7#gA@sd_hV+eK#gt
zP>2v|FX7=ID+OOa=_}v^sdpxlhjFNRc`Yhme@-;_9^0_UjJNqayTsyZ+i{&xHYibR
z*Z!s=%&d@V1CUY_u~W%7wg~1K7ttJpLl9u;>eaWV{cuXi&rFkJ(9{%(@7VX{>Bf3@
zBt7TG;}Vx~aVUNpGv-8#<mBPy8TLZ>TE*y3=Gqh>et(kkTfBfxtoxR7ZkzgV+4oM5
zcng=2FR<X3;y#)Bl<<cn?7Px&!PWco4iyy@9jf!=<@%fNna0%pp9M}$fX#x1@&Vw$
z>qQ3><nxufL+*a~OWaWz_1P35lb>)+FJN#|+U!Ek2$J)rw(6THBLMjq-S15xC4pN$
znfqp%2S|og@}`RbyNgqzlB+Uyz*Rdp;RRm9mPr6xWw9evFK*E)$Z`WzLBR<FE}>>~
z2wR7(#rH!MWADW-xDPGb>Hr#P63|dVat55-FL#0f^&pGo-GTWmX3?qb&D#B^Qwh<>
zhSV;_%l-F@=Uv}pc=NdACo!7x2a{~!?)g7KttRVNa40HY7cPtWG`!;t_37}lQBxZX
zpN<HPQp|kF0A$z4n08oHTz3hKj7M)Le@BXDc?5mC%?`TdyZjJj-scB0AA36}m171W
ziw?@`4;#KhTpFU3T8<D=iLju25cN`S8(D1`-gY0R9hJ7b9=@-|T72sVR5i(Fxe07F
z;3J&#HTQ|eN^?wrJ|yPs<6~G%gIBKZxIwOh15eESjd6F&&Dt$1o_=yfUnoL^9<@ae
z#xJVtsh@N`uV_p^%mncNg#qb4pUqGM+aX58Ekm<2oqtV?bY&iCh8(R)zzH-J)vFkW
zA~NQtzDf<Bx*zu=TIQrS21RkB-^v?ZY>uzn&S<V5{pXPNGd&UsU&~J#Q%EZ#sqAL8
zxs_HPp!-3HB4_>}ka_G3XwYt0+irVBu{`uXELA+f<&&t2-(fYOdUNg2SxH>P$VZ3>
zU+&++tfhWW&iV=BC?7D_X4wtKQKk>&d^4gWSY6qCqd4n|?4}lOLbuei9ZW%Iu%GqD
z_v=caS;^H0R>tho6@L2;tp@Y1t@-mjgNxe5I`1{@s0Sr>1#Ese6$?~mU)=qU6HBBH
zepGK&G^4x*3s!KzBcEcSFn;t|T@#QDB$`oukZY^iYqy;-MEQIk*o;}Pyp@$(b+w?0
z)BO{Nl)I;vrl;7;7B!=T@qKB3LZDSJ+)1x8SQ`fDry~YT{*uYL>k;fQn!c2r7dk3^
zIM_~PP~3svU)?8jaHMv<%WCrcr;5jU??AY%Q#F)7dO^sZ+6d<XcWS0pE~2o8+4L$t
zLPiUi8sFlN)~|YJ0sw$L&Di=8EENht=p{)4E;YYPmM)2B^RaNn1hl9yM}u}&KkjxK
zdshwyCvr`<49RWg3&)r1RfZQ2XQ7pvnPmQr_dNq;&6Sp$Gh+Vj2Nw(fN}*`JKjJTx
z?37!|5rmA2ohw$Bln{n~O6mH>@z`D2#mXq86ZpR3d;KW}h+5Joc_@cDKDA+l?vXB<
z_AG`Y@wbE$+S$hM0e6nSC8<n|Y-#A1C$n5%-OZiGQ$1b;R~d+U5~nLBoD$3j#o-|p
z)tz_k(HEZ*zpd92UWXG(jR?8*{-5;#xi-YlxHKV8BQs8S<ErEiXimO9v9Cj(E(N{l
zE2=$lB}IBD7HW<K%+5v=DLt-wQrz&oA^w|pA9~c=o_CZQ_jj)4^JA#1MM!_0^5~<j
zSZ?f7#RPYLa#Qd4<t%ePapP!Z<86+`_vZ&T$DL}~9s^U&_nY{^+#gb~#bdpe4$g_1
zFvLC(p1f}yskmqB3codFdXTQV2v*J{v&yqOvdFTb^fszF4c6LxMrU}wu)<Exc-oLi
zDAIgNT>nZl3eq*Cb&7%gRfs6Bu;~fvU3PW^=hk;m<%i1C%_*1n=PzT2Ay;qA#h_XB
zM742UKP_22yb9?(UWt>n?K&SD?0)A!5};Pt3=zFI)wJd%COf6-xZRdFK5gDk!lP>m
z`y{WL?6U7Z;JQk<uFj~{z(Fx8_@CY$w|8heP{`2OvGDB<tQ3U<p=JH)R0v$YYIxK{
zRoAGEx)L&zoAq~0d<lj6zNE>E{2=Rj+hx8utIcN7;M<w8k^tI~9qz*1Tm6jrTa%%C
zY;6MLx6ZGJ?I}Yp?Vk+l#p>$JYr9PaA3%<jt74vl@MFC6w@Kor)c4lG_v!wM8=lV+
z&85TA<+1A==T$(M)58nwotI}%aD#FhF&*%JnCFtOHSQl8wpj?W&?_<w0)hUr4qvPa
zw0=!T%oNH(yI^jI<F;MTd+xNG{_r0P&<U10>pcv_F9Wv4pKmQyv7HXA1>F36mM=Jx
zoUC8=Ta`o!uOm!`JI_T7m#tgj`GEu*%0`14No`2Z8Nemmub`T$yT>!3W2x3rlyCCz
zUUmn?_K$Zk2~XS&yu6<wQO7ynTdpr(@fc<kxx{ZiHX&X`#UvWXg<-~&^+$r^gkde3
zY#9}El?RqC7jV05mEbwc^Sifc&>e)iu%WNtV7<+C*0%lHkW%;`nATDLkxcll$}!K6
zS(*v8AIN|T$C0x&;S4mVRi4E85$jP7|EFbE5fRBi^@R9cEbXTcJ;ZufYYx87x_)S`
z2OC!n4M?@;!1aOJ6MFaIHrY(vt=RC`olWA4`k|}!l508#!><!9u#z$llbcVE4dU41
zAHEb@Ce)(o2S9s>s@#Ml<?*#d&^69lhy90(vNZtsc|SR{zintBT?@G)kVE>j+<u@R
z(Fj&7Y6fBZmf!Z<Dnfz*F>J4_LiC=VPO&}E(LlSZTUg^!@klacX8KjFdTZ&m5>4I9
zmqQ`PGY9KtJ9b3OD!10vEn@yjPVAjs1PFaLA>%~Zh6BF`h$xxW6~|9XPxpdA26a8#
zuTN^q*z(0ArPt&T&<CIN4>2HZs%b-iEta;}v9`{@r+!q?Vloycf+=p7=0fNG>ld>g
zgJb1q9Vf4GE)2&-Ovd&zb{Xdb9amz)@Nt5h71m5@e%oYmGCEt(@0@5>@BHmN7QGQ)
zCu<q65`HvdUC;o)#@1+zD+J7>>pu-T;cfetwDM-qU)`aEh;F?+y*s?2il-_i*jd2_
za{rO4`sADqU^`9!bp-YKQ5~1`diMXEweatRO;Xxy``$|f0bj0F6Rq+j!6Yj#LrL{5
z<?i<%B8d;fH(LAq`;v)eTyLU2`mw5J38=#}g#5M6ge(Ri@VvpYKbAIq^jQO^)oDLu
zWMV9!UM9{3zpV|;9e*BBvA;iZv&E_kv)JV;=&zbSo%pBS^=!M}AEIl+?tFmFyHZy>
zefj5kE1E;=*)CkNPz~laZ_FeWfn937-sE*ixHvR%bT4XcVexCokaEOvX^H&?xMUzY
zl%A9U8jQf9KACo0GBKr2(kd+b2kCLLcz4kZYcm9iJ<xyNPah9M%VHw*OPuBbTj~>=
zW4rhbdd8~Xj@84x2xhs)F^&^Oj_yg%*u)N7%BAl3%gc!y;^&n#D$lWh#uRwt!TU0-
zkBa{?=VP*D%D=L+pF%RP8O%s%+e9(lGg1dlXoQ=-t*pePq$(|@Br@C%Id&r1Use2d
zy$NK2T+D*1A~1^$g}Rfjo1UGmkAO4|5-N$I=S%=Ij9h4}YU)6ZvJAUNN|=O&>F1Wo
zo9szs!P;#n7C*<Dvxv_up|?v*v^^4L;`(wpM+Wdv*y<NcsIMFxjsO6Bnaal!P}}tB
z9Z8l9Fmn8U7~Xv|8F=9aC?-m@A$Nt(-u~43+_JP@Hk9(hw!+-*#MYUl$<x^8|JMlP
zOD+BRGj~LdcxS(k$8hnECGOuiPTd-9Gbq;Sk9I1PP0F7UJC2`oHIQgUA=MWOkj0|I
z9YtSG`qb1<(r8fOaGODyR>N+LxpuZ0<8MYQ{H=nj1rwGTed?U3eg(Zll!div)hhD+
zLpBR|1Grk|v-mq=!N{9S;H>6RNP6_eB{KCYy98BBLt2e;s2qRn6FPJaKeROX{80Ox
z{<u<}qNSXg3dP^gMjqiBB#xOE4f<V-D;5Orh5}ig3+Zd?jS-1)Bi;2;iyivzhh>eN
ztQi<^HFx6D?53bm28o{!#hpxQG;LI;;ZlaLSdtJJ@%%uMnlSpWcGLgM7>kJgs6NTQ
z*y!A>1uKsFAXgf|^!xcNYVm183I99P=ofhDQbxza6kToXZ+DaF1>v5?KbAL+)|l6*
z1@)TaZGF`HS|cQHLNq5;HjZWEr?ad1Wkyp5BPh0<U@`V#TdVdV@1fZ6*$>*p^!f&_
zH|P?H)&Jzk=Ti_7`o@9N1!%TUP7NFx0#pN8xsP}~)p>%e8S{W!)64uUEXAp9@J1>Y
zm>H#5*70<6s!1O!EY7HwWvRwKspWdfky)g+P4?H_AQL#5FceZ=JW>4s_hpIjcQTX;
zl>x!~>dtAs(Zm?tUXD{~p#g;|*;;E3_;<MjmT{ttM`>>@x=&Y`vOgt4a(<TIuX}KH
z!XJa*eR^qhjDVXbkZDpwp=}#)jR?>tJcJ2E@s~4<dGPR`6y1p_7SOYw4&$g1THgE;
zNy=m6%L&iiH`_D(>Bq_bE52#N`$G_tjgeT2kJLNK_~@>Ita>yxh0pOtNKHJG&$+<E
zq`xP=Wlnq@VOGGSj1*z{PA{S7*<um<?&dpIRRDRUcDVT*!msPNK<=IJ*T;Zf<(Vcb
z2cx(*BOFFC!NZ;k5cK%?Y5t@0`1p*Jdpjpr!A8x$baq$taXRJo0gM{7J%t^J<lYn2
zdgygtHyv*W$KIb&N<NZ>HmP7KnN1fk$GgNSnb31gH`gQFZGooPxM-xu_mWwCJ`0>=
zeLH<<ke$k8_#_T2(>Y$|5LWoRV~Q5{EaUw;*`k*kW5ZNP^4&UeIfQG~K*7T)&u?Jj
zKF~SsMEcJ$4)NeG<AE?|q8j0M)s04h7U5Y_V)}^^*nFLWELfu}+eXb@{u>|S=f<C!
z^!cD6Kf_`aDg-M769YbhsDa<3(CjdvBgNJHL#Bo1C4}rQmJ@fEd3L1H73=(1BgwY|
z1#gH&ZEFT=!WMQqQi*L`lKI~vnr>}<AMW|}?^yt;@TK;L@AbK{?^@Ge$a)bKNDqP<
z>@r0z;7wsX(zV(Zw_D6RlT_36l#P`H$;vJ3x#$c>Vlc3O#exN#CP@zub~Vd=&6rpG
zY8i#}W!_xhW4vN4h}gX0O=kNu?gq(j8Z2RVauYv1B8*1rKs83t^*{S)UJ|dyJ5W!9
ze&GVV$F7gRC2v*>fg^sOnN(Q&u|$3RAV<u#{f65mu-)8Mqy1^4+V~yhPyhVIPh)T`
zMfrXOG&i}pXxo7YU;VNn{$39^G3QynYrW}Z#Cc?5>XK}3tu!kFfbvh%1CIIOj0%1U
zrm!gJ+$z^)%pAW_Utj;*=@rCU*m{WY?x=cP<T&Y50hG5j09e%|hT6$S)2Z}pF~c-H
zZ+}`<oNaJ$u#@QaeWiIfN=k5gQw<yD`qRVuiwXRmX%^4Hsnn9lPolei60bbXrVkg3
z+$f%J`^D*u?b997`m+>$R@@`b4ZQzGKpH-Vz<acQhV2M<7eC<F$&@?eFBsN2smu7r
z=Is1qq`}(?7j<htL-!J%PnuyvRo25I0{Ha1?_l%grlo2}I@WM}<=6S%JXAh_U_60d
z?m+Xud&3a^83FNNhe=zezNH*Eo+FC(zDoCcYyGv)AFM#*m%Kb%Z(qG1Ln)(b+kmRT
zmw=KYP~P!Bw<<E>`$|;B3<pj5*)T?3bEI^ZCy-!2<vRtZu|_^W+SJ_WZ>=_(aO1yS
zk4tC>pFGXeS;ObA(hlF`f8WV1dt-BYPN1>#GA6}-+%WJNrLs|D=80c^4eBgdF)oFD
zWP7j|QtDwQ&V|_My-NG*HhBk|cI`D%9p5b2y~l=15-5JYZE%;q(-OpOiEQMEj_<E!
z^Xt+s?cMS8R=%}lvUC`~c8~kwT40s4Zdl23qaoi3YYLG3+lEXGwR46bvbmr6&4<RP
z8ElV$VF99{q51z+-B(9N)$jen2$B+tgtTDLBGO%gv{KR|D4>*x3^{~=pp=42x6(0y
zGy_9RH#0Ca(%mt{-NWN^&hwn(dEdL%eeYWL`UkF=+51=f8=r4(`S`(771tlNNBv@M
zubc1d2SLVg!Hfa0b$O#^6+Zu)>%I9EPtz4t<?w=RQLR5n#3A}^FZ>~9!QMG{?4_Ud
z;y!*lNIu9X@#Dk>C@}_^?sv#DC<nD4y}4WA(#LuD{DNF=8XvJ51c!tL=+olyw1v8_
z#4h6L%@&pG%Bt?s+e}UP<Y<aYPp+kckw@GU*BK6*$5?`$&63%!O*Kf@<d((z9~o0o
zC?9O7$a#7=1{~9qu#C%+LBc|nm~GotLOw_>67}V%zuEEW`a(o`Ian9D05=pVyyAu^
zo@$cxfcoZgW`4LN<biN+Zw#(#oh0LsZQYS6Wm-mxIEW~UkpHLL?*=xRyU*vt5{jx^
zqvnT0@N{3a-+Q29%R2RuP(Q><NUG$}<64<%@fyKW=M=A%16Q<{)Y~l+CEE}Cj1c(T
z!3G=RxNAr2=Avrq=XJu=w6td((fn-!LIcz>HZh`M-Y2{IyJR6p%8&Mz232R(_h!}A
z3yj}>x#`%sZ^_>GK+P10NvFy~#0&kV3KxChjKOmEdessFFU!T3as#G5Wrg7_J8UVf
zu~VmDNsy!YkC(_K&#iY>y&z}nBEb()h4AnlpBf<B>1r*a^xZR{KST(<u!h@fbUT4q
zI#Qs)*9ari7ya}*-xH(Z;bG1US1+klMLE|nrb8fSFp4;MJ;yZJ$(S(NDbGfnHX@!~
z%qbreOFz(p`RUXxnX%<!bCT=AC|6!g{;M+L3LdHND#4BMQ94floX&$A=?(7%U@rwe
z=yb}MMk-ss`7E!LLkkk$`PeBRwDgY9;RKmQ|1o!Xx~VM#=R*)<THFJ!4__%!4sQ>&
zL)}W#UM1x!keSOWKR>>n<yQH8NTI7+EyCNU+Wu1bD7c=P@*Cle6H-FuDuTU3Le^8A
z1i}>@CX{}|1#x9UM>_$@*T}1%+S;{?W)pM|aGS0e`uLj5esqEp{8*X$LBmAkxe#hX
z_DY1{>v1fx|D5>%f!Fv>KXZ0Mzo-F|%;P0)ifaF`>^DU1pXKKxrSm`1P`}B%u~x}c
zrt@Y`V9IxlH#<h53cl!;MNttr|Jhdt8BD6BLG;Perw7iMGTlPf)(8qf{4B}2e42W+
z#du@u-pd&OCX)B0NSZO)l2>TT3tM8OtFR_0>|?oYeKm)y>C-`6JCfy@CtD4{RqZ{z
z&%>0VdMOlf$9?dvo-(f4hR<&K{xkWiydgoC*1X52b`#%kzDK;=7+^xhk}&9hwqSA^
z?^ovZ^H1~47<laVb7$`^-a|asiF^N{BJZLv8y#+hfZ`82;}4o3$J*=fl+ABlAGib8
zWMfMU`?+@bJz-sO0eB7C8sz1ML0h~nAs&Q{90T7U&lxw_hO$gN5g6SE66{t5%$)DU
z(UJjS4timG9X#G$M<#)!J;1rNRD7h$2WKa(>MWufaqLy3X{ED@pDM1djbxu%oNE7o
zjAfava=rYOJ^qC9>*>y>r-nwD=kB0>o1X!4em;_E>AR+ffequ%7t3;+%LkPbjHAk8
z1lxCq#L90Z!7MwkxdC+vR|~tH8Uqu%g-2h?E3@{xg&zW?M6-aUvfUQfsYtw2w130a
zW^J_M{Z9G`)ktGS9X{?gGc#vs0vn&)<&(U!@vWb^=~8czw;nJDq2_7tsNl%=*`(YO
zhxbF>5X*Oesfpb4tSCcy;O)4Z<MW#Lav13GLRevW7Hgih45kf{UsgW}TiJa&Jp6%C
zdmH{WA@cbQuLtNy53WJyY-zt3=c_N$WN(Px)kR3Mz&=JE!YEPJukGyLk)F~>z+`K<
zQaCoq60d-KJKb|jb5d2tF@TvY;F4p@d4YrRvXGX{BN_!eaaWq~wFquu($(8BJgGQ&
zS6)6>4E^MDobcT|L(Q~C+tO}~r)d^OV7eKmlQo54KECe|Mlk_6?K$<6;yFwvKQaXb
zR%rRxZkI*r{m@s1v>6o(wO(|O){gI@`|?ea1#?9f;^OaeEaFebJjgSa5GelI_7#Qt
zq~D>@)43vsm{zT2!V8+Cj21IY$o#@5uVib7Udc`jMNl&1n=&5!LZ|;?1z>&jl;J*H
zjjtRZvdZn*D_J{A2yYkCb#E0q7WQR3)4vM=vr9@v=0p)Hd4>;@)NBXs%7eGZK3{pW
z@)GcgfPvt+^I*%%t{wYt6i)8kk=8dv(S;T*A6Qr)9PXhV+N<5RKOwnyHyPx0`;B6L
zX7I7Hs;ocf*~y0kNfyr*dQy%)ja*!2gV!#0c2bPTN?2k5t6H}*6UkY(Q@957Soka~
z*pQ<AIPfs{!G?flnXaoH4Ru=eMm<Tkv^!ZOj@Lp;;8&oD=Fl|mepawQA4dLiqMQe|
z0|GV!YDVklfIOdvdsBC;VEdKyI6Oi5WNqev+3#?PIcN8#;zX%#yKp8+9my|5><}m^
zv&Q=JRW3bO+eu?K-!9|%y#O4&F=Hy(={TlEL)y0E0`HR?|5$BfB>DFe^jd+f1>b!4
z+L6cvHr?M03ZBu-z`A2afvdu|1K%rN8AXvT(O<i?lhNZY1!U=&31}Q#&X`@gP$ECL
zIIuU7J({xDZ8AZDa=vi6mPF^?f|)RNn`VyFvsNL0-7cP}_bL=$si-$Dt5J1brSQMd
zZ$5j^KXh$ERIxInMfDL?y@se~&(wT06H1dcaML@(;pHsCQk&&cH;_s_-?e;`XyYZB
zqV>{9{Mg-t%_Isr2Is8s(!2f#&bLJO5JPrFM@+eDiD3+_wfUWqO=#7py!Gn(btGe-
zZ4HGHJWZjONZ<Litha9blw_jaRT^JO0entOQ*bwj>|lxl_yJ|brIZ4zyw*rME~i<N
zw|E&!0u67pKhaSexvsH<O<e+d!W~TxZQi9#nbN%An?OZp>S1j#?n^PR@}W^0D6s6#
z*?!7V|E+*K`_uJ^;p$BFjg@l86nkUMdQ^NO(=B<hQg><RH~PKBI#d0Jdy`Y*;<aj)
z*Q0iGOAhZ=;8fJSyF^3js<WJM&Et_p!woK8ISrt>4x829JG)rb5bf}t%4@bufAqNf
zcaQ3%u#AuYLw+ao`IS#g@cex#PfbdiB$mZI!v{LA$Ee57E}9UHuus&Wh-{ZJ)5*k3
zSw7fUuKm#8V!3Ni#i2*9XG=PYJZlw;K9Bd2PN{b<g3U{OnX0bb($Fz&k7Y}?S`DIP
zqhej-Vh|xRdwYlP_U(;(ijTI^<Ln<P6zbMy;6RQOtfW@MrB6{aot>|4CAy|wFQoHA
zn@Q%_uN5=MTldGsRwTQrJ*myF-cniY&j<%12ANJn(mNn;AR}AytAI4JBkClo_PEI}
z16U;gZs9lqvnPvyWyi)&z<ki`Rej?pc13zgD@O};5<o|)*{hoCj;6QhPJt>-jEN4y
zO{RMrn}X`5R8023(!Cdr-oM-2f(t)heA3M@N@cDLjVJd$!%Td;y5`{P-zeVV>n|Lz
zTR}TngIUYb?HqT(6y}4a9i+)h;&}_oPIngnt%^CP>~v%hNpEh0O%||&{ViEQxN*qd
z@8mI1_3R*>U+~J1ugrN1Ar<QcHmVaH+P~y$vQH;t=)fFQW5SYWb-i};Y|;cX({-7K
z|Gj7ZL34|AnU{Zuo@vt$M^E`A(u!OcE(#>!L(On~lV)Fz14fov!Wjv#E=xBmrml{(
z_@KxsR3gEnV}}Ou2gd&B;1ORb>Z$PH(8F(i>m_b)9uHg&oOy&Azj_!l!PKula-(J7
zWek~f5$!h5{3o5(aF)VbzOaNuZ9TtBM30DNpJq^~!149=_f*2Zw7Wl32AYEO5#zc0
zZ@s>+kwtOmkvoq!3NLF(k|liJjwgZ$aPz*PC|Q2Mu<mDmdDu{z#DX8^-3~8{H9=ii
zAr!5g@rhPpuvyWhD^+^HWV%RuDuPeDXnR(i;<)XidA)b#rRjR1O}<MOVj^s&`RhnM
z_b0s<wjc7(3)X6pDT?f2dT-7T+@n}#X3a)q*Aox7y<9Uk$X{jUjnDZ`AX?b%5O@1l
zU7|p^W~9)&hf94as3%L;{_H&ar#JM^Mzs8FGMeq9qYNAhT?XC7B8jcq{YEM(ud-F-
zQPHa|FX$)kMpH~{^kz@IVt=ssJSSi({lTNY3*u~S@=0WsgQhhkG`u?F$V$2PFFd-Z
zR_<Jp5+&_GN#y62qQR;!i6dC6@oReeNHBq<C5Dtu!@-v4`;<#9CV^PE(?v(a|LNj<
zl6U@=fsKRx<sTimP|RvsPJCz04oun`gPzb~ax}F{=;%-zWWGGC5r8MK?Iv(mT0^tY
zGrMyN$h-dHZYi!CH*T~Em@KDdxISf16Ca><S99VA!4VJc6>HG;CBL%o5tpJe0BYIl
z9f8C>CQX~^dQr#ayt0d=%w3T3A8!&T8~x=NfGPzJY=}-w$or3zxq=6T)T-+aF|PpU
z=oRbm<GZ~qdg?VJ^7KkxVaiX}37*68nM!iDRKDSezU9i^_q{h(;B{c-8{LW1i>z_`
zuHYA9dof~QxclxzsjlB^pn^~QkZz8tg`d>4!MWE12a0N^=%dJdDD9OQbwu{eONNKA
z?}D;js05P54v9bu6O8vXv7GGJd0jr>F|WTt#{`!&Y<_S}*~a~<up*P~h+}!cQ=V0h
zY}73@&?8u&*~aE!Kv965VqLJvrAB9cR0OV3s?1PCxfB(t=dQ0S+ntLCQ+|(pJkcam
zb+tL>1uGYMb!iFq5=bZwaiU(M96?(j1m=YtSuF$UuXQn-%g1f0<~$`S94;ZS(94<=
zYGD@@0+lK#3xoTw-;^6xzI^i1Vupn8+=2YtCZ)B4Aa)9uwNZ|5gi%%A(y1L&1~GVs
z#IKtA1Oz4GDfhx`r!c-p?63>);L__zRJ(y)Ucm`g_Q<_?NW#zBtD<QZ@fq@_pT?4%
znywTRsVVibP33F26UqE4;0WZVF@EXW*xbC-Z)!C6{^nHfHuv^*`6z+fLcyP91rRt@
z1$n9MaOw^}K^gUB*v@TcDbMn>9<;G?k^PvMB1^wVbp76g%Dbb2yD(`;+~aV60r{eq
z#>Px{31Qa_u;S5O2Pc-orJU^S2&+W5`$D#;$t*dCJyqU{uX@(bnz-n&p)#)%_iw>s
zgBX*YfUl<)$R#EN3QNQuIJkq0rH3}gjg4Od?I={2hw}28PS&)^ogDZKzyFYzDZb-;
zH}@ELmE>o22>u8p(F)qd)5$j&fGAxmE_)^kG3+wTsM${#89=oU3Tx<{oTuIyV00!h
z=c#v~sh?2@%NyQ|0(cM><?uwC<_nmOup6xsV%b})Z&QU1b~RHY$Z-i|DPDPS6Bn{F
z9Y2&0!6AaNT+6@UeskMcp6DI;S!9*$Nb}r-X->4Y<^@`LiSNQC38a>oYNbeH``fsr
zyb<v}G6NSo`^gMb=3X_#q}ks2#S{Nt#Ia$-hM0GWKn36*9_T06&uuZBHL>Eoa?c4d
z{5tJm=ajj+?O2<SS6b`R0ig%bQWbpHje(G4j<PGTMS4Dp`f?w;NFXWQ^|bq;TR*vn
zka<Q!CM7REU#D$vYm+b?V)sCQf1W0hM0b58d&2N6&Y|Wbw(<#1KJ6AUPTH;BJzDvG
zk70<Z<$HktHc-X<?dWH?nw7^6-mR4G(f*z3AxPx$v)fxhNw;3Kw?Xh#qlgXZ;|)LJ
zCmKaD_3oq<P#~ihb0fHf*FGz}9yW`0O96fCvMf%g>fc~f(6GI+x+}oqJM318PH&$h
zPi``GK-EJQbyrmvTDvx;I@IlV<8LtmZaCsVnUq?f6cv1nsH+;O$7mdV2gnfpd?uzv
zW)r%)KjB?)XE1T&dx}XFgp;6DYCea2&&e?9<k-P&WgxxTG@q5<7TeFK8?5IiBmac@
z10tWl#iN@7HSWp~s&ssKB~+_bP}(o7*)&$_FF=Qa5cg!#foH~D$fev6pFAt&_{$_Y
z_||kgulH{JlFqI?OqdFuGOVh~FiG(yPbE#>)?g?Lz8fH`yl4JKG~pSsZlH)r6AQbM
zJYp|%h`xSQSM?`@Cqod8gODY!$l0;bTTx3Xs~2-(HS<I3(HyJ1exQ>}vrcmxUhQ)o
z#w>(g>rWb2v}*n2v%2>}MC5&(+$L&Xr0ma!*`al$OcL3~JF3u?cH*^-k;T+r=tsvY
zrSphWJ03A1BdgOzC1$8jz?aeqX;GwdeV_an|GN!tkvKP|9z6Q;5Cqina<xbC=_}lj
z3>AEkwNA1#unKq&5gr2a(VB@827}Ke-Bly+F$-9#B@hIZ@*nSGjSthk(62mp%=fH6
zVEdl!2AcLs=s$dz^6F(j+HH4fuwDzfLf-mtq|pxOjMfJ2%#&vIJsUD|$bty+cQ|!;
ztJkBldh{X|fDY{Gl>{)RRhB2mfP-{AQWgD9>he~H`e`%0gDaseuNJl%jmp<{H@|er
z0%gz3&GOFS);oh*pwS!9!ijIOC&iCgu=}*{zPW*p&exSB0jn_EiJGzC^7+xJdcwDv
z<o@3Vzr+YwWqVGF5eOB0#H&4+4*pCQ#<zR7MU=)AHD9C~b>28M%mi@ZKv%+`J&8cI
zWimy~KfB3eyNX=+xi)=SpHRVt&yhHhjq-pd&-DAnC5c48Z`Vw3U%fi(dV8+@xfBAE
z<r_>%Ih7c1;iZz*?}^CI&HMy+MeSy2Am`;h6m-72r@WcFK)idGz<76j$#8b1^h#X$
zfk=0IqTyBUTL;(Es&kFuYP}h&M+0uOsPC<gm#@$)EAyGjeER$)D@x3;pq9t%wjztB
zq4rk}qMJJm*^MNm_Zc^D=%LwsdCo*xFC9MeucUKq=G22ONxTdS3Q{<zvEIGgWLUZ<
zWm>AU_L1|r|B-4E29%bgproWqYZ{9ed7|v_NEaXOk^SbXF5lbKCplijS2OFqs<mc}
z!EKk#%uC>Ok3K6*%^C}IOo3lbBD{7DfOy<eKeE1|fxTJtwzA1vBQL%EwY&&P!dr_D
zysm9&DKD#!-@kJXFy4vZFK+1O+uZe#mXRFkic`~h{mH%e>Pi5$bgc&`S2B%LhsS#k
zu9e3d+vr@nZPAN00xxshHbrOVJ2@fH>LIG5gL+h}F)Fa9D4vAi4)V<dk+sl=2ggkF
zY-eBfO=z1i*IPoIP)XM4LxahAZral7E?BC^hA3MXr1xmYZT!mj{RpXfr*{6ClZ*Rf
z;2DHa>c-(LoI84DqRi}x)+JDDV1?=6A!EvXG!mbgpgJmQ>KW{cZCr&k`5vK))Ihd}
z>+=U$-5wRfrRsOA!nIA~Yj#lGn_lgh{F6i*zLw2P4NzYTOYdGgzIKamWeo0qzQIMu
zQZsPYXf2v|pmDa7tP=A&k5EC0Jn578yl+*9;?0=t<s6B1f{z-~wXn#f;ul}>Np7;!
z9*B}09R{Sr%e{L^?nJ|@Z3Jf-8eH&d(QjI%TsJAr#7o~jcqc1;R`)t?`IK_f#fkq%
zG6Th{*y<@O?n_TRwYB??^=*IZA5>)L5J1@0M<|~$NdoPaX0H~&CaRO1!ZtDQ_jQ?W
z-MYn9U@<Q6;8Eo04Stn%oAGXg`y~5rW={xTy&WOjG3mm;22^4vb^uL27oT}>cQdrJ
zi!O6kpYgvI+!hv4sMxJ~yZ7zu*USKxO6ipg>kcdA@3e9@_8$X{H;jSw;JzBD2av?y
z?wEA)L>c1Fv%D^6e|9`C&6SEY9iQOeOc1pBvh@PF2O}rV)3`hOzD%0&NP_4m5b^r~
z<l-+Mp2&OJS1#Tx5s^h_wL7&h@6@EbRNDs|O)H{Z6ZSTuMe7zncCB+H_Se$Ca&m$v
z8f8T}y{ElPDpq#-!|6*2@#a&hm+s$;0=7f!iph`j>J`b2lfi#<**K6xP@rnG_~)rW
z{FEIr#@iOnh`)FU5ruv@%n%ZwoCc?gCL#Y&5<%0#U+9<^Xzm@fAzS`J^G1-M{wcrr
ztns`o?=(JN1J}hXOFT`g7i|)Syqyosqe}13u!=_a^gHe6ryBA}dk$S-;MtJ7olB>e
zi{aLCiXcC3+K7PFvS)KVEN%%rV6RpqJbi>SE;;D&jc{)Z|2azvl%K_mitarPRh>WI
ztg0$R-eEG)k-=P6LnMYIUiAipz`prguuzYlJ4azC;!Oq4bVhYG_kx4o2nBfp`|USG
zW*a?vmqT)Cl*1ZQeGU%rN@Q5HL``R=32QpY#4Hzt-&Km#=n&sGOr6K<>#KU%n`#*y
ziq6S>RQNeXdb2Sxsjw*GGQmubLV$ug4&7z715q|s<FKgEI)7B8byx_oMp2*XNLv)x
zH-P$HovOH}SHqzjp+t>muf9$>39UF=wz(sDFyeX=SorM=Pp?5^LYaHT%quizge}?l
z0%TIPEVXSavZyCy1k3a^CG6lZ{J5EFpb;*yr>Lc`bmaFaQ`17C@+bn{wyr1DoQ@8+
z-&~TVevw&0v2OkGN~@SV<Hrti?r5SGLER*?h_yQZ{Vfwhq7G3?wQ>;_$|*T0+;}fL
z*ozS{%GtrE#TN+AKNOj)|3b}?MJ&!71=rX#?ZT|p`5AXidM7%0idtFBO#AUBPJdXd
zvkIA@q^<F(ulRtHAlhf(ekDim>;Ehma&_8+=OFdM(gU2*j~}a}2>3xah-F@sOOfX;
znjOIDgBdc8!;&R^{e!p5qkVrWz-k=%FGyR})gZGnK*6G%FcF`LbuES;c-)_e2*}>O
zm{Pk#dtflRj?ZPfV)4RIm10>|$^$p=4vilx0;6bZPd3N8#&`g4R^(~jc-#5In(=ia
ze*cM)X*l9dxX3Q!k&&UJZCPja;EvA_J4bIa0T+?!Tkkmsk#T*`v2M??vd3XxuRBRa
zR;M$0*B7ztPq@u9es+kRaF%FUKQpqgP^^TkE)`^_Z67n%IL_Q3thEa=N35IWbZQ(a
z#EH4A<-YRm_Y*@&Ml-FFOC2O1F=jbhjN8m@+oMbK*s2VIf^})%$FmoeMjkC}dIGki
z_)x<e()BKeKGiSW4};mV!jI_0YE+|mO{1-w*rylkq!>+6?y@&98ssl|f0n$(L+&kB
zy4y6x&Id`$Ckq6Mz_${(Zw^$tl6xM`M(HxItZqizZPLXsodp(c$d*}^VqK`c?||kA
z;a!fB-0-_hs6hPtX@9UXHkDS19pst(ni4h6)aqG5=||v>X~=^|dfs+c0c+F1mzD2~
zFxvx#Z%VqeOcc7l6Z~9TKI7>Rd--9={8led3Izc|TM2~*LX2X8#R!@K4+4u~&HYMb
zy9%B8_5(5*kD%t}&b(Mi9Z&k~%0m{MG+)2=nSr6q)ejc;``AQ$dF(;rLJP($rUMlO
zUtdy~8XGj=6KKqVZutZwF810k2cCQo`k=t_oJLKK4feq2<9E{hC?}s__JhmfA3m6e
z5{BdXCQ!s9?zcSX+9<vTbl-=72v@+an@*X8wjb^hY)*5+$_ZCo72jUuZPnBheSLSt
zH?n(p>K4qT>{1C1{!=iEc{odNTp2B%kYu_a(K-P;jE$e>%~*>vc%7G%Oy%vPa0?K<
zQ~cL&!3Z*m7F5&`3rsY9r}75fH65+eH|$&zbq9x}?CLPucVdI)>;v1k24=3)w7=w2
zdV29p@iXzb(+mvf#@%a46D-f<Sz(LA3K5@={D^42i?G3_d>MyyB4oh?ab~to<m<CR
zsngeCCq}XzA~Yf{aTg1tC@Kh7CP|E}*@#w@DNuDaJ_cV%;0G0icd8RLZiC{2KQd71
zHL}+5`4PRd<ZXM;!q22oa1(aeCmF592k(fTKBZ6=Av5;T_6^WJ_9DOe9&xe8_$QUE
z)QaLg%C`X}8n3*%zXGj}#A&+%-4kY1-h3QS`S$qfRZ6=2>H-T3bHA}yWF?F4Io%i^
zB34s}4KY25=Y8ylN6Q6piDwTY5`kp;dWIi6rd_SAMAw?$?vszVD-RX%%D?Wn5&r-q
zlmCEidXs(03%SYQO+Mj$c5?Z)@GBPhkj+5Lad*c%IAD(MLa(iee|A=mjsP<X*+#bw
zs!e~HXsz@2#}==zaojhdqoaF;wkvyJ7`q)r1Qf{JeP?yb`<3Z3U{HG&g#i6<(k5ET
zT0Wjk6g>;J@}<M<mDQf8Q})%o?(s-)eym^CJ@Jt}@WD^b1TkG|?#CNv=&Cr12N9Lw
zGs7m5CgNiP;#K+GQW8b=#7nYNb#S7}TZ#&jgsB;WlCCDQ;Y<>)l$zrT65Oc=Z~2mg
z)(OaYr|bM!)?}Z`Qja>CSq{Z4?l0Z)GPp7(${$1dagZ|Rq3*;dK`ySe)5CC0OXvN8
z+UrGK?HcqG)JomMd}Nttv8BeysMF@vss-1H<2m8w<=oZiBhjpfvz2>~DP(Ybg>Z>c
zcAOIWBwC0$IT;*WJPH{w6n`hYl^x$!yPOoNl0j$2yMN8sju>TGIoXp+E@n=GI`WxH
zKlBmK_c4h9?Vas#f|oB`iaDa5jdRB<4a;yCH4|d~(o6-@-RU9RF9$(Jtvz==q(I{K
zBg0bow(4p5Sr<-<(!I=y!Q`eMx0n2)xbbZ*UOr6^92BJH2FKa(d*PL)lQ5L|<n0%v
zFBOoWYDh*``Y~#01ueYb+PaR{$%_MUeA{AuX8%1Na7&0<R&7VJO<m(Y&(?chaGXr6
zSGgvx(LDFP&XVz3xDq&yywERr(a09p2w$tSCnf1w%C|&b12$M^xAEa2^+`t_3F@5^
zxSIqOdriB4026fpf;<DZi?>JLNx0I(r`vBX1*%fQMN{#ONIixWBYAG>1d$BCIBN*6
zuJfd8=}-oDB{@7qkK?Zdk@q?unYp=*QldDG7ogNf^>O1?R%|c}64W(cMt;?nl*5Xb
zX1M4{i5K^6-V)2zY&X}q;$D5Tvu)}fY^#23W*HPNrlESG^Ne`YaV1V^@Oww4bkT`H
z?*0~FW1UtiXHR&?UNi&}RrW`ncGnx`Y~sRoKX^`H#`S3)3oea>jo6hsGA~D0D($+|
zY$tf!=eEH+09@g-leI3Rt~GX{A*cOGr#j)2*mk{3sV|H?PDpj9jv3col_Zyr&N_5X
z_ycU3`X){|i`oWwm(tdlPl31P3^qYA)ed>cg?3JC1BhiO3z8p4Tl-4VN4@SZAQQ`D
z&`R$ea_@tN;@%yx76-5GxaHoq$^tiZs)x^V`=v7=LuP4$yzN1;;bLMxC&lu*LV~!7
zc;!U7W$s54bSkJOKhy=jY_bb&L60;_WF&dFsqU*HS8&?0Di$27GLpP1k?`iYf_Z)r
zVQvjV4u2)eo>9j`>D1e{Kx%Ah{{v`mDapb;2nR&S;v*B~bIKo=>LZy{|9$!P_PjH9
zC?YILuG$J0?fJ6jRC9Z~dMA7l1u8Y&4LbaAYsdHd$RhLVHu+owTlP+!f*L?04Qv7>
z9xH?^Zn)@CP0>3bwZkqYaP9aZ?#B`9UMq6_{N}4L-RjH?Xn4GwXA2dY`JOgK74yrE
z@X=#$0lU{LuLqj-m?!=3t1KJpIP!ZDZ!!U-Ml}s5M%|o<c1rJczg9<AY@I!6L*5dK
zNDQB|s}*+Getb95Mh6#N4fCf&ZQ`%Y96^Vyh&Q7FIx2cWys3q2)b$?eO}x2LJGd$u
z%n6XRh~TqV*1aBjWieV0MZRa|-9J<cMF=6Iy`&d{HBv@a3oH0GJ>9JYK;kg#rLDs}
z{Cc+cJH-+Z;>`ol-WW$?=fKbR%!WJL$xObCG9Olb4?h?df9_b~XslvGuA6BiYG2$h
zCkT$2qghOhnHB?X3smQSf)1hX=l2ZWf_74I+;c8<qczT#e0m>3Q2RPL%gHL;$1dGo
zsEO$8)r<SeW_OX%F><3J2tBEqZ%n>+H9=Pr(`)XQaia1I3VXU=(amyzyVvvbI7Z@v
z-48P9bZI2!J83p)aoLh>T7*GH>q$KxrXRA<9CKQ#OQV?Pe%niCXGh2!z!iR)+kRAp
zR+8HKxa7xt3_giB&;F1Hf~`q=0;%~&BuTq8939>Cp<ld!%0<qr?t({RbL}ox1Ont#
zMzswk3y*<aD>$OR)Ie>UdkS@22WcGdIaj>jhAh3F&0EY9>?|94)#J2xEkrUuj8z}r
zwWrkNc|zaT{9<lk_i=6eY+t;aYJt9KA-*%c5mcwvPU?Lc9}kD#cDrh))aewtcjmhH
z3D6ohepj){KJ?`fJf|eHpiQ%B+52=m>GKGu)oi}S!Z@l@XeWt#7n9SO>}DM2gfv7Z
zgr!hUA8EGeN$)raKsTp0TbS3Omr5DAgfjZjCA&OGr*B4$b5ZRXJodI4-^y&3jq>Mt
zUleNLVk$Yn6Dd9k%kk7#qCgP-j5i8^WCtWUt+C9R7F-J#-EkG}o!nbJ_0jMe{_r{z
z>Lb3blS7M%HYxC?9;vfr@lqc<>-$W;lC>>{CZv}_CI<r2$jx-95~Dq+-%?+dG_7@O
z@C=$d!!dKbS7~xq0?q7|2;1vc0#67r*WcU!F1-dNw+%%MJ=o+PIZKX-@bH?@(UdQ4
z4?_&WmE~1Wo+Qwj&}`QwNpbhx+fDHRFGnAJ$U#bCnZgTKF#%Y^67T`P#{sW2CIan=
zxGGj#3A@L6&y#%RFmT<$^=D|a$en=l%|}y}`@n^P(3EFh<hI*VB-P%#;ZP_d<9S6{
z={@}PvrFG=%N#t-YQ?{0p1tqUX+?H;Tz{H#eVyk@F)*FD6hg{{zanOO+7oZ@@jfEa
z8<lU4nN4`v`-r(_{0{RL6G2zZQD$nMy#a+b!-^zGe2)pWOEq1oUvd`Pd${gRzOu3I
z47}Kq*b!=-KVfx5N7trbv+V&61jxU%cOM}X@V-VnD5spl{fzrH`N|JU)RrzK+tVSj
zfO1jkGn@Ap>C-10-^eiIo+v!QgD@9<x7+p7=RcyKIJm!yjii(1&+Am53}i^@N&+nf
z_ZhXkkxMSor`kp*qkY|ES8v7z^k7UMqpq}#?KOEVw(lu6&GmH8)K^B{U{q&5T?W`A
z@!`IM2O6E-?KvlqmzFuZc)B(y$l4uU`{W2BhfEH$3Z4~fE;0#(V$?kFD9X!m@K=KA
z^KJ2nQCuD`7HP2p6bB@W3nhG+0xCWHzzQ4yAhyzB*>Z!4hcCxHiDy<@(t$4!l@@Eh
zt7bY2y~>M&33A4OR+r8p_x5bB_Q~UDY)Hcm<^acGI54Wtv2}X*D_+_`6BU3;^+m+Z
z$uosc<|;ujww;@TFC4rkW_qj<Zn!Ji_rY-o<4ui{{7qp|W|jFG?s3@y{aF>l4h0_I
zI62*ocG5W@Ye;TXt(QqfcHab^_i!6Hat$*Cb;Q~eZw^;-+u?B@a85GSLcEkuKh!^?
