Merge branch 'main' into add-wiki-current-events-portal-source

Makefile

@@ -45,7 +45,7 @@ IMAGE_URI ?= $(REGION)-docker.pkg.dev/$(PROJECT_ID)/$(REPOSITORY_NAME)/$(IMAGE_N
 
 .PHONY: build
 build:
-	docker build -t $(IMAGE_URI) -f Dockerfile .
+	docker build --network host -t $(IMAGE_URI) -f Dockerfile .
 
 .PHONY: push
 push:

VERSION

@@ -1 +1 @@
-0.4.0
+0.3.2

news_signals/signals.py

@@ -3,7 +3,7 @@
 import sys
 from abc import abstractmethod
 from collections import Counter, defaultdict
-from typing import List
+from typing import List, Optional
 import json
 import base64
 from pathlib import Path
@@ -918,8 +918,13 @@ def add_wikimedia_pageviews_timeseries(
                 granularity='daily',
                 wikidata_client=wikidata_client,
                 wikimedia_endpoint=wikimedia_endpoint,
-        )        
-        self.timeseries_df['wikimedia_pageviews'] = pageviews_df['wikimedia_pageviews'].values
+        ) 
+        try:
+            self.timeseries_df['wikimedia_pageviews'] = pageviews_df['wikimedia_pageviews'].values
+        except TypeError as e:
+            logger.error(e)
+            logger.warning('Retrieved wikimedia pageviews dataframe is None, not adding to signal')
+
         return self
 
     def add_wikipedia_current_events(
@@ -1214,7 +1219,7 @@ def __init__(
         self,
         name: str,
         components: List[Signal],
-        metadata: dict = None
+        metadata: Optional[dict] = None
     ):
         super().__init__(name, metadata=metadata)
         self.components = components

news_signals/signals_dataset.py

@@ -25,7 +25,7 @@
 
 logger = create_logger(__name__, level=logging.INFO)
 
-MAX_BODY_TOKENS = 500
+MAX_BODY_TOKENS = 1000
 DEFAULT_METADATA = {
     'name': 'News Signals Dataset'
 }
@@ -328,6 +328,7 @@ def reduce_aylien_story(
             "categories": s["categories"],
             "industries": s["industries"],
             "smart_tagger_categories": smart_cats,
+            "media": s["media"],
             "clusters": s["clusters"]
         }, **{f: s[f] for f in additional_fields}
     )