za^4}iCzpKPyv8SWKZh8#O^J%#8a+E5R8Byj>Eeo8;+1lMrEP&#1sks?m}&>@R}kzm
z3V_g_Pi`Kh6t|{yQ%zqRPTMzHm`{)r3FQGVTTR<o-dlIYUpbV7MZF+)+%8T(eH9)f
zH0-jtBrzPLA~&zLK4H3#7^0fC)bX+|+((1~f-k>t_BPl-CqX;tjBg)7W<s-4Yvu#m
zJGm(@l{TRm%Y^Otm1inc8~J@<KK=f%*Isc_-uTv+GSV2b7EZNE0_g);Xp#+PUF0Mo
zEvrX*^PT1ZTMmh8)N3q!yY1TrD)qvJoXAH|q8^s2PFP!c2)Jr~|F+J~_y?N9@zujU
zVLUKww-yN1LF>wjXm#{XM8ukE2jHEVpU&k9ErX(oQ&SWj1>a}~#P+0mf0uqaJbG5a
zRJZ7mR~}2f+oQJ4P%QEEzE~#ss^Jc9{qoJV&6Go&F(WgHyxK*tPAc46l#bz=VI2ci
zQ&}~ilhRp+s8;#YxK~6Ym8^S9lKBW;pTwmcx{qZTnpp=C8ikh+zJ8m+@ZC&1e}scH
zuBB?f@YJOE`F&)Q`p3*(T#s~EI>tAU+Qi~Ly#P;gy3kMwSXyVRY3N6w$ra(p2qD=|
z>p3nN#5+v`yJdFM**y$~PpiCPW~*bBcIhl+W+a@Cnnpn~in-#JoNi+cQdD}#&ujFV
zy(wbeHHwGLr^Z?gp<$P(Vd~<;FZOWq<by3mdP2w_BHwTX)p}|77h`6!&D_u0WX-;8
z@HQ_aLJ(gH?hp8nM_s#-OPLXeYt+<BLv}CEYayP)tTUZ8aMFBaEhcFuc1`?Kj@Q_N
zs>qQ6li`k3KQxw{j2xuCD}am_*!H+(c@ZD(;4lvBukrAXj<ad+NCu~;4Rr_hq?X5O
zNX^aHf2n9@x?6@5WI)z|(Jy=4Bf7~`G$CYW+bc>kpTn2VI56MA?Chm*qSg~-393C_
zVsKc3rc<J7p02OOwOs0com?}~BN}`l*U~r|<%%?}ayRp*hT<v;yk1bfd^)kbjrk&-
zxi~DpZsw!~^;MImw*a_;@2jM>B{^Y|fPh7Cjl?sr{ZT&~TPA+9pMZ2)uq~dr+T(F=
zp3oNVAYXgsllJ1OahVSK=;SoMqM&3g)Hb8{OH<tVPWsYz(wXP(heJ%dje&l#P<S7Y
z`L@m8D)0uC!Ts8^lO;ay!}j5IuM=n*#-Vl}z8+c=<sx9Y?Rc)l_Bi%uus5`?4umkr
z0dvGc+fK?~2cwQXH4^Y`;hD=gR50k*oZ@gwRE`txBGDfrDwyy!)aPq0rA3t3+iY?}
zP(g4g`ZMyFW&h0E))>B3gJP`UV<=`z?5uP#ZXzk?guFJs(j|VqU)0od^h+jFFsgh%
zo=>yd^kVv^q~or6ck08YeX5HRcH%RoKbdTj8^&nLZX~L6iRHg@)JxZ{MmU#h8OMF9
zvS;8ET7zB`vJ6(VRSPKRvsaby$Ruehku@A@offN5Y7s0DucD4U6y03O518RJZE@B+
z8(KtJh1($a69lT#o-romxaA4wTR2XYYe>W$1<Z`dbcqS&BN5HjahNQQt`^?$@-Ri)
zd@+cPag$BO7hZ<4wBkafy-mJ-$J-{w?bU5};Y+1ReHr9-D!#3W6F<^zarLbPCO?1}
zReLm8kezh=^OlvF+UG^_iNmg?7^_60^tt9`ciAO~NlWXd=hA4)dn9r$fW|bl1~ay<
zXUJv6a|HUj=6wQ~v11tg_QIp+qjG2`Z=aO&nnhDpkn999Y_MgZWfi?1{9Q8v^LUfj
zJQ?Xuf-<S&0K2s2c5NT|m|KZGKT9@uu6&^R2k!VKc(AtQxk1Uol3av(CP3f#d3A$Z
z!B=a=R7OU?=wv5DI?rprYGSZBFq1YV#w^iut92<7*a3*rtu>qQc-6r>r{-kUK_7Ur
z!jb_l=A#l*R?S8LpN_*o2>l+3R!-lOBVQbl4I2!=K;MiA2FE{oR~DC08c&7Y1$hK$
z@|wt>Pya7}SRVO+AnZDCEjjhD5v9RyK;kZSd!H6pm^6K>)XGBJc&=D5=SXe%l0OyU
zWQ|gkK{l4v?D9B?b31vE6fd~DEgMZ3Lf4#<f-CXci*_O<C}qGWc+{L=0g#6z@29cF
zbZ_F;U2t;Lxi<aRKe=$kKnS9yjc}!|;KEgQ0L>mQ%}bxT?E=gK9~l|Rww%+uq@Qn`
zmFT{dwe;>*BCW(Nvd3Wv2O!tOo!HTY#t>*Wu;gLpebk?nU%R5Ek^b}zVEOD^N%k4*
zw#hRD7LlhRoS^tGyFUMztm{LG%H#Hv$O7##17_J0P3?iAGXR=jI5`#bJ~FL><Q+}Z
z)IQd{BB^|x;j0L4Xyp*J7vNvB#Ifb0g@w!`z?3tE9yBmxXEyF*_XlQHAl5>J`5U%a
zQo)d$x&2muKJ~vXVC^dKjBxLRCxY+xe7SFw<9ZS|ahyJaM(3<Q&%M3}_+CsW<{$6Z
zTN-Hk0%d`E-BK9aU1aTPNvzsU!TP#7+1(C-Zk=_vZjU8C@!Skj_<bB-U5aEMII$JK
zHuG(;)tejKe~5;EId4^VEgG1*@5<6^DfP|E-)eTUUZ^arL)Se9fj$nH;_Khu53K`4
zm;+duN47s#`&>Se?NA>8&%y+{YGWo&A5(yX!2qRHHCvyT3R{?#8N+F?80iL)7J_%t
zmg)-${z|<D;Uf0Cf0>z(<<1)%Hkgu3t;Y{+mN8&0hi79uBF1WD>=sv^?E>i<2G9k0
zX7kvEO^wb4C2TvboC2@ioX_`7Y;<bU<7zz*U+sS5gJ$(nZF2*oBklrT8q>MrC{Y(;
zRE#fC_;2R)L0l%(LNKkfFD5p}avCbCR9ROpjE#*gEe+=EaSYkqPWAv}+j%)p8IDb$
zP?;pdz>9Nk_XieC?aD+^J{H8{iJH7LW>>rRU4W$$(jrSuB+BGK;^??R%N9YH+}KyH
zf7|H!<|iqCm>~9`)MIKnJCIIeCu!cX>q51o8l4A)qMyg}hciaQRm*}p`O35fy#ABy
zf5q%V4I6~zSWv-Cng>S#7tD)9efNOh?Nf%?H<$`mC1n8#H**&ko`^@En@9aWLK(4o
zWYlYR^n3yF_w83u@hT4UdtCF7Me4%C*?X@nTIO`z#mmf~WpMnJd@wM~Q@we&P&krs
zCI8_;PQSK0mZg64Or@}MLLG+#|NQwo5w{M)t5uD)riE(SE|F`Ofeog=^tzAkw_HGd
z8Q_Agp|6!i@RFWoH_AWf=urZvfP~m!7z#cI^jH?1uia_Oe^|jX&mBShl^lTAFxQj!
z|5zg_y9}-e%fd4WQZ`uBX6{FUKgazWQ9j3~^-UOxwEnfPNt$Gz)qR6ycvU~qOabzU
zIC=YbJ0{e*if@|_`n4TeKz8nZ7*U0A-v{#z6S0~6cAIwzGks1FjcTk?vrH&DSE6^n
zPAd!t-<H6?(+O}n&QnXs{Y^mpQWanm{3VdkKEf2)uUSiZu0CjufjK9CwmZoXquQPF
ze^C8Hul!{IUt<sz%&GM<&95Iv@We<^85tRa&CU1GHrdN%WZB|MxAHKqublB9rNTL#
zVK!iApwr0j6&B*whbo^{40mOWK7dFn=Ddt66>YsL5@I(koH*eo#hMk&QTCK^X7(AD
z$5`4zu4u2doRc3qi%uGAuM1*+O098&t1iW%`2Zg&6aLE0Ruw6Lo>da{#C}iDe7>Xj
zE2gbIw7<5XXZi<G=$}?*-rF-}gH_$1DyYGcRQSub|8f+7K$;W#6Y{g-|Dle4Xx|$z
zr1h0dlvKL^yW5hfaIeDn9MrS9&_>_BDlq+`09ZYU<SLBL7hbQgkY-b*EdZ)w3LbNf
z2>dODu)rW7ZEZx23a-a$mpJ1e6M;#6Lx*i4Bk?yYm*0+e&coSYyUr4w(u{`Y=z!^v
zLTRD>!jT(y0i=0dx&AktM#utd1RX{9Gy0AlSG(rsSBRFQh4}W}T*URo>5=(WVK)2Z
zg%GEgWp-q1+iQ?~h)})>&^Kj`*?e_w(jLt)IzP5*ylQJ)wI-H7CfzvFSvOWHTHalS
zt{53Ehm1KJS9u4O#Q9a(IlWABDRU|*c3iDoE{B{^*a#E2pywu~*6rtCSj~($BOO;y
zS%>V&)=E3;pyOj*gI&#ov6LwFOH31NF!2NhD9uzNO(B3oes?xpH+u9CrdeeFkO#L_
z0PyiJBQ5tSt*^<@Aq)L)8K{{8kQ2V|V&pvABw;NZ^U;5dof`+(>(1gcl1HkzRRq-P
zfAGh@psmLbeJg!BC1LtID}6kFjNS%d1Cw!4j+<=6sHS-0%d)>C1@Jw9Ez^3Pz<=B`
zgynR@#JzG2ZA6$;>eb(FFNYftJfFLHp??6aWC>pYhw_byW$uM(v@KtvyKyzyC*e~h
z_;*J`$o&o2%{K`<_R#aQL-Y{?qO``I80B&s%(wa={DEh~y^-PZTp^*IApE}&`d?0L
zNEcXSWun{BuOUSKAX~|<)@w(ZAqXbJ_!-u}$-rL=(Mt3oMn#p~9Q?eRw-v$0_G`$;
zqS*C^AWSQklS|~WWRE@$|4&mJfnN}#^f|x<`-(E_za*71*g_~SK`7&$l%%9xf>7mQ
zwF_QpO&E<fkklYdp-|A!I47l7Wty;P-$A%yPrvtjy662T-E$M1Q{(J^MU5rP`7<LY
z^0{LSqjqtAnChy8oP<ytq)AH&Ch&0;mTd_Bhu8+dmQ`7GGah)5jL6LO>!aJ(g?=wT
z!Qb#k0GHCBteD>~zEH&CrKQ)OJ_9PMU5xqZpWpwDG7z*NPO$VHe(fMlz!gJ#%5dfU
zvHLn-d7SC9njq&nhNoX%O9j(E_}M4;7c%(gBCg}#1}n++lKvu>2tt6QPP_B`mb!X#
z#RShR4AgPQnmq$K8uhn-_Hq4o2zO%RntNFgK7LTuE&9%1EC$BKQbxh3=so<E$fc+E
zaXoOizqt_s>UK`&<L%$_#xS0ckWhBJQRIsjszQT5m*_i+B~gc!EFZ}nGEK{hZMM@O
zKyxWI&iDN7DE<^DRM%l{KpgBL3XA{Fzm&1ZF6B&8B!7*8N#COdP<u(qh0{pwAO{D%
zBghSs->!x*w*c{l7=<<(d|gygL%756>x3?1j}~r~PCg+XuPaO*&e;~{mylCO-Uyf{
zt)eWy9ZQjD;bugCtLM>1p1RZ}-ld-(NdLHwixaTKJRt6dnyjZuMg*4rYwz27^4^1l
zXmEc!&@xXls1#_2YI7g#0;TeNA;lLK)5`B$#oDcgHB<wKD`ZwgMZ6VHm9Pksw|3x=
zQUySe>(xq!%uN7-t2%x(uXG!WGPjv#y{Y1~^{dWx&hboYP73C@CPwNQaE<3gL%UNu
z^J|~RUa~+7nrXMqphjC>ME!UN`ma<Er)f^mnDJ<La2)2X`wCX2Nc`VaDgXOStiK#^
zAG{o(bSq<!%Y`iGWaf%pQEoC~?e*nd6aT9#VkUpZRT4?YIyg9#77r9;b^ank*^=ut
z_0@4b;xJ}TlE3H*Ap;_zfeCIL&7D4UZa8i!iUk3Zuyb3b;7{89%K&&3L06-rF{Wm5
zrrAOjQBS$)R)1YJ`35u&f*?P<^yq-Isk7CXJx{m&G6*OjxcmGkrr%2A9;L~x8%}4-
zszy-LSDNS^tBz8~#Z}ux`CP2M;6TXb5rPF-Rv3ZCN~~w-Ev%>|s55H^QY!s#mlQ9I
z7$L~;6JpwSz3rKSAfXx?Q;}imNoD9EnRXB#`Z=4~Z@E5#$a-jlAOk(<--~QjBf{FX
z*mHT}3vlbz)7_I9ch1c#og7_7kFckb)&Ici_oqVLgF`d6fAl^^$~gR=OrL_sYYY_e
zbY@36FzXSt6sNqJ6S)9LL?*)6hTrM{t!ST$U1v$>gV#2X9$S9>Q(ETmzMKM;`q-~A
z>z+wUWQA#1deS>6=102|Z?1N}@LTlN`x(-A>9-r4#@mURd|mW#)z&P#gM0cP;slWl
zdO(a45P%+B68L=Jt<D<vu8}~}+km{8a3!}IPw971Az9fsPhG(IU{9~)+mnAvCIkXI
zRNJ97#1LNUqBABc!BkGdz#9dqdE=tS-#aAKeBW`Hnh7>71OLU5hjR%iTHNW??&r0<
z*)p*D=G_h34`Ta;(>K_O-d@WI&kCnsv97<E<;2WoDi;CHJa<;rh28w!MKqnzb6tE^
zzCgSn<vmflIB?qg04lw5CA5L;Qq$1s#)K46*5$YFJE6Idca<u*;*{)9e)~)xU-nH{
zeNa+da<yaV#8c$p4e6y(K-GFE{|{-^|5Gq3HcV)XuMOWzrwf2?PnJKjdi=!*>>_3`
zvQEogCNa$RI{<X28u7ZwU^TWE8Aq9xAJ`qncKwgb5S4_lsbN*wi@YIdjc3dH0KJxV
z=uCftndHrmj1Q0h3?G~od-gQ=PyO^)>P5(bDkpXa7r(u@E4Mjv5>k<K;ak~qn|RCi
zxoOmp&Si<U`<Iu@!sgFC0fFlJ9p~oLU@b%Uxj{u1mmhuZEy(e9uPOG)O9+OnG=|i8
zt#&;u@%DVg80TS8%$<P)aPqtxo6T<t-5kFoAwNeoGW5lKpWe6Ykfmq6-O*KBv|Oh8
zQlg2>SSpIcF6G=6ik&zd>)-5r(tlL|>+(EbUCWZi(##xe@qG-I`kZ(9_S_tB#f6lO
zs$%H8;V$=pN<?N|f);`FRC{^O1JRSL$_A=w{3O;#zjZiHnuF*qzwcg%x;Z%8+St+P
zTKKs7+MmKy`NoGlpFBLlJ@@sEua!^Uqm*uOJ}>5u1UyGki?$!ReycbgLD{;?Qpo<i
z2fGOBGRhF`v_J7V>5UI1K6x6M)==~)xe;4oxTXe&Ik~gt6^1|%1f(GM*XQ+*X9!3e
ztV<3$e4HUC^bh=?6(O?+DqYP@r<@SjZyXmF%i8N4m_9an`mD`F8^OCj`ho7Zi-Y3b
z>&awtwp(+7+oh9EB--w|iQ4^Xs`%F7UE~7iC>^ma0%Bt9Zd##3jND$9TgUflzl7y+
zf>fWexR7%X+n;#npR0`bS#y>Y#J35pOpKmFe##FPS3Gw>V)t~(_X*bCxBR(#|Jl;e
zU`oZUJzu(yz#*_<9tYSKujfYR5p)>$i>S{?eD#qG6Ic4JIGRia9l1I;Q04I1Mc0?m
zIm4e3qac+pq0~9u&*fDYSLF^fM$NFU01kDzLUXJuATx<$|D1AFBDHd{v`n*P)pl}D
z?|tl4=b`(6qnKA`iabJ=#lz3H!$0Ty0QPF>>L&oGkJ$<f&~utH$1XOlM`NBcM|(W&
z?qJqz;w!}81B_$PSDysJh&%sw5LixqdfD0w?>VzW)=PMWft^E4Y6%_3R}WKR=?nJh
z(_UW7_QGf(=9W^Mnwz^8SQCP&`<LzdzYjG2uNc`SOt0}E+lbDVY=fleGi`p1wNOr&
zIx7Yh(lc}$=#-?pK!z;Yosap_BTV4*-9|k9#K|#HcmHhPj^>2}hNg(HOq>ghPPZGg
zh2&~hl|FWdqPIIO791xaDSJW=O=OG{rHpANXnuL9cGCA9>`jbRRXlu>1vg4fiu3qP
zCm?o(a?pPM^WUUP1C9pfQs!6t1_SfwSC>~R9jdfER?C9<_EC`c&(KHHyLL!wt^r|m
zLCvb%s@)k?Ef1fEah@HTf24#_U_dy|13T}K>?V5a1rt(guSA=kUf>z9tdyEzPfRIE
zN!f|^acT96V@au9i!~C`p*3N<21N&q7!fo%wCA`PV^By9JKK&=t0U$CN0!^GXx@SG
zJxyS^qL#stKX+1GalU&-VD^GJ$5jh6@sdJ*?NXcIt4nd6W!NR7rT?4Z{;So+wv>cJ
zR|lG#)rRwV^fDcUCjPP6|C#AJYhqfiDdasIjEkinl^DOPXd{3cInJ=tuCxz+f@%DR
zD7>wcT<kNZtL}pNSeY>09XuGy(bP?rU0c3l=4Yi@DiLIhJoYQN(Kf8TU$PvuZijSL
zUxHY@E0wfnD6`uib{bn1TB`lo-5qF{c;+m<9<nzA84;ea|6WCZT5ZDEIHBSFgsa-X
zw)(KQW4qI}4&q*Nnk*fzPmq_!`RUE`%`$Bgbj!BucF6bc8q@ID)j9P1ne_T6!Admc
zHZfWWUTM0v7Ax)GBvEE<hg?Vt6;3y(;<21Zb|33>pLu(oEIMu<#h8^vA9j>?y&mi+
zNP>{8mXRQ1J}=k|*|&T2?~i{Hl<q!)3s(R??+DC@Eju(jzm0K=8w?(|n?Jk8AHEb?
zwO-d;%hhFsPFlXyV7ht$U&O#$J&GMo25BrT;+8v(-G9OjYL(Vbd$PT~1oKI}66t6$
zY<kSu*XSP}enE9RLHQr82V=2umWk&k)>!IBwhUm-Mb!%$vubRpb$ge8w+1V!qDGo%
z3eH2VP&GIo)@yla40ui5A1;L*UVZ@Q!>U8o)#+eiJ!WlYBzBxf^HO@%1H3h^7Qh&e
zQ0#60Utd&}8C>QA?VYS<`cobMi!<%7AAKJK4zj{W-+!y){|=LXaLV9{bMe*vsrCMg
z(aiymMT<%=*?)1#FI38s0|4p&1j+wCf6!Ih?><m>cSAc{TRbN&xC!h65&b8u#~KTp
z&(#zbNS~{z5A3v@*k_L~9t)Gj#z(MV`#qL~u*g~qtL=y$yh%AHAz0e~Yr_~1nD>An
z^<Gzq|N8?3H{sm|7odo~iT_rm{uLS_iiiOi%qR98jsGrmzJ_-mf<><F&wn;!V@lFx
zW};>1kpK-g;q#bM#YtPmdD!XIwTeP4w(VZk?^-=)G9Vyjf;~k0+E@12uu~N_Acf^a
zC(LIjaqP)s15#KFk%TUYN1uB-GCtB*VHpz}#0tYc3enI}6!vMI2eJOgmLi_?&<fXx
zPi2hX?0VMzw=QU|D^_eYW6|%?QM9`m79(S&ghQ@Wx>SBa`2X)0(Kv4cPxV?n^0!9O
szZgmsTmTd6y@&j_#DW7dz&~XsxoZabx-ndb1N?iSsG*QA2Y&Ov07QEtFaQ7m

literal 0
HcmV?d00001

diff --git a/task-guide.md b/docs/task_guide.md
similarity index 88%
rename from task-guide.md
rename to docs/task_guide.md
index c750ab5a1c..f3b2c986ba 100644
--- a/task-guide.md
+++ b/docs/task_guide.md
@@ -142,41 +142,6 @@ def doc_to_target(self, doc):
 
 Understand that the strings from `doc_to_text` and `doc_to_target` will be concatenated together to build up labeled examples in the k-shot setting where k > 0. Design with that in mind 👍.
 
-### Formatting Prompts
-
-If you'd like to prepend your few-shot examples with a natural language description or provide a lone custom prompt for a zero-shot task, you can do so on a per-task basis via the `description_dict` arg of `evaluator.evaluate` which is accessible from the `evaluator` module. This `description_dict` must adhere to the following key-value structure:
-
-- **key**: the task name as specified in the lm-eval-harness task registry (see the following section on task registry).
-- **value**: the corresponding description/prompt for the task identified by **key**.
-
-E.g.
-
-```python
-description_dict = {
-    "task_name_1": "fewshot description",
-    "task_name_2": "fewshot description",
-    ...
-}
-```
-
-One can also interface with `evaluator.evaluate`/`evaluator.simple_evaluate` from a higher level by simply passing a JSON file path to the `description_dict_path` arg of the command-line interface (CLI) programs, `main.py` and `write_out.py` . The JSON file pointed to should be structured the same way as the aforementioned `description_dict`. E.g. for some file at `/your/path/descriptions.json` you might have:
-
-```json
-{
-    "cycle_letters": "Please unscramble the letters into a word, and write that word:",
-    "copa": "Given a premise and one alternative with a causal relation to the premise and another without, choose the more plausible alternative"
-}
-```
-
-which can then be used, for example, in the `main.py` CLI as:
-
-```python
-python main.py  \
---tasks cycle_letters,copa \
---description_dict_path /your/path/descriptions.json \
-...
-```
-
 ### Registering Your Task
 
 Now's a good time to register your task to expose it for usage. All you'll need to do is import your task module in `lm_eval/tasks/__init__.py` and provide an entry in the `TASK_REGISTRY`  dictionary with the key as the name of your benchmark task (in the form it'll be referred to in the command line) and the value as the task class. See how it's done for other tasks in the [file](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/lm_eval/tasks/__init__.py).

From 70f92733af76ae1c8ad1524f4c1712f7ff9c05ea Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Fri, 17 Dec 2021 01:41:36 -0500
Subject: [PATCH 16/65] Fix doc reference

---
 docs/description_guide.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/description_guide.md b/docs/description_guide.md
index c97a82ae0c..3627e23e6b 100644
--- a/docs/description_guide.md
+++ b/docs/description_guide.md
@@ -5,7 +5,7 @@
 
 Task descriptions provide in-context task instruction for your language model. If you'd like to prepend a natural language description to your few-shot examples and prompt, you can do so on a per-task basis via the `description_dict` arg of [`evaluator.evaluate`](../lm_eval/evaluator.py). This `description_dict` must adhere to the following key-value structure:
 
-- **key**: the task name (`str`) as specified in the lm-eval-harness task registry (see the following section on task registry).
+- **key**: the task name (`str`) as specified in the lm-eval-harness [task registry](../lm_eval/tasks/__init__.py).
 - **value**: the corresponding (`str`) description/prompt for the task identified by **key**.
 
 ```python

From d34ae3cfc9fba7f17da48c8a33ea87f2a9d27995 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Tue, 21 Dec 2021 15:38:49 -0500
Subject: [PATCH 17/65] Add `description_dict` to results config

---
 lm_eval/evaluator.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index 6feb356788..a838dcd853 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -62,7 +62,8 @@ def simple_evaluate(model, model_args, task_names,
         "device": device,
         "no_cache": no_cache,
         "limit": limit,
-        "bootstrap_iters": bootstrap_iters
+        "bootstrap_iters": bootstrap_iters,
+        "description_dict": description_dict
     }
 
     return results

From 8ebe36b25f81be9a9cd237156a2037bf750d6189 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Tue, 21 Dec 2021 15:48:03 -0500
Subject: [PATCH 18/65] Add positional arg deprecation decorator

---
 lm_eval/base.py              |  1 +
 lm_eval/evaluator.py         | 12 +++++++++++-
 lm_eval/tasks/prost.py       |  8 +++++++-
 lm_eval/tasks/truthfulqa.py  | 15 +++++++++++++--
 lm_eval/utils.py             | 17 ++++++++++++++++-
 scripts/cost_estimate.py     | 10 +++++++++-
 tests/test_evaluator.py      | 20 ++++++++++++++++++--
 tests/test_version_stable.py | 11 ++++++++++-
 8 files changed, 85 insertions(+), 9 deletions(-)

diff --git a/lm_eval/base.py b/lm_eval/base.py
index 442d8d5d8e..5492ba9cfb 100644
--- a/lm_eval/base.py
+++ b/lm_eval/base.py
@@ -457,6 +457,7 @@ def fewshot_description(self):
             DeprecationWarning)
         return ""
 
+    @utils.positional_deprecated
     def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert not provide_description, (
             "The `provide_description` arg will be removed in future versions. To prepend "
diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index a838dcd853..02fb71285c 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -6,8 +6,10 @@
 import lm_eval.tasks
 import lm_eval.base
 import numpy as np
+from lm_eval.utils import positional_deprecated
 
 
+@positional_deprecated
 def simple_evaluate(model, model_args, task_names,
                     num_fewshot=0, batch_size=None, device=None,
                     no_cache=False, limit=None, bootstrap_iters=100000,
@@ -51,7 +53,14 @@ def simple_evaluate(model, model_args, task_names,
     
     task_dict = lm_eval.tasks.get_task_dict(task_names)
 
-    results = evaluate(lm, task_dict, False, num_fewshot, limit, description_dict=description_dict)
+    results = evaluate(
+        lm=lm,
+        task_dict=task_dict,
+        provide_description=False,
+        num_fewshot=num_fewshot,
+        limit=limit,
+        description_dict=description_dict
+    )
 
     # add info about the model and few shot config
     results["config"] = {
@@ -69,6 +78,7 @@ def simple_evaluate(model, model_args, task_names,
     return results
 
 
+@positional_deprecated
 def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_iters=100000, description_dict=None):
     """Instantiate and evaluate a model on a list of tasks.
 
diff --git a/lm_eval/tasks/prost.py b/lm_eval/tasks/prost.py
index d53ece825b..7bcbf1e917 100644
--- a/lm_eval/tasks/prost.py
+++ b/lm_eval/tasks/prost.py
@@ -38,7 +38,13 @@ def has_test_docs(self):
 
     def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert num_fewshot == 0, 'PROST is designed to probe models in a zero-shot fashion only.'
-        return super().fewshot_context(doc, num_fewshot, provide_description, rnd, description)
+        return super().fewshot_context(
+            doc=doc,
+            num_fewshot=num_fewshot,
+            provide_description=provide_description,
+            rnd=rnd,
+            description=description
+        )
 
     def _convert_standard(self, doc):
         out_doc = {
diff --git a/lm_eval/tasks/truthfulqa.py b/lm_eval/tasks/truthfulqa.py
index 9fe941f614..6322625401 100644
--- a/lm_eval/tasks/truthfulqa.py
+++ b/lm_eval/tasks/truthfulqa.py
@@ -87,7 +87,13 @@ def doc_to_target(self, doc):
 
     def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert num_fewshot == 0, "TruthfulQA is intended only for the zero-shot setting."
-        return super().fewshot_context(doc, num_fewshot, provide_description, rnd, description)
+        return super().fewshot_context(
+            doc=doc,
+            num_fewshot=num_fewshot,
+            provide_description=provide_description,
+            rnd=rnd,
+            description=description
+        )
 
     def construct_requests(self, doc, ctx):
         """ Uses RequestFactory to construct Requests and returns an iterable of
@@ -219,7 +225,12 @@ def doc_to_target(self, doc):
 
     def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
         assert num_fewshot == 0, "TruthfulQA is intended only for the zero-shot setting."
-        return super().fewshot_context(doc, num_fewshot, provide_description, rnd, description)
+        return super().fewshot_context(
+            doc=doc,
+            num_fewshot=num_fewshot,
+            provide_description=provide_description,
+            rnd=rnd,
+            description=description)
 
     def construct_requests(self, doc, ctx):
         """ Uses RequestFactory to construct Requests and returns an iterable of
diff --git a/lm_eval/utils.py b/lm_eval/utils.py
index c3d718a500..aca7791181 100644
--- a/lm_eval/utils.py
+++ b/lm_eval/utils.py
@@ -1,6 +1,7 @@
 import os
 import re
 import collections
+import functools
 
 
 class ExitCodeError(Exception):
@@ -138,4 +139,18 @@ def get_original(self, newarr):
         
         assert all(cov)
         
-        return res
\ No newline at end of file
+        return res
+
+def positional_deprecated(fn):
+    """
+    A decorator to nudge users into passing only keyword args (`kwargs`) to the 
+    wrapped function, `fn`.
+    """
+    @functools.wraps(fn)
+    def _wrapper(*args, **kwargs):
+        if len(args) != 0: 
+            print(f"WARNING: using {fn.__name__} with positional arguments is "
+                "deprecated and will be disallowed in a future version of "
+                "lm-evaluation-harness!")
+        return fn(*args, **kwargs)
+    return _wrapper
diff --git a/scripts/cost_estimate.py b/scripts/cost_estimate.py
index 4339b8dbd2..f08d2f37f2 100644
--- a/scripts/cost_estimate.py
+++ b/scripts/cost_estimate.py
@@ -51,7 +51,15 @@ def main():
     values = []
     for taskname in task_list.split(","):
         lm.tokencost = 0
-        evaluator.evaluate(lm, {taskname: tasks.get_task(taskname)()}, False, 0, None, bootstrap_iters=10)
+        evaluator.evaluate(
+            lm=lm,
+            task_dict={taskname: tasks.get_task(taskname)()},
+            provide_description=False,
+            num_fewshot=0,
+            limit=None,
+            bootstrap_iters=10,
+            description_dict=None
+        )
 
         print(taskname, lm.tokencost)
         values.append([taskname, lm.tokencost, lm.tokencost / 1000 * 0.0008, lm.tokencost / 1000 * 0.0012, lm.tokencost / 1000 * 0.006, lm.tokencost / 1000 * 0.06])
diff --git a/tests/test_evaluator.py b/tests/test_evaluator.py
index 85e1449f08..ad12b053da 100644
--- a/tests/test_evaluator.py
+++ b/tests/test_evaluator.py
@@ -48,8 +48,24 @@ def ll_perp_fn(reqs):
     lm.loglikelihood_rolling = ll_perp_fn
 
     limit = 10
-    e1 = evaluator.evaluate(lm, task_dict, False, 0, limit, bootstrap_iters=10, description_dict=None)
-    e2 = evaluator.evaluate(lm, task_dict, False, 0, limit, bootstrap_iters=10, description_dict=None)
+    e1 = evaluator.evaluate(
+            lm=lm,
+            task_dict=task_dict,
+            provide_description=False,
+            num_fewshot=0,
+            limit=limit,
+            bootstrap_iters=10,
+            description_dict=None
+    )
+    e2 = evaluator.evaluate(
+            lm=lm,
+            task_dict=task_dict,
+            provide_description=False,
+            num_fewshot=0,
+            limit=limit,
+            bootstrap_iters=10,
+            description_dict=None
+    )
 
     # check that caching is working
     assert e1 == e2
diff --git a/tests/test_version_stable.py b/tests/test_version_stable.py
index d230112de1..a19cbf3f90 100644
--- a/tests/test_version_stable.py
+++ b/tests/test_version_stable.py
@@ -99,5 +99,14 @@ def greedy_until(reqs):
     lm.greedy_until = greedy_until
 
     limit = None
-    result = evaluator.evaluate(lm, task_dict, False, 0, limit, bootstrap_iters=10)
+    result = evaluator.evaluate(
+            lm=lm,
+            task_dict=task_dict,
+            provide_description=False,
+            num_fewshot=0,
+            limit=limit,
+            bootstrap_iters=10,
+            description_dict=None
+    )
+
     assert_target(f"{taskname}-v{task_class.VERSION}-res", result)

From aea963a10c0d4baa43c6585ab11771e1ebce4415 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Tue, 21 Dec 2021 15:49:55 -0500
Subject: [PATCH 19/65] Format for consistency

---
 lm_eval/tasks/truthfulqa.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/lm_eval/tasks/truthfulqa.py b/lm_eval/tasks/truthfulqa.py
index 6322625401..0eeb364e0f 100644
--- a/lm_eval/tasks/truthfulqa.py
+++ b/lm_eval/tasks/truthfulqa.py
@@ -230,7 +230,8 @@ def fewshot_context(self, doc, num_fewshot, provide_description, rnd, descriptio
             num_fewshot=num_fewshot,
             provide_description=provide_description,
             rnd=rnd,
-            description=description)
+            description=description
+        )
 
     def construct_requests(self, doc, ctx):
         """ Uses RequestFactory to construct Requests and returns an iterable of

From 5855f482c48f7f23efac13bcd38370e7ffc17470 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Thu, 23 Dec 2021 18:01:53 +0100
Subject: [PATCH 20/65] Allow users to specify en headqa or es

---
 README.md                 | 3 ++-
 lm_eval/tasks/__init__.py | 3 ++-
 lm_eval/tasks/headqa.py   | 9 +++++++--
 3 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/README.md b/README.md
index fc6c4e4ba7..53cb81b019 100644
--- a/README.md
+++ b/README.md
@@ -128,7 +128,8 @@ To implement a new task in eval harness, see [this guide](https://github.com/Ele
 |openbookqa                                       |✓    |✓  |✓   |          500|acc, acc_norm                                                                 |
 |squad2                                           |✓    |✓  |    |        11873|exact, f1, HasAns_exact, HasAns_f1, NoAns_exact, NoAns_f1, best_exact, best_f1|
 |race                                             |✓    |✓  |✓   |         1045|acc                                                                           |
-|headqa                                           |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
+|headqa_es                                        |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
+|headqa_en                                        |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
 |mathqa                                           |✓    |✓  |✓   |         2985|acc, acc_norm                                                                 |
 |webqs                                            |✓    |   |✓   |         2032|acc                                                                           |
 |wsc273                                           |     |   |✓   |          273|acc                                                                           |
diff --git a/lm_eval/tasks/__init__.py b/lm_eval/tasks/__init__.py
index 561c7fbde0..eb75aea500 100644
--- a/lm_eval/tasks/__init__.py
+++ b/lm_eval/tasks/__init__.py
@@ -132,7 +132,8 @@
     "squad2": squad.SQuAD2,
     "race": race.RACE,
     # "naturalqs": naturalqs.NaturalQs, # not implemented yet
-    "headqa": headqa.HeadQA,
+    "headqa_es": headqa.HeadQAEs,
+    "headqa_en": headqa.HeadQAEn,
     "mathqa": mathqa.MathQA,
     "webqs": webqs.WebQs,
     "wsc273": wsc273.WinogradSchemaChallenge273,
diff --git a/lm_eval/tasks/headqa.py b/lm_eval/tasks/headqa.py
index 3c66dc064b..d59a88df23 100644
--- a/lm_eval/tasks/headqa.py
+++ b/lm_eval/tasks/headqa.py
@@ -2,10 +2,9 @@
 from lm_eval.base import MultipleChoiceTask
 
 
-class HeadQA(HFTask, MultipleChoiceTask):
+class HeadQABase(HFTask, MultipleChoiceTask):
     VERSION = 0
     DATASET_PATH = "head_qa"
-    DATASET_NAME = None
 
     def has_training_docs(self):
         return True
@@ -31,3 +30,9 @@ def fewshot_description(self):
 
     def doc_to_text(self, doc):
         return doc["query"]
+
+class HeadQAEn(HeadQABase):
+    DATASET_NAME = "en"
+
+class HeadQAEs(HeadQABase):
+    DATASET_NAME = "es"
\ No newline at end of file

From cdab2c033c75f22c5a0dc2b1c5e004f8ae4413cc Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Thu, 23 Dec 2021 19:33:00 -0700
Subject: [PATCH 21/65] headqa: maintain backwards compatibility

---
 lm_eval/tasks/__init__.py | 1 +
 lm_eval/tasks/headqa.py   | 8 +++++++-
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/lm_eval/tasks/__init__.py b/lm_eval/tasks/__init__.py
index eb75aea500..af09702de2 100644
--- a/lm_eval/tasks/__init__.py
+++ b/lm_eval/tasks/__init__.py
@@ -132,6 +132,7 @@
     "squad2": squad.SQuAD2,
     "race": race.RACE,
     # "naturalqs": naturalqs.NaturalQs, # not implemented yet
+    "headqa": headqa.HeadQAEsDeprecated, # for backwards compat - headqa used to default to es
     "headqa_es": headqa.HeadQAEs,
     "headqa_en": headqa.HeadQAEn,
     "mathqa": mathqa.MathQA,
diff --git a/lm_eval/tasks/headqa.py b/lm_eval/tasks/headqa.py
index d59a88df23..18bb77519f 100644
--- a/lm_eval/tasks/headqa.py
+++ b/lm_eval/tasks/headqa.py
@@ -35,4 +35,10 @@ class HeadQAEn(HeadQABase):
     DATASET_NAME = "en"
 
 class HeadQAEs(HeadQABase):
-    DATASET_NAME = "es"
\ No newline at end of file
+    DATASET_NAME = "es"
+
+# for backwards compatibility
+class HeadQAEsDeprecated(HeadQABase):
+    DATASET_NAME = "es"
+
+    print("WARNING: headqa is deprecated. Please use headqa_es or headqa_en instead. See https://github.com/EleutherAI/lm-evaluation-harness/pull/240 for more info.")
\ No newline at end of file

From 22c4124fcc91d9552d1b71d9ae9655da7ef8d390 Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Thu, 23 Dec 2021 19:36:22 -0700
Subject: [PATCH 22/65] Add new testdata

---
 tests/testdata/headqa_en-v0-loglikelihood | 1 +
 tests/testdata/headqa_en-v0-res.json      | 1 +
 tests/testdata/headqa_es-v0-loglikelihood | 1 +
 tests/testdata/headqa_es-v0-res.json      | 1 +
 4 files changed, 4 insertions(+)
 create mode 100644 tests/testdata/headqa_en-v0-loglikelihood
 create mode 100644 tests/testdata/headqa_en-v0-res.json
 create mode 100644 tests/testdata/headqa_es-v0-loglikelihood
 create mode 100644 tests/testdata/headqa_es-v0-res.json

diff --git a/tests/testdata/headqa_en-v0-loglikelihood b/tests/testdata/headqa_en-v0-loglikelihood
new file mode 100644
index 0000000000..11f07878fb
--- /dev/null
+++ b/tests/testdata/headqa_en-v0-loglikelihood
@@ -0,0 +1 @@
+09da45119b12a0144e3081f8fb790c2a22af7b9c3aac42f54423d348a711fbf5
\ No newline at end of file
diff --git a/tests/testdata/headqa_en-v0-res.json b/tests/testdata/headqa_en-v0-res.json
new file mode 100644
index 0000000000..6ac5a9c0b8
--- /dev/null
+++ b/tests/testdata/headqa_en-v0-res.json
@@ -0,0 +1 @@
+{"results": {"headqa_en": {"acc": 0.23559445660102116, "acc_norm": 0.2447118891320204, "acc_norm_stderr": 0.008211629406841468, "acc_stderr": 0.008105688874297972}}, "versions": {"headqa_en": 0}}
\ No newline at end of file
diff --git a/tests/testdata/headqa_es-v0-loglikelihood b/tests/testdata/headqa_es-v0-loglikelihood
new file mode 100644
index 0000000000..9129d834b6
--- /dev/null
+++ b/tests/testdata/headqa_es-v0-loglikelihood
@@ -0,0 +1 @@
+767ca34d9714edd9fb030ddbcc35a64e5180d1e247b0cb557fbb22fdf971ad1f
\ No newline at end of file
diff --git a/tests/testdata/headqa_es-v0-res.json b/tests/testdata/headqa_es-v0-res.json
new file mode 100644
index 0000000000..0964db9bbb
--- /dev/null
+++ b/tests/testdata/headqa_es-v0-res.json
@@ -0,0 +1 @@
+{"results": {"headqa_es": {"acc": 0.23559445660102116, "acc_norm": 0.25018234865062, "acc_norm_stderr": 0.008272783230806014, "acc_stderr": 0.008105688874297972}}, "versions": {"headqa_es": 0}}
\ No newline at end of file

From 7b2b2a23689cab7faf90fdf05105a235713d57ea Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Thu, 23 Dec 2021 20:10:16 -0700
Subject: [PATCH 23/65] Make simple_evaluate take LM and Task objects directly
 too

---
 lm_eval/evaluator.py      | 30 +++++++++++++++++++-----------
 lm_eval/tasks/__init__.py | 18 ++++++++++++++++--
 2 files changed, 35 insertions(+), 13 deletions(-)

diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index 02fb71285c..10648dcfb1 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -10,18 +10,19 @@
 
 
 @positional_deprecated
-def simple_evaluate(model, model_args, task_names,
+def simple_evaluate(model, model_args=None, tasks=[],
                     num_fewshot=0, batch_size=None, device=None,
                     no_cache=False, limit=None, bootstrap_iters=100000,
                     description_dict=None):
     """Instantiate and evaluate a model on a list of tasks.
 
-    :param model: str
-        Name of model, see lm_eval.models.get_model
-    :param model_args: str
-        String arguments for each model class, see LM.create_from_arg_string
-    :param task_names: list[str]
-        List of task names
+    :param model: Union[str, LM]
+        Name of model or LM object, see lm_eval.models.get_model
+    :param model_args: Optional[str]
+        String arguments for each model class, see LM.create_from_arg_string. 
+        Ignored if `model` argument is a LM object.
+    :param tasks: list[Union[str, Task]]
+        List of task names or Task objects
     :param num_fewshot: int
         Number of examples in few-shot context
     :param batch_size: int, optional
@@ -42,16 +43,23 @@ def simple_evaluate(model, model_args, task_names,
     random.seed(1234)
     np.random.seed(1234)
 
-    lm = lm_eval.models.get_model(model).create_from_arg_string(model_args, {
-        'batch_size': batch_size, 'device': device
-    })
+    assert tasks != [], "No tasks specified"
+
+    if isinstance(model, str):
+        if model_args is None: model_args = ""
+        lm = lm_eval.models.get_model(model).create_from_arg_string(model_args, {
+            'batch_size': batch_size, 'device': device
+        })
+    else:
+        assert isinstance(model, lm_eval.base.LM)
+        lm = model
 
     if not no_cache:
         lm = lm_eval.base.CachingLM(
             lm, 'lm_cache/' + model + '_' + model_args.replace('=', '-').replace(',', '_').replace('/', '-') + '.db'
         )
     
-    task_dict = lm_eval.tasks.get_task_dict(task_names)
+    task_dict = lm_eval.tasks.get_task_dict(tasks)
 
     results = evaluate(
         lm=lm,
diff --git a/lm_eval/tasks/__init__.py b/lm_eval/tasks/__init__.py
index 17736f982e..1f8b9b36f0 100644
--- a/lm_eval/tasks/__init__.py
+++ b/lm_eval/tasks/__init__.py
@@ -1,6 +1,8 @@
 from pprint import pprint
+from typing import List, Union
 
 import sacrebleu
+import lm_eval.base
 
 from . import superglue
 from . import glue
@@ -232,8 +234,20 @@ def get_task(task_name):
         raise KeyError(f"Missing task {task_name}")
 
 
-def get_task_dict(task_name_list):
+def get_task_name_from_object(task_object):
+    for name, class_ in TASK_REGISTRY.items():
+        if class_ is task_object:
+            return name
+    
+    # this gives a mechanism for non-registered tasks to have a custom name anyways when reporting
+    return task_object.EVAL_HARNESS_NAME if hasattr(task_object, "EVAL_HARNESS_NAME") else task_object.__name__
+
+
+def get_task_dict(task_name_list: List[Union[str, lm_eval.base.Task]]):
     return {
         task_name: get_task(task_name)()
-        for task_name in task_name_list
+        for task_name in task_name_list if isinstance(task_name, str)
+    } + {
+        get_task_name_from_object(task_object): task_object
+        for task_object in task_name_list if not isinstance(task_object, str)
     }

From d86aabc4aa78406ee22114752790287d4fa18fce Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Thu, 23 Dec 2021 20:22:07 -0700
Subject: [PATCH 24/65] more changes

---
 lm_eval/evaluator.py      | 10 ++++++----
 lm_eval/tasks/__init__.py |  2 +-
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index 10648dcfb1..95fe42d9e0 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -22,7 +22,7 @@ def simple_evaluate(model, model_args=None, tasks=[],
         String arguments for each model class, see LM.create_from_arg_string. 
         Ignored if `model` argument is a LM object.
     :param tasks: list[Union[str, Task]]
-        List of task names or Task objects
+        List of task names or Task objects. Task objects will be taken to have name task.EVAL_HARNESS_NAME if defined and type(task).__name__ otherwise.
     :param num_fewshot: int
         Number of examples in few-shot context
     :param batch_size: int, optional
@@ -64,7 +64,6 @@ def simple_evaluate(model, model_args=None, tasks=[],
     results = evaluate(
         lm=lm,
         task_dict=task_dict,
-        provide_description=False,
         num_fewshot=num_fewshot,
         limit=limit,
         description_dict=description_dict
@@ -87,13 +86,13 @@ def simple_evaluate(model, model_args=None, tasks=[],
 
 
 @positional_deprecated
-def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_iters=100000, description_dict=None):
+def evaluate(lm, task_dict, provide_description=None, num_fewshot=0, limit=None, bootstrap_iters=100000, description_dict=None):
     """Instantiate and evaluate a model on a list of tasks.
 
     :param lm: obj
         Language Model
     :param task_dict: dict[str, Task]
-        Dictionary of tasks
+        Dictionary of tasks. Tasks will be taken to have name task.EVAL_HARNESS_NAME if defined and type(task).__name__ otherwise.
     :param provide_description: bool
         Not implemented, and this option is deprecated and will be removed in a future version in favor of a different description providing method
     :param num_fewshot: int
@@ -111,6 +110,9 @@ def evaluate(lm, task_dict, provide_description, num_fewshot, limit, bootstrap_i
 
     # TODO: todo: implement proper description-providing system
     assert not provide_description  # not implemented.
+    if provide_description is not None:
+        # nudge people to not specify it at all
+        print("WARNING: provide_description is deprecated and will be removed in a future version in favor of description_dict")
 
     task_dict_items = [
         (name, task)
diff --git a/lm_eval/tasks/__init__.py b/lm_eval/tasks/__init__.py
index 1f8b9b36f0..1759149b43 100644
--- a/lm_eval/tasks/__init__.py
+++ b/lm_eval/tasks/__init__.py
@@ -240,7 +240,7 @@ def get_task_name_from_object(task_object):
             return name
     
     # this gives a mechanism for non-registered tasks to have a custom name anyways when reporting
-    return task_object.EVAL_HARNESS_NAME if hasattr(task_object, "EVAL_HARNESS_NAME") else task_object.__name__
+    return task_object.EVAL_HARNESS_NAME if hasattr(task_object, "EVAL_HARNESS_NAME") else type(task_object).__name__
 
 
 def get_task_dict(task_name_list: List[Union[str, lm_eval.base.Task]]):

From a34bbe695d8da0eaf8e488b082efd8ad1ffe1745 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Fri, 24 Dec 2021 02:36:34 -0500
Subject: [PATCH 25/65] Remove more `provide_description` uses

---
 lm_eval/base.py                | 43 +++++++++++++++++++++++++++++++---
 lm_eval/evaluator.py           |  1 -
 scripts/write_out.py           |  1 -
 tests/test_description_dict.py |  1 -
 tests/test_evaluator.py        |  2 --
 tests/test_version_stable.py   |  1 -
 6 files changed, 40 insertions(+), 9 deletions(-)

diff --git a/lm_eval/base.py b/lm_eval/base.py
index 5492ba9cfb..c3f897c711 100644
--- a/lm_eval/base.py
+++ b/lm_eval/base.py
@@ -1,6 +1,7 @@
 import abc
 from typing import Iterable
 import numpy as np
+import random
 import re
 import os
 import json
@@ -458,14 +459,41 @@ def fewshot_description(self):
         return ""
 
     @utils.positional_deprecated
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None, description=None):
+        """ Returns a fewshot context string that is made up of a prepended description
+        (if provided), the `num_fewshot` number of examples, and an appended prompt example.
+
+        :param doc: str
+            The document as returned from training_docs, validation_docs, or test_docs.
+        :param num_fewshot: int
+            The number of fewshot examples to provide in the returned context string.
+        :param provide_description: bool
+            Not implemented, and this option is deprecated and will be removed in a future version in favor of a different description providing method
+        :param rnd: random.Random
+            The pseudo-random number generator used to randomly sample examples.
+            WARNING: If you do not provide a `rnd` arg, a default `random.Random`
+            object will be created and seeded with this Task's name attribute, `__name__`.
+        :param description: str
+            The task's description that will be prepended to the fewshot examples.
+        :returns: str
+            The fewshot context.
+        """
         assert not provide_description, (
             "The `provide_description` arg will be removed in future versions. To prepend "
-            "a custom description to the context, supply the corresponding string via the  "
+            "a custom description to the context, supply the corresponding string via the "
             "`description` arg."
         )
+        if provide_description is not None:
+            # nudge people to not specify it at all
+            print("WARNING: provide_description is deprecated and will be removed in a future version in favor of description_dict")
+
         description = description + "\n\n" if description else ""
 
+        # TODO (jon-tow): Remove this default `rand` behaviour after `provide_description` is removed and remove the respective `rand` arg warning in the docs above.
+        if rnd is None:
+            rnd = random.Random()
+            rnd.seed(self.__name__)
+
         if num_fewshot == 0:
             labeled_examples = ""
         else:
@@ -537,13 +565,22 @@ def fewshot_examples(self, k, rnd):
         assert k == 0
         return []
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None, description=None):
         assert num_fewshot == 0
         assert not provide_description, (
             "The `provide_description` arg will be removed in future versions. To prepend "
             "a custom description to the context, supply the corresponding string via the  "
             "`description` arg."
         )
+        if provide_description is not None:
+            # nudge people to not specify it at all
+            print("WARNING: provide_description is deprecated and will be removed in a future version in favor of description_dict")
+
+        # TODO (jon-tow): Remove this default `rand` behaviour after `provide_description` is removed and remove the respective `rand` arg warning in the docs above.
+        if rnd is None:
+            rnd = random.Random()
+            rnd.seed(self.__name__)
+
         return ""
 
     def higher_is_better(self):
diff --git a/lm_eval/evaluator.py b/lm_eval/evaluator.py
index 95fe42d9e0..087f2a1313 100644
--- a/lm_eval/evaluator.py
+++ b/lm_eval/evaluator.py
@@ -159,7 +159,6 @@ def evaluate(lm, task_dict, provide_description=None, num_fewshot=0, limit=None,
             ctx = task.fewshot_context(
                 doc=doc,
                 num_fewshot=num_fewshot,
-                provide_description=provide_description,
                 rnd=rnd,
                 description=description
             )
diff --git a/scripts/write_out.py b/scripts/write_out.py
index b39fd64541..2039d3934f 100644
--- a/scripts/write_out.py
+++ b/scripts/write_out.py
@@ -63,7 +63,6 @@ def main():
                 ctx = task.fewshot_context(
                     doc=doc,
                     num_fewshot=args.num_fewshot,
-                    provide_description=args.provide_description,
                     rnd=rnd,
                     description=description
                 )
diff --git a/tests/test_description_dict.py b/tests/test_description_dict.py
index aaec13fbce..f80f529063 100644
--- a/tests/test_description_dict.py
+++ b/tests/test_description_dict.py
@@ -36,7 +36,6 @@ def test_description_dict():
             ctx = task.fewshot_context(
                 doc=doc,
                 num_fewshot=1,
-                provide_description=False,
                 rnd=rnd,
                 description=description,
             )
diff --git a/tests/test_evaluator.py b/tests/test_evaluator.py
index ad12b053da..363384a05c 100644
--- a/tests/test_evaluator.py
+++ b/tests/test_evaluator.py
@@ -51,7 +51,6 @@ def ll_perp_fn(reqs):
     e1 = evaluator.evaluate(
             lm=lm,
             task_dict=task_dict,
-            provide_description=False,
             num_fewshot=0,
             limit=limit,
             bootstrap_iters=10,
@@ -60,7 +59,6 @@ def ll_perp_fn(reqs):
     e2 = evaluator.evaluate(
             lm=lm,
             task_dict=task_dict,
-            provide_description=False,
             num_fewshot=0,
             limit=limit,
             bootstrap_iters=10,
diff --git a/tests/test_version_stable.py b/tests/test_version_stable.py
index a19cbf3f90..7dd36a94b6 100644
--- a/tests/test_version_stable.py
+++ b/tests/test_version_stable.py
@@ -102,7 +102,6 @@ def greedy_until(reqs):
     result = evaluator.evaluate(
             lm=lm,
             task_dict=task_dict,
-            provide_description=False,
             num_fewshot=0,
             limit=limit,
             bootstrap_iters=10,

From 57d0718a718243d6ee64a995f3da8fc2146bbd29 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Fri, 24 Dec 2021 02:38:30 -0500
Subject: [PATCH 26/65] Remove all `provide_description` argument uses

---
 lm_eval/tasks/truthfulqa.py | 6 ++----
 scripts/cost_estimate.py    | 1 -
 2 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/lm_eval/tasks/truthfulqa.py b/lm_eval/tasks/truthfulqa.py
index 0eeb364e0f..67259a5a04 100644
--- a/lm_eval/tasks/truthfulqa.py
+++ b/lm_eval/tasks/truthfulqa.py
@@ -85,12 +85,11 @@ def doc_to_text(self, doc):
     def doc_to_target(self, doc):
         return " "
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None, description=None):
         assert num_fewshot == 0, "TruthfulQA is intended only for the zero-shot setting."
         return super().fewshot_context(
             doc=doc,
             num_fewshot=num_fewshot,
-            provide_description=provide_description,
             rnd=rnd,
             description=description
         )
@@ -223,12 +222,11 @@ def doc_to_text(self, doc):
     def doc_to_target(self, doc):
         return " "
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None, description=None):
         assert num_fewshot == 0, "TruthfulQA is intended only for the zero-shot setting."
         return super().fewshot_context(
             doc=doc,
             num_fewshot=num_fewshot,
-            provide_description=provide_description,
             rnd=rnd,
             description=description
         )
diff --git a/scripts/cost_estimate.py b/scripts/cost_estimate.py
index f08d2f37f2..d2e60bfa0d 100644
--- a/scripts/cost_estimate.py
+++ b/scripts/cost_estimate.py
@@ -54,7 +54,6 @@ def main():
         evaluator.evaluate(
             lm=lm,
             task_dict={taskname: tasks.get_task(taskname)()},
-            provide_description=False,
             num_fewshot=0,
             limit=None,
             bootstrap_iters=10,

From 0e232f7af6ff6a48bd78b3ff3752549e723056de Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Fri, 24 Dec 2021 02:40:11 -0500
Subject: [PATCH 27/65] Update new `task` arg and task dict getter

---
 lm_eval/tasks/__init__.py | 7 +++++--
 lm_eval/tasks/prost.py    | 3 +--
 main.py                   | 2 +-
 3 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/lm_eval/tasks/__init__.py b/lm_eval/tasks/__init__.py
index 1759149b43..99513d3202 100644
--- a/lm_eval/tasks/__init__.py
+++ b/lm_eval/tasks/__init__.py
@@ -244,10 +244,13 @@ def get_task_name_from_object(task_object):
 
 
 def get_task_dict(task_name_list: List[Union[str, lm_eval.base.Task]]):
-    return {
+    task_name_dict = {
         task_name: get_task(task_name)()
         for task_name in task_name_list if isinstance(task_name, str)
-    } + {
+    }
+    task_name_from_object_dict = {
         get_task_name_from_object(task_object): task_object
         for task_object in task_name_list if not isinstance(task_object, str)
     }
+    assert set(task_name_dict.keys()).isdisjoint(set(task_name_from_object_dict.keys()))
+    return {**task_name_dict, **task_name_from_object_dict}
diff --git a/lm_eval/tasks/prost.py b/lm_eval/tasks/prost.py
index 7bcbf1e917..e972d39ac0 100644
--- a/lm_eval/tasks/prost.py
+++ b/lm_eval/tasks/prost.py
@@ -36,12 +36,11 @@ def has_validation_docs(self):
     def has_test_docs(self):
         return True
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd, description=None):
+    def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None, description=None):
         assert num_fewshot == 0, 'PROST is designed to probe models in a zero-shot fashion only.'
         return super().fewshot_context(
             doc=doc,
             num_fewshot=num_fewshot,
-            provide_description=provide_description,
             rnd=rnd,
             description=description
         )
diff --git a/main.py b/main.py
index c7f9ce6d8b..cf13fe1315 100644
--- a/main.py
+++ b/main.py
@@ -43,7 +43,7 @@ def main():
     results = evaluator.simple_evaluate(
         model=args.model,
         model_args=args.model_args,
-        task_names=task_names,
+        tasks=task_names,
         num_fewshot=args.num_fewshot,
         batch_size=args.batch_size,
         device=args.device,

From 666b615f2f80e213354dc5dc78aa85e639d48ff9 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Fri, 24 Dec 2021 13:04:59 +0100
Subject: [PATCH 28/65] Fix README

---
 README.md | 292 +++++++++++++++++++++++++++---------------------------
 1 file changed, 146 insertions(+), 146 deletions(-)

diff --git a/README.md b/README.md
index 987b79573e..18ec4dc342 100644
--- a/README.md
+++ b/README.md
@@ -128,152 +128,152 @@ To implement a new task in eval harness, see [this guide](https://github.com/Ele
 |openbookqa                                               |✓    |✓  |✓   |          500|acc, acc_norm                                                                 |
 |squad2                                                   |✓    |✓  |    |        11873|exact, f1, HasAns_exact, HasAns_f1, NoAns_exact, NoAns_f1, best_exact, best_f1|
 |race                                                     |✓    |✓  |✓   |         1045|acc                                                                           |
-|headqa_es                                        |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
-|headqa_en                                        |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
-|mathqa                                           |✓    |✓  |✓   |         2985|acc, acc_norm                                                                 |
-|webqs                                            |✓    |   |✓   |         2032|acc                                                                           |
-|wsc273                                           |     |   |✓   |          273|acc                                                                           |
-|winogrande                                       |✓    |✓  |    |         1267|acc                                                                           |
-|anli_r1                                          |✓    |✓  |✓   |         1000|acc                                                                           |
-|anli_r2                                          |✓    |✓  |✓   |         1000|acc                                                                           |
-|anli_r3                                          |✓    |✓  |✓   |         1200|acc                                                                           |
-|ethics_cm                                        |✓    |   |✓   |         3885|acc                                                                           |
-|ethics_deontology                                |✓    |   |✓   |         3596|acc, em                                                                       |
-|ethics_justice                                   |✓    |   |✓   |         2704|acc, em                                                                       |
-|ethics_utilitarianism_original                   |     |   |✓   |         4808|acc                                                                           |
-|ethics_utilitarianism                            |✓    |   |✓   |         4808|acc                                                                           |
-|ethics_virtue                                    |✓    |   |✓   |         4975|acc, em                                                                       |
-|math_algebra                                     |✓    |   |✓   |         1187|acc                                                                           |
-|math_counting_and_prob                           |✓    |   |✓   |          474|acc                                                                           |
-|math_geometry                                    |✓    |   |✓   |          479|acc                                                                           |
-|math_intermediate_algebra                        |✓    |   |✓   |          903|acc                                                                           |
-|math_num_theory                                  |✓    |   |✓   |          540|acc                                                                           |
-|math_prealgebra                                  |✓    |   |✓   |          871|acc                                                                           |
-|math_precalc                                     |✓    |   |✓   |          546|acc                                                                           |
-|arithmetic_2da                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_2ds                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_3da                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_3ds                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_4da                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_4ds                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_5da                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_5ds                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_2dm                                   |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_1dc                                   |     |✓  |    |         2000|acc                                                                           |
-|hendrycksTest-abstract_algebra                   |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-anatomy                            |✓    |✓  |✓   |          135|acc, acc_norm                                                                 |
-|hendrycksTest-astronomy                          |✓    |✓  |✓   |          152|acc, acc_norm                                                                 |
-|hendrycksTest-business_ethics                    |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-clinical_knowledge                 |✓    |✓  |✓   |          265|acc, acc_norm                                                                 |
-|hendrycksTest-college_biology                    |✓    |✓  |✓   |          144|acc, acc_norm                                                                 |
-|hendrycksTest-college_chemistry                  |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-college_computer_science           |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-college_mathematics                |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-college_medicine                   |✓    |✓  |✓   |          173|acc, acc_norm                                                                 |
-|hendrycksTest-college_physics                    |✓    |✓  |✓   |          102|acc, acc_norm                                                                 |
-|hendrycksTest-computer_security                  |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-conceptual_physics                 |✓    |✓  |✓   |          235|acc, acc_norm                                                                 |
-|hendrycksTest-econometrics                       |✓    |✓  |✓   |          114|acc, acc_norm                                                                 |
-|hendrycksTest-electrical_engineering             |✓    |✓  |✓   |          145|acc, acc_norm                                                                 |
-|hendrycksTest-elementary_mathematics             |✓    |✓  |✓   |          378|acc, acc_norm                                                                 |
-|hendrycksTest-formal_logic                       |✓    |✓  |✓   |          126|acc, acc_norm                                                                 |
-|hendrycksTest-global_facts                       |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_biology                |✓    |✓  |✓   |          310|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_chemistry              |✓    |✓  |✓   |          203|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_computer_science       |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_european_history       |✓    |✓  |✓   |          165|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_geography              |✓    |✓  |✓   |          198|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_government_and_politics|✓    |✓  |✓   |          193|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_macroeconomics         |✓    |✓  |✓   |          390|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_mathematics            |✓    |✓  |✓   |          270|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_microeconomics         |✓    |✓  |✓   |          238|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_physics                |✓    |✓  |✓   |          151|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_psychology             |✓    |✓  |✓   |          545|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_statistics             |✓    |✓  |✓   |          216|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_us_history             |✓    |✓  |✓   |          204|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_world_history          |✓    |✓  |✓   |          237|acc, acc_norm                                                                 |
-|hendrycksTest-human_aging                        |✓    |✓  |✓   |          223|acc, acc_norm                                                                 |
-|hendrycksTest-human_sexuality                    |✓    |✓  |✓   |          131|acc, acc_norm                                                                 |
-|hendrycksTest-international_law                  |✓    |✓  |✓   |          121|acc, acc_norm                                                                 |
-|hendrycksTest-jurisprudence                      |✓    |✓  |✓   |          108|acc, acc_norm                                                                 |
-|hendrycksTest-logical_fallacies                  |✓    |✓  |✓   |          163|acc, acc_norm                                                                 |
-|hendrycksTest-machine_learning                   |✓    |✓  |✓   |          112|acc, acc_norm                                                                 |
-|hendrycksTest-management                         |✓    |✓  |✓   |          103|acc, acc_norm                                                                 |
-|hendrycksTest-marketing                          |✓    |✓  |✓   |          234|acc, acc_norm                                                                 |
-|hendrycksTest-medical_genetics                   |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-miscellaneous                      |✓    |✓  |✓   |          783|acc, acc_norm                                                                 |
-|hendrycksTest-moral_disputes                     |✓    |✓  |✓   |          346|acc, acc_norm                                                                 |
-|hendrycksTest-moral_scenarios                    |✓    |✓  |✓   |          895|acc, acc_norm                                                                 |
-|hendrycksTest-nutrition                          |✓    |✓  |✓   |          306|acc, acc_norm                                                                 |
-|hendrycksTest-philosophy                         |✓    |✓  |✓   |          311|acc, acc_norm                                                                 |
-|hendrycksTest-prehistory                         |✓    |✓  |✓   |          324|acc, acc_norm                                                                 |
-|hendrycksTest-professional_accounting            |✓    |✓  |✓   |          282|acc, acc_norm                                                                 |
-|hendrycksTest-professional_law                   |✓    |✓  |✓   |         1534|acc, acc_norm                                                                 |
-|hendrycksTest-professional_medicine              |✓    |✓  |✓   |          272|acc, acc_norm                                                                 |
-|hendrycksTest-professional_psychology            |✓    |✓  |✓   |          612|acc, acc_norm                                                                 |
-|hendrycksTest-public_relations                   |✓    |✓  |✓   |          110|acc, acc_norm                                                                 |
-|hendrycksTest-security_studies                   |✓    |✓  |✓   |          245|acc, acc_norm                                                                 |
-|hendrycksTest-sociology                          |✓    |✓  |✓   |          201|acc, acc_norm                                                                 |
-|hendrycksTest-us_foreign_policy                  |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-virology                           |✓    |✓  |✓   |          166|acc, acc_norm                                                                 |
-|hendrycksTest-world_religions                    |✓    |✓  |✓   |          171|acc, acc_norm                                                                 |
-|wmt14-en-fr                                      |     |   |✓   |         3003|bleu, chrf, ter                                                               |
-|wmt14-fr-en                                      |     |   |✓   |         3003|bleu, chrf, ter                                                               |
-|wmt16-en-ro                                      |     |   |✓   |         1999|bleu, chrf, ter                                                               |
-|wmt16-ro-en                                      |     |   |✓   |         1999|bleu, chrf, ter                                                               |
-|wmt16-de-en                                      |     |   |✓   |         2999|bleu, chrf, ter                                                               |
-|wmt16-en-de                                      |     |   |✓   |         2999|bleu, chrf, ter                                                               |
-|wmt20-cs-en                                      |     |   |✓   |          664|bleu, chrf, ter                                                               |
-|wmt20-de-en                                      |     |   |✓   |          785|bleu, chrf, ter                                                               |
-|wmt20-de-fr                                      |     |   |✓   |         1619|bleu, chrf, ter                                                               |
-|wmt20-en-cs                                      |     |   |✓   |         1418|bleu, chrf, ter                                                               |
-|wmt20-en-de                                      |     |   |✓   |         1418|bleu, chrf, ter                                                               |
-|wmt20-en-iu                                      |     |   |✓   |         2971|bleu, chrf, ter                                                               |
-|wmt20-en-ja                                      |     |   |✓   |         1000|bleu, chrf, ter                                                               |
-|wmt20-en-km                                      |     |   |✓   |         2320|bleu, chrf, ter                                                               |
-|wmt20-en-pl                                      |     |   |✓   |         1000|bleu, chrf, ter                                                               |
-|wmt20-en-ps                                      |     |   |✓   |         2719|bleu, chrf, ter                                                               |
-|wmt20-en-ru                                      |     |   |✓   |         2002|bleu, chrf, ter                                                               |
-|wmt20-en-ta                                      |     |   |✓   |         1000|bleu, chrf, ter                                                               |
-|wmt20-en-zh                                      |     |   |✓   |         1418|bleu, chrf, ter                                                               |
-|wmt20-fr-de                                      |     |   |✓   |         1619|bleu, chrf, ter                                                               |
-|wmt20-iu-en                                      |     |   |✓   |         2971|bleu, chrf, ter                                                               |
-|wmt20-ja-en                                      |     |   |✓   |          993|bleu, chrf, ter                                                               |
-|wmt20-km-en                                      |     |   |✓   |         2320|bleu, chrf, ter                                                               |
-|wmt20-pl-en                                      |     |   |✓   |         1001|bleu, chrf, ter                                                               |
-|wmt20-ps-en                                      |     |   |✓   |         2719|bleu, chrf, ter                                                               |
-|wmt20-ru-en                                      |     |   |✓   |          991|bleu, chrf, ter                                                               |
-|wmt20-ta-en                                      |     |   |✓   |          997|bleu, chrf, ter                                                               |
-|wmt20-zh-en                                      |     |   |✓   |         2000|bleu, chrf, ter                                                               |
-|iwslt17-en-ar                                    |     |   |✓   |         1460|bleu, chrf, ter                                                               |
-|iwslt17-ar-en                                    |     |   |✓   |         1460|bleu, chrf, ter                                                               |
-|anagrams1                                        |     |✓  |    |        10000|acc                                                                           |
-|anagrams2                                        |     |✓  |    |        10000|acc                                                                           |
-|cycle_letters                                    |     |✓  |    |        10000|acc                                                                           |
-|random_insertion                                 |     |✓  |    |        10000|acc                                                                           |
-|reversed_words                                   |     |✓  |    |        10000|acc                                                                           |
-|pile_arxiv                                       |     |✓  |✓   |         2407|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_books3                                      |     |✓  |✓   |          269|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_bookcorpus2                                 |     |✓  |✓   |           28|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_dm-mathematics                              |     |✓  |✓   |         1922|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_enron                                       |     |✓  |✓   |         1010|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_europarl                                    |     |✓  |✓   |          157|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_freelaw                                     |     |✓  |✓   |         5101|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_github                                      |     |✓  |✓   |        18195|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_gutenberg                                   |     |✓  |✓   |           80|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_hackernews                                  |     |✓  |✓   |         1632|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_nih-exporter                                |     |✓  |✓   |         1884|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_opensubtitles                               |     |✓  |✓   |          642|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_openwebtext2                                |     |✓  |✓   |        32925|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_philpapers                                  |     |✓  |✓   |           68|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_pile-cc                                     |     |✓  |✓   |        52790|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_pubmed-abstracts                            |     |✓  |✓   |        29895|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_pubmed-central                              |     |✓  |✓   |         5911|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_stackexchange                               |     |✓  |✓   |        30378|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_uspto                                       |     |✓  |✓   |        11415|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_ubuntu-irc                                  |     |✓  |✓   |           22|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_wikipedia                                   |     |✓  |✓   |        17511|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_youtubesubtitles                            |     |✓  ||         1000|acc
+|mathqa                                                   |✓    |✓  |✓   |         2985|acc, acc_norm                                                                 |
+|headqa_es                                                |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
+|headqa_en                                                |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
+|webqs                                                    |✓    |   |✓   |         2032|acc                                                                           |
+|wsc273                                                   |     |   |✓   |          273|acc                                                                           |
+|winogrande                                               |✓    |✓  |    |         1267|acc                                                                           |
+|anli_r1                                                  |✓    |✓  |✓   |         1000|acc                                                                           |
+|anli_r2                                                  |✓    |✓  |✓   |         1000|acc                                                                           |
+|anli_r3                                                  |✓    |✓  |✓   |         1200|acc                                                                           |
+|ethics_cm                                                |✓    |   |✓   |         3885|acc                                                                           |
+|ethics_deontology                                        |✓    |   |✓   |         3596|acc, em                                                                       |
+|ethics_justice                                           |✓    |   |✓   |         2704|acc, em                                                                       |
+|ethics_utilitarianism_original                           |     |   |✓   |         4808|acc                                                                           |
+|ethics_utilitarianism                                    |✓    |   |✓   |         4808|acc                                                                           |
+|ethics_virtue                                            |✓    |   |✓   |         4975|acc, em                                                                       |
+|math_algebra                                             |✓    |   |✓   |         1187|acc                                                                           |
+|math_counting_and_prob                                   |✓    |   |✓   |          474|acc                                                                           |
+|math_geometry                                            |✓    |   |✓   |          479|acc                                                                           |
+|math_intermediate_algebra                                |✓    |   |✓   |          903|acc                                                                           |
+|math_num_theory                                          |✓    |   |✓   |          540|acc                                                                           |
+|math_prealgebra                                          |✓    |   |✓   |          871|acc                                                                           |
+|math_precalc                                             |✓    |   |✓   |          546|acc                                                                           |
+|arithmetic_2da                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_2ds                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_3da                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_3ds                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_4da                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_4ds                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_5da                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_5ds                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_2dm                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_1dc                                           |     |✓  |    |         2000|acc                                                                           |
+|hendrycksTest-abstract_algebra                           |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-anatomy                                    |✓    |✓  |✓   |          135|acc, acc_norm                                                                 |
+|hendrycksTest-astronomy                                  |✓    |✓  |✓   |          152|acc, acc_norm                                                                 |
+|hendrycksTest-business_ethics                            |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-clinical_knowledge                         |✓    |✓  |✓   |          265|acc, acc_norm                                                                 |
+|hendrycksTest-college_biology                            |✓    |✓  |✓   |          144|acc, acc_norm                                                                 |
+|hendrycksTest-college_chemistry                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-college_computer_science                   |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-college_mathematics                        |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-college_medicine                           |✓    |✓  |✓   |          173|acc, acc_norm                                                                 |
+|hendrycksTest-college_physics                            |✓    |✓  |✓   |          102|acc, acc_norm                                                                 |
+|hendrycksTest-computer_security                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-conceptual_physics                         |✓    |✓  |✓   |          235|acc, acc_norm                                                                 |
+|hendrycksTest-econometrics                               |✓    |✓  |✓   |          114|acc, acc_norm                                                                 |
+|hendrycksTest-electrical_engineering                     |✓    |✓  |✓   |          145|acc, acc_norm                                                                 |
+|hendrycksTest-elementary_mathematics                     |✓    |✓  |✓   |          378|acc, acc_norm                                                                 |
+|hendrycksTest-formal_logic                               |✓    |✓  |✓   |          126|acc, acc_norm                                                                 |
+|hendrycksTest-global_facts                               |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_biology                        |✓    |✓  |✓   |          310|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_chemistry                      |✓    |✓  |✓   |          203|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_computer_science               |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_european_history               |✓    |✓  |✓   |          165|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_geography                      |✓    |✓  |✓   |          198|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_government_and_politics        |✓    |✓  |✓   |          193|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_macroeconomics                 |✓    |✓  |✓   |          390|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_mathematics                    |✓    |✓  |✓   |          270|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_microeconomics                 |✓    |✓  |✓   |          238|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_physics                        |✓    |✓  |✓   |          151|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_psychology                     |✓    |✓  |✓   |          545|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_statistics                     |✓    |✓  |✓   |          216|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_us_history                     |✓    |✓  |✓   |          204|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_world_history                  |✓    |✓  |✓   |          237|acc, acc_norm                                                                 |
+|hendrycksTest-human_aging                                |✓    |✓  |✓   |          223|acc, acc_norm                                                                 |
+|hendrycksTest-human_sexuality                            |✓    |✓  |✓   |          131|acc, acc_norm                                                                 |
+|hendrycksTest-international_law                          |✓    |✓  |✓   |          121|acc, acc_norm                                                                 |
+|hendrycksTest-jurisprudence                              |✓    |✓  |✓   |          108|acc, acc_norm                                                                 |
+|hendrycksTest-logical_fallacies                          |✓    |✓  |✓   |          163|acc, acc_norm                                                                 |
+|hendrycksTest-machine_learning                           |✓    |✓  |✓   |          112|acc, acc_norm                                                                 |
+|hendrycksTest-management                                 |✓    |✓  |✓   |          103|acc, acc_norm                                                                 |
+|hendrycksTest-marketing                                  |✓    |✓  |✓   |          234|acc, acc_norm                                                                 |
+|hendrycksTest-medical_genetics                           |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-miscellaneous                              |✓    |✓  |✓   |          783|acc, acc_norm                                                                 |
+|hendrycksTest-moral_disputes                             |✓    |✓  |✓   |          346|acc, acc_norm                                                                 |
+|hendrycksTest-moral_scenarios                            |✓    |✓  |✓   |          895|acc, acc_norm                                                                 |
+|hendrycksTest-nutrition                                  |✓    |✓  |✓   |          306|acc, acc_norm                                                                 |
+|hendrycksTest-philosophy                                 |✓    |✓  |✓   |          311|acc, acc_norm                                                                 |
+|hendrycksTest-prehistory                                 |✓    |✓  |✓   |          324|acc, acc_norm                                                                 |
+|hendrycksTest-professional_accounting                    |✓    |✓  |✓   |          282|acc, acc_norm                                                                 |
+|hendrycksTest-professional_law                           |✓    |✓  |✓   |         1534|acc, acc_norm                                                                 |
+|hendrycksTest-professional_medicine                      |✓    |✓  |✓   |          272|acc, acc_norm                                                                 |
+|hendrycksTest-professional_psychology                    |✓    |✓  |✓   |          612|acc, acc_norm                                                                 |
+|hendrycksTest-public_relations                           |✓    |✓  |✓   |          110|acc, acc_norm                                                                 |
+|hendrycksTest-security_studies                           |✓    |✓  |✓   |          245|acc, acc_norm                                                                 |
+|hendrycksTest-sociology                                  |✓    |✓  |✓   |          201|acc, acc_norm                                                                 |
+|hendrycksTest-us_foreign_policy                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-virology                                   |✓    |✓  |✓   |          166|acc, acc_norm                                                                 |
+|hendrycksTest-world_religions                            |✓    |✓  |✓   |          171|acc, acc_norm                                                                 |
+|wmt14-en-fr                                              |     |   |✓   |         3003|bleu, chrf, ter                                                               |
+|wmt14-fr-en                                              |     |   |✓   |         3003|bleu, chrf, ter                                                               |
+|wmt16-en-ro                                              |     |   |✓   |         1999|bleu, chrf, ter                                                               |
+|wmt16-ro-en                                              |     |   |✓   |         1999|bleu, chrf, ter                                                               |
+|wmt16-de-en                                              |     |   |✓   |         2999|bleu, chrf, ter                                                               |
+|wmt16-en-de                                              |     |   |✓   |         2999|bleu, chrf, ter                                                               |
+|wmt20-cs-en                                              |     |   |✓   |          664|bleu, chrf, ter                                                               |
+|wmt20-de-en                                              |     |   |✓   |          785|bleu, chrf, ter                                                               |
+|wmt20-de-fr                                              |     |   |✓   |         1619|bleu, chrf, ter                                                               |
+|wmt20-en-cs                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
+|wmt20-en-de                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
+|wmt20-en-iu                                              |     |   |✓   |         2971|bleu, chrf, ter                                                               |
+|wmt20-en-ja                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
+|wmt20-en-km                                              |     |   |✓   |         2320|bleu, chrf, ter                                                               |
+|wmt20-en-pl                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
+|wmt20-en-ps                                              |     |   |✓   |         2719|bleu, chrf, ter                                                               |
+|wmt20-en-ru                                              |     |   |✓   |         2002|bleu, chrf, ter                                                               |
+|wmt20-en-ta                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
+|wmt20-en-zh                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
+|wmt20-fr-de                                              |     |   |✓   |         1619|bleu, chrf, ter                                                               |
+|wmt20-iu-en                                              |     |   |✓   |         2971|bleu, chrf, ter                                                               |
+|wmt20-ja-en                                              |     |   |✓   |          993|bleu, chrf, ter                                                               |
+|wmt20-km-en                                              |     |   |✓   |         2320|bleu, chrf, ter                                                               |
+|wmt20-pl-en                                              |     |   |✓   |         1001|bleu, chrf, ter                                                               |
+|wmt20-ps-en                                              |     |   |✓   |         2719|bleu, chrf, ter                                                               |
+|wmt20-ru-en                                              |     |   |✓   |          991|bleu, chrf, ter                                                               |
+|wmt20-ta-en                                              |     |   |✓   |          997|bleu, chrf, ter                                                               |
+|wmt20-zh-en                                              |     |   |✓   |         2000|bleu, chrf, ter                                                               |
+|iwslt17-en-ar                                            |     |   |✓   |         1460|bleu, chrf, ter                                                               |
+|iwslt17-ar-en                                            |     |   |✓   |         1460|bleu, chrf, ter                                                               |
+|anagrams1                                                |     |✓  |    |        10000|acc                                                                           |
+|anagrams2                                                |     |✓  |    |        10000|acc                                                                           |
+|cycle_letters                                            |     |✓  |    |        10000|acc                                                                           |
+|random_insertion                                         |     |✓  |    |        10000|acc                                                                           |
+|reversed_words                                           |     |✓  |    |        10000|acc                                                                           |
+|pile_arxiv                                               |     |✓  |✓   |         2407|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_books3                                              |     |✓  |✓   |          269|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_bookcorpus2                                         |     |✓  |✓   |           28|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_dm-mathematics                                      |     |✓  |✓   |         1922|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_enron                                               |     |✓  |✓   |         1010|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_europarl                                            |     |✓  |✓   |          157|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_freelaw                                             |     |✓  |✓   |         5101|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_github                                              |     |✓  |✓   |        18195|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_gutenberg                                           |     |✓  |✓   |           80|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_hackernews                                          |     |✓  |✓   |         1632|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_nih-exporter                                        |     |✓  |✓   |         1884|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_opensubtitles                                       |     |✓  |✓   |          642|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_openwebtext2                                        |     |✓  |✓   |        32925|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_philpapers                                          |     |✓  |✓   |           68|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_pile-cc                                             |     |✓  |✓   |        52790|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_pubmed-abstracts                                    |     |✓  |✓   |        29895|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_pubmed-central                                      |     |✓  |✓   |         5911|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_stackexchange                                       |     |✓  |✓   |        30378|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_uspto                                               |     |✓  |✓   |        11415|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_ubuntu-irc                                          |     |✓  |✓   |           22|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_wikipedia                                           |     |✓  |✓   |        17511|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_youtubesubtitles                                    |     |✓  ||         1000|acc
 |blimp_adjunct_island                                     |     |✓  |    |         1000|acc
 |blimp_anaphor_gender_agreement                           |     |✓  |    |         1000|acc
 |blimp_anaphor_number_agreement                           |     |✓  |    |         1000|acc

From e34c6bd606ab68d3b73f19dc3732bc28465a6fa0 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Fri, 24 Dec 2021 13:06:02 +0100
Subject: [PATCH 29/65] Fix README

---
 README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/README.md b/README.md
index 18ec4dc342..f0d910cca8 100644
--- a/README.md
+++ b/README.md
@@ -273,7 +273,7 @@ To implement a new task in eval harness, see [this guide](https://github.com/Ele
 |pile_uspto                                               |     |✓  |✓   |        11415|word_perplexity, byte_perplexity, bits_per_byte                               |
 |pile_ubuntu-irc                                          |     |✓  |✓   |           22|word_perplexity, byte_perplexity, bits_per_byte                               |
 |pile_wikipedia                                           |     |✓  |✓   |        17511|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_youtubesubtitles                                    |     |✓  ||         1000|acc
+|pile_youtubesubtitles                                    |     |✓  |    |         1000|acc
 |blimp_adjunct_island                                     |     |✓  |    |         1000|acc
 |blimp_anaphor_gender_agreement                           |     |✓  |    |         1000|acc
 |blimp_anaphor_number_agreement                           |     |✓  |    |         1000|acc

From 383605122c428437dae452e543e5a73deadbe63f Mon Sep 17 00:00:00 2001
From: Igor Ostrovsky <igoros@gmail.com>
Date: Sat, 25 Dec 2021 17:49:47 +0000
Subject: [PATCH 30/65] Fix bits_per_byte metric in PerplexityTask

bits_per_byte was calculated as average per-byte loglikelihood, which would work if loglikelihood was base-2 log,
but it is natural log. To correct for that, bits_per_byte should be divided by math.log(2).

Also, it should be true that 2^bits_per_byte == byte_perplexity, which is true after the fix.
---
 lm_eval/base.py    | 6 +++---
 lm_eval/metrics.py | 3 +++
 2 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/lm_eval/base.py b/lm_eval/base.py
index 927ecb49f0..65d4ae15d7 100644
--- a/lm_eval/base.py
+++ b/lm_eval/base.py
@@ -10,7 +10,7 @@
 import torch
 import torch.nn.functional as F
 
-from lm_eval.metrics import mean, weighted_perplexity, weighted_mean
+from lm_eval.metrics import mean, weighted_perplexity, weighted_mean, bits_per_byte
 from lm_eval import utils
 from abc import abstractmethod
 
@@ -560,14 +560,14 @@ def process_results(self, doc, results):
         return {
             "word_perplexity": (loglikelihood, words),
             "byte_perplexity": (loglikelihood, bytes_),
-            "bits_per_byte": (-loglikelihood, self.count_bytes(doc))
+            "bits_per_byte": (loglikelihood, bytes_),
         }
 
     def aggregation(self):
         return {
             "word_perplexity": weighted_perplexity,
             "byte_perplexity": weighted_perplexity,
-            "bits_per_byte": weighted_mean
+            "bits_per_byte": bits_per_byte,
         }
 
     @classmethod
diff --git a/lm_eval/metrics.py b/lm_eval/metrics.py
index c95d4cd61c..76bfdf7364 100644
--- a/lm_eval/metrics.py
+++ b/lm_eval/metrics.py
@@ -102,6 +102,9 @@ def weighted_mean(items):
 def weighted_perplexity(items):
     return math.exp(-weighted_mean(items))
 
+def bits_per_byte(items):
+    return -weighted_mean(items) / math.log(2)
+
 
 def bleu(items):
     """The Bilingual Evaluation Understudy Score, or BLEU for short, is a metric

From df1fc6c355c771eb090efabb230bb267bbb1e885 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Sun, 26 Dec 2021 12:51:30 +0100
Subject: [PATCH 31/65] Fix multirc

---
 lm_eval/tasks/superglue.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/lm_eval/tasks/superglue.py b/lm_eval/tasks/superglue.py
index 33598f2301..ae21bcedfa 100644
--- a/lm_eval/tasks/superglue.py
+++ b/lm_eval/tasks/superglue.py
@@ -228,7 +228,7 @@ def doc_to_target(self, doc):
     @staticmethod
     def format_answer(answer, label):
         label_str = "yes" if label else "no"
-        return f"{label_str}, {answer}"
+        return f"{answer}\nIs the answer correct? {label_str}"
 
     def construct_requests(self, doc, ctx):
         true_choice = self.format_answer(answer=doc["answer"], label=True)
@@ -240,7 +240,8 @@ def construct_requests(self, doc, ctx):
         return ll_true_choice, ll_false_choice
 
     def process_results(self, doc, results):
-        pred = np.argmax(results)
+        ll_true_choice, ll_false_choice = results
+        pred = ll_true_choice > ll_false_choice
         return {
             "acc": (pred, doc)
         }

From 23a420674b49476c5be66b0bf4e0d487ca38f1c7 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Sun, 26 Dec 2021 13:00:33 +0100
Subject: [PATCH 32/65] Add capital letters

---
 lm_eval/tasks/superglue.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lm_eval/tasks/superglue.py b/lm_eval/tasks/superglue.py
index ae21bcedfa..f3d07b3483 100644
--- a/lm_eval/tasks/superglue.py
+++ b/lm_eval/tasks/superglue.py
@@ -227,7 +227,7 @@ def doc_to_target(self, doc):
 
     @staticmethod
     def format_answer(answer, label):
-        label_str = "yes" if label else "no"
+        label_str = "Yes" if label else "No"
         return f"{answer}\nIs the answer correct? {label_str}"
 
     def construct_requests(self, doc, ctx):

From b0a123102305a8b55ef3af24b535929a60642b52 Mon Sep 17 00:00:00 2001
From: rokosbasilisk <thecomplexitytheorist@gmail.com>
Date: Sun, 26 Dec 2021 21:20:30 +0530
Subject: [PATCH 33/65] add asdiv task

---
 eval.sh                               |   1 +
 lm_eval.egg-info/PKG-INFO             | 448 ++++++++++++++++++++++++++
 lm_eval.egg-info/SOURCES.txt          |  74 +++++
 lm_eval.egg-info/dependency_links.txt |   1 +
 lm_eval.egg-info/requires.txt         |  24 ++
 lm_eval.egg-info/top_level.txt        |   2 +
 lm_eval/tasks/__init__.py             |   2 +
 lm_eval/tasks/apps.py                 | 130 ++++++++
 lm_eval/tasks/asdiv.py                | 135 ++++++++
 9 files changed, 817 insertions(+)
 create mode 100644 eval.sh
 create mode 100644 lm_eval.egg-info/PKG-INFO
 create mode 100644 lm_eval.egg-info/SOURCES.txt
 create mode 100644 lm_eval.egg-info/dependency_links.txt
 create mode 100644 lm_eval.egg-info/requires.txt
 create mode 100644 lm_eval.egg-info/top_level.txt
 create mode 100644 lm_eval/tasks/apps.py
 create mode 100644 lm_eval/tasks/asdiv.py

diff --git a/eval.sh b/eval.sh
new file mode 100644
index 0000000000..68c468f4ba
--- /dev/null
+++ b/eval.sh
@@ -0,0 +1 @@
+python main.py --model gpt2 --model_args pretrained=EleutherAI/gpt-neo-125M --device cuda:0 --tasks math_asdiv
diff --git a/lm_eval.egg-info/PKG-INFO b/lm_eval.egg-info/PKG-INFO
new file mode 100644
index 0000000000..f4d1cbda8d
--- /dev/null
+++ b/lm_eval.egg-info/PKG-INFO
@@ -0,0 +1,448 @@
+Metadata-Version: 2.1
+Name: lm-eval
+Version: 0.1.0
+Summary: A framework for evaluating autoregressive language models
+Home-page: https://github.com/EleutherAI/lm-evaluation-harness
+Author: Leo Gao
+Author-email: lg@eleuther.ai
+License: UNKNOWN
+Platform: UNKNOWN
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.6
+Description-Content-Type: text/markdown
+License-File: LICENSE.md
+
+# Language Model Evaluation Harness
+
+![](https://github.com/EleutherAI/lm-evaluation-harness/workflows/Build/badge.svg)
+[![codecov](https://codecov.io/gh/EleutherAI/lm-evaluation-harness/branch/master/graph/badge.svg?token=JSG3O2427J)](https://codecov.io/gh/EleutherAI/lm-evaluation-harness)
+
+## Overview 
+
+This project provides a unified framework to test autoregressive language models (GPT-2, GPT-3, GPTNeo, etc) on a large number of different evaluation tasks.
+
+Features:
+
+- 100+ tasks implemented
+- Support for GPT-2, GPT-3, GPT-Neo, GPT-NeoX, and GPT-J, with flexible tokenization-agnostic interface
+- Task versioning to ensure reproducibility
+
+## Install
+
+```bash
+pip install lm-eval
+```
+
+## Basic Usage
+
+To evaluate a model, (e.g. GPT-2) on NLU tasks (e.g. LAMBADA, HellaSwag), you can run the following command.
+
+```bash
+python main.py \
+	--model gpt2 \
+	--device cuda:0 \
+	--tasks lambada,hellaswag
+```
+(This uses gpt2-117M by default as per HF defaults, use --model_args to specify other gpt2 sizes)
+
+Additional arguments can be provided to the model constructor using the `--model_args` flag. Most importantly, the `gpt2` model can be used to load an arbitrary HuggingFace model. For example, to run GPTNeo use the following:
+
+```bash
+python main.py \
+	--model gpt2 \
+	--model_args pretrained=EleutherAI/gpt-neo-2.7B \
+	--device cuda:0 \
+	--tasks lambada,hellaswag
+```
+
+If you have access to the OpenAI API, you can also evaluate GPT-3:
+
+```bash
+export OPENAI_API_SECRET_KEY=YOUR_KEY_HERE
+python main.py \
+	--model gpt3 \
+	--model_args engine=davinci \
+	--tasks lambada,hellaswag
+```
+
+To evaluate mesh-transformer-jax models that are not available on HF, please invoke eval harness through [this script](https://github.com/kingoflolz/mesh-transformer-jax/blob/master/eval_harness.py).
+
+## Implementing new tasks
+
+To implement a new task in eval harness, see [this guide](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/task-guide.md).
+
+## Cite as
+
+```
+@software{eval-harness,
+  author       = {Gao, Leo and
+                  Tow, Jonathan and
+                  Biderman, Stella and
+                  Black, Sid and
+                  DiPofi, Anthony and
+                  Foster, Charles and
+                  Golding, Laurence and
+                  Hsu, Jeffrey and
+                  McDonell, Kyle and
+                  Muennighoff, Niklas and
+                  Phang, Jason and
+                  Reynolds, Laria and
+                  Tang, Eric and
+                  Thite, Anish and
+                  Wang, Ben and
+                  Wang, Kevin and
+                  Zou, Andy},
+  title        = {A framework for few-shot language model evaluation},
+  month        = sep,
+  year         = 2021,
+  publisher    = {Zenodo},
+  version      = {v0.0.1},
+  doi          = {10.5281/zenodo.5371628},
+  url          = {https://doi.org/10.5281/zenodo.5371628}
+}
+```
+
+### Full Task List
+
+|                    Task Name                            |Train|Val|Test|Val/Test Docs|                                   Metrics                                    |
+|---------------------------------------------------------|-----|---|----|------------:|------------------------------------------------------------------------------|
+|cola                                                     |✓    |✓  |    |         1043|mcc                                                                           |
+|mnli                                                     |✓    |✓  |    |         9815|acc                                                                           |
+|mnli_mismatched                                          |✓    |✓  |    |         9832|acc                                                                           |
+|mrpc                                                     |✓    |✓  |    |          408|acc, f1                                                                       |
+|rte                                                      |✓    |✓  |    |          277|acc                                                                           |
+|qnli                                                     |✓    |✓  |    |         5463|acc                                                                           |
+|qqp                                                      |✓    |✓  |    |        40430|acc, f1                                                                       |
+|sst                                                      |✓    |✓  |    |          872|acc                                                                           |
+|wnli                                                     |✓    |✓  |    |           71|acc                                                                           |
+|boolq                                                    |✓    |✓  |    |         3270|acc                                                                           |
+|cb                                                       |✓    |✓  |    |           56|acc, f1                                                                       |
+|copa                                                     |✓    |✓  |    |          100|acc                                                                           |
+|multirc                                                  |✓    |✓  |    |         4848|acc                                                                           |
+|record                                                   |✓    |✓  |    |        10000|f1, em                                                                        |
+|wic                                                      |✓    |✓  |    |          638|acc                                                                           |
+|wsc                                                      |✓    |✓  |    |          104|acc                                                                           |
+|coqa                                                     |✓    |✓  |    |          500|f1, em                                                                        |
+|drop                                                     |✓    |✓  |    |         9536|em, f1                                                                        |
+|lambada                                                  |     |✓  |    |         5153|ppl, acc                                                                      |
+|lambada_cloze                                            |     |✓  |    |         5153|ppl, acc                                                                      |
+|wikitext                                                 |     |✓  |✓   |           62|word_perplexity, byte_perplexity, bits_per_byte                               |
+|piqa                                                     |✓    |✓  |    |         1838|acc, acc_norm                                                                 |
+|prost                                                    |     |   |✓   |        18736|acc, acc_norm                                                                 |
+|pubmedqa                                                 |     |   |✓   |         1000|acc                                                                           |
+|sciq                                                     |✓    |✓  |✓   |         1000|acc, acc_norm                                                                 |
+|qa4mre_2011                                              |     |   |✓   |          120|acc, acc_norm                                                                 |
+|qa4mre_2012                                              |     |   |✓   |          160|acc, acc_norm                                                                 |
+|qa4mre_2013                                              |     |   |✓   |          284|acc, acc_norm                                                                 |
+|triviaqa                                                 |✓    |✓  |    |        11313|acc                                                                           |
+|arc_easy                                                 |✓    |✓  |✓   |         2376|acc, acc_norm                                                                 |
+|arc_challenge                                            |✓    |✓  |✓   |         1172|acc, acc_norm                                                                 |
+|logiqa                                                   |✓    |✓  |✓   |          651|acc, acc_norm                                                                 |
+|hellaswag                                                |✓    |✓  |    |        10042|acc, acc_norm                                                                 |
+|openbookqa                                               |✓    |✓  |✓   |          500|acc, acc_norm                                                                 |
+|squad2                                                   |✓    |✓  |    |        11873|exact, f1, HasAns_exact, HasAns_f1, NoAns_exact, NoAns_f1, best_exact, best_f1|
+|race                                                     |✓    |✓  |✓   |         1045|acc                                                                           |
+|headqa                                                   |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
+|mathqa                                                   |✓    |✓  |✓   |         2985|acc, acc_norm                                                                 |
+|webqs                                                    |✓    |   |✓   |         2032|acc                                                                           |
+|wsc273                                                   |     |   |✓   |          273|acc                                                                           |
+|winogrande                                               |✓    |✓  |    |         1267|acc                                                                           |
+|anli_r1                                                  |✓    |✓  |✓   |         1000|acc                                                                           |
+|anli_r2                                                  |✓    |✓  |✓   |         1000|acc                                                                           |
+|anli_r3                                                  |✓    |✓  |✓   |         1200|acc                                                                           |
+|ethics_cm                                                |✓    |   |✓   |         3885|acc                                                                           |
+|ethics_deontology                                        |✓    |   |✓   |         3596|acc, em                                                                       |
+|ethics_justice                                           |✓    |   |✓   |         2704|acc, em                                                                       |
+|ethics_utilitarianism_original                           |     |   |✓   |         4808|acc                                                                           |
+|ethics_utilitarianism                                    |✓    |   |✓   |         4808|acc                                                                           |
+|ethics_virtue                                            |✓    |   |✓   |         4975|acc, em                                                                       |
+|math_algebra                                             |✓    |   |✓   |         1187|acc                                                                           |
+|math_counting_and_prob                                   |✓    |   |✓   |          474|acc                                                                           |
+|math_geometry                                            |✓    |   |✓   |          479|acc                                                                           |
+|math_intermediate_algebra                                |✓    |   |✓   |          903|acc                                                                           |
+|math_num_theory                                          |✓    |   |✓   |          540|acc                                                                           |
+|math_prealgebra                                          |✓    |   |✓   |          871|acc                                                                           |
+|math_precalc                                             |✓    |   |✓   |          546|acc                                                                           |
+|arithmetic_2da                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_2ds                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_3da                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_3ds                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_4da                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_4ds                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_5da                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_5ds                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_2dm                                           |     |✓  |    |         2000|acc                                                                           |
+|arithmetic_1dc                                           |     |✓  |    |         2000|acc                                                                           |
+|hendrycksTest-abstract_algebra                           |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-anatomy                                    |✓    |✓  |✓   |          135|acc, acc_norm                                                                 |
+|hendrycksTest-astronomy                                  |✓    |✓  |✓   |          152|acc, acc_norm                                                                 |
+|hendrycksTest-business_ethics                            |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-clinical_knowledge                         |✓    |✓  |✓   |          265|acc, acc_norm                                                                 |
+|hendrycksTest-college_biology                            |✓    |✓  |✓   |          144|acc, acc_norm                                                                 |
+|hendrycksTest-college_chemistry                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-college_computer_science                   |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-college_mathematics                        |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-college_medicine                           |✓    |✓  |✓   |          173|acc, acc_norm                                                                 |
+|hendrycksTest-college_physics                            |✓    |✓  |✓   |          102|acc, acc_norm                                                                 |
+|hendrycksTest-computer_security                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-conceptual_physics                         |✓    |✓  |✓   |          235|acc, acc_norm                                                                 |
+|hendrycksTest-econometrics                               |✓    |✓  |✓   |          114|acc, acc_norm                                                                 |
+|hendrycksTest-electrical_engineering                     |✓    |✓  |✓   |          145|acc, acc_norm                                                                 |
+|hendrycksTest-elementary_mathematics                     |✓    |✓  |✓   |          378|acc, acc_norm                                                                 |
+|hendrycksTest-formal_logic                               |✓    |✓  |✓   |          126|acc, acc_norm                                                                 |
+|hendrycksTest-global_facts                               |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_biology                        |✓    |✓  |✓   |          310|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_chemistry                      |✓    |✓  |✓   |          203|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_computer_science               |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_european_history               |✓    |✓  |✓   |          165|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_geography                      |✓    |✓  |✓   |          198|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_government_and_politics        |✓    |✓  |✓   |          193|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_macroeconomics                 |✓    |✓  |✓   |          390|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_mathematics                    |✓    |✓  |✓   |          270|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_microeconomics                 |✓    |✓  |✓   |          238|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_physics                        |✓    |✓  |✓   |          151|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_psychology                     |✓    |✓  |✓   |          545|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_statistics                     |✓    |✓  |✓   |          216|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_us_history                     |✓    |✓  |✓   |          204|acc, acc_norm                                                                 |
+|hendrycksTest-high_school_world_history                  |✓    |✓  |✓   |          237|acc, acc_norm                                                                 |
+|hendrycksTest-human_aging                                |✓    |✓  |✓   |          223|acc, acc_norm                                                                 |
+|hendrycksTest-human_sexuality                            |✓    |✓  |✓   |          131|acc, acc_norm                                                                 |
+|hendrycksTest-international_law                          |✓    |✓  |✓   |          121|acc, acc_norm                                                                 |
+|hendrycksTest-jurisprudence                              |✓    |✓  |✓   |          108|acc, acc_norm                                                                 |
+|hendrycksTest-logical_fallacies                          |✓    |✓  |✓   |          163|acc, acc_norm                                                                 |
+|hendrycksTest-machine_learning                           |✓    |✓  |✓   |          112|acc, acc_norm                                                                 |
+|hendrycksTest-management                                 |✓    |✓  |✓   |          103|acc, acc_norm                                                                 |
+|hendrycksTest-marketing                                  |✓    |✓  |✓   |          234|acc, acc_norm                                                                 |
+|hendrycksTest-medical_genetics                           |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-miscellaneous                              |✓    |✓  |✓   |          783|acc, acc_norm                                                                 |
+|hendrycksTest-moral_disputes                             |✓    |✓  |✓   |          346|acc, acc_norm                                                                 |
+|hendrycksTest-moral_scenarios                            |✓    |✓  |✓   |          895|acc, acc_norm                                                                 |
+|hendrycksTest-nutrition                                  |✓    |✓  |✓   |          306|acc, acc_norm                                                                 |
+|hendrycksTest-philosophy                                 |✓    |✓  |✓   |          311|acc, acc_norm                                                                 |
+|hendrycksTest-prehistory                                 |✓    |✓  |✓   |          324|acc, acc_norm                                                                 |
+|hendrycksTest-professional_accounting                    |✓    |✓  |✓   |          282|acc, acc_norm                                                                 |
+|hendrycksTest-professional_law                           |✓    |✓  |✓   |         1534|acc, acc_norm                                                                 |
+|hendrycksTest-professional_medicine                      |✓    |✓  |✓   |          272|acc, acc_norm                                                                 |
+|hendrycksTest-professional_psychology                    |✓    |✓  |✓   |          612|acc, acc_norm                                                                 |
+|hendrycksTest-public_relations                           |✓    |✓  |✓   |          110|acc, acc_norm                                                                 |
+|hendrycksTest-security_studies                           |✓    |✓  |✓   |          245|acc, acc_norm                                                                 |
+|hendrycksTest-sociology                                  |✓    |✓  |✓   |          201|acc, acc_norm                                                                 |
+|hendrycksTest-us_foreign_policy                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
+|hendrycksTest-virology                                   |✓    |✓  |✓   |          166|acc, acc_norm                                                                 |
+|hendrycksTest-world_religions                            |✓    |✓  |✓   |          171|acc, acc_norm                                                                 |
+|wmt14-en-fr                                              |     |   |✓   |         3003|bleu, chrf, ter                                                               |
+|wmt14-fr-en                                              |     |   |✓   |         3003|bleu, chrf, ter                                                               |
+|wmt16-en-ro                                              |     |   |✓   |         1999|bleu, chrf, ter                                                               |
+|wmt16-ro-en                                              |     |   |✓   |         1999|bleu, chrf, ter                                                               |
+|wmt16-de-en                                              |     |   |✓   |         2999|bleu, chrf, ter                                                               |
+|wmt16-en-de                                              |     |   |✓   |         2999|bleu, chrf, ter                                                               |
+|wmt20-cs-en                                              |     |   |✓   |          664|bleu, chrf, ter                                                               |
+|wmt20-de-en                                              |     |   |✓   |          785|bleu, chrf, ter                                                               |
+|wmt20-de-fr                                              |     |   |✓   |         1619|bleu, chrf, ter                                                               |
+|wmt20-en-cs                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
+|wmt20-en-de                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
+|wmt20-en-iu                                              |     |   |✓   |         2971|bleu, chrf, ter                                                               |
+|wmt20-en-ja                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
+|wmt20-en-km                                              |     |   |✓   |         2320|bleu, chrf, ter                                                               |
+|wmt20-en-pl                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
+|wmt20-en-ps                                              |     |   |✓   |         2719|bleu, chrf, ter                                                               |
+|wmt20-en-ru                                              |     |   |✓   |         2002|bleu, chrf, ter                                                               |
+|wmt20-en-ta                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
+|wmt20-en-zh                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
+|wmt20-fr-de                                              |     |   |✓   |         1619|bleu, chrf, ter                                                               |
+|wmt20-iu-en                                              |     |   |✓   |         2971|bleu, chrf, ter                                                               |
+|wmt20-ja-en                                              |     |   |✓   |          993|bleu, chrf, ter                                                               |
+|wmt20-km-en                                              |     |   |✓   |         2320|bleu, chrf, ter                                                               |
+|wmt20-pl-en                                              |     |   |✓   |         1001|bleu, chrf, ter                                                               |
+|wmt20-ps-en                                              |     |   |✓   |         2719|bleu, chrf, ter                                                               |
+|wmt20-ru-en                                              |     |   |✓   |          991|bleu, chrf, ter                                                               |
+|wmt20-ta-en                                              |     |   |✓   |          997|bleu, chrf, ter                                                               |
+|wmt20-zh-en                                              |     |   |✓   |         2000|bleu, chrf, ter                                                               |
+|iwslt17-en-ar                                            |     |   |✓   |         1460|bleu, chrf, ter                                                               |
+|iwslt17-ar-en                                            |     |   |✓   |         1460|bleu, chrf, ter                                                               |
+|anagrams1                                                |     |✓  |    |        10000|acc                                                                           |
+|anagrams2                                                |     |✓  |    |        10000|acc                                                                           |
+|cycle_letters                                            |     |✓  |    |        10000|acc                                                                           |
+|random_insertion                                         |     |✓  |    |        10000|acc                                                                           |
+|reversed_words                                           |     |✓  |    |        10000|acc                                                                           |
+|pile_arxiv                                               |     |✓  |✓   |         2407|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_books3                                              |     |✓  |✓   |          269|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_bookcorpus2                                         |     |✓  |✓   |           28|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_dm-mathematics                                      |     |✓  |✓   |         1922|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_enron                                               |     |✓  |✓   |         1010|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_europarl                                            |     |✓  |✓   |          157|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_freelaw                                             |     |✓  |✓   |         5101|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_github                                              |     |✓  |✓   |        18195|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_gutenberg                                           |     |✓  |✓   |           80|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_hackernews                                          |     |✓  |✓   |         1632|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_nih-exporter                                        |     |✓  |✓   |         1884|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_opensubtitles                                       |     |✓  |✓   |          642|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_openwebtext2                                        |     |✓  |✓   |        32925|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_philpapers                                          |     |✓  |✓   |           68|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_pile-cc                                             |     |✓  |✓   |        52790|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_pubmed-abstracts                                    |     |✓  |✓   |        29895|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_pubmed-central                                      |     |✓  |✓   |         5911|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_stackexchange                                       |     |✓  |✓   |        30378|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_uspto                                               |     |✓  |✓   |        11415|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_ubuntu-irc                                          |     |✓  |✓   |           22|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_wikipedia                                           |     |✓  |✓   |        17511|word_perplexity, byte_perplexity, bits_per_byte                               |
+|pile_youtubesubtitles                                    |     |✓  |    |         1000|acc
+|blimp_adjunct_island                                     |     |✓  |    |         1000|acc
+|blimp_anaphor_gender_agreement                           |     |✓  |    |         1000|acc
+|blimp_anaphor_number_agreement                           |     |✓  |    |         1000|acc
+|blimp_animate_subject_passive                            |     |✓  |    |         1000|acc
+|blimp_animate_subject_trans                              |     |✓  |    |         1000|acc
+|blimp_causative                                          |     |✓  |    |         1000|acc
+|blimp_complex_NP_island                                  |     |✓  |    |         1000|acc
+|blimp_coordinate_structure_constraint_complex_left_branch|     |✓  |    |         1000|acc
+|blimp_coordinate_structure_constraint_object_extraction  |     |✓  |    |         1000|acc
+|blimp_determiner_noun_agreement_1                        |     |✓  |    |         1000|acc
+|blimp_determiner_noun_agreement_2                        |     |✓  |    |         1000|acc
+|blimp_determiner_noun_agreement_irregular_1              |     |✓  |    |         1000|acc
+|blimp_determiner_noun_agreement_irregular_2              |     |✓  |    |         1000|acc
+|blimp_determiner_noun_agreement_with_adj_2               |     |✓  |    |         1000|acc
+|blimp_determiner_noun_agreement_with_adj_irregular_1     |     |✓  |    |         1000|acc
+|blimp_determiner_noun_agreement_with_adj_irregular_2     |     |✓  |    |         1000|acc
+|blimp_determiner_noun_agreement_with_adjective_1         |     |✓  |    |         1000|acc
+|blimp_distractor_agreement_relational_noun               |     |✓  |    |         1000|acc
+|blimp_distractor_agreement_relative_clause               |     |✓  |    |         1000|acc
+|blimp_drop_argument                                      |     |✓  |    |         1000|acc
+|blimp_ellipsis_n_bar_1                                   |     |✓  |    |         1000|acc
+|blimp_ellipsis_n_bar_2                                   |     |✓  |    |         1000|acc
+|blimp_existential_there_object_raising                   |     |✓  |    |         1000|acc
+|blimp_existential_there_quantifiers_1                    |     |✓  |    |         1000|acc
+|blimp_existential_there_quantifiers_2                    |     |✓  |    |         1000|acc
+|blimp_existential_there_subject_raising                  |     |✓  |    |         1000|acc
+|blimp_expletive_it_object_raising                        |     |✓  |    |         1000|acc
+|blimp_inchoative                                         |     |✓  |    |         1000|acc
+|blimp_intransitive                                       |     |✓  |    |         1000|acc
+|blimp_irregular_past_participle_adjectives               |     |✓  |    |         1000|acc
+|blimp_irregular_past_participle_verbs                    |     |✓  |    |         1000|acc
+|blimp_irregular_plural_subject_verb_agreement_1          |     |✓  |    |         1000|acc
+|blimp_irregular_plural_subject_verb_agreement_2          |     |✓  |    |         1000|acc
+|blimp_left_branch_island_echo_question                   |     |✓  |    |         1000|acc
+|blimp_left_branch_island_simple_question                 |     |✓  |    |         1000|acc
+|blimp_matrix_question_npi_licensor_present               |     |✓  |    |         1000|acc
+|blimp_npi_present_1                                      |     |✓  |    |         1000|acc
+|blimp_npi_present_2                                      |     |✓  |    |         1000|acc
+|blimp_only_npi_licensor_present                          |     |✓  |    |         1000|acc
+|blimp_only_npi_scope                                     |     |✓  |    |         1000|acc
+|blimp_passive_1                                          |     |✓  |    |         1000|acc
+|blimp_passive_2                                          |     |✓  |    |         1000|acc
+|blimp_principle_A_c_command                              |     |✓  |    |         1000|acc
+|blimp_principle_A_case_1                                 |     |✓  |    |         1000|acc
+|blimp_principle_A_case_2                                 |     |✓  |    |         1000|acc
+|blimp_principle_A_domain_1                               |     |✓  |    |         1000|acc
+|blimp_principle_A_domain_2                               |     |✓  |    |         1000|acc
+|blimp_principle_A_domain_3                               |     |✓  |    |         1000|acc
+|blimp_principle_A_reconstruction                         |     |✓  |    |         1000|acc
+|blimp_regular_plural_subject_verb_agreement_1            |     |✓  |    |         1000|acc
+|blimp_regular_plural_subject_verb_agreement_2            |     |✓  |    |         1000|acc
+|blimp_sentential_negation_npi_licensor_present           |     |✓  |    |         1000|acc
+|blimp_sentential_negation_npi_scope                      |     |✓  |    |         1000|acc
+|blimp_sentential_subject_island                          |     |✓  |    |         1000|acc
+|blimp_superlative_quantifiers_1                          |     |✓  |    |         1000|acc
+|blimp_superlative_quantifiers_2                          |     |✓  |    |         1000|acc
+|blimp_tough_vs_raising_1                                 |     |✓  |    |         1000|acc
+|blimp_tough_vs_raising_2                                 |     |✓  |    |         1000|acc
+|blimp_transitive                                         |     |✓  |    |         1000|acc
+|blimp_wh_island                                          |     |✓  |    |         1000|acc
+|blimp_wh_questions_object_gap                            |     |✓  |    |         1000|acc
+|blimp_wh_questions_subject_gap                           |     |✓  |    |         1000|acc
+|blimp_wh_questions_subject_gap_long_distance             |     |✓  |    |         1000|acc
+|blimp_wh_vs_that_no_gap                                  |     |✓  |    |         1000|acc
+|blimp_wh_vs_that_no_gap_long_distance                    |     |✓  |    |         1000|acc
+|blimp_wh_vs_that_with_gap                                |     |✓  |    |         1000|acc
+|blimp_wh_vs_that_with_gap_long_distance                  |     |✓  |    |         1000|acc
+
+
+## Usage
+
+### Evaluate a task
+
+Additional arguments can be provided to the model constructor using the `--model_args` flag. Most importantly, the `gpt2` model can be used to load an arbitrary HuggingFace model as follows:
+
+
+```bash
+python main.py \
+	--model gpt2 \
+	--model_args pretrained=EleutherAI/gpt-neo-1.3B \
+	--device cuda:0 \
+	--tasks lambada,hellaswag \
+	--num_fewshot 2
+```
+
+To inspect what the LM inputs look like, you can run the following command:
+
+```bash
+python write_out.py \
+	--tasks all_tasks \
+	--provide_description \
+	--num_fewshot 5 \
+	--num_examples 10 \
+	--output_base_path /path/to/output/folder
+```
+
+This will write out one text file for each task.
+
+### Code Structure
+
+There are two major components of the library:
+
+1. LMs (language models), e.g. GPT-2, GPT-3
+2. Tasks, e.g. MNLI, RTE, SQuAD (coming soon)
+
+Both LMs (`lm_eval.models`) and Tasks (`lm_eval.tasks`) are kept in a registry data structure, for easy CLI instantiation.
+
+**If you want to extend either models or tasks, simply add a new LM or Task subclass, and decorate with the registry decorator**.
+
+The [GPT-3 Evaluations Project](https://github.com/EleutherAI/lm_evaluation_harness/projects/1) tracks our progress implementing new tasks. Right now, we are focused on getting all the datasets loaded so that we can dedupe against the training data. Implementing the actual evaluations is nice but not necessary at the current moment.
+
+### Task Versioning 
+
+To help improve reproducibility, all tasks have a VERSION field. When run from the command line, this is reported in a column in the table, or in the "version" field in the evaluator return dict. The purpose of the version is so that if the task definition changes (i.e to fix a bug), then we can know exactly which metrics were computed using the old buggy implementation to avoid unfair comparisons. To enforce this, there are unit tests that make sure the behavior of all tests remains the same as when they were first implemented. Task versions start at 0, and each time a breaking change is made, the version is incremented by one. 
+
+When reporting eval harness results, please also report the version of each task. This can be done either with a separate column in the table, or by reporting the task name with the version appended as such: taskname-v0.
+
+## Description
+
+### 1. LM Evaluation
+Given an LM, we want to evaluate it on a wide range of NLU tasks. We should at least cover the set of tasks in the GPT-3 paper, and any other tasks/benchmarks that are relevant. We will follow the GPT-3 format of a) zero-shot, b) one-shot, c) few-shot evaluation.
+
+To do this, we need 3 components:
+* Data downloader (shared with later sections, potentially needs to be directly linked to the latter 2 components)
+* Task formatter
+* Task evaluator
+
+The **data downloader** should download data for the relevant tasks.
+* We should heavily rely on Hugging Face's NLP for this. They are already doing most of the work with handling data scripts/caching.
+* Optionally, we can rely directly on HF-NLP's caching, but that makes it awkward to handle non-HF-NLP datasets. Otherwise, we can just write them out to .jsonl. My feeling is that NLU data storage will be a drop in the bucket compared to LM data.
+* Where we're not using HF-NLP, we can keep the data in the raw format (.jsonl, tsv, etc) and let the other components handle transforming it.
+
+The **task formatter** formats the task input data into an LM-usable format.
+* We should potentially support multiple formats for a given task, e.g. some formats may be better or worse suited for LM evaluation. See also: prompt-engineering
+* The task formatter should also support zero/one/few-shot packing of training examples into an input. This may require weird interactions with the tokenizer for dealing with max-token issues.
+
+The **task evaluator** scores a task.
+* In essence, we want to generation output predictions for all our input examples, and feed them into some function that pops out a score (or scores)
+An alternative approach is to collect the output logits and score them against the expected set of outputs.
+* Some tasks have weird evaluation schemes, so we should make this as general as possible.
+* Will thus likely have to be closely tied with the formatter.
+* Likewise, we should take advantage of HF-NLP's metrics.
+We might as well provide a sufficiently general API for the model to support OpenAI API as well. This can double up as an effort to reproduce the OpenAI NLU results.
+
+### 2. Removing val/test data from LM training set
+With the data downloader in place, we simply need to (1) expose the val/test examples, and (2) remove them from the training set.
+
+* Arguably, (2) should be handled by LM preprocessing in a more general way. There are probably non-NLU-eval cases where we want to remove some specific data from training.
+* Depending on how exactly we do the val/test removal, we may want to format the same example multiple ways to ensure that they don't get leaked into the training set in a slightly tweaked format.
+* Thought experiment: SQuAD is based largely on Wikipedia. What exactly would we want to remove from the LM?
+* [GPT-3]: In GPT-3, they attempted to remove val/test from their LM set, but there was a bug that caused leakage. So they ended up doing the opposite: removing overlaps from the LM set from the val/test. Funky.
+* [GPT-3]: See page 30 and Appendix C for details. They do some funky n-gram based search and removal. We should think about whether we want to follow their protocol exactly
+
+### 3. Adding task training data to LM training set
+This part is the easiest. I guess we just write out some text files containing the training data? We can let the usual LM preprocessing pipeline handle it from there.
+
+
diff --git a/lm_eval.egg-info/SOURCES.txt b/lm_eval.egg-info/SOURCES.txt
new file mode 100644
index 0000000000..71a6a2fdb8
--- /dev/null
+++ b/lm_eval.egg-info/SOURCES.txt
@@ -0,0 +1,74 @@
+LICENSE.md
+README.md
+setup.py
+lm_eval/__init__.py
+lm_eval/base.py
+lm_eval/evaluator.py
+lm_eval/metrics.py
+lm_eval/utils.py
+lm_eval.egg-info/PKG-INFO
+lm_eval.egg-info/SOURCES.txt
+lm_eval.egg-info/dependency_links.txt
+lm_eval.egg-info/requires.txt
+lm_eval.egg-info/top_level.txt
+lm_eval/models/__init__.py
+lm_eval/models/dummy.py
+lm_eval/models/gpt2.py
+lm_eval/models/gpt3.py
+lm_eval/tasks/__init__.py
+lm_eval/tasks/anli.py
+lm_eval/tasks/arc.py
+lm_eval/tasks/arithmetic.py
+lm_eval/tasks/blimp.py
+lm_eval/tasks/cbt.py
+lm_eval/tasks/common.py
+lm_eval/tasks/coqa.py
+lm_eval/tasks/drop.py
+lm_eval/tasks/glue.py
+lm_eval/tasks/headqa.py
+lm_eval/tasks/hellaswag.py
+lm_eval/tasks/hendrycks_ethics.py
+lm_eval/tasks/hendrycks_math.py
+lm_eval/tasks/hendrycks_test.py
+lm_eval/tasks/lambada.py
+lm_eval/tasks/lambada_cloze.py
+lm_eval/tasks/lambada_multilingual.py
+lm_eval/tasks/logiqa.py
+lm_eval/tasks/mathqa.py
+lm_eval/tasks/mc_taco.py
+lm_eval/tasks/mutual.py
+lm_eval/tasks/naturalqs.py
+lm_eval/tasks/openbookqa.py
+lm_eval/tasks/pile.py
+lm_eval/tasks/piqa.py
+lm_eval/tasks/prost.py
+lm_eval/tasks/pubmedqa.py
+lm_eval/tasks/qa4mre.py
+lm_eval/tasks/quac.py
+lm_eval/tasks/race.py
+lm_eval/tasks/sat.py
+lm_eval/tasks/sciq.py
+lm_eval/tasks/squad.py
+lm_eval/tasks/storycloze.py
+lm_eval/tasks/superglue.py
+lm_eval/tasks/translation.py
+lm_eval/tasks/triviaqa.py
+lm_eval/tasks/truthfulqa.py
+lm_eval/tasks/unscramble.py
+lm_eval/tasks/webqs.py
+lm_eval/tasks/wikitext.py
+lm_eval/tasks/winogrande.py
+lm_eval/tasks/wsc273.py
+scripts/__init__.py
+scripts/cost_estimate.py
+scripts/fewshot_description_experiment.py
+scripts/get_prompts.py
+scripts/make_gpt2_test_cases.py
+scripts/make_table_tasks.py
+scripts/write_out.py
+scripts/clean_training_data/__init__.py
+scripts/clean_training_data/archiver.py
+scripts/clean_training_data/generate_13_grams.py
+scripts/clean_training_data/janitor.py
+scripts/clean_training_data/process_sorted_buckets.py
+scripts/clean_training_data/sort_13_gram_buckets.py
\ No newline at end of file
diff --git a/lm_eval.egg-info/dependency_links.txt b/lm_eval.egg-info/dependency_links.txt
new file mode 100644
index 0000000000..8b13789179
--- /dev/null
+++ b/lm_eval.egg-info/dependency_links.txt
@@ -0,0 +1 @@
+
diff --git a/lm_eval.egg-info/requires.txt b/lm_eval.egg-info/requires.txt
new file mode 100644
index 0000000000..c0f5b007a4
--- /dev/null
+++ b/lm_eval.egg-info/requires.txt
@@ -0,0 +1,24 @@
+black
+best_download>=0.0.6
+datasets==1.15.1
+click>=7.1
+scikit-learn>=0.24.1
+torch>=1.7
+transformers>=4.1
+sqlitedict==1.6.0
+pytablewriter==0.58.0
+sacrebleu==1.5.0
+rouge-score==0.0.4
+bleurt@ https://github.com/google-research/bleurt/archive/b610120347ef22b494b6d69b4316e303f5932516.zip#egg=bleurt
+pycountry==20.7.3
+numexpr==2.7.2
+lm_dataformat==0.0.20
+pytest==6.2.3
+pybind11==2.6.2
+tqdm-multiprocess==0.0.11
+zstandard==0.15.2
+jsonlines==2.0.0
+mock==4.0.3
+openai==0.6.4
+jieba==0.42.1
+nagisa==0.2.7
diff --git a/lm_eval.egg-info/top_level.txt b/lm_eval.egg-info/top_level.txt
new file mode 100644
index 0000000000..3c734591d0
--- /dev/null
+++ b/lm_eval.egg-info/top_level.txt
@@ -0,0 +1,2 @@
+lm_eval
+scripts
diff --git a/lm_eval/tasks/__init__.py b/lm_eval/tasks/__init__.py
index 53d7e88f16..86b61f68d8 100644
--- a/lm_eval/tasks/__init__.py
+++ b/lm_eval/tasks/__init__.py
@@ -45,6 +45,7 @@
 from . import mutual
 from . import truthfulqa
 from . import blimp
+from . import asdiv
 
 ########################################
 # Translation tasks
@@ -164,6 +165,7 @@
     "math_num_theory": hendrycks_math.MathNumberTheory,
     "math_prealgebra": hendrycks_math.MathPrealgebra,
     "math_precalc": hendrycks_math.MathPrecalculus,
+    "math_asdiv": asdiv.Asdiv,
 
     # arithmetic
     "arithmetic_2da": arithmetic.Arithmetic2DPlus,
diff --git a/lm_eval/tasks/apps.py b/lm_eval/tasks/apps.py
new file mode 100644
index 0000000000..f5d6a42cce
--- /dev/null
+++ b/lm_eval/tasks/apps.py
@@ -0,0 +1,130 @@
+"""
+Measuring Coding Challenge Competence With APPS
+https://arxiv.org/pdf/2105.09938
+
+@article{hendrycksapps2021,
+  title={Measuring Coding Challenge Competence With APPS},
+  author={Dan Hendrycks and Steven Basart and Saurav Kadavath and Mantas Mazeika and Akul Arora and Ethan Guo and Collin Burns and Samir Puranik and Horace He and Dawn Song and Jacob Steinhardt},
+  journal={NeurIPS},
+  year={2021}
+}
+"""
+from lm_eval.base import Task
+from pathlib import Path
+from best_download import download_file 
+import xml.etree.ElementTree as ET
+from lm_eval.base import rf
+from lm_eval.metrics import mean,perplexity
+import numpy as np
+from zipfile import ZipFile
+import os 
+
+class Apps(Task):
+    VERSION = 0
+    DATASET_PATH = Path("data/asdiv")
+
+    def download(self):
+        if self.DATASET_PATH.exists():
+            return
+        Path.mkdir(self.DATASET_PATH)
+        url = "https://people.eecs.berkeley.edu/~hendrycks/APPS.tar.gz"
+        checksum = "2f71f8003929d605369ad924be4b95c15879fc2bfac0d4d01a81f8aabceaad5c"
+        zip_path = self.DATASET_PATH / "master.zip"
+        download_file(url, str(zip_path), checksum)
+        with ZipFile(zip_path, "r") as zip:
+            zip.extractall(self.DATASET_PATH)
+        os.remove(zip_path)
+
+    def _convert_standard(self, problem):
+        #TODO: include solution-type and formula
+        out_doc = {
+            "question" : problem.find('Question').text,
+            "body" : problem.find('Body').text,
+            "answer": problem.find('Answer').text
+        }
+        return out_doc
+
+    def load_docs(self, textfilename, tfds=False):
+        tree = ET.parse(textfilename)
+        root = tree.getroot()
+        for pid, problem in enumerate(root.iter('Problem')):
+            out_doc = self._convert_standard(problem)
+            yield out_doc
+
+    def _strip_bracket(self,test_str):
+        ret = ''
+        skip1c = 0
+        skip2c = 0
+        for i in test_str:
+            if i == '(':
+                skip2c += 1
+            elif i == ')'and skip2c > 0:
+                skip2c -= 1
+            elif skip1c == 0 and skip2c == 0:
+                ret += i
+        return ret
+
+    def has_training_docs(self):
+        return False
+    
+    def has_validation_docs(self):
+        return True
+
+    def has_test_docs(self):
+        return False
+
+    def training_docs(self):
+        raise NotImplementedError("This dataset has no training docs")
+
+    def test_docs(self):
+        raise NotImplementedError("This dataset has no test docs")
+
+    def validation_docs(self):
+        data_xml_path = self.DATASET_PATH / "nlu-asdiv-dataset-master/dataset/ASDiv.xml"
+        return self.load_docs(data_xml_path)
+
+    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
+        assert num_fewshot == 0, "ASDiv is intended only for the zero-shot setting."
+        return super().fewshot_context(doc, num_fewshot, provide_description, rnd)
+
+    
+    def fewshot_description(self):
+        # TODO: add solution-type and formula
+        desc = "information containing the context of the question\nQuestion: Text of a question.\nAnswer: Answer to the question, based on the passage.\n"
+        return desc
+
+    def doc_to_text(self, doc):
+        # TODO: add solution-type
+        return doc['body'] + '\n' + 'Question:' + doc['question'] + '\n' + 'Answer:'
+
+    def doc_to_target(self, doc):
+        # TODO: add formula
+
+        answer = self._strip_bracket(doc['answer'])
+        if len(answer)>0: # check if answer is present only in brackets
+            return answer
+        else:
+            return doc['answer']
+
+    def construct_requests(self, doc, ctx):
+        ll, is_greedy = rf.loglikelihood(ctx, self.doc_to_target(doc))
+
+        return ll, is_greedy
+    
+    def process_results(self, doc, results):
+        ll, is_greedy = results
+
+        return {
+            'acc': int(is_greedy)
+        }
+        
+    def aggregation(self):
+        return {
+            'acc': mean
+        }
+
+    def higher_is_better(self):
+        return {
+            'acc': True
+        }
+
diff --git a/lm_eval/tasks/asdiv.py b/lm_eval/tasks/asdiv.py
new file mode 100644
index 0000000000..dcde1f1e23
--- /dev/null
+++ b/lm_eval/tasks/asdiv.py
@@ -0,0 +1,135 @@
+"""
+ASDiv: A Diverse Corpus for Evaluating and Developing English Math Word Problem Solvers
+https://arxiv.org/abs/2106.15772
+
+@misc{miao2021diverse,
+      title={A Diverse Corpus for Evaluating and Developing English Math Word Problem Solvers},
+      author={Shen-Yun Miao and Chao-Chun Liang and Keh-Yih Su},
+      year={2021},
+      eprint={2106.15772},
+      archivePrefix={arXiv},
+      primaryClass={cs.AI}
+}
+"""
+from lm_eval.base import Task
+from pathlib import Path
+from best_download import download_file 
+import xml.etree.ElementTree as ET
+from lm_eval.base import rf
+from lm_eval.metrics import mean,perplexity
+import numpy as np
+from zipfile import ZipFile
+import os 
+
+#currently ignoring formula for answer generation
+
+# given a subset, splits return the docs 
+class Asdiv(Task):
+    VERSION = 0
+    DATASET_PATH = Path("data/asdiv")
+
+    def download(self):
+        if self.DATASET_PATH.exists():
+            return
+        Path.mkdir(self.DATASET_PATH)
+        url = "https://github.com/chaochun/nlu-asdiv-dataset/archive/refs/heads/master.zip"
+        checksum = "2f71f8003929d605369ad924be4b95c15879fc2bfac0d4d01a81f8aabceaad5c"
+        zip_path = self.DATASET_PATH / "master.zip"
+        download_file(url, str(zip_path), checksum)
+        with ZipFile(zip_path, "r") as zip:
+            zip.extractall(self.DATASET_PATH)
+        os.remove(zip_path)
+
+    def _convert_standard(self, problem):
+        #TODO: include solution-type and formula
+        out_doc = {
+            "question" : problem.find('Question').text,
+            "body" : problem.find('Body').text,
+            "answer": problem.find('Answer').text
+        }
+        return out_doc
+
+    def load_docs(self, textfilename, tfds=False):
+        tree = ET.parse(textfilename)
+        root = tree.getroot()
+        for pid, problem in enumerate(root.iter('Problem')):
+            out_doc = self._convert_standard(problem)
+            yield out_doc
+
+    def _strip_bracket(self,test_str):
+        ret = ''
+        skip1c = 0
+        skip2c = 0
+        for i in test_str:
+            if i == '(':
+                skip2c += 1
+            elif i == ')'and skip2c > 0:
+                skip2c -= 1
+            elif skip1c == 0 and skip2c == 0:
+                ret += i
+        return ret
+
+    def has_training_docs(self):
+        return False
+    
+    def has_validation_docs(self):
+        return True
+
+    def has_test_docs(self):
+        return False
+
+    def training_docs(self):
+        raise NotImplementedError("This dataset has no training docs")
+
+    def test_docs(self):
+        raise NotImplementedError("This dataset has no test docs")
+
+    def validation_docs(self):
+        data_xml_path = self.DATASET_PATH / "nlu-asdiv-dataset-master/dataset/ASDiv.xml"
+        return self.load_docs(data_xml_path)
+
+    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
+        assert num_fewshot == 0, "ASDiv is intended only for the zero-shot setting."
+        return super().fewshot_context(doc, num_fewshot, provide_description, rnd)
+
+    
+    def fewshot_description(self):
+        # TODO: add solution-type and formula
+        desc = "information containing the context of the question\nQuestion: Text of a question.\nAnswer: Answer to the question, based on the passage.\n"
+        return desc
+
+    def doc_to_text(self, doc):
+        # TODO: add solution-type
+        return doc['body'] + '\n' + 'Question:' + doc['question'] + '\n' + 'Answer:'
+
+    def doc_to_target(self, doc):
+        # TODO: add formula
+
+        answer = self._strip_bracket(doc['answer'])
+        if len(answer)>0: # check if answer is present only in brackets
+            return answer
+        else:
+            return doc['answer']
+
+    def construct_requests(self, doc, ctx):
+        ll, is_greedy = rf.loglikelihood(ctx, self.doc_to_target(doc))
+
+        return ll, is_greedy
+    
+    def process_results(self, doc, results):
+        ll, is_greedy = results
+
+        return {
+            'acc': int(is_greedy)
+        }
+        
+    def aggregation(self):
+        return {
+            'acc': mean
+        }
+
+    def higher_is_better(self):
+        return {
+            'acc': True
+        }
+

From bce9f289f8eee207fbc56288492f1db0807961eb Mon Sep 17 00:00:00 2001
From: rokosbasilisk <thecomplexitytheorist@gmail.com>
Date: Sun, 26 Dec 2021 21:30:05 +0530
Subject: [PATCH 34/65] remove apps

---
 lm_eval/tasks/apps.py | 130 ------------------------------------------
 1 file changed, 130 deletions(-)
 delete mode 100644 lm_eval/tasks/apps.py

diff --git a/lm_eval/tasks/apps.py b/lm_eval/tasks/apps.py
deleted file mode 100644
index f5d6a42cce..0000000000
--- a/lm_eval/tasks/apps.py
+++ /dev/null
@@ -1,130 +0,0 @@
-"""
-Measuring Coding Challenge Competence With APPS
-https://arxiv.org/pdf/2105.09938
-
-@article{hendrycksapps2021,
-  title={Measuring Coding Challenge Competence With APPS},
-  author={Dan Hendrycks and Steven Basart and Saurav Kadavath and Mantas Mazeika and Akul Arora and Ethan Guo and Collin Burns and Samir Puranik and Horace He and Dawn Song and Jacob Steinhardt},
-  journal={NeurIPS},
-  year={2021}
-}
-"""
-from lm_eval.base import Task
-from pathlib import Path
-from best_download import download_file 
-import xml.etree.ElementTree as ET
-from lm_eval.base import rf
-from lm_eval.metrics import mean,perplexity
-import numpy as np
-from zipfile import ZipFile
-import os 
-
-class Apps(Task):
-    VERSION = 0
-    DATASET_PATH = Path("data/asdiv")
-
-    def download(self):
-        if self.DATASET_PATH.exists():
-            return
-        Path.mkdir(self.DATASET_PATH)
-        url = "https://people.eecs.berkeley.edu/~hendrycks/APPS.tar.gz"
-        checksum = "2f71f8003929d605369ad924be4b95c15879fc2bfac0d4d01a81f8aabceaad5c"
-        zip_path = self.DATASET_PATH / "master.zip"
-        download_file(url, str(zip_path), checksum)
-        with ZipFile(zip_path, "r") as zip:
-            zip.extractall(self.DATASET_PATH)
-        os.remove(zip_path)
-
-    def _convert_standard(self, problem):
-        #TODO: include solution-type and formula
-        out_doc = {
-            "question" : problem.find('Question').text,
-            "body" : problem.find('Body').text,
-            "answer": problem.find('Answer').text
-        }
-        return out_doc
-
-    def load_docs(self, textfilename, tfds=False):
-        tree = ET.parse(textfilename)
-        root = tree.getroot()
-        for pid, problem in enumerate(root.iter('Problem')):
-            out_doc = self._convert_standard(problem)
-            yield out_doc
-
-    def _strip_bracket(self,test_str):
-        ret = ''
-        skip1c = 0
-        skip2c = 0
-        for i in test_str:
-            if i == '(':
-                skip2c += 1
-            elif i == ')'and skip2c > 0:
-                skip2c -= 1
-            elif skip1c == 0 and skip2c == 0:
-                ret += i
-        return ret
-
-    def has_training_docs(self):
-        return False
-    
-    def has_validation_docs(self):
-        return True
-
-    def has_test_docs(self):
-        return False
-
-    def training_docs(self):
-        raise NotImplementedError("This dataset has no training docs")
-
-    def test_docs(self):
-        raise NotImplementedError("This dataset has no test docs")
-
-    def validation_docs(self):
-        data_xml_path = self.DATASET_PATH / "nlu-asdiv-dataset-master/dataset/ASDiv.xml"
-        return self.load_docs(data_xml_path)
-
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
-        assert num_fewshot == 0, "ASDiv is intended only for the zero-shot setting."
-        return super().fewshot_context(doc, num_fewshot, provide_description, rnd)
-
-    
-    def fewshot_description(self):
-        # TODO: add solution-type and formula
-        desc = "information containing the context of the question\nQuestion: Text of a question.\nAnswer: Answer to the question, based on the passage.\n"
-        return desc
-
-    def doc_to_text(self, doc):
-        # TODO: add solution-type
-        return doc['body'] + '\n' + 'Question:' + doc['question'] + '\n' + 'Answer:'
-
-    def doc_to_target(self, doc):
-        # TODO: add formula
-
-        answer = self._strip_bracket(doc['answer'])
-        if len(answer)>0: # check if answer is present only in brackets
-            return answer
-        else:
-            return doc['answer']
-
-    def construct_requests(self, doc, ctx):
-        ll, is_greedy = rf.loglikelihood(ctx, self.doc_to_target(doc))
-
-        return ll, is_greedy
-    
-    def process_results(self, doc, results):
-        ll, is_greedy = results
-
-        return {
-            'acc': int(is_greedy)
-        }
-        
-    def aggregation(self):
-        return {
-            'acc': mean
-        }
-
-    def higher_is_better(self):
-        return {
-            'acc': True
-        }
-

From 440216db7b1a3faebc0b1cb085e20192e13ce386 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Mon, 27 Dec 2021 11:46:45 +0100
Subject: [PATCH 35/65] Questions in BoolQ don't have interrogation punctuation
 at the end

---
 lm_eval/tasks/superglue.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lm_eval/tasks/superglue.py b/lm_eval/tasks/superglue.py
index 33598f2301..c26988107a 100644
--- a/lm_eval/tasks/superglue.py
+++ b/lm_eval/tasks/superglue.py
@@ -13,7 +13,7 @@
 
 
 class BoolQ(HFTask):
-    VERSION = 0
+    VERSION = 1
     DATASET_PATH = "super_glue"
     DATASET_NAME = "boolq"
 
@@ -31,7 +31,7 @@ def fewshot_description(self):
         return "Read the following passages and answer each question with a yes or a no."
 
     def doc_to_text(self, doc):
-        return f"{doc['passage']}\nQuestion: {doc['question']}\nAnswer:"
+        return f"{doc['passage']}\nQuestion: {doc['question']}?\nAnswer:"
     
     def doc_to_target(self, doc):
         return " " + yesno(doc['label']) 

From 97ca18e0854dbf01b68d0f77c249f5ebe64ec057 Mon Sep 17 00:00:00 2001
From: Leo Gao <54557097+leogao2@users.noreply.github.com>
Date: Tue, 28 Dec 2021 18:57:52 -0700
Subject: [PATCH 36/65] Update README.md

---
 README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/README.md b/README.md
index f0d910cca8..8a2d43998f 100644
--- a/README.md
+++ b/README.md
@@ -21,7 +21,7 @@ pip install lm-eval
 
 ## Basic Usage
 
-To evaluate a model, (e.g. GPT-2) on NLU tasks (e.g. LAMBADA, HellaSwag), you can run the following command.
+To evaluate a model, (e.g. GPT-2) on NLU tasks (e.g. LAMBADA, HellaSwag), you can run the following command. **When reporting results from eval harness, please include the task versions (shown in `results["versions"]`) for reproducibility.** This allows bug fixes to tasks while also ensuring that previously reported scores are reproducible.
 
 ```bash
 python main.py \

From 0d9d47da05170e3c9629b76360f614104f654056 Mon Sep 17 00:00:00 2001
From: Leo Gao <54557097+leogao2@users.noreply.github.com>
Date: Tue, 28 Dec 2021 19:00:01 -0700
Subject: [PATCH 37/65] Update README.md

---
 README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/README.md b/README.md
index 8a2d43998f..949cbdb74b 100644
--- a/README.md
+++ b/README.md
@@ -21,7 +21,7 @@ pip install lm-eval
 
 ## Basic Usage
 
-To evaluate a model, (e.g. GPT-2) on NLU tasks (e.g. LAMBADA, HellaSwag), you can run the following command. **When reporting results from eval harness, please include the task versions (shown in `results["versions"]`) for reproducibility.** This allows bug fixes to tasks while also ensuring that previously reported scores are reproducible.
+To evaluate a model, (e.g. GPT-2) on NLU tasks (e.g. LAMBADA, HellaSwag), you can run the following command. **When reporting results from eval harness, please include the task versions (shown in `results["versions"]`) for reproducibility.** This allows bug fixes to tasks while also ensuring that previously reported scores are reproducible. See the [Task Versioning](https://github.com/EleutherAI/lm-evaluation-harness#task-versioning) section for more info.
 
 ```bash
 python main.py \

From 5a53c3684523cd873bdef18cb8264fecd82a3f21 Mon Sep 17 00:00:00 2001
From: Leo Gao <54557097+leogao2@users.noreply.github.com>
Date: Tue, 28 Dec 2021 19:20:10 -0700
Subject: [PATCH 38/65] pile: Switch download over to backup host temporarily

We can have fallback instead after @researcher2 implements it in best_download
---
 lm_eval/tasks/pile.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lm_eval/tasks/pile.py b/lm_eval/tasks/pile.py
index 68ff7ed9a8..6936905f0c 100644
--- a/lm_eval/tasks/pile.py
+++ b/lm_eval/tasks/pile.py
@@ -19,8 +19,8 @@ class PilePerplexityTask(PerplexityTask, abc.ABC):
     def download(self):
         # TODO: separate pile val/test out by component so we don't have to scan the entire file once per set
         os.makedirs("data/pile/", exist_ok=True)
-        download_file("https://the-eye.eu/public/AI/pile/val.jsonl.zst", self.VAL_PATH, "264c875d8bbd355d8daa9d032b75fd8fb91606218bb84dd1155b203fcd5fab92")
-        download_file("https://the-eye.eu/public/AI/pile/test.jsonl.zst", self.TEST_PATH, "0bb28c52d0b5596d389bf179ce2d43bf7f7ffae76b0d2d20b180c97f62e0975e")
+        download_file("http://eaidata.bmk.sh/data/pile/val.jsonl.zst", self.VAL_PATH, "264c875d8bbd355d8daa9d032b75fd8fb91606218bb84dd1155b203fcd5fab92")
+        download_file("http://eaidata.bmk.sh/data/pile/test.jsonl.zst", self.TEST_PATH, "0bb28c52d0b5596d389bf179ce2d43bf7f7ffae76b0d2d20b180c97f62e0975e")
 
     def validation_docs(self):
         rdr = lm_dataformat.Reader(self.VAL_PATH)

From f42a8d6619a321ca270b1159702f2d9874cbb375 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Wed, 29 Dec 2021 02:27:47 +0000
Subject: [PATCH 39/65] Add testdata for boolq v1

---
 tests/testdata/boolq-v1-loglikelihood | 1 +
 tests/testdata/boolq-v1-res.json      | 1 +
 2 files changed, 2 insertions(+)
 create mode 100644 tests/testdata/boolq-v1-loglikelihood
 create mode 100644 tests/testdata/boolq-v1-res.json

diff --git a/tests/testdata/boolq-v1-loglikelihood b/tests/testdata/boolq-v1-loglikelihood
new file mode 100644
index 0000000000..7811121c9f
--- /dev/null
+++ b/tests/testdata/boolq-v1-loglikelihood
@@ -0,0 +1 @@
+6577e0d88572772ef08e64f624c0e3df0953286ae1f118ccef15623b59ffeabf
\ No newline at end of file
diff --git a/tests/testdata/boolq-v1-res.json b/tests/testdata/boolq-v1-res.json
new file mode 100644
index 0000000000..291b9f122d
--- /dev/null
+++ b/tests/testdata/boolq-v1-res.json
@@ -0,0 +1 @@
+{"results": {"boolq": {"acc": 0.5048929663608562, "acc_stderr": 0.00874463623355505}}, "versions": {"boolq": 1}}
\ No newline at end of file

From 0bde758966b7787d368412cc8ab02869d3463523 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Wed, 29 Dec 2021 02:32:49 +0000
Subject: [PATCH 40/65] Revert "Add capital letters"

This reverts commit 23a420674b49476c5be66b0bf4e0d487ca38f1c7.
---
 lm_eval/tasks/superglue.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lm_eval/tasks/superglue.py b/lm_eval/tasks/superglue.py
index f3d07b3483..ae21bcedfa 100644
--- a/lm_eval/tasks/superglue.py
+++ b/lm_eval/tasks/superglue.py
@@ -227,7 +227,7 @@ def doc_to_target(self, doc):
 
     @staticmethod
     def format_answer(answer, label):
-        label_str = "Yes" if label else "No"
+        label_str = "yes" if label else "no"
         return f"{answer}\nIs the answer correct? {label_str}"
 
     def construct_requests(self, doc, ctx):

From d7e3248990acc20e48c2facc34840384acfdc1e0 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Wed, 29 Dec 2021 04:40:09 +0000
Subject: [PATCH 41/65] Pretty sure questions need to be paragraph dependent
 also

---
 lm_eval/metrics.py         | 7 ++++---
 lm_eval/tasks/superglue.py | 2 +-
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/lm_eval/metrics.py b/lm_eval/metrics.py
index c95d4cd61c..749159b09a 100644
--- a/lm_eval/metrics.py
+++ b/lm_eval/metrics.py
@@ -52,13 +52,14 @@ def acc_all(items):
     docs = list(zip(*items))[1]
 
     for doc, pred in zip(docs, preds):
+        paragraph_id = doc["idx"]["paragraph"]
         question_id = doc["idx"]["question"]
-        if question_id not in question_scoring_dict:
-            question_scoring_dict[question_id] = []
+        if (paragraph_id, question_id) not in question_scoring_dict:
+            question_scoring_dict[(paragraph_id, question_id)] = []
 
         gold_label = doc["label"] == 1
-        question_scoring_dict[question_id].append(gold_label == pred)
 
+        question_scoring_dict[(paragraph_id, question_id)].append(gold_label == pred)
     acc = np.mean([int(all(x)) for x in question_scoring_dict.values()])
     return acc
 
diff --git a/lm_eval/tasks/superglue.py b/lm_eval/tasks/superglue.py
index ae21bcedfa..2ce5810352 100644
--- a/lm_eval/tasks/superglue.py
+++ b/lm_eval/tasks/superglue.py
@@ -202,7 +202,7 @@ def convert_choice(choice):
 
 
 class MultiRC(HFTask):
-    VERSION = 0
+    VERSION = 1
     DATASET_PATH = "super_glue"
     DATASET_NAME = "multirc"
 

From 73d0ae5e448c36ff663fa82376c78f6b43a6cee8 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Wed, 29 Dec 2021 04:40:56 +0000
Subject: [PATCH 42/65] Add testdata for multirc v1

---
 tests/testdata/multirc-v1-loglikelihood | 1 +
 tests/testdata/multirc-v1-res.json      | 1 +
 2 files changed, 2 insertions(+)
 create mode 100644 tests/testdata/multirc-v1-loglikelihood
 create mode 100644 tests/testdata/multirc-v1-res.json

diff --git a/tests/testdata/multirc-v1-loglikelihood b/tests/testdata/multirc-v1-loglikelihood
new file mode 100644
index 0000000000..52a89c6f9e
--- /dev/null
+++ b/tests/testdata/multirc-v1-loglikelihood
@@ -0,0 +1 @@
+0e793bd6f637a70a04c6f2cda080188fc037961b2f909095fe63f7bdbc4a90c6
\ No newline at end of file
diff --git a/tests/testdata/multirc-v1-res.json b/tests/testdata/multirc-v1-res.json
new file mode 100644
index 0000000000..938141bbb8
--- /dev/null
+++ b/tests/testdata/multirc-v1-res.json
@@ -0,0 +1 @@
+{"results": {"multirc": {"acc": 0.046169989506820566, "acc_stderr": 0.006801377886208738}}, "versions": {"multirc": 1}}
\ No newline at end of file

From 04635731923251fd79500dc0e1e356f43e666df5 Mon Sep 17 00:00:00 2001
From: rokosbasilisk <thecomplexitytheorist@gmail.com>
Date: Wed, 29 Dec 2021 22:55:23 +0530
Subject: [PATCH 43/65] remove unrequired files&add pin commit hash

---
 eval.sh                               |   1 -
 lm_eval.egg-info/PKG-INFO             | 448 --------------------------
 lm_eval.egg-info/SOURCES.txt          |  74 -----
 lm_eval.egg-info/dependency_links.txt |   1 -
 lm_eval.egg-info/requires.txt         |  24 --
 lm_eval.egg-info/top_level.txt        |   2 -
 lm_eval/tasks/asdiv.py                |  11 +-
 7 files changed, 5 insertions(+), 556 deletions(-)
 delete mode 100644 eval.sh
 delete mode 100644 lm_eval.egg-info/PKG-INFO
 delete mode 100644 lm_eval.egg-info/SOURCES.txt
 delete mode 100644 lm_eval.egg-info/dependency_links.txt
 delete mode 100644 lm_eval.egg-info/requires.txt
 delete mode 100644 lm_eval.egg-info/top_level.txt

diff --git a/eval.sh b/eval.sh
deleted file mode 100644
index 68c468f4ba..0000000000
--- a/eval.sh
+++ /dev/null
@@ -1 +0,0 @@
-python main.py --model gpt2 --model_args pretrained=EleutherAI/gpt-neo-125M --device cuda:0 --tasks math_asdiv
diff --git a/lm_eval.egg-info/PKG-INFO b/lm_eval.egg-info/PKG-INFO
deleted file mode 100644
index f4d1cbda8d..0000000000
--- a/lm_eval.egg-info/PKG-INFO
+++ /dev/null
@@ -1,448 +0,0 @@
-Metadata-Version: 2.1
-Name: lm-eval
-Version: 0.1.0
-Summary: A framework for evaluating autoregressive language models
-Home-page: https://github.com/EleutherAI/lm-evaluation-harness
-Author: Leo Gao
-Author-email: lg@eleuther.ai
-License: UNKNOWN
-Platform: UNKNOWN
-Classifier: Programming Language :: Python :: 3
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Operating System :: OS Independent
-Requires-Python: >=3.6
-Description-Content-Type: text/markdown
-License-File: LICENSE.md
-
-# Language Model Evaluation Harness
-
-![](https://github.com/EleutherAI/lm-evaluation-harness/workflows/Build/badge.svg)
-[![codecov](https://codecov.io/gh/EleutherAI/lm-evaluation-harness/branch/master/graph/badge.svg?token=JSG3O2427J)](https://codecov.io/gh/EleutherAI/lm-evaluation-harness)
-
-## Overview 
-
-This project provides a unified framework to test autoregressive language models (GPT-2, GPT-3, GPTNeo, etc) on a large number of different evaluation tasks.
-
-Features:
-
-- 100+ tasks implemented
-- Support for GPT-2, GPT-3, GPT-Neo, GPT-NeoX, and GPT-J, with flexible tokenization-agnostic interface
-- Task versioning to ensure reproducibility
-
-## Install
-
-```bash
-pip install lm-eval
-```
-
-## Basic Usage
-
-To evaluate a model, (e.g. GPT-2) on NLU tasks (e.g. LAMBADA, HellaSwag), you can run the following command.
-
-```bash
-python main.py \
-	--model gpt2 \
-	--device cuda:0 \
-	--tasks lambada,hellaswag
-```
-(This uses gpt2-117M by default as per HF defaults, use --model_args to specify other gpt2 sizes)
-
-Additional arguments can be provided to the model constructor using the `--model_args` flag. Most importantly, the `gpt2` model can be used to load an arbitrary HuggingFace model. For example, to run GPTNeo use the following:
-
-```bash
-python main.py \
-	--model gpt2 \
-	--model_args pretrained=EleutherAI/gpt-neo-2.7B \
-	--device cuda:0 \
-	--tasks lambada,hellaswag
-```
-
-If you have access to the OpenAI API, you can also evaluate GPT-3:
-
-```bash
-export OPENAI_API_SECRET_KEY=YOUR_KEY_HERE
-python main.py \
-	--model gpt3 \
-	--model_args engine=davinci \
-	--tasks lambada,hellaswag
-```
-
-To evaluate mesh-transformer-jax models that are not available on HF, please invoke eval harness through [this script](https://github.com/kingoflolz/mesh-transformer-jax/blob/master/eval_harness.py).
-
-## Implementing new tasks
-
-To implement a new task in eval harness, see [this guide](https://github.com/EleutherAI/lm-evaluation-harness/blob/master/task-guide.md).
-
-## Cite as
-
-```
-@software{eval-harness,
-  author       = {Gao, Leo and
-                  Tow, Jonathan and
-                  Biderman, Stella and
-                  Black, Sid and
-                  DiPofi, Anthony and
-                  Foster, Charles and
-                  Golding, Laurence and
-                  Hsu, Jeffrey and
-                  McDonell, Kyle and
-                  Muennighoff, Niklas and
-                  Phang, Jason and
-                  Reynolds, Laria and
-                  Tang, Eric and
-                  Thite, Anish and
-                  Wang, Ben and
-                  Wang, Kevin and
-                  Zou, Andy},
-  title        = {A framework for few-shot language model evaluation},
-  month        = sep,
-  year         = 2021,
-  publisher    = {Zenodo},
-  version      = {v0.0.1},
-  doi          = {10.5281/zenodo.5371628},
-  url          = {https://doi.org/10.5281/zenodo.5371628}
-}
-```
-
-### Full Task List
-
-|                    Task Name                            |Train|Val|Test|Val/Test Docs|                                   Metrics                                    |
-|---------------------------------------------------------|-----|---|----|------------:|------------------------------------------------------------------------------|
-|cola                                                     |✓    |✓  |    |         1043|mcc                                                                           |
-|mnli                                                     |✓    |✓  |    |         9815|acc                                                                           |
-|mnli_mismatched                                          |✓    |✓  |    |         9832|acc                                                                           |
-|mrpc                                                     |✓    |✓  |    |          408|acc, f1                                                                       |
-|rte                                                      |✓    |✓  |    |          277|acc                                                                           |
-|qnli                                                     |✓    |✓  |    |         5463|acc                                                                           |
-|qqp                                                      |✓    |✓  |    |        40430|acc, f1                                                                       |
-|sst                                                      |✓    |✓  |    |          872|acc                                                                           |
-|wnli                                                     |✓    |✓  |    |           71|acc                                                                           |
-|boolq                                                    |✓    |✓  |    |         3270|acc                                                                           |
-|cb                                                       |✓    |✓  |    |           56|acc, f1                                                                       |
-|copa                                                     |✓    |✓  |    |          100|acc                                                                           |
-|multirc                                                  |✓    |✓  |    |         4848|acc                                                                           |
-|record                                                   |✓    |✓  |    |        10000|f1, em                                                                        |
-|wic                                                      |✓    |✓  |    |          638|acc                                                                           |
-|wsc                                                      |✓    |✓  |    |          104|acc                                                                           |
-|coqa                                                     |✓    |✓  |    |          500|f1, em                                                                        |
-|drop                                                     |✓    |✓  |    |         9536|em, f1                                                                        |
-|lambada                                                  |     |✓  |    |         5153|ppl, acc                                                                      |
-|lambada_cloze                                            |     |✓  |    |         5153|ppl, acc                                                                      |
-|wikitext                                                 |     |✓  |✓   |           62|word_perplexity, byte_perplexity, bits_per_byte                               |
-|piqa                                                     |✓    |✓  |    |         1838|acc, acc_norm                                                                 |
-|prost                                                    |     |   |✓   |        18736|acc, acc_norm                                                                 |
-|pubmedqa                                                 |     |   |✓   |         1000|acc                                                                           |
-|sciq                                                     |✓    |✓  |✓   |         1000|acc, acc_norm                                                                 |
-|qa4mre_2011                                              |     |   |✓   |          120|acc, acc_norm                                                                 |
-|qa4mre_2012                                              |     |   |✓   |          160|acc, acc_norm                                                                 |
-|qa4mre_2013                                              |     |   |✓   |          284|acc, acc_norm                                                                 |
-|triviaqa                                                 |✓    |✓  |    |        11313|acc                                                                           |
-|arc_easy                                                 |✓    |✓  |✓   |         2376|acc, acc_norm                                                                 |
-|arc_challenge                                            |✓    |✓  |✓   |         1172|acc, acc_norm                                                                 |
-|logiqa                                                   |✓    |✓  |✓   |          651|acc, acc_norm                                                                 |
-|hellaswag                                                |✓    |✓  |    |        10042|acc, acc_norm                                                                 |
-|openbookqa                                               |✓    |✓  |✓   |          500|acc, acc_norm                                                                 |
-|squad2                                                   |✓    |✓  |    |        11873|exact, f1, HasAns_exact, HasAns_f1, NoAns_exact, NoAns_f1, best_exact, best_f1|
-|race                                                     |✓    |✓  |✓   |         1045|acc                                                                           |
-|headqa                                                   |✓    |✓  |✓   |         2742|acc, acc_norm                                                                 |
-|mathqa                                                   |✓    |✓  |✓   |         2985|acc, acc_norm                                                                 |
-|webqs                                                    |✓    |   |✓   |         2032|acc                                                                           |
-|wsc273                                                   |     |   |✓   |          273|acc                                                                           |
-|winogrande                                               |✓    |✓  |    |         1267|acc                                                                           |
-|anli_r1                                                  |✓    |✓  |✓   |         1000|acc                                                                           |
-|anli_r2                                                  |✓    |✓  |✓   |         1000|acc                                                                           |
-|anli_r3                                                  |✓    |✓  |✓   |         1200|acc                                                                           |
-|ethics_cm                                                |✓    |   |✓   |         3885|acc                                                                           |
-|ethics_deontology                                        |✓    |   |✓   |         3596|acc, em                                                                       |
-|ethics_justice                                           |✓    |   |✓   |         2704|acc, em                                                                       |
-|ethics_utilitarianism_original                           |     |   |✓   |         4808|acc                                                                           |
-|ethics_utilitarianism                                    |✓    |   |✓   |         4808|acc                                                                           |
-|ethics_virtue                                            |✓    |   |✓   |         4975|acc, em                                                                       |
-|math_algebra                                             |✓    |   |✓   |         1187|acc                                                                           |
-|math_counting_and_prob                                   |✓    |   |✓   |          474|acc                                                                           |
-|math_geometry                                            |✓    |   |✓   |          479|acc                                                                           |
-|math_intermediate_algebra                                |✓    |   |✓   |          903|acc                                                                           |
-|math_num_theory                                          |✓    |   |✓   |          540|acc                                                                           |
-|math_prealgebra                                          |✓    |   |✓   |          871|acc                                                                           |
-|math_precalc                                             |✓    |   |✓   |          546|acc                                                                           |
-|arithmetic_2da                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_2ds                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_3da                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_3ds                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_4da                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_4ds                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_5da                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_5ds                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_2dm                                           |     |✓  |    |         2000|acc                                                                           |
-|arithmetic_1dc                                           |     |✓  |    |         2000|acc                                                                           |
-|hendrycksTest-abstract_algebra                           |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-anatomy                                    |✓    |✓  |✓   |          135|acc, acc_norm                                                                 |
-|hendrycksTest-astronomy                                  |✓    |✓  |✓   |          152|acc, acc_norm                                                                 |
-|hendrycksTest-business_ethics                            |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-clinical_knowledge                         |✓    |✓  |✓   |          265|acc, acc_norm                                                                 |
-|hendrycksTest-college_biology                            |✓    |✓  |✓   |          144|acc, acc_norm                                                                 |
-|hendrycksTest-college_chemistry                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-college_computer_science                   |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-college_mathematics                        |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-college_medicine                           |✓    |✓  |✓   |          173|acc, acc_norm                                                                 |
-|hendrycksTest-college_physics                            |✓    |✓  |✓   |          102|acc, acc_norm                                                                 |
-|hendrycksTest-computer_security                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-conceptual_physics                         |✓    |✓  |✓   |          235|acc, acc_norm                                                                 |
-|hendrycksTest-econometrics                               |✓    |✓  |✓   |          114|acc, acc_norm                                                                 |
-|hendrycksTest-electrical_engineering                     |✓    |✓  |✓   |          145|acc, acc_norm                                                                 |
-|hendrycksTest-elementary_mathematics                     |✓    |✓  |✓   |          378|acc, acc_norm                                                                 |
-|hendrycksTest-formal_logic                               |✓    |✓  |✓   |          126|acc, acc_norm                                                                 |
-|hendrycksTest-global_facts                               |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_biology                        |✓    |✓  |✓   |          310|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_chemistry                      |✓    |✓  |✓   |          203|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_computer_science               |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_european_history               |✓    |✓  |✓   |          165|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_geography                      |✓    |✓  |✓   |          198|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_government_and_politics        |✓    |✓  |✓   |          193|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_macroeconomics                 |✓    |✓  |✓   |          390|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_mathematics                    |✓    |✓  |✓   |          270|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_microeconomics                 |✓    |✓  |✓   |          238|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_physics                        |✓    |✓  |✓   |          151|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_psychology                     |✓    |✓  |✓   |          545|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_statistics                     |✓    |✓  |✓   |          216|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_us_history                     |✓    |✓  |✓   |          204|acc, acc_norm                                                                 |
-|hendrycksTest-high_school_world_history                  |✓    |✓  |✓   |          237|acc, acc_norm                                                                 |
-|hendrycksTest-human_aging                                |✓    |✓  |✓   |          223|acc, acc_norm                                                                 |
-|hendrycksTest-human_sexuality                            |✓    |✓  |✓   |          131|acc, acc_norm                                                                 |
-|hendrycksTest-international_law                          |✓    |✓  |✓   |          121|acc, acc_norm                                                                 |
-|hendrycksTest-jurisprudence                              |✓    |✓  |✓   |          108|acc, acc_norm                                                                 |
-|hendrycksTest-logical_fallacies                          |✓    |✓  |✓   |          163|acc, acc_norm                                                                 |
-|hendrycksTest-machine_learning                           |✓    |✓  |✓   |          112|acc, acc_norm                                                                 |
-|hendrycksTest-management                                 |✓    |✓  |✓   |          103|acc, acc_norm                                                                 |
-|hendrycksTest-marketing                                  |✓    |✓  |✓   |          234|acc, acc_norm                                                                 |
-|hendrycksTest-medical_genetics                           |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-miscellaneous                              |✓    |✓  |✓   |          783|acc, acc_norm                                                                 |
-|hendrycksTest-moral_disputes                             |✓    |✓  |✓   |          346|acc, acc_norm                                                                 |
-|hendrycksTest-moral_scenarios                            |✓    |✓  |✓   |          895|acc, acc_norm                                                                 |
-|hendrycksTest-nutrition                                  |✓    |✓  |✓   |          306|acc, acc_norm                                                                 |
-|hendrycksTest-philosophy                                 |✓    |✓  |✓   |          311|acc, acc_norm                                                                 |
-|hendrycksTest-prehistory                                 |✓    |✓  |✓   |          324|acc, acc_norm                                                                 |
-|hendrycksTest-professional_accounting                    |✓    |✓  |✓   |          282|acc, acc_norm                                                                 |
-|hendrycksTest-professional_law                           |✓    |✓  |✓   |         1534|acc, acc_norm                                                                 |
-|hendrycksTest-professional_medicine                      |✓    |✓  |✓   |          272|acc, acc_norm                                                                 |
-|hendrycksTest-professional_psychology                    |✓    |✓  |✓   |          612|acc, acc_norm                                                                 |
-|hendrycksTest-public_relations                           |✓    |✓  |✓   |          110|acc, acc_norm                                                                 |
-|hendrycksTest-security_studies                           |✓    |✓  |✓   |          245|acc, acc_norm                                                                 |
-|hendrycksTest-sociology                                  |✓    |✓  |✓   |          201|acc, acc_norm                                                                 |
-|hendrycksTest-us_foreign_policy                          |✓    |✓  |✓   |          100|acc, acc_norm                                                                 |
-|hendrycksTest-virology                                   |✓    |✓  |✓   |          166|acc, acc_norm                                                                 |
-|hendrycksTest-world_religions                            |✓    |✓  |✓   |          171|acc, acc_norm                                                                 |
-|wmt14-en-fr                                              |     |   |✓   |         3003|bleu, chrf, ter                                                               |
-|wmt14-fr-en                                              |     |   |✓   |         3003|bleu, chrf, ter                                                               |
-|wmt16-en-ro                                              |     |   |✓   |         1999|bleu, chrf, ter                                                               |
-|wmt16-ro-en                                              |     |   |✓   |         1999|bleu, chrf, ter                                                               |
-|wmt16-de-en                                              |     |   |✓   |         2999|bleu, chrf, ter                                                               |
-|wmt16-en-de                                              |     |   |✓   |         2999|bleu, chrf, ter                                                               |
-|wmt20-cs-en                                              |     |   |✓   |          664|bleu, chrf, ter                                                               |
-|wmt20-de-en                                              |     |   |✓   |          785|bleu, chrf, ter                                                               |
-|wmt20-de-fr                                              |     |   |✓   |         1619|bleu, chrf, ter                                                               |
-|wmt20-en-cs                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
-|wmt20-en-de                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
-|wmt20-en-iu                                              |     |   |✓   |         2971|bleu, chrf, ter                                                               |
-|wmt20-en-ja                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
-|wmt20-en-km                                              |     |   |✓   |         2320|bleu, chrf, ter                                                               |
-|wmt20-en-pl                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
-|wmt20-en-ps                                              |     |   |✓   |         2719|bleu, chrf, ter                                                               |
-|wmt20-en-ru                                              |     |   |✓   |         2002|bleu, chrf, ter                                                               |
-|wmt20-en-ta                                              |     |   |✓   |         1000|bleu, chrf, ter                                                               |
-|wmt20-en-zh                                              |     |   |✓   |         1418|bleu, chrf, ter                                                               |
-|wmt20-fr-de                                              |     |   |✓   |         1619|bleu, chrf, ter                                                               |
-|wmt20-iu-en                                              |     |   |✓   |         2971|bleu, chrf, ter                                                               |
-|wmt20-ja-en                                              |     |   |✓   |          993|bleu, chrf, ter                                                               |
-|wmt20-km-en                                              |     |   |✓   |         2320|bleu, chrf, ter                                                               |
-|wmt20-pl-en                                              |     |   |✓   |         1001|bleu, chrf, ter                                                               |
-|wmt20-ps-en                                              |     |   |✓   |         2719|bleu, chrf, ter                                                               |
-|wmt20-ru-en                                              |     |   |✓   |          991|bleu, chrf, ter                                                               |
-|wmt20-ta-en                                              |     |   |✓   |          997|bleu, chrf, ter                                                               |
-|wmt20-zh-en                                              |     |   |✓   |         2000|bleu, chrf, ter                                                               |
-|iwslt17-en-ar                                            |     |   |✓   |         1460|bleu, chrf, ter                                                               |
-|iwslt17-ar-en                                            |     |   |✓   |         1460|bleu, chrf, ter                                                               |
-|anagrams1                                                |     |✓  |    |        10000|acc                                                                           |
-|anagrams2                                                |     |✓  |    |        10000|acc                                                                           |
-|cycle_letters                                            |     |✓  |    |        10000|acc                                                                           |
-|random_insertion                                         |     |✓  |    |        10000|acc                                                                           |
-|reversed_words                                           |     |✓  |    |        10000|acc                                                                           |
-|pile_arxiv                                               |     |✓  |✓   |         2407|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_books3                                              |     |✓  |✓   |          269|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_bookcorpus2                                         |     |✓  |✓   |           28|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_dm-mathematics                                      |     |✓  |✓   |         1922|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_enron                                               |     |✓  |✓   |         1010|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_europarl                                            |     |✓  |✓   |          157|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_freelaw                                             |     |✓  |✓   |         5101|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_github                                              |     |✓  |✓   |        18195|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_gutenberg                                           |     |✓  |✓   |           80|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_hackernews                                          |     |✓  |✓   |         1632|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_nih-exporter                                        |     |✓  |✓   |         1884|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_opensubtitles                                       |     |✓  |✓   |          642|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_openwebtext2                                        |     |✓  |✓   |        32925|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_philpapers                                          |     |✓  |✓   |           68|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_pile-cc                                             |     |✓  |✓   |        52790|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_pubmed-abstracts                                    |     |✓  |✓   |        29895|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_pubmed-central                                      |     |✓  |✓   |         5911|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_stackexchange                                       |     |✓  |✓   |        30378|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_uspto                                               |     |✓  |✓   |        11415|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_ubuntu-irc                                          |     |✓  |✓   |           22|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_wikipedia                                           |     |✓  |✓   |        17511|word_perplexity, byte_perplexity, bits_per_byte                               |
-|pile_youtubesubtitles                                    |     |✓  |    |         1000|acc
-|blimp_adjunct_island                                     |     |✓  |    |         1000|acc
-|blimp_anaphor_gender_agreement                           |     |✓  |    |         1000|acc
-|blimp_anaphor_number_agreement                           |     |✓  |    |         1000|acc
-|blimp_animate_subject_passive                            |     |✓  |    |         1000|acc
-|blimp_animate_subject_trans                              |     |✓  |    |         1000|acc
-|blimp_causative                                          |     |✓  |    |         1000|acc
-|blimp_complex_NP_island                                  |     |✓  |    |         1000|acc
-|blimp_coordinate_structure_constraint_complex_left_branch|     |✓  |    |         1000|acc
-|blimp_coordinate_structure_constraint_object_extraction  |     |✓  |    |         1000|acc
-|blimp_determiner_noun_agreement_1                        |     |✓  |    |         1000|acc
-|blimp_determiner_noun_agreement_2                        |     |✓  |    |         1000|acc
-|blimp_determiner_noun_agreement_irregular_1              |     |✓  |    |         1000|acc
-|blimp_determiner_noun_agreement_irregular_2              |     |✓  |    |         1000|acc
-|blimp_determiner_noun_agreement_with_adj_2               |     |✓  |    |         1000|acc
-|blimp_determiner_noun_agreement_with_adj_irregular_1     |     |✓  |    |         1000|acc
-|blimp_determiner_noun_agreement_with_adj_irregular_2     |     |✓  |    |         1000|acc
-|blimp_determiner_noun_agreement_with_adjective_1         |     |✓  |    |         1000|acc
-|blimp_distractor_agreement_relational_noun               |     |✓  |    |         1000|acc
-|blimp_distractor_agreement_relative_clause               |     |✓  |    |         1000|acc
-|blimp_drop_argument                                      |     |✓  |    |         1000|acc
-|blimp_ellipsis_n_bar_1                                   |     |✓  |    |         1000|acc
-|blimp_ellipsis_n_bar_2                                   |     |✓  |    |         1000|acc
-|blimp_existential_there_object_raising                   |     |✓  |    |         1000|acc
-|blimp_existential_there_quantifiers_1                    |     |✓  |    |         1000|acc
-|blimp_existential_there_quantifiers_2                    |     |✓  |    |         1000|acc
-|blimp_existential_there_subject_raising                  |     |✓  |    |         1000|acc
-|blimp_expletive_it_object_raising                        |     |✓  |    |         1000|acc
-|blimp_inchoative                                         |     |✓  |    |         1000|acc
-|blimp_intransitive                                       |     |✓  |    |         1000|acc
-|blimp_irregular_past_participle_adjectives               |     |✓  |    |         1000|acc
-|blimp_irregular_past_participle_verbs                    |     |✓  |    |         1000|acc
-|blimp_irregular_plural_subject_verb_agreement_1          |     |✓  |    |         1000|acc
-|blimp_irregular_plural_subject_verb_agreement_2          |     |✓  |    |         1000|acc
-|blimp_left_branch_island_echo_question                   |     |✓  |    |         1000|acc
-|blimp_left_branch_island_simple_question                 |     |✓  |    |         1000|acc
-|blimp_matrix_question_npi_licensor_present               |     |✓  |    |         1000|acc
-|blimp_npi_present_1                                      |     |✓  |    |         1000|acc
-|blimp_npi_present_2                                      |     |✓  |    |         1000|acc
-|blimp_only_npi_licensor_present                          |     |✓  |    |         1000|acc
-|blimp_only_npi_scope                                     |     |✓  |    |         1000|acc
-|blimp_passive_1                                          |     |✓  |    |         1000|acc
-|blimp_passive_2                                          |     |✓  |    |         1000|acc
-|blimp_principle_A_c_command                              |     |✓  |    |         1000|acc
-|blimp_principle_A_case_1                                 |     |✓  |    |         1000|acc
-|blimp_principle_A_case_2                                 |     |✓  |    |         1000|acc
-|blimp_principle_A_domain_1                               |     |✓  |    |         1000|acc
-|blimp_principle_A_domain_2                               |     |✓  |    |         1000|acc
-|blimp_principle_A_domain_3                               |     |✓  |    |         1000|acc
-|blimp_principle_A_reconstruction                         |     |✓  |    |         1000|acc
-|blimp_regular_plural_subject_verb_agreement_1            |     |✓  |    |         1000|acc
-|blimp_regular_plural_subject_verb_agreement_2            |     |✓  |    |         1000|acc
-|blimp_sentential_negation_npi_licensor_present           |     |✓  |    |         1000|acc
-|blimp_sentential_negation_npi_scope                      |     |✓  |    |         1000|acc
-|blimp_sentential_subject_island                          |     |✓  |    |         1000|acc
-|blimp_superlative_quantifiers_1                          |     |✓  |    |         1000|acc
-|blimp_superlative_quantifiers_2                          |     |✓  |    |         1000|acc
-|blimp_tough_vs_raising_1                                 |     |✓  |    |         1000|acc
-|blimp_tough_vs_raising_2                                 |     |✓  |    |         1000|acc
-|blimp_transitive                                         |     |✓  |    |         1000|acc
-|blimp_wh_island                                          |     |✓  |    |         1000|acc
-|blimp_wh_questions_object_gap                            |     |✓  |    |         1000|acc
-|blimp_wh_questions_subject_gap                           |     |✓  |    |         1000|acc
-|blimp_wh_questions_subject_gap_long_distance             |     |✓  |    |         1000|acc
-|blimp_wh_vs_that_no_gap                                  |     |✓  |    |         1000|acc
-|blimp_wh_vs_that_no_gap_long_distance                    |     |✓  |    |         1000|acc
-|blimp_wh_vs_that_with_gap                                |     |✓  |    |         1000|acc
-|blimp_wh_vs_that_with_gap_long_distance                  |     |✓  |    |         1000|acc
-
-
-## Usage
-
-### Evaluate a task
-
-Additional arguments can be provided to the model constructor using the `--model_args` flag. Most importantly, the `gpt2` model can be used to load an arbitrary HuggingFace model as follows:
-
-
-```bash
-python main.py \
-	--model gpt2 \
-	--model_args pretrained=EleutherAI/gpt-neo-1.3B \
-	--device cuda:0 \
-	--tasks lambada,hellaswag \
-	--num_fewshot 2
-```
-
-To inspect what the LM inputs look like, you can run the following command:
-
-```bash
-python write_out.py \
-	--tasks all_tasks \
-	--provide_description \
-	--num_fewshot 5 \
-	--num_examples 10 \
-	--output_base_path /path/to/output/folder
-```
-
-This will write out one text file for each task.
-
-### Code Structure
-
-There are two major components of the library:
-
-1. LMs (language models), e.g. GPT-2, GPT-3
-2. Tasks, e.g. MNLI, RTE, SQuAD (coming soon)
-
-Both LMs (`lm_eval.models`) and Tasks (`lm_eval.tasks`) are kept in a registry data structure, for easy CLI instantiation.
-
-**If you want to extend either models or tasks, simply add a new LM or Task subclass, and decorate with the registry decorator**.
-
-The [GPT-3 Evaluations Project](https://github.com/EleutherAI/lm_evaluation_harness/projects/1) tracks our progress implementing new tasks. Right now, we are focused on getting all the datasets loaded so that we can dedupe against the training data. Implementing the actual evaluations is nice but not necessary at the current moment.
-
-### Task Versioning 
-
-To help improve reproducibility, all tasks have a VERSION field. When run from the command line, this is reported in a column in the table, or in the "version" field in the evaluator return dict. The purpose of the version is so that if the task definition changes (i.e to fix a bug), then we can know exactly which metrics were computed using the old buggy implementation to avoid unfair comparisons. To enforce this, there are unit tests that make sure the behavior of all tests remains the same as when they were first implemented. Task versions start at 0, and each time a breaking change is made, the version is incremented by one. 
-
-When reporting eval harness results, please also report the version of each task. This can be done either with a separate column in the table, or by reporting the task name with the version appended as such: taskname-v0.
-
-## Description
-
-### 1. LM Evaluation
-Given an LM, we want to evaluate it on a wide range of NLU tasks. We should at least cover the set of tasks in the GPT-3 paper, and any other tasks/benchmarks that are relevant. We will follow the GPT-3 format of a) zero-shot, b) one-shot, c) few-shot evaluation.
-
-To do this, we need 3 components:
-* Data downloader (shared with later sections, potentially needs to be directly linked to the latter 2 components)
-* Task formatter
-* Task evaluator
-
-The **data downloader** should download data for the relevant tasks.
-* We should heavily rely on Hugging Face's NLP for this. They are already doing most of the work with handling data scripts/caching.
-* Optionally, we can rely directly on HF-NLP's caching, but that makes it awkward to handle non-HF-NLP datasets. Otherwise, we can just write them out to .jsonl. My feeling is that NLU data storage will be a drop in the bucket compared to LM data.
-* Where we're not using HF-NLP, we can keep the data in the raw format (.jsonl, tsv, etc) and let the other components handle transforming it.
-
-The **task formatter** formats the task input data into an LM-usable format.
-* We should potentially support multiple formats for a given task, e.g. some formats may be better or worse suited for LM evaluation. See also: prompt-engineering
-* The task formatter should also support zero/one/few-shot packing of training examples into an input. This may require weird interactions with the tokenizer for dealing with max-token issues.
-
-The **task evaluator** scores a task.
-* In essence, we want to generation output predictions for all our input examples, and feed them into some function that pops out a score (or scores)
-An alternative approach is to collect the output logits and score them against the expected set of outputs.
-* Some tasks have weird evaluation schemes, so we should make this as general as possible.
-* Will thus likely have to be closely tied with the formatter.
-* Likewise, we should take advantage of HF-NLP's metrics.
-We might as well provide a sufficiently general API for the model to support OpenAI API as well. This can double up as an effort to reproduce the OpenAI NLU results.
-
-### 2. Removing val/test data from LM training set
-With the data downloader in place, we simply need to (1) expose the val/test examples, and (2) remove them from the training set.
-
-* Arguably, (2) should be handled by LM preprocessing in a more general way. There are probably non-NLU-eval cases where we want to remove some specific data from training.
-* Depending on how exactly we do the val/test removal, we may want to format the same example multiple ways to ensure that they don't get leaked into the training set in a slightly tweaked format.
-* Thought experiment: SQuAD is based largely on Wikipedia. What exactly would we want to remove from the LM?
-* [GPT-3]: In GPT-3, they attempted to remove val/test from their LM set, but there was a bug that caused leakage. So they ended up doing the opposite: removing overlaps from the LM set from the val/test. Funky.
-* [GPT-3]: See page 30 and Appendix C for details. They do some funky n-gram based search and removal. We should think about whether we want to follow their protocol exactly
-
-### 3. Adding task training data to LM training set
-This part is the easiest. I guess we just write out some text files containing the training data? We can let the usual LM preprocessing pipeline handle it from there.
-
-
diff --git a/lm_eval.egg-info/SOURCES.txt b/lm_eval.egg-info/SOURCES.txt
deleted file mode 100644
index 71a6a2fdb8..0000000000
--- a/lm_eval.egg-info/SOURCES.txt
+++ /dev/null
@@ -1,74 +0,0 @@
-LICENSE.md
-README.md
-setup.py
-lm_eval/__init__.py
-lm_eval/base.py
-lm_eval/evaluator.py
-lm_eval/metrics.py
-lm_eval/utils.py
-lm_eval.egg-info/PKG-INFO
-lm_eval.egg-info/SOURCES.txt
-lm_eval.egg-info/dependency_links.txt
-lm_eval.egg-info/requires.txt
-lm_eval.egg-info/top_level.txt
-lm_eval/models/__init__.py
-lm_eval/models/dummy.py
-lm_eval/models/gpt2.py
-lm_eval/models/gpt3.py
-lm_eval/tasks/__init__.py
-lm_eval/tasks/anli.py
-lm_eval/tasks/arc.py
-lm_eval/tasks/arithmetic.py
-lm_eval/tasks/blimp.py
-lm_eval/tasks/cbt.py
-lm_eval/tasks/common.py
-lm_eval/tasks/coqa.py
-lm_eval/tasks/drop.py
-lm_eval/tasks/glue.py
-lm_eval/tasks/headqa.py
-lm_eval/tasks/hellaswag.py
-lm_eval/tasks/hendrycks_ethics.py
-lm_eval/tasks/hendrycks_math.py
-lm_eval/tasks/hendrycks_test.py
-lm_eval/tasks/lambada.py
-lm_eval/tasks/lambada_cloze.py
-lm_eval/tasks/lambada_multilingual.py
-lm_eval/tasks/logiqa.py
-lm_eval/tasks/mathqa.py
-lm_eval/tasks/mc_taco.py
-lm_eval/tasks/mutual.py
-lm_eval/tasks/naturalqs.py
-lm_eval/tasks/openbookqa.py
-lm_eval/tasks/pile.py
-lm_eval/tasks/piqa.py
-lm_eval/tasks/prost.py
-lm_eval/tasks/pubmedqa.py
-lm_eval/tasks/qa4mre.py
-lm_eval/tasks/quac.py
-lm_eval/tasks/race.py
-lm_eval/tasks/sat.py
-lm_eval/tasks/sciq.py
-lm_eval/tasks/squad.py
-lm_eval/tasks/storycloze.py
-lm_eval/tasks/superglue.py
-lm_eval/tasks/translation.py
-lm_eval/tasks/triviaqa.py
-lm_eval/tasks/truthfulqa.py
-lm_eval/tasks/unscramble.py
-lm_eval/tasks/webqs.py
-lm_eval/tasks/wikitext.py
-lm_eval/tasks/winogrande.py
-lm_eval/tasks/wsc273.py
-scripts/__init__.py
-scripts/cost_estimate.py
-scripts/fewshot_description_experiment.py
-scripts/get_prompts.py
-scripts/make_gpt2_test_cases.py
-scripts/make_table_tasks.py
-scripts/write_out.py
-scripts/clean_training_data/__init__.py
-scripts/clean_training_data/archiver.py
-scripts/clean_training_data/generate_13_grams.py
-scripts/clean_training_data/janitor.py
-scripts/clean_training_data/process_sorted_buckets.py
-scripts/clean_training_data/sort_13_gram_buckets.py
\ No newline at end of file
diff --git a/lm_eval.egg-info/dependency_links.txt b/lm_eval.egg-info/dependency_links.txt
deleted file mode 100644
index 8b13789179..0000000000
--- a/lm_eval.egg-info/dependency_links.txt
+++ /dev/null
@@ -1 +0,0 @@
-
diff --git a/lm_eval.egg-info/requires.txt b/lm_eval.egg-info/requires.txt
deleted file mode 100644
index c0f5b007a4..0000000000
--- a/lm_eval.egg-info/requires.txt
+++ /dev/null
@@ -1,24 +0,0 @@
-black
-best_download>=0.0.6
-datasets==1.15.1
-click>=7.1
-scikit-learn>=0.24.1
-torch>=1.7
-transformers>=4.1
-sqlitedict==1.6.0
-pytablewriter==0.58.0
-sacrebleu==1.5.0
-rouge-score==0.0.4
-bleurt@ https://github.com/google-research/bleurt/archive/b610120347ef22b494b6d69b4316e303f5932516.zip#egg=bleurt
-pycountry==20.7.3
-numexpr==2.7.2
-lm_dataformat==0.0.20
-pytest==6.2.3
-pybind11==2.6.2
-tqdm-multiprocess==0.0.11
-zstandard==0.15.2
-jsonlines==2.0.0
-mock==4.0.3
-openai==0.6.4
-jieba==0.42.1
-nagisa==0.2.7
diff --git a/lm_eval.egg-info/top_level.txt b/lm_eval.egg-info/top_level.txt
deleted file mode 100644
index 3c734591d0..0000000000
--- a/lm_eval.egg-info/top_level.txt
+++ /dev/null
@@ -1,2 +0,0 @@
-lm_eval
-scripts
diff --git a/lm_eval/tasks/asdiv.py b/lm_eval/tasks/asdiv.py
index dcde1f1e23..d84fc81a54 100644
--- a/lm_eval/tasks/asdiv.py
+++ b/lm_eval/tasks/asdiv.py
@@ -32,9 +32,9 @@ def download(self):
         if self.DATASET_PATH.exists():
             return
         Path.mkdir(self.DATASET_PATH)
-        url = "https://github.com/chaochun/nlu-asdiv-dataset/archive/refs/heads/master.zip"
-        checksum = "2f71f8003929d605369ad924be4b95c15879fc2bfac0d4d01a81f8aabceaad5c"
-        zip_path = self.DATASET_PATH / "master.zip"
+        url = "https://github.com/chaochun/nlu-asdiv-dataset/archive/55790e5270bb91ccfa5053194b25732534696b50.zip"
+        checksum = "8f1fe4f6d5f170ec1e24ab78c244153c14c568b1bb2b1dad0324e71f37939a2d"
+        zip_path = self.DATASET_PATH / "55790e5270bb91ccfa5053194b25732534696b50.zip"
         download_file(url, str(zip_path), checksum)
         with ZipFile(zip_path, "r") as zip:
             zip.extractall(self.DATASET_PATH)
@@ -85,7 +85,7 @@ def test_docs(self):
         raise NotImplementedError("This dataset has no test docs")
 
     def validation_docs(self):
-        data_xml_path = self.DATASET_PATH / "nlu-asdiv-dataset-master/dataset/ASDiv.xml"
+        data_xml_path = self.DATASET_PATH / "nlu-asdiv-dataset-55790e5270bb91ccfa5053194b25732534696b50/dataset/ASDiv.xml"
         return self.load_docs(data_xml_path)
 
     def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
@@ -109,11 +109,10 @@ def doc_to_target(self, doc):
         if len(answer)>0: # check if answer is present only in brackets
             return answer
         else:
-            return doc['answer']
+            return " "+doc['answer']
 
     def construct_requests(self, doc, ctx):
         ll, is_greedy = rf.loglikelihood(ctx, self.doc_to_target(doc))
-
         return ll, is_greedy
     
     def process_results(self, doc, results):

From 6653cc509e810752eeda2349c20c315a18c7fd5d Mon Sep 17 00:00:00 2001
From: Igor Ostrovsky <igoros@gmail.com>
Date: Thu, 30 Dec 2021 12:46:12 -0800
Subject: [PATCH 44/65] Bump the version number for all tasks based on
 PerplexityTask

This is due to the change in the bits_per_byte calculation.
---
 lm_eval/tasks/pile.py     | 2 +-
 lm_eval/tasks/wikitext.py | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/lm_eval/tasks/pile.py b/lm_eval/tasks/pile.py
index 68ff7ed9a8..de16caa3db 100644
--- a/lm_eval/tasks/pile.py
+++ b/lm_eval/tasks/pile.py
@@ -10,7 +10,7 @@
 
 
 class PilePerplexityTask(PerplexityTask, abc.ABC):
-    VERSION = 0
+    VERSION = 1
 
     PILE_SET_NAME = None
     VAL_PATH = 'data/pile/val.jsonl.zst'
diff --git a/lm_eval/tasks/wikitext.py b/lm_eval/tasks/wikitext.py
index 24f9ec3507..b84cd3838b 100644
--- a/lm_eval/tasks/wikitext.py
+++ b/lm_eval/tasks/wikitext.py
@@ -41,7 +41,7 @@ def wikitext_detokenizer(string):
 
 
 class WikiText(PerplexityTask):
-    VERSION = 0
+    VERSION = 1
 
     def download(self):
         if not os.path.exists('data/wikitext/wikitext-2-raw/wiki.valid.raw'):
@@ -87,4 +87,4 @@ def doc_to_target(self, doc):
     
     def count_words(self, doc):
         # count number of words in *original doc before detokenization*
-        return len(re.split(r"\s+", doc))
\ No newline at end of file
+        return len(re.split(r"\s+", doc))

From 72d7cc0cc545b187b9cb243bc1fb73d350657298 Mon Sep 17 00:00:00 2001
From: rokosbasilisk <thecomplexitytheorist@gmail.com>
Date: Fri, 31 Dec 2021 19:48:50 +0530
Subject: [PATCH 45/65] remove _strip_bracket function

---
 lm_eval/tasks/asdiv.py | 18 +++---------------
 1 file changed, 3 insertions(+), 15 deletions(-)

diff --git a/lm_eval/tasks/asdiv.py b/lm_eval/tasks/asdiv.py
index d84fc81a54..beee30e813 100644
--- a/lm_eval/tasks/asdiv.py
+++ b/lm_eval/tasks/asdiv.py
@@ -56,19 +56,6 @@ def load_docs(self, textfilename, tfds=False):
             out_doc = self._convert_standard(problem)
             yield out_doc
 
-    def _strip_bracket(self,test_str):
-        ret = ''
-        skip1c = 0
-        skip2c = 0
-        for i in test_str:
-            if i == '(':
-                skip2c += 1
-            elif i == ')'and skip2c > 0:
-                skip2c -= 1
-            elif skip1c == 0 and skip2c == 0:
-                ret += i
-        return ret
-
     def has_training_docs(self):
         return False
     
@@ -105,11 +92,12 @@ def doc_to_text(self, doc):
     def doc_to_target(self, doc):
         # TODO: add formula
 
-        answer = self._strip_bracket(doc['answer'])
+        answer = doc['answer'].split(' (')[0]
+        print(answer)
         if len(answer)>0: # check if answer is present only in brackets
             return answer
         else:
-            return " "+doc['answer']
+            return doc['answer']
 
     def construct_requests(self, doc, ctx):
         ll, is_greedy = rf.loglikelihood(ctx, self.doc_to_target(doc))

From 83e1a1166cb2df68accf353737e1d0450a719517 Mon Sep 17 00:00:00 2001
From: rokosbasilisk <thecomplexitytheorist@gmail.com>
Date: Fri, 31 Dec 2021 19:50:15 +0530
Subject: [PATCH 46/65] removed strip_bracket function

---
 lm_eval/tasks/asdiv.py | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/lm_eval/tasks/asdiv.py b/lm_eval/tasks/asdiv.py
index beee30e813..0ff563538e 100644
--- a/lm_eval/tasks/asdiv.py
+++ b/lm_eval/tasks/asdiv.py
@@ -93,11 +93,7 @@ def doc_to_target(self, doc):
         # TODO: add formula
 
         answer = doc['answer'].split(' (')[0]
-        print(answer)
-        if len(answer)>0: # check if answer is present only in brackets
-            return answer
-        else:
-            return doc['answer']
+        return answer
 
     def construct_requests(self, doc, ctx):
         ll, is_greedy = rf.loglikelihood(ctx, self.doc_to_target(doc))

From 33315a1f50b163856e3c83f1ea514654c7367deb Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Fri, 31 Dec 2021 17:13:20 -0700
Subject: [PATCH 47/65] pile/wikitext: add testdata

---
 tests/testdata/pile_arxiv-v1-loglikelihood_rolling            | 1 +
 tests/testdata/pile_arxiv-v1-res.json                         | 1 +
 tests/testdata/pile_bookcorpus2-v1-loglikelihood_rolling      | 1 +
 tests/testdata/pile_bookcorpus2-v1-res.json                   | 1 +
 tests/testdata/pile_books3-v1-loglikelihood_rolling           | 1 +
 tests/testdata/pile_books3-v1-res.json                        | 1 +
 tests/testdata/pile_dm-mathematics-v1-loglikelihood_rolling   | 1 +
 tests/testdata/pile_dm-mathematics-v1-res.json                | 1 +
 tests/testdata/pile_enron-v1-loglikelihood_rolling            | 1 +
 tests/testdata/pile_enron-v1-res.json                         | 1 +
 tests/testdata/pile_europarl-v1-loglikelihood_rolling         | 1 +
 tests/testdata/pile_europarl-v1-res.json                      | 1 +
 tests/testdata/pile_freelaw-v1-loglikelihood_rolling          | 1 +
 tests/testdata/pile_freelaw-v1-res.json                       | 1 +
 tests/testdata/pile_github-v1-loglikelihood_rolling           | 1 +
 tests/testdata/pile_github-v1-res.json                        | 1 +
 tests/testdata/pile_gutenberg-v1-loglikelihood_rolling        | 1 +
 tests/testdata/pile_gutenberg-v1-res.json                     | 1 +
 tests/testdata/pile_hackernews-v1-loglikelihood_rolling       | 1 +
 tests/testdata/pile_hackernews-v1-res.json                    | 1 +
 tests/testdata/pile_nih-exporter-v1-loglikelihood_rolling     | 1 +
 tests/testdata/pile_nih-exporter-v1-res.json                  | 1 +
 tests/testdata/pile_opensubtitles-v1-loglikelihood_rolling    | 1 +
 tests/testdata/pile_opensubtitles-v1-res.json                 | 1 +
 tests/testdata/pile_openwebtext2-v1-loglikelihood_rolling     | 1 +
 tests/testdata/pile_openwebtext2-v1-res.json                  | 1 +
 tests/testdata/pile_philpapers-v1-loglikelihood_rolling       | 1 +
 tests/testdata/pile_philpapers-v1-res.json                    | 1 +
 tests/testdata/pile_pile-cc-v1-loglikelihood_rolling          | 1 +
 tests/testdata/pile_pile-cc-v1-res.json                       | 1 +
 tests/testdata/pile_pubmed-abstracts-v1-loglikelihood_rolling | 1 +
 tests/testdata/pile_pubmed-abstracts-v1-res.json              | 1 +
 tests/testdata/pile_pubmed-central-v1-loglikelihood_rolling   | 1 +
 tests/testdata/pile_pubmed-central-v1-res.json                | 1 +
 tests/testdata/pile_stackexchange-v1-loglikelihood_rolling    | 1 +
 tests/testdata/pile_stackexchange-v1-res.json                 | 1 +
 tests/testdata/pile_ubuntu-irc-v1-loglikelihood_rolling       | 1 +
 tests/testdata/pile_ubuntu-irc-v1-res.json                    | 1 +
 tests/testdata/pile_uspto-v1-loglikelihood_rolling            | 1 +
 tests/testdata/pile_uspto-v1-res.json                         | 1 +
 tests/testdata/pile_wikipedia-v1-loglikelihood_rolling        | 1 +
 tests/testdata/pile_wikipedia-v1-res.json                     | 1 +
 tests/testdata/pile_youtubesubtitles-v1-loglikelihood_rolling | 1 +
 tests/testdata/pile_youtubesubtitles-v1-res.json              | 1 +
 tests/testdata/wikitext-v1-loglikelihood_rolling              | 1 +
 tests/testdata/wikitext-v1-res.json                           | 1 +
 46 files changed, 46 insertions(+)
 create mode 100644 tests/testdata/pile_arxiv-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_arxiv-v1-res.json
 create mode 100644 tests/testdata/pile_bookcorpus2-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_bookcorpus2-v1-res.json
 create mode 100644 tests/testdata/pile_books3-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_books3-v1-res.json
 create mode 100644 tests/testdata/pile_dm-mathematics-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_dm-mathematics-v1-res.json
 create mode 100644 tests/testdata/pile_enron-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_enron-v1-res.json
 create mode 100644 tests/testdata/pile_europarl-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_europarl-v1-res.json
 create mode 100644 tests/testdata/pile_freelaw-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_freelaw-v1-res.json
 create mode 100644 tests/testdata/pile_github-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_github-v1-res.json
 create mode 100644 tests/testdata/pile_gutenberg-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_gutenberg-v1-res.json
 create mode 100644 tests/testdata/pile_hackernews-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_hackernews-v1-res.json
 create mode 100644 tests/testdata/pile_nih-exporter-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_nih-exporter-v1-res.json
 create mode 100644 tests/testdata/pile_opensubtitles-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_opensubtitles-v1-res.json
 create mode 100644 tests/testdata/pile_openwebtext2-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_openwebtext2-v1-res.json
 create mode 100644 tests/testdata/pile_philpapers-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_philpapers-v1-res.json
 create mode 100644 tests/testdata/pile_pile-cc-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_pile-cc-v1-res.json
 create mode 100644 tests/testdata/pile_pubmed-abstracts-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_pubmed-abstracts-v1-res.json
 create mode 100644 tests/testdata/pile_pubmed-central-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_pubmed-central-v1-res.json
 create mode 100644 tests/testdata/pile_stackexchange-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_stackexchange-v1-res.json
 create mode 100644 tests/testdata/pile_ubuntu-irc-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_ubuntu-irc-v1-res.json
 create mode 100644 tests/testdata/pile_uspto-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_uspto-v1-res.json
 create mode 100644 tests/testdata/pile_wikipedia-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_wikipedia-v1-res.json
 create mode 100644 tests/testdata/pile_youtubesubtitles-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/pile_youtubesubtitles-v1-res.json
 create mode 100644 tests/testdata/wikitext-v1-loglikelihood_rolling
 create mode 100644 tests/testdata/wikitext-v1-res.json

diff --git a/tests/testdata/pile_arxiv-v1-loglikelihood_rolling b/tests/testdata/pile_arxiv-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..3aa1d8c734
--- /dev/null
+++ b/tests/testdata/pile_arxiv-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+814f9954e44368559602c00f7e85fa3971acdfd0315f508ec7df6318a79c55ec
\ No newline at end of file
diff --git a/tests/testdata/pile_arxiv-v1-res.json b/tests/testdata/pile_arxiv-v1-res.json
new file mode 100644
index 0000000000..05cbab3873
--- /dev/null
+++ b/tests/testdata/pile_arxiv-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_arxiv": {"bits_per_byte": 1.55095665856779e-05, "byte_perplexity": 1.0000107504701365, "word_perplexity": 1.0000819333090385}}, "versions": {"pile_arxiv": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_bookcorpus2-v1-loglikelihood_rolling b/tests/testdata/pile_bookcorpus2-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..b37a91cc2d
--- /dev/null
+++ b/tests/testdata/pile_bookcorpus2-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+5c17ddfebeab8c41dabadb6fc216ceda91e3fe5dc95aaf1b2c843d7f11828b03
\ No newline at end of file
diff --git a/tests/testdata/pile_bookcorpus2-v1-res.json b/tests/testdata/pile_bookcorpus2-v1-res.json
new file mode 100644
index 0000000000..967c14934b
--- /dev/null
+++ b/tests/testdata/pile_bookcorpus2-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_bookcorpus2": {"bits_per_byte": 1.6780040419457868e-06, "byte_perplexity": 1.000001163104447, "word_perplexity": 1.0000066499426599}}, "versions": {"pile_bookcorpus2": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_books3-v1-loglikelihood_rolling b/tests/testdata/pile_books3-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..b483d3b45b
--- /dev/null
+++ b/tests/testdata/pile_books3-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+0f8f36f705b999b6d55fa72ff89a82793dd1cb568ab1f8727a6a2086a12b9410
\ No newline at end of file
diff --git a/tests/testdata/pile_books3-v1-res.json b/tests/testdata/pile_books3-v1-res.json
new file mode 100644
index 0000000000..6ff7a51711
--- /dev/null
+++ b/tests/testdata/pile_books3-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_books3": {"bits_per_byte": 1.2901280503011222e-06, "byte_perplexity": 1.0000008942490204, "word_perplexity": 1.0000052870063607}}, "versions": {"pile_books3": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_dm-mathematics-v1-loglikelihood_rolling b/tests/testdata/pile_dm-mathematics-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..2fb27786c5
--- /dev/null
+++ b/tests/testdata/pile_dm-mathematics-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+d5b7967c0ece8b816f3921a8bd0fad23365349e935b491595e2ad1135af42da6
\ No newline at end of file
diff --git a/tests/testdata/pile_dm-mathematics-v1-res.json b/tests/testdata/pile_dm-mathematics-v1-res.json
new file mode 100644
index 0000000000..192e9066a4
--- /dev/null
+++ b/tests/testdata/pile_dm-mathematics-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_dm-mathematics": {"bits_per_byte": 8.910951449933553e-05, "byte_perplexity": 1.0000617679162955, "word_perplexity": 1.0002875035042451}}, "versions": {"pile_dm-mathematics": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_enron-v1-loglikelihood_rolling b/tests/testdata/pile_enron-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..57dbe76460
--- /dev/null
+++ b/tests/testdata/pile_enron-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+4baa6ccdc9e3aa9921675ab4400d5e89d7b546b844a8ea28f6461d649066418a
\ No newline at end of file
diff --git a/tests/testdata/pile_enron-v1-res.json b/tests/testdata/pile_enron-v1-res.json
new file mode 100644
index 0000000000..abe7b45f9a
--- /dev/null
+++ b/tests/testdata/pile_enron-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_enron": {"bits_per_byte": 0.0004564546920781453, "byte_perplexity": 1.000316440339552, "word_perplexity": 1.00224668051869}}, "versions": {"pile_enron": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_europarl-v1-loglikelihood_rolling b/tests/testdata/pile_europarl-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..8027260755
--- /dev/null
+++ b/tests/testdata/pile_europarl-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+e67d3dbccd47d308bfc5b0e66b76d0dfc5e386ebfa94e056562c2281c395543f
\ No newline at end of file
diff --git a/tests/testdata/pile_europarl-v1-res.json b/tests/testdata/pile_europarl-v1-res.json
new file mode 100644
index 0000000000..b948f0d369
--- /dev/null
+++ b/tests/testdata/pile_europarl-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_europarl": {"bits_per_byte": 1.2477664839621123e-05, "byte_perplexity": 1.000008648895605, "word_perplexity": 1.000063506523818}}, "versions": {"pile_europarl": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_freelaw-v1-loglikelihood_rolling b/tests/testdata/pile_freelaw-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..7b5771f491
--- /dev/null
+++ b/tests/testdata/pile_freelaw-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+d77f3f68aadd6cbf1290c2f6737b2ed5d5c2a60e4c81a65c280f207783caabe1
\ No newline at end of file
diff --git a/tests/testdata/pile_freelaw-v1-res.json b/tests/testdata/pile_freelaw-v1-res.json
new file mode 100644
index 0000000000..dd0e0bac36
--- /dev/null
+++ b/tests/testdata/pile_freelaw-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_freelaw": {"bits_per_byte": 4.5623635481434923e-05, "byte_perplexity": 1.0000316243943415, "word_perplexity": 1.000203169094218}}, "versions": {"pile_freelaw": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_github-v1-loglikelihood_rolling b/tests/testdata/pile_github-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..cf8251e4f6
--- /dev/null
+++ b/tests/testdata/pile_github-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+df384c3df3d8f53273e97127c5bb84c17e638acad7d6bc9c91f6dee96d43b639
\ No newline at end of file
diff --git a/tests/testdata/pile_github-v1-res.json b/tests/testdata/pile_github-v1-res.json
new file mode 100644
index 0000000000..cc06a45501
--- /dev/null
+++ b/tests/testdata/pile_github-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_github": {"bits_per_byte": 0.00013764216145332133, "byte_perplexity": 1.0000954108274611, "word_perplexity": 1.0009643183931227}}, "versions": {"pile_github": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_gutenberg-v1-loglikelihood_rolling b/tests/testdata/pile_gutenberg-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..bd7b15927f
--- /dev/null
+++ b/tests/testdata/pile_gutenberg-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+02a559f74a9105145e7d4d9c5ddea372b5b4938f5368dc8ffafc39cbe3b4c7ef
\ No newline at end of file
diff --git a/tests/testdata/pile_gutenberg-v1-res.json b/tests/testdata/pile_gutenberg-v1-res.json
new file mode 100644
index 0000000000..6d22ed3ff5
--- /dev/null
+++ b/tests/testdata/pile_gutenberg-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_gutenberg": {"bits_per_byte": 1.7952329146458065e-06, "byte_perplexity": 1.0000012443614075, "word_perplexity": 1.0000072174665404}}, "versions": {"pile_gutenberg": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_hackernews-v1-loglikelihood_rolling b/tests/testdata/pile_hackernews-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..48b767bfe7
--- /dev/null
+++ b/tests/testdata/pile_hackernews-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+ec1082ee5a5326e0d57aa4e73b634937140c1de9af95f154e8ab57b05d9b422b
\ No newline at end of file
diff --git a/tests/testdata/pile_hackernews-v1-res.json b/tests/testdata/pile_hackernews-v1-res.json
new file mode 100644
index 0000000000..ea135278b7
--- /dev/null
+++ b/tests/testdata/pile_hackernews-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_hackernews": {"bits_per_byte": 0.00014672607267878518, "byte_perplexity": 1.0001017079354932, "word_perplexity": 1.0006273924348839}}, "versions": {"pile_hackernews": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_nih-exporter-v1-loglikelihood_rolling b/tests/testdata/pile_nih-exporter-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..5f76588a81
--- /dev/null
+++ b/tests/testdata/pile_nih-exporter-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+520ea6e04e8a39dc0b5f63a837429a78a40e63d39d109096101feb8c5b2cf8d8
\ No newline at end of file
diff --git a/tests/testdata/pile_nih-exporter-v1-res.json b/tests/testdata/pile_nih-exporter-v1-res.json
new file mode 100644
index 0000000000..0e40fc8268
--- /dev/null
+++ b/tests/testdata/pile_nih-exporter-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_nih-exporter": {"bits_per_byte": 0.00035193728014978225, "byte_perplexity": 1.0002439740903082, "word_perplexity": 1.0016712202288802}}, "versions": {"pile_nih-exporter": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_opensubtitles-v1-loglikelihood_rolling b/tests/testdata/pile_opensubtitles-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..47805d3b5f
--- /dev/null
+++ b/tests/testdata/pile_opensubtitles-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+0f1c23a1f4ddec0c2b1ff34de8d1505b0eb9e2868d8edbcc1b6de13d02f32036
\ No newline at end of file
diff --git a/tests/testdata/pile_opensubtitles-v1-res.json b/tests/testdata/pile_opensubtitles-v1-res.json
new file mode 100644
index 0000000000..1468294732
--- /dev/null
+++ b/tests/testdata/pile_opensubtitles-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_opensubtitles": {"bits_per_byte": 2.1948356082685497e-05, "byte_perplexity": 1.0000152135568616, "word_perplexity": 1.0000856162053249}}, "versions": {"pile_opensubtitles": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_openwebtext2-v1-loglikelihood_rolling b/tests/testdata/pile_openwebtext2-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..22046e4405
--- /dev/null
+++ b/tests/testdata/pile_openwebtext2-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+5d6c19665f429ab1ccbe027da67f42bdaf219f819ab093673976eee55e015ff4
\ No newline at end of file
diff --git a/tests/testdata/pile_openwebtext2-v1-res.json b/tests/testdata/pile_openwebtext2-v1-res.json
new file mode 100644
index 0000000000..ca433e3c85
--- /dev/null
+++ b/tests/testdata/pile_openwebtext2-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_openwebtext2": {"bits_per_byte": 0.000184802319359215, "byte_perplexity": 1.000128103411166, "word_perplexity": 1.0007951516532847}}, "versions": {"pile_openwebtext2": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_philpapers-v1-loglikelihood_rolling b/tests/testdata/pile_philpapers-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..4fbbc241ba
--- /dev/null
+++ b/tests/testdata/pile_philpapers-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+339ba5d8c044c4a3ff9b9a8eaa24da1d6c01b72972074eb671a7da049eeb7047
\ No newline at end of file
diff --git a/tests/testdata/pile_philpapers-v1-res.json b/tests/testdata/pile_philpapers-v1-res.json
new file mode 100644
index 0000000000..5a2f77678a
--- /dev/null
+++ b/tests/testdata/pile_philpapers-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_philpapers": {"bits_per_byte": 9.004690592465457e-06, "byte_perplexity": 1.0000062415953748, "word_perplexity": 1.0000409888564146}}, "versions": {"pile_philpapers": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_pile-cc-v1-loglikelihood_rolling b/tests/testdata/pile_pile-cc-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..d5369ed3c9
--- /dev/null
+++ b/tests/testdata/pile_pile-cc-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+731fdef4a43949b179ba0c540148ebc2fa41583dd583ef580dd812076c66a451
\ No newline at end of file
diff --git a/tests/testdata/pile_pile-cc-v1-res.json b/tests/testdata/pile_pile-cc-v1-res.json
new file mode 100644
index 0000000000..bd2772e32a
--- /dev/null
+++ b/tests/testdata/pile_pile-cc-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_pile-cc": {"bits_per_byte": 0.0001620742639125056, "byte_perplexity": 1.0001123476295946, "word_perplexity": 1.0006738958554477}}, "versions": {"pile_pile-cc": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_pubmed-abstracts-v1-loglikelihood_rolling b/tests/testdata/pile_pubmed-abstracts-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..de5660d60a
--- /dev/null
+++ b/tests/testdata/pile_pubmed-abstracts-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+66436569a43163afb2caf422d32c5f329899e74c49865d4d13881fd465fd9976
\ No newline at end of file
diff --git a/tests/testdata/pile_pubmed-abstracts-v1-res.json b/tests/testdata/pile_pubmed-abstracts-v1-res.json
new file mode 100644
index 0000000000..21b6bb451f
--- /dev/null
+++ b/tests/testdata/pile_pubmed-abstracts-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_pubmed-abstracts": {"bits_per_byte": 0.0005417858444030858, "byte_perplexity": 1.0003756078534862, "word_perplexity": 1.0025884332779}}, "versions": {"pile_pubmed-abstracts": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_pubmed-central-v1-loglikelihood_rolling b/tests/testdata/pile_pubmed-central-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..283109f32e
--- /dev/null
+++ b/tests/testdata/pile_pubmed-central-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+40b39d120d99a145690444e86acc3e3e24d41e6e0538a75e26929ad84926e5e0
\ No newline at end of file
diff --git a/tests/testdata/pile_pubmed-central-v1-res.json b/tests/testdata/pile_pubmed-central-v1-res.json
new file mode 100644
index 0000000000..4d4a241ace
--- /dev/null
+++ b/tests/testdata/pile_pubmed-central-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_pubmed-central": {"bits_per_byte": 2.2812488135667854e-05, "byte_perplexity": 1.0000158125368497, "word_perplexity": 1.000123107107861}}, "versions": {"pile_pubmed-central": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_stackexchange-v1-loglikelihood_rolling b/tests/testdata/pile_stackexchange-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..dcf0e64cf0
--- /dev/null
+++ b/tests/testdata/pile_stackexchange-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+e524bfb3e21cbdaddc117403a50df598520c7bf5b2c60ad8f2372cfa564e79be
\ No newline at end of file
diff --git a/tests/testdata/pile_stackexchange-v1-res.json b/tests/testdata/pile_stackexchange-v1-res.json
new file mode 100644
index 0000000000..2773302990
--- /dev/null
+++ b/tests/testdata/pile_stackexchange-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_stackexchange": {"bits_per_byte": 0.0003302063346758449, "byte_perplexity": 1.0002289077852733, "word_perplexity": 1.0016993562258851}}, "versions": {"pile_stackexchange": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_ubuntu-irc-v1-loglikelihood_rolling b/tests/testdata/pile_ubuntu-irc-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..ce04199863
--- /dev/null
+++ b/tests/testdata/pile_ubuntu-irc-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+4eb69e314f0864ec8890e2323d7e76f8a8309692c4f090e2b41bf4be681a811d
\ No newline at end of file
diff --git a/tests/testdata/pile_ubuntu-irc-v1-res.json b/tests/testdata/pile_ubuntu-irc-v1-res.json
new file mode 100644
index 0000000000..0e3b1b2597
--- /dev/null
+++ b/tests/testdata/pile_ubuntu-irc-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_ubuntu-irc": {"bits_per_byte": 2.3513498942121155e-06, "byte_perplexity": 1.0000016298328778, "word_perplexity": 1.0000108866656874}}, "versions": {"pile_ubuntu-irc": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_uspto-v1-loglikelihood_rolling b/tests/testdata/pile_uspto-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..4649d3b9b7
--- /dev/null
+++ b/tests/testdata/pile_uspto-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+789b2bdb31564d512b70f801316f49320a26c83ba361226bac0afb255341d477
\ No newline at end of file
diff --git a/tests/testdata/pile_uspto-v1-res.json b/tests/testdata/pile_uspto-v1-res.json
new file mode 100644
index 0000000000..599ae44ef4
--- /dev/null
+++ b/tests/testdata/pile_uspto-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_uspto": {"bits_per_byte": 0.000174024142670342, "byte_perplexity": 1.00012063161925, "word_perplexity": 1.0007716198916954}}, "versions": {"pile_uspto": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_wikipedia-v1-loglikelihood_rolling b/tests/testdata/pile_wikipedia-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..e44bd27628
--- /dev/null
+++ b/tests/testdata/pile_wikipedia-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+ef9ec0dd408316ca6537228a6812e839f14b30608973081d41efc47c138338da
\ No newline at end of file
diff --git a/tests/testdata/pile_wikipedia-v1-res.json b/tests/testdata/pile_wikipedia-v1-res.json
new file mode 100644
index 0000000000..4f2314e66b
--- /dev/null
+++ b/tests/testdata/pile_wikipedia-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_wikipedia": {"bits_per_byte": 0.00024287370359008176, "byte_perplexity": 1.0001683613940646, "word_perplexity": 1.001084677949439}}, "versions": {"pile_wikipedia": 1}}
\ No newline at end of file
diff --git a/tests/testdata/pile_youtubesubtitles-v1-loglikelihood_rolling b/tests/testdata/pile_youtubesubtitles-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..81c2e5ed06
--- /dev/null
+++ b/tests/testdata/pile_youtubesubtitles-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+68263c52adc0086011e2220b619983935cabb1cc1f5f9f8ee1a74ab2a7457967
\ No newline at end of file
diff --git a/tests/testdata/pile_youtubesubtitles-v1-res.json b/tests/testdata/pile_youtubesubtitles-v1-res.json
new file mode 100644
index 0000000000..fcf2faa8bc
--- /dev/null
+++ b/tests/testdata/pile_youtubesubtitles-v1-res.json
@@ -0,0 +1 @@
+{"results": {"pile_youtubesubtitles": {"bits_per_byte": 3.3827117222045906e-05, "byte_perplexity": 1.000023447445816, "word_perplexity": 1.0001529192262875}}, "versions": {"pile_youtubesubtitles": 1}}
\ No newline at end of file
diff --git a/tests/testdata/wikitext-v1-loglikelihood_rolling b/tests/testdata/wikitext-v1-loglikelihood_rolling
new file mode 100644
index 0000000000..f09af45a38
--- /dev/null
+++ b/tests/testdata/wikitext-v1-loglikelihood_rolling
@@ -0,0 +1 @@
+b6f83e6cf7535ee41b0057c3e2ec2cf7f2fa5a9119b305c479a83091d1142b2c
\ No newline at end of file
diff --git a/tests/testdata/wikitext-v1-res.json b/tests/testdata/wikitext-v1-res.json
new file mode 100644
index 0000000000..122098aec2
--- /dev/null
+++ b/tests/testdata/wikitext-v1-res.json
@@ -0,0 +1 @@
+{"results": {"wikitext": {"bits_per_byte": 3.202519859941674e-05, "byte_perplexity": 1.0000221984224973, "word_perplexity": 1.000118710696617}}, "versions": {"wikitext": 1}}
\ No newline at end of file

From 4b3dee673a776de7be988b40a1a9c40b27e04e4d Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Sat, 1 Jan 2022 21:44:56 -0700
Subject: [PATCH 48/65] asdiv: space convention

---
 lm_eval/tasks/asdiv.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lm_eval/tasks/asdiv.py b/lm_eval/tasks/asdiv.py
index 0ff563538e..ed46445968 100644
--- a/lm_eval/tasks/asdiv.py
+++ b/lm_eval/tasks/asdiv.py
@@ -93,7 +93,7 @@ def doc_to_target(self, doc):
         # TODO: add formula
 
         answer = doc['answer'].split(' (')[0]
-        return answer
+        return " " + answer
 
     def construct_requests(self, doc, ctx):
         ll, is_greedy = rf.loglikelihood(ctx, self.doc_to_target(doc))

From cb3babd3620cc597f3399c17622db7bddec99a39 Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Mon, 3 Jan 2022 00:35:41 -0700
Subject: [PATCH 49/65] Improve pile test efficiency

---
 lm_eval/tasks/pile.py | 7 ++++---
 tests/test_tasks.py   | 2 +-
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/lm_eval/tasks/pile.py b/lm_eval/tasks/pile.py
index 68ff7ed9a8..c7cea5ac97 100644
--- a/lm_eval/tasks/pile.py
+++ b/lm_eval/tasks/pile.py
@@ -18,9 +18,10 @@ class PilePerplexityTask(PerplexityTask, abc.ABC):
 
     def download(self):
         # TODO: separate pile val/test out by component so we don't have to scan the entire file once per set
-        os.makedirs("data/pile/", exist_ok=True)
-        download_file("https://the-eye.eu/public/AI/pile/val.jsonl.zst", self.VAL_PATH, "264c875d8bbd355d8daa9d032b75fd8fb91606218bb84dd1155b203fcd5fab92")
-        download_file("https://the-eye.eu/public/AI/pile/test.jsonl.zst", self.TEST_PATH, "0bb28c52d0b5596d389bf179ce2d43bf7f7ffae76b0d2d20b180c97f62e0975e")
+        if not os.path.exists("data/pile/test.jsonl.zst"):
+            os.makedirs("data/pile/", exist_ok=True)
+            download_file("https://the-eye.eu/public/AI/pile/val.jsonl.zst", self.VAL_PATH, "264c875d8bbd355d8daa9d032b75fd8fb91606218bb84dd1155b203fcd5fab92")
+            download_file("https://the-eye.eu/public/AI/pile/test.jsonl.zst", self.TEST_PATH, "0bb28c52d0b5596d389bf179ce2d43bf7f7ffae76b0d2d20b180c97f62e0975e")
 
     def validation_docs(self):
         rdr = lm_dataformat.Reader(self.VAL_PATH)
diff --git a/tests/test_tasks.py b/tests/test_tasks.py
index 97baeacf8a..46812798a9 100644
--- a/tests/test_tasks.py
+++ b/tests/test_tasks.py
@@ -32,7 +32,7 @@ def test_basic_interface(taskname, task_class):
 
     limit = None
 
-    if taskname in ["triviaqa"]:
+    if taskname in ["triviaqa"] or taskname.startswith("pile_"):
         limit = 10000
     if task.has_validation_docs():
         arr = list(islice(task.validation_docs(), limit))

From 9d87d47696f73c18597e12c74c16f001f0497ffe Mon Sep 17 00:00:00 2001
From: Leo Gao <54557097+leogao2@users.noreply.github.com>
Date: Mon, 3 Jan 2022 00:36:42 -0700
Subject: [PATCH 50/65] Delete test_cache.db

---
 tests/test_cache.db | Bin 12288 -> 0 bytes
 1 file changed, 0 insertions(+), 0 deletions(-)
 delete mode 100644 tests/test_cache.db

diff --git a/tests/test_cache.db b/tests/test_cache.db
deleted file mode 100644
index 7477f429bf681485fd985975b0c979298d0a181c..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 12288
zcmeI#&r8EF6u|MM$`E06w;g(U?4+QA_z&1hL9s*a3c6E?EW)tsP*c&9f1m%6N1M@|
zcQ4-u`SB7G$S0>ad&ri~%FFtFWt?27k<waSiYTRY+xFVF4WI9ClXmp^uF&f2@p>5k
zsUUPJbm8x=1MDM!00IagfB*srAb<b@2q4fGftenRqe%Za^IX}cuFc9Wx*@pA)3`{b
zh?6XpNw+$YqgVSaMfz08eSR~I^SRun^Aq_rRbwT|?vmrvU@(p&t(Nt|eto=EZB<{I
z#;rGb*<DngcN1UNA9&B9??%2O5I_I{1Q0*~0R#|0009ILK%lPzhyG#I|9$;l-h}`H
U2q1s}0tg_000IagfWSuJA6ODIt^fc4


From d09561fe8cee57775b31bc1adfe6696c3c0f514e Mon Sep 17 00:00:00 2001
From: Leo Gao <54557097+leogao2@users.noreply.github.com>
Date: Mon, 3 Jan 2022 22:58:42 -0700
Subject: [PATCH 51/65] Update description_guide.md

---
 docs/description_guide.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/description_guide.md b/docs/description_guide.md
index 3627e23e6b..b3fea0834f 100644
--- a/docs/description_guide.md
+++ b/docs/description_guide.md
@@ -41,7 +41,7 @@ One can also interface with the aforementioned [`evaluator.evaluate`](../lm_eval
 
 which can then be supplied to the CLI as:
 
-```python
+```bash
 python main.py  \
 --tasks cycle_letters,copa \
 --description_dict_path /your/path/descriptions.json \

From d2636b4e71d2eb23faf6af547c53ed139b79d376 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Tue, 4 Jan 2022 02:38:18 -0500
Subject: [PATCH 52/65] Enforce `rnd` args with assertions

---
 lm_eval/base.py | 15 +++------------
 1 file changed, 3 insertions(+), 12 deletions(-)

diff --git a/lm_eval/base.py b/lm_eval/base.py
index c3f897c711..9313fcffba 100644
--- a/lm_eval/base.py
+++ b/lm_eval/base.py
@@ -471,13 +471,13 @@ def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None,
             Not implemented, and this option is deprecated and will be removed in a future version in favor of a different description providing method
         :param rnd: random.Random
             The pseudo-random number generator used to randomly sample examples.
-            WARNING: If you do not provide a `rnd` arg, a default `random.Random`
-            object will be created and seeded with this Task's name attribute, `__name__`.
+            WARNING: This is currently a required arg although it's optionalized with a default `None`.
         :param description: str
             The task's description that will be prepended to the fewshot examples.
         :returns: str
             The fewshot context.
         """
+        assert rnd is not None, "A `random.Random` generator argument must be provided to `rnd`"
         assert not provide_description, (
             "The `provide_description` arg will be removed in future versions. To prepend "
             "a custom description to the context, supply the corresponding string via the "
@@ -489,11 +489,6 @@ def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None,
 
         description = description + "\n\n" if description else ""
 
-        # TODO (jon-tow): Remove this default `rand` behaviour after `provide_description` is removed and remove the respective `rand` arg warning in the docs above.
-        if rnd is None:
-            rnd = random.Random()
-            rnd.seed(self.__name__)
-
         if num_fewshot == 0:
             labeled_examples = ""
         else:
@@ -567,6 +562,7 @@ def fewshot_examples(self, k, rnd):
 
     def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None, description=None):
         assert num_fewshot == 0
+        assert rnd is not None, "A `random.Random` generator argument must be provided to `rnd`"
         assert not provide_description, (
             "The `provide_description` arg will be removed in future versions. To prepend "
             "a custom description to the context, supply the corresponding string via the  "
@@ -576,11 +572,6 @@ def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None,
             # nudge people to not specify it at all
             print("WARNING: provide_description is deprecated and will be removed in a future version in favor of description_dict")
 
-        # TODO (jon-tow): Remove this default `rand` behaviour after `provide_description` is removed and remove the respective `rand` arg warning in the docs above.
-        if rnd is None:
-            rnd = random.Random()
-            rnd.seed(self.__name__)
-
         return ""
 
     def higher_is_better(self):

From 76dc60938b43ec471149640e7ed954f3229fbc3e Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Sat, 8 Jan 2022 21:37:02 +0100
Subject: [PATCH 53/65] Best-download have backward compatibility issue

---
 setup.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/setup.py b/setup.py
index fdb74c4d79..b6ba841b97 100644
--- a/setup.py
+++ b/setup.py
@@ -21,7 +21,7 @@
     python_requires='>=3.6',
     install_requires=[
         "black",
-        "best_download>=0.0.6",
+        "best_download>=0.0.6,<0.0.8",
         "datasets==1.15.1",
         "click>=7.1",
         "scikit-learn>=0.24.1",

From b2f6bce86c62a40498d863d103bd3c25c6d7f672 Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Sat, 8 Jan 2022 13:37:42 -0700
Subject: [PATCH 54/65] Fix fewshot_context method handling

---
 lm_eval/utils.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/lm_eval/utils.py b/lm_eval/utils.py
index aca7791181..2a8c6d17fe 100644
--- a/lm_eval/utils.py
+++ b/lm_eval/utils.py
@@ -2,6 +2,7 @@
 import re
 import collections
 import functools
+import inspect
 
 
 class ExitCodeError(Exception):
@@ -148,7 +149,7 @@ def positional_deprecated(fn):
     """
     @functools.wraps(fn)
     def _wrapper(*args, **kwargs):
-        if len(args) != 0: 
+        if len(args) != 1 if inspect.ismethod(fn) else 0: 
             print(f"WARNING: using {fn.__name__} with positional arguments is "
                 "deprecated and will be disallowed in a future version of "
                 "lm-evaluation-harness!")

From c65412e530c1780ec7b2b970373e9e83f796ab20 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Sat, 8 Jan 2022 21:45:40 +0100
Subject: [PATCH 55/65] Actually it shouldn't be hard to fix it to be
 compatible with future version

---
 lm_eval/tasks/arithmetic.py           | 2 +-
 lm_eval/tasks/coqa.py                 | 4 ++--
 lm_eval/tasks/drop.py                 | 2 +-
 lm_eval/tasks/hendrycks_ethics.py     | 2 +-
 lm_eval/tasks/hendrycks_math.py       | 2 +-
 lm_eval/tasks/hendrycks_test.py       | 2 +-
 lm_eval/tasks/lambada.py              | 4 ++--
 lm_eval/tasks/lambada_multilingual.py | 4 ++--
 lm_eval/tasks/logiqa.py               | 2 +-
 lm_eval/tasks/mutual.py               | 4 ++--
 lm_eval/tasks/pile.py                 | 4 ++--
 lm_eval/tasks/qa4mre.py               | 4 ++--
 lm_eval/tasks/sciq.py                 | 4 ++--
 lm_eval/tasks/triviaqa.py             | 2 +-
 lm_eval/tasks/truthfulqa.py           | 4 ++--
 lm_eval/tasks/unscramble.py           | 2 +-
 lm_eval/tasks/wikitext.py             | 2 +-
 setup.py                              | 2 +-
 18 files changed, 26 insertions(+), 26 deletions(-)

diff --git a/lm_eval/tasks/arithmetic.py b/lm_eval/tasks/arithmetic.py
index 147b66a175..b3256b5c87 100644
--- a/lm_eval/tasks/arithmetic.py
+++ b/lm_eval/tasks/arithmetic.py
@@ -21,7 +21,7 @@ def download(self):
         url = 'https://raw.githubusercontent.com/openai/gpt-3/master/data/' + file_name
         if not os.path.exists(self.directory):
             os.makedirs(self.directory)
-        download_file(url, self.directory+file_name, checksum)
+        download_file(url, local_file=self.directory+file_name, expected_checksum=checksum)
         self.set_docs()
 
     @abc.abstractmethod
diff --git a/lm_eval/tasks/coqa.py b/lm_eval/tasks/coqa.py
index beba53a663..d9f6194580 100644
--- a/lm_eval/tasks/coqa.py
+++ b/lm_eval/tasks/coqa.py
@@ -16,8 +16,8 @@ def download(self):
 
         sh ("""mkdir -p data/coqa""")
 
-        download_file("http://downloads.cs.stanford.edu/nlp/data/coqa/coqa-train-v1.0.json", coqa_train_filepath, "b0fdb2bc1bd38dd3ca2ce5fa2ac3e02c6288ac914f241ac409a655ffb6619fa6")
-        download_file("http://downloads.cs.stanford.edu/nlp/data/coqa/coqa-dev-v1.0.json", coqa_dev_filepath, "dfa367a9733ce53222918d0231d9b3bedc2b8ee831a2845f62dfc70701f2540a")
+        download_file("http://downloads.cs.stanford.edu/nlp/data/coqa/coqa-train-v1.0.json", local_file=coqa_train_filepath, expected_checksum="b0fdb2bc1bd38dd3ca2ce5fa2ac3e02c6288ac914f241ac409a655ffb6619fa6")
+        download_file("http://downloads.cs.stanford.edu/nlp/data/coqa/coqa-dev-v1.0.json", local_file=coqa_dev_filepath, expected_checksum="dfa367a9733ce53222918d0231d9b3bedc2b8ee831a2845f62dfc70701f2540a")
 
     def has_training_docs(self):
         return True
diff --git a/lm_eval/tasks/drop.py b/lm_eval/tasks/drop.py
index 97d1098327..3b89b7d1ec 100644
--- a/lm_eval/tasks/drop.py
+++ b/lm_eval/tasks/drop.py
@@ -27,7 +27,7 @@ def download(self):
         url = "https://s3-us-west-2.amazonaws.com/allennlp/datasets/drop/drop_dataset.zip"
         checksum = "39d2278a29fd729de301b111a45f434c24834f40df8f4ff116d864589e3249d6"
         zip_path = self.DATASET_PATH / "drop_dataset.zip"
-        download_file(url, str(zip_path), checksum)
+        download_file(url, local_file=str(zip_path), expected_checksum=checksum)
         with ZipFile(zip_path, "r") as zip:
             zip.extractall(self.DATASET_PATH)
 
diff --git a/lm_eval/tasks/hendrycks_ethics.py b/lm_eval/tasks/hendrycks_ethics.py
index 50e94a508c..55d50ffc5d 100644
--- a/lm_eval/tasks/hendrycks_ethics.py
+++ b/lm_eval/tasks/hendrycks_ethics.py
@@ -20,7 +20,7 @@ class Ethics(Task):
     def download(self):
         if not os.path.exists('data/ethics/done'):
             sh("mkdir -p data")
-            download_file("https://people.eecs.berkeley.edu/~hendrycks/ethics.tar", "data/ethics.tar", "40acbf1ac0da79a2aabef394d58889136b8d38b05be09482006de2453fb06333")
+            download_file("https://people.eecs.berkeley.edu/~hendrycks/ethics.tar", local_file="data/ethics.tar", expected_checksum="40acbf1ac0da79a2aabef394d58889136b8d38b05be09482006de2453fb06333")
             sh("""
             tar -xf data/ethics.tar -C data/
             rm data/ethics.tar
diff --git a/lm_eval/tasks/hendrycks_math.py b/lm_eval/tasks/hendrycks_math.py
index 379e727d61..0c96ba52dc 100644
--- a/lm_eval/tasks/hendrycks_math.py
+++ b/lm_eval/tasks/hendrycks_math.py
@@ -18,7 +18,7 @@ class Math(Task):
     def download(self):
         if not (self.DATASET_PATH / 'test').exists() or not (self.DATASET_PATH / 'done').exists():
             sh(f"mkdir -p {self.DATASET_PATH}")
-            download_file("https://people.eecs.berkeley.edu/~hendrycks/MATH.tar", f"{self.DATASET_PATH}.tar", "01256fd7cd5430596fdf07e6e6a5827111b5235b7ffed679c662a12f898932da")
+            download_file("https://people.eecs.berkeley.edu/~hendrycks/MATH.tar", local_file=f"{self.DATASET_PATH}.tar", expected_checksum="01256fd7cd5430596fdf07e6e6a5827111b5235b7ffed679c662a12f898932da")
             sh(f"""
             tar -xf {self.DATASET_PATH}.tar -C data/ && touch {self.DATASET_PATH / 'done'}
             rm {self.DATASET_PATH}.tar
diff --git a/lm_eval/tasks/hendrycks_test.py b/lm_eval/tasks/hendrycks_test.py
index 46c0306fcd..0546c9d9e1 100644
--- a/lm_eval/tasks/hendrycks_test.py
+++ b/lm_eval/tasks/hendrycks_test.py
@@ -45,7 +45,7 @@ def __init__(self, subject):
     def download(self):
         if not (self.DATASET_PATH / 'done').exists():
             sh("mkdir -p data")
-            download_file("https://people.eecs.berkeley.edu/~hendrycks/data.tar", "data/data.tar", "78a804365a59028188fb19bd1adcadc5e0c260b220a9d8b2e33a5ea7d5fbe3b4")
+            download_file("https://people.eecs.berkeley.edu/~hendrycks/data.tar", local_file="data/data.tar", expected_checksum="78a804365a59028188fb19bd1adcadc5e0c260b220a9d8b2e33a5ea7d5fbe3b4")
             sh("""
             tar -xf data/data.tar -C data/
             rm data/data.tar
diff --git a/lm_eval/tasks/lambada.py b/lm_eval/tasks/lambada.py
index bcb4ae019c..a73602f7f5 100644
--- a/lm_eval/tasks/lambada.py
+++ b/lm_eval/tasks/lambada.py
@@ -14,8 +14,8 @@ def download(self):
             if not os.path.exists("data/lambada/lambada_test.jsonl"):
                 download_file(
                     "http://eaidata.bmk.sh/data/lambada_test.jsonl", 
-                    "data/lambada/lambada_test.jsonl", 
-                    "4aa8d02cd17c719165fc8a7887fddd641f43fcafa4b1c806ca8abc31fabdb226"
+                    local_file="data/lambada/lambada_test.jsonl",
+                    expected_checksum="4aa8d02cd17c719165fc8a7887fddd641f43fcafa4b1c806ca8abc31fabdb226"
                 )
         except:
             # fallback - for some reason best_download doesnt work all the time here
diff --git a/lm_eval/tasks/lambada_multilingual.py b/lm_eval/tasks/lambada_multilingual.py
index dd6da10bef..7123ecf01a 100644
--- a/lm_eval/tasks/lambada_multilingual.py
+++ b/lm_eval/tasks/lambada_multilingual.py
@@ -32,8 +32,8 @@ def download(self):
         if not os.path.exists(f):
           download_file(
               url, 
-              f, 
-              CHECKSUMS[self.LANG]
+              local_file=f,
+              expected_checksum=CHECKSUMS[self.LANG]
           )
       except:
         # fallback - for some reason best_download doesnt work all the time here
diff --git a/lm_eval/tasks/logiqa.py b/lm_eval/tasks/logiqa.py
index e403623beb..97412738dd 100644
--- a/lm_eval/tasks/logiqa.py
+++ b/lm_eval/tasks/logiqa.py
@@ -19,7 +19,7 @@ def download(self):
         ]
         for split in splits:
             file = self.DATASET_PATH / f"{split['name']}.txt"
-            download_file(f"{base_url}/{split['name']}.txt", str(file), split["checksum"])
+            download_file(f"{base_url}/{split['name']}.txt", local_file=str(file), expected_checksum=split["checksum"])
 
     def has_training_docs(self):
         return True
diff --git a/lm_eval/tasks/mutual.py b/lm_eval/tasks/mutual.py
index 17274a46fd..221c28070f 100644
--- a/lm_eval/tasks/mutual.py
+++ b/lm_eval/tasks/mutual.py
@@ -36,8 +36,8 @@ def download(self):
         master_zip = Path("data/master.zip")
         download_file(
             "https://github.com/Nealcly/MuTual/archive/master.zip",
-            str(master_zip),
-            "bb325cf6c672f0f02699993a37138b0fa0af6fcfc77ec81dfbe46add4d7b29f9")
+            local_file=str(master_zip),
+            expected_checksum="bb325cf6c672f0f02699993a37138b0fa0af6fcfc77ec81dfbe46add4d7b29f9")
         with zipfile.ZipFile(master_zip, 'r') as zip:
             zip.extractall("data")
         Path("data/MuTual-master/data").rename(str(self.BASE_PATH))
diff --git a/lm_eval/tasks/pile.py b/lm_eval/tasks/pile.py
index 68ff7ed9a8..2ffd8b833a 100644
--- a/lm_eval/tasks/pile.py
+++ b/lm_eval/tasks/pile.py
@@ -19,8 +19,8 @@ class PilePerplexityTask(PerplexityTask, abc.ABC):
     def download(self):
         # TODO: separate pile val/test out by component so we don't have to scan the entire file once per set
         os.makedirs("data/pile/", exist_ok=True)
-        download_file("https://the-eye.eu/public/AI/pile/val.jsonl.zst", self.VAL_PATH, "264c875d8bbd355d8daa9d032b75fd8fb91606218bb84dd1155b203fcd5fab92")
-        download_file("https://the-eye.eu/public/AI/pile/test.jsonl.zst", self.TEST_PATH, "0bb28c52d0b5596d389bf179ce2d43bf7f7ffae76b0d2d20b180c97f62e0975e")
+        download_file("https://the-eye.eu/public/AI/pile/val.jsonl.zst", local_file=self.VAL_PATH, expected_checksum="264c875d8bbd355d8daa9d032b75fd8fb91606218bb84dd1155b203fcd5fab92")
+        download_file("https://the-eye.eu/public/AI/pile/test.jsonl.zst", local_file=self.TEST_PATH, expected_checksum="0bb28c52d0b5596d389bf179ce2d43bf7f7ffae76b0d2d20b180c97f62e0975e")
 
     def validation_docs(self):
         rdr = lm_dataformat.Reader(self.VAL_PATH)
diff --git a/lm_eval/tasks/qa4mre.py b/lm_eval/tasks/qa4mre.py
index 67810ad747..de3d2e56d6 100644
--- a/lm_eval/tasks/qa4mre.py
+++ b/lm_eval/tasks/qa4mre.py
@@ -32,8 +32,8 @@ def download(self):
         if not os.path.isfile(f"data/qa4mre/QA4MRE-{year}-{lang}"):
             download_file(
                 url_path,
-                f"data/qa4mre/QA4MRE-{year}-{lang}_GS.xml",
-                sha256sums[year],
+                local_file=f"data/qa4mre/QA4MRE-{year}-{lang}_GS.xml",
+                expected_checksum=sha256sums[year],
                 )
 
     def has_training_docs(self):
diff --git a/lm_eval/tasks/sciq.py b/lm_eval/tasks/sciq.py
index b750354a7b..021a79d78f 100644
--- a/lm_eval/tasks/sciq.py
+++ b/lm_eval/tasks/sciq.py
@@ -13,8 +13,8 @@ def download(self):
             os.makedirs('data/sciq', exist_ok=True)
             download_file(
                 'https://ai2-public-datasets.s3.amazonaws.com/sciq/SciQ.zip',
-                'data/sciq/SciQ.zip', 
-                '7f3312f6ac6b09970b32942d106a8c44ec0dad46a0369f17d635aff8e348a87c',
+                local_file='data/sciq/SciQ.zip',
+                expected_checksum='7f3312f6ac6b09970b32942d106a8c44ec0dad46a0369f17d635aff8e348a87c',
             )
             with zipfile.ZipFile("data/sciq/SciQ.zip", "r") as zf:
                 zf.extractall("data/sciq/")
diff --git a/lm_eval/tasks/triviaqa.py b/lm_eval/tasks/triviaqa.py
index e61a40bdde..1b23a1f908 100644
--- a/lm_eval/tasks/triviaqa.py
+++ b/lm_eval/tasks/triviaqa.py
@@ -12,7 +12,7 @@ class TriviaQA(Task):
     def download(self):
         if not os.path.exists('data/triviaqa/unfiltered-web-train.jsonl'):
             os.makedirs("data/triviaqa/", exist_ok=True)
-            download_file("http://eaidata.bmk.sh/data/triviaqa-unfiltered.tar.gz", "data/triviaqa/triviaqa-unfiltered.tar.gz", "adc19b42769062d241a8fbe834c56e58598d9322eb6c614e9f33a68a2cf5523e")
+            download_file("http://eaidata.bmk.sh/data/triviaqa-unfiltered.tar.gz", local_file="data/triviaqa/triviaqa-unfiltered.tar.gz", expected_checksum="adc19b42769062d241a8fbe834c56e58598d9322eb6c614e9f33a68a2cf5523e")
             sh("""
             cd data/triviaqa/
             tar -xf triviaqa-unfiltered.tar.gz
diff --git a/lm_eval/tasks/truthfulqa.py b/lm_eval/tasks/truthfulqa.py
index f0b46196bc..854b63d475 100644
--- a/lm_eval/tasks/truthfulqa.py
+++ b/lm_eval/tasks/truthfulqa.py
@@ -58,7 +58,7 @@ def download(self):
         Path.mkdir(self.DATASET_PATH, parents=True)
         mc_url = "https://raw.githubusercontent.com/sylinrl/TruthfulQA/013686a06be7a7bde5bf8223943e106c7250123c/data/mc_task.json"
         checksum = "6eb4125d25750c0145c4be2dce00440736684ab6f74ce6bff2139571cc758954"
-        download_file(mc_url, str(self.DATASET_PATH / "mc_task.json"), checksum)
+        download_file(mc_url, local_file=str(self.DATASET_PATH / "mc_task.json"), expected_checksum=checksum)
 
     def has_training_docs(self):
         return False
@@ -163,7 +163,7 @@ def download(self):
         Path.mkdir(self.DATASET_PATH, parents=True)
         url = "https://raw.githubusercontent.com/sylinrl/TruthfulQA/013686a06be7a7bde5bf8223943e106c7250123c/TruthfulQA.csv"
         checksum = "8d7dd15f033196140f032d97d30f037da7a7b1192c3f36f9937c1850925335a2"
-        download_file(url, str(self.DATASET_PATH / "TruthfulQA.csv"), checksum)
+        download_file(url, local_file=str(self.DATASET_PATH / "TruthfulQA.csv"), expected_checksum=checksum)
 
     def has_training_docs(self):
         return False
diff --git a/lm_eval/tasks/unscramble.py b/lm_eval/tasks/unscramble.py
index dc742a2cee..542df9b4eb 100644
--- a/lm_eval/tasks/unscramble.py
+++ b/lm_eval/tasks/unscramble.py
@@ -29,7 +29,7 @@ def download(self):
         if not file.exists():
             rawfile = file.parent / (file.name + ".gz")
             base_url = "https://raw.githubusercontent.com/openai/gpt-3/master/data"
-            download_file(f"{base_url}/{self.FILENAME}.gz", str(rawfile), self.CHECKSUM)
+            download_file(f"{base_url}/{self.FILENAME}.gz", local_file=str(rawfile), expected_checksum=self.CHECKSUM)
             extract_gzip(gz=rawfile, to=file)
 
     def has_training_docs(self):
diff --git a/lm_eval/tasks/wikitext.py b/lm_eval/tasks/wikitext.py
index 24f9ec3507..94a93c0899 100644
--- a/lm_eval/tasks/wikitext.py
+++ b/lm_eval/tasks/wikitext.py
@@ -46,7 +46,7 @@ class WikiText(PerplexityTask):
     def download(self):
         if not os.path.exists('data/wikitext/wikitext-2-raw/wiki.valid.raw'):
             os.makedirs("data/wikitext/", exist_ok=True)
-            download_file("https://s3.amazonaws.com/research.metamind.io/wikitext/wikitext-2-raw-v1.zip", "data/wikitext/wikitext-2-raw-v1.zip", "ef7edb566e3e2b2d31b29c1fdb0c89a4cc683597484c3dc2517919c615435a11")
+            download_file("https://s3.amazonaws.com/research.metamind.io/wikitext/wikitext-2-raw-v1.zip", local_file="data/wikitext/wikitext-2-raw-v1.zip", expected_checksum="ef7edb566e3e2b2d31b29c1fdb0c89a4cc683597484c3dc2517919c615435a11")
             sh("cd data/wikitext/ && unzip wikitext-2-raw-v1.zip")
 
     def fewshot_description(self):
diff --git a/setup.py b/setup.py
index b6ba841b97..fdb74c4d79 100644
--- a/setup.py
+++ b/setup.py
@@ -21,7 +21,7 @@
     python_requires='>=3.6',
     install_requires=[
         "black",
-        "best_download>=0.0.6,<0.0.8",
+        "best_download>=0.0.6",
         "datasets==1.15.1",
         "click>=7.1",
         "scikit-learn>=0.24.1",

From 02a4def21b4de5f29c6314fb73abe1810d194128 Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Sat, 8 Jan 2022 13:53:50 -0700
Subject: [PATCH 56/65] Update blimp fewshot_context

---
 lm_eval/tasks/blimp.py | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/lm_eval/tasks/blimp.py b/lm_eval/tasks/blimp.py
index e8e7bd9f2b..8a52d888ca 100644
--- a/lm_eval/tasks/blimp.py
+++ b/lm_eval/tasks/blimp.py
@@ -29,9 +29,18 @@ def download(self):
         self.data["validation"] = self.data["train"]
         del self.data["train"]
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
+    def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None, description=None):
         assert num_fewshot == 0
-        assert not provide_description
+        assert rnd is not None, "A `random.Random` generator argument must be provided to `rnd`"
+        assert not provide_description, (
+            "The `provide_description` arg will be removed in future versions. To prepend "
+            "a custom description to the context, supply the corresponding string via the  "
+            "`description` arg."
+        )
+        if provide_description is not None:
+            # nudge people to not specify it at all
+            print("WARNING: provide_description is deprecated and will be removed in a future version in favor of description_dict")
+
         return ""
 
     def doc_to_text(self, doc):

From 2d9fc25632fb0e48c8d6f7d183252800ce02afcd Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Sun, 9 Jan 2022 00:26:43 +0100
Subject: [PATCH 57/65] Missed asdiv

---
 lm_eval/tasks/asdiv.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lm_eval/tasks/asdiv.py b/lm_eval/tasks/asdiv.py
index ed46445968..732d0f1d6c 100644
--- a/lm_eval/tasks/asdiv.py
+++ b/lm_eval/tasks/asdiv.py
@@ -35,7 +35,7 @@ def download(self):
         url = "https://github.com/chaochun/nlu-asdiv-dataset/archive/55790e5270bb91ccfa5053194b25732534696b50.zip"
         checksum = "8f1fe4f6d5f170ec1e24ab78c244153c14c568b1bb2b1dad0324e71f37939a2d"
         zip_path = self.DATASET_PATH / "55790e5270bb91ccfa5053194b25732534696b50.zip"
-        download_file(url, str(zip_path), checksum)
+        download_file(url, local_file=str(zip_path), expected_checksum=checksum)
         with ZipFile(zip_path, "r") as zip:
             zip.extractall(self.DATASET_PATH)
         os.remove(zip_path)

From ea3fd79bd2309fec6e2fea74aaa072a967173dc3 Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Tue, 11 Jan 2022 10:32:12 +0100
Subject: [PATCH 58/65] Fix CB

---
 lm_eval/tasks/superglue.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/lm_eval/tasks/superglue.py b/lm_eval/tasks/superglue.py
index 33598f2301..849fb0781b 100644
--- a/lm_eval/tasks/superglue.py
+++ b/lm_eval/tasks/superglue.py
@@ -93,14 +93,14 @@ def doc_to_target(self, doc):
         # True = entailment
         # False = contradiction
         # Neither = neutral
-        return " {}".format({0: "True", 1: "Neither", 2: "False"}[doc["label"]])
+        return " {}".format({0: "True", 1: "False", 2: "Neither"}[doc["label"]])
 
     def construct_requests(self, doc, ctx):
         ll_true, _ = rf.loglikelihood(ctx, ' True')
-        ll_neither, _ = rf.loglikelihood(ctx, ' Neither')
         ll_false, _ = rf.loglikelihood(ctx, ' False')
+        ll_neither, _ = rf.loglikelihood(ctx, ' Neither')
 
-        return ll_true, ll_neither, ll_false
+        return ll_true, ll_false, ll_neither
 
     def process_results(self, doc, results):
         gold = doc["label"]

From b421f0529a4d2bf1ef8e68cd74671be44010c37a Mon Sep 17 00:00:00 2001
From: thomasw21 <24695242+thomasw21@users.noreply.github.com>
Date: Tue, 11 Jan 2022 12:50:07 +0100
Subject: [PATCH 59/65] Bump version

---
 lm_eval/tasks/superglue.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lm_eval/tasks/superglue.py b/lm_eval/tasks/superglue.py
index 849fb0781b..489d18a6ae 100644
--- a/lm_eval/tasks/superglue.py
+++ b/lm_eval/tasks/superglue.py
@@ -65,7 +65,7 @@ def aggregation(self):
 
 
 class CommitmentBank(HFTask):
-    VERSION = 0
+    VERSION = 1
     DATASET_PATH = "super_glue"
     DATASET_NAME = "cb"
 

From caff4f1099349f599a3bdeb7725bb2add06d741b Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Thu, 13 Jan 2022 02:36:29 -0500
Subject: [PATCH 60/65] Conform `asidv` to the new description api

---
 lm_eval/tasks/asdiv.py | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/lm_eval/tasks/asdiv.py b/lm_eval/tasks/asdiv.py
index 732d0f1d6c..58bcdcd250 100644
--- a/lm_eval/tasks/asdiv.py
+++ b/lm_eval/tasks/asdiv.py
@@ -31,7 +31,7 @@ class Asdiv(Task):
     def download(self):
         if self.DATASET_PATH.exists():
             return
-        Path.mkdir(self.DATASET_PATH)
+        Path.mkdir(self.DATASET_PATH, parents=True)
         url = "https://github.com/chaochun/nlu-asdiv-dataset/archive/55790e5270bb91ccfa5053194b25732534696b50.zip"
         checksum = "8f1fe4f6d5f170ec1e24ab78c244153c14c568b1bb2b1dad0324e71f37939a2d"
         zip_path = self.DATASET_PATH / "55790e5270bb91ccfa5053194b25732534696b50.zip"
@@ -75,10 +75,14 @@ def validation_docs(self):
         data_xml_path = self.DATASET_PATH / "nlu-asdiv-dataset-55790e5270bb91ccfa5053194b25732534696b50/dataset/ASDiv.xml"
         return self.load_docs(data_xml_path)
 
-    def fewshot_context(self, doc, num_fewshot, provide_description, rnd):
+    def fewshot_context(self, doc, num_fewshot, provide_description=None, rnd=None, description=None):
         assert num_fewshot == 0, "ASDiv is intended only for the zero-shot setting."
-        return super().fewshot_context(doc, num_fewshot, provide_description, rnd)
-
+        return super().fewshot_context(
+            doc=doc,
+            num_fewshot=num_fewshot,
+            rnd=rnd,
+            description=description
+        )
     
     def fewshot_description(self):
         # TODO: add solution-type and formula
@@ -115,4 +119,3 @@ def higher_is_better(self):
         return {
             'acc': True
         }
-

From 1c52e917cf9196a72dee111ff3a107b115e9da96 Mon Sep 17 00:00:00 2001
From: Leo Gao <leogao31@gmail.com>
Date: Fri, 14 Jan 2022 21:29:31 -0700
Subject: [PATCH 61/65] cb: add testdata

---
 tests/testdata/cb-v1-loglikelihood | 1 +
 tests/testdata/cb-v1-res.json      | 1 +
 2 files changed, 2 insertions(+)
 create mode 100644 tests/testdata/cb-v1-loglikelihood
 create mode 100644 tests/testdata/cb-v1-res.json

diff --git a/tests/testdata/cb-v1-loglikelihood b/tests/testdata/cb-v1-loglikelihood
new file mode 100644
index 0000000000..ad7e928fe6
--- /dev/null
+++ b/tests/testdata/cb-v1-loglikelihood
@@ -0,0 +1 @@
+77b11f4348eb8a7f57faf95c531fda01ab4bf0e729f91a82451ed8e71ec8e66d
\ No newline at end of file
diff --git a/tests/testdata/cb-v1-res.json b/tests/testdata/cb-v1-res.json
new file mode 100644
index 0000000000..1cff410b2c
--- /dev/null
+++ b/tests/testdata/cb-v1-res.json
@@ -0,0 +1 @@
+{"results": {"cb": {"acc": 0.3392857142857143, "acc_stderr": 0.06384226561930825, "f1": 0.2819143819143819}}, "versions": {"cb": 1}}
\ No newline at end of file

From 03c15e065e6d4c68e9ecbc9b78c45450bf8a9976 Mon Sep 17 00:00:00 2001
From: Jonathan Tow <jonathantow1@gmail.com>
Date: Sun, 23 Jan 2022 16:11:51 -0500
Subject: [PATCH 62/65] Update `headqa` deprecation warning to display on init
 only

---
 lm_eval/tasks/headqa.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lm_eval/tasks/headqa.py b/lm_eval/tasks/headqa.py
index 7551f380ac..d9ac2d87c1 100644
--- a/lm_eval/tasks/headqa.py
+++ b/lm_eval/tasks/headqa.py
@@ -37,4 +37,6 @@ class HeadQAEs(HeadQABase):
 class HeadQAEsDeprecated(HeadQABase):
     DATASET_NAME = "es"
 
-    print("WARNING: headqa is deprecated. Please use headqa_es or headqa_en instead. See https://github.com/EleutherAI/lm-evaluation-harness/pull/240 for more info.")
\ No newline at end of file
+    def __init__(self):
+        super().__init__()
+        print("WARNING: headqa is deprecated. Please use headqa_es or headqa_en instead. See https://github.com/EleutherAI/lm-evaluation-harness/pull/240 for more info.")
\ No newline at end of file

From 1dc6eb013b33cae6e52d8756ff9dbae188e1d69a Mon Sep 17 00:00:00 2001
From: hadyelsahar <hadyelsahar@gmail.com>
Date: Wed, 2 Feb 2022 00:45:41 +0100
Subject: [PATCH 63/65] adding support of XGLM model

---
 lm_eval/models/__init__.py |  2 +
 lm_eval/models/xglm.py     | 84 ++++++++++++++++++++++++++++++++++++++
 2 files changed, 86 insertions(+)
 create mode 100644 lm_eval/models/xglm.py

diff --git a/lm_eval/models/__init__.py b/lm_eval/models/__init__.py
index a12f68a513..1fba7ce81c 100644
--- a/lm_eval/models/__init__.py
+++ b/lm_eval/models/__init__.py
@@ -1,12 +1,14 @@
 from . import gpt2
 from . import gpt3
 from . import dummy
+from . import xglm
 
 MODEL_REGISTRY = {
     "hf": gpt2.HFLM,
     "gpt2": gpt2.GPT2LM,
     "gpt3": gpt3.GPT3LM,
     "dummy": dummy.DummyLM,
+    "XGLM": xglm.XGLM,
 }
 
 
diff --git a/lm_eval/models/xglm.py b/lm_eval/models/xglm.py
new file mode 100644
index 0000000000..1c0da62030
--- /dev/null
+++ b/lm_eval/models/xglm.py
@@ -0,0 +1,84 @@
+import transformers
+import torch
+from lm_eval.base import BaseLM
+
+
+class XGLM(BaseLM):
+    def __init__(self, device='cuda', pretrained='facebook/xglm-1.7B', revision='main', subfolder=None, tokenizer=None, batch_size=1):
+        super().__init__()
+        assert isinstance(device, str)
+        assert isinstance(pretrained, str)
+        assert isinstance(batch_size, int)
+        if device:
+            self._device = torch.device(device)
+        else:
+            self._device = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')
+        # TODO: update this to be less of a hack once subfolder is fixed in HF
+        self.xglm = transformers.AutoModelForCausalLM.from_pretrained(
+            pretrained,
+            # cache_dir="/users/zyong2/data/zyong2/huggingface/xglm"
+        ).to(self.device)
+        print(f"🤖 Loading model {pretrained}")
+        self.xglm.eval()
+        # pretrained tokenizer for neo is broken for now so just hard-coding this to gpt2
+        self.tokenizer = transformers.AutoTokenizer.from_pretrained(
+            pretrained if tokenizer is None else tokenizer, revision=revision, subfolder=subfolder)
+        # assert isinstance(self.tokenizer, (
+        #     transformers.GPT2Tokenizer, transformers.GPT2TokenizerFast,
+        #     transformers.T5Tokenizer, transformers.T5TokenizerFast,
+        # )), "this tokenizer has not been checked for compatibility yet!"
+        self.vocab_size = self.tokenizer.vocab_size
+        # if isinstance(self.tokenizer, (transformers.GPT2Tokenizer, transformers.GPT2TokenizerFast)):
+        #     assert self.tokenizer.encode('hello\n\nhello') == [31373, 198, 198, 31373], \
+        #         self.tokenizer.encode('hello\n\nhello')
+        # multithreading and batching
+        self.batch_size_per_gpu = batch_size  # todo: adaptive batch size
+        # TODO: fix multi-gpu
+        # gpus = torch.cuda.device_count()
+        # if gpus > 1:
+        #     self.gpt2 = nn.DataParallel(self.gpt2)
+    @property
+    def eot_token_id(self):
+        # we use EOT because end of *text* is more accurate for what we're doing than end of *sentence*
+        return self.tokenizer.eos_token_id
+    @property
+    def max_length(self):
+        try:
+            return self.xglm.config.n_ctx
+        except AttributeError:
+            # gptneoconfig doesn't have n_ctx apparently
+            return self.xglm.config.max_position_embeddings
+    @property
+    def max_gen_toks(self):
+        return 256
+    @property
+    def batch_size(self):
+        # TODO: fix multi-gpu
+        return self.batch_size_per_gpu  # * gpus
+    @property
+    def device(self):
+        # TODO: fix multi-gpu
+        return self._device
+    def tok_encode(self, string: str):
+        return self.tokenizer.encode(string, add_special_tokens=False)
+    
+    def tok_decode(self, tokens):
+        return self.tokenizer.decode(tokens)
+    def _model_call(self, inps):
+        """
+        inps: a torch tensor of shape [batch, sequence]
+        the size of sequence may vary from call to call
+        returns: a torch tensor of shape [batch, sequence, vocab] with the
+        logits returned from the model
+        """
+        with torch.no_grad():
+            return self.xglm(inps)[0][:, :, :256008]
+    
+    def _model_generate(self, context, max_length, eos_token_id):
+        result = self.xglm.generate(
+            context,
+            max_length=max_length,
+            eos_token_id=eos_token_id,
+            do_sample=False
+        )
+        return result
\ No newline at end of file

From c3b72670d23d013c91b6a0397ac2289fc57c506f Mon Sep 17 00:00:00 2001
From: hadyelsahar <hadyelsahar@gmail.com>
Date: Thu, 10 Feb 2022 12:21:22 +0100
Subject: [PATCH 64/65] Fix new lines tokenizer issue in XGLM

---
 lm_eval/models/xglm.py       | 20 +++++++++++++++++---
 lm_eval/tasks/translation.py |  3 +--
 2 files changed, 18 insertions(+), 5 deletions(-)

diff --git a/lm_eval/models/xglm.py b/lm_eval/models/xglm.py
index 1c0da62030..3aac8f9ead 100644
--- a/lm_eval/models/xglm.py
+++ b/lm_eval/models/xglm.py
@@ -1,6 +1,8 @@
 import transformers
 import torch
+from lm_eval import utils
 from lm_eval.base import BaseLM
+from tqdm import tqdm
 
 
 class XGLM(BaseLM):
@@ -23,6 +25,7 @@ def __init__(self, device='cuda', pretrained='facebook/xglm-1.7B', revision='mai
         # pretrained tokenizer for neo is broken for now so just hard-coding this to gpt2
         self.tokenizer = transformers.AutoTokenizer.from_pretrained(
             pretrained if tokenizer is None else tokenizer, revision=revision, subfolder=subfolder)
+
         # assert isinstance(self.tokenizer, (
         #     transformers.GPT2Tokenizer, transformers.GPT2TokenizerFast,
         #     transformers.T5Tokenizer, transformers.T5TokenizerFast,
@@ -60,10 +63,21 @@ def device(self):
         # TODO: fix multi-gpu
         return self._device
     def tok_encode(self, string: str):
+        # HACK: to overcome problem of  XGLM tokenizer removing new lines
+        # we replace newline with SEP token
+        # WARNING: Since typical SEP token == EOS token
+        # Generation will stop after the first appearance of SEP token prevnting XGLM from 
+        # outputting Multi line generations
+        string = string.replace("\n", self.tokenizer.sep_token)
         return self.tokenizer.encode(string, add_special_tokens=False)
-    
+
     def tok_decode(self, tokens):
-        return self.tokenizer.decode(tokens)
+        # HACK: to overcome problem of  XGLM tokenizer removing new lines
+        # replace back the generated sep_tokens with newlines
+        output =  self.tokenizer.decode(tokens)
+        output = output.replace(self.tokenizer.sep_token, "\n")
+        print(output)
+        return output
     def _model_call(self, inps):
         """
         inps: a torch tensor of shape [batch, sequence]
@@ -81,4 +95,4 @@ def _model_generate(self, context, max_length, eos_token_id):
             eos_token_id=eos_token_id,
             do_sample=False
         )
-        return result
\ No newline at end of file
+        return result
diff --git a/lm_eval/tasks/translation.py b/lm_eval/tasks/translation.py
index 2e70b03a84..4d65de43fb 100644
--- a/lm_eval/tasks/translation.py
+++ b/lm_eval/tasks/translation.py
@@ -107,7 +107,7 @@ def doc_to_text(self, doc):
         language_codes = self.sacrebleu_language_pair.split("-")
         src_lang = code_to_language(language_codes[0])
         tar_lang = code_to_language(language_codes[1])
-        return f"{src_lang} phrase: " + doc["src"] + f"\n{tar_lang} phrase:"
+        return f"\nTranslate {src_lang} to {tar_lang}:\n [{src_lang}] " + doc["src"] + f"\n[{tar_lang}]"
 
     def doc_to_target(self, doc):
         # This shows a single target, though there may be multiple targets in a lang test
@@ -132,7 +132,6 @@ def process_results(self, doc, results):
         if tar_lang_code in NO_SPACE_LANG:
             doc["ref"] = NO_SPACE_LANG[tar_lang_code]([doc["ref"]])[0]
             results = NO_SPACE_LANG[tar_lang_code](results)
-
         # These metrics are corpus-level not sentence level, so we'll hide the
         # results in this dict and compute the corpus score in the aggregate method
         ref_pred = (doc["ref"], results)

From fbdc61c66a86ba12f359df84757043ec743d283e Mon Sep 17 00:00:00 2001
From: hadyelsahar <hadyelsahar@gmail.com>
Date: Thu, 17 Feb 2022 14:42:01 +0100
Subject: [PATCH 65/65] integration of bigscience model

---
 lm_eval/models/__init__.py   |  2 +
 lm_eval/models/bigscience.py | 84 ++++++++++++++++++++++++++++++++++++
 2 files changed, 86 insertions(+)
 create mode 100644 lm_eval/models/bigscience.py

diff --git a/lm_eval/models/__init__.py b/lm_eval/models/__init__.py
index 1fba7ce81c..9ffd1ceffb 100644
--- a/lm_eval/models/__init__.py
+++ b/lm_eval/models/__init__.py
@@ -2,6 +2,7 @@
 from . import gpt3
 from . import dummy
 from . import xglm
+from . import bigscience
 
 MODEL_REGISTRY = {
     "hf": gpt2.HFLM,
@@ -9,6 +10,7 @@
     "gpt3": gpt3.GPT3LM,
     "dummy": dummy.DummyLM,
     "XGLM": xglm.XGLM,
+    "bigscience":bigscience.BigScience,
 }
 
 
diff --git a/lm_eval/models/bigscience.py b/lm_eval/models/bigscience.py
new file mode 100644
index 0000000000..e54b062716
--- /dev/null
+++ b/lm_eval/models/bigscience.py
@@ -0,0 +1,84 @@
+import transformers
+import torch
+from lm_eval.base import BaseLM
+#
+# 
+# 
+# ​
+class BigScience(BaseLM):
+
+    def __init__(self, device='cuda', pretrained='bigscience/tr5b-1B3-multilingual-alpha-checkpoints', revision='global_step118500', subfolder=None, tokenizer=None, batch_size=1):
+        super().__init__()
+
+        assert isinstance(device, str)
+        assert isinstance(pretrained, str)
+        assert isinstance(batch_size, int)
+        
+        if device:
+            self._device = torch.device(device)
+        else:
+            self._device = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')
+        # TODO: update this to be less of a hack once subfolder is fixed in HF
+        self.bigscience = transformers.AutoModelForCausalLM.from_pretrained(
+            pretrained, revision=revision
+        ).to(self.device)
+        self.bigscience.eval()
+        # pretrained tokenizer for neo is broken for now so just hard-coding this to gpt2
+        self.tokenizer = transformers.AutoTokenizer.from_pretrained(
+            pretrained if tokenizer is None else tokenizer, revision=revision, subfolder=subfolder)
+        # assert isinstance(self.tokenizer, (
+        #     transformers.GPT2Tokenizer, transformers.GPT2TokenizerFast,
+        #     transformers.T5Tokenizer, transformers.T5TokenizerFast,
+        # )), "this tokenizer has not been checked for compatibility yet!"
+        self.vocab_size = self.tokenizer.vocab_size
+        # if isinstance(self.tokenizer, (transformers.GPT2Tokenizer, transformers.GPT2TokenizerFast)):
+        #     assert self.tokenizer.encode('hello\n\nhello') == [31373, 198, 198, 31373], \
+        #         self.tokenizer.encode('hello\n\nhello')
+        # multithreading and batching
+        self.batch_size_per_gpu = batch_size  # todo: adaptive batch size
+        # TODO: fix multi-gpu
+        # gpus = torch.cuda.device_count()
+        # if gpus > 1:
+        #     self.gpt2 = nn.DataParallel(self.gpt2)
+    @property
+    def eot_token_id(self):
+        # we use EOT because end of *text* is more accurate for what we're doing than end of *sentence*
+        return self.tokenizer.eos_token_id
+    @property
+    def max_length(self):
+        try:
+            return self.bigscience.config.n_ctx
+        except AttributeError:
+            # gptneoconfig doesn't have n_ctx apparently
+            return self.bigscience.config.max_position_embeddings
+    @property
+    def max_gen_toks(self):
+        return 256
+    @property
+    def batch_size(self):
+        # TODO: fix multi-gpu
+        return self.batch_size_per_gpu  # * gpus
+    @property
+    def device(self):
+        # TODO: fix multi-gpu
+        return self._device
+    def tok_encode(self, string: str):
+        return self.tokenizer.encode(string, add_special_tokens=False)
+    def tok_decode(self, tokens):
+        return self.tokenizer.decode(tokens)
+    def _model_call(self, inps):
+        """
+        inps: a torch tensor of shape [batch, sequence]
+        the size of sequence may vary from call to call
+        returns: a torch tensor of shape [batch, sequence, vocab] with the
+        logits returned from the model
+        """
+        with torch.no_grad():
+            return self.bigscience(inps)[0][:, :, :130000]
+    def _model_generate(self, context, max_length, eos_token_id):
+        result = self.bigscience.generate(
+            context,
+            max_length=max_length,
+            eos_token_id=eos_token_id,
+            do_sample=False)
+        return result