need to find a way to make the ssh key integration tests seperate

conf/config.yaml

@@ -1,16 +1,19 @@
 defaults:
   - server: serenity
-
+  
 comet:
   project: BOEM
   workspace: bw4sz
 
-check_annotations: false
+check_annotations: true
+force_upload: false
 label_studio:
   project_name: "Bureau of Ocean Energy Management"
   url: "https://labelstudio.naturecast.org/"
-  api_key: "${oc.env:LABEL_STUDIO_API_KEY}"
   folder_name: "/pgsql/retrieverdash/everglades-label-studio/everglades-data"
+  images_to_annotate_dir: /blue/ewhite/b.weinstein/BOEM/sample_flight/JPG_2024_Jan27
+  annotated_images_dir: /blue/ewhite/b.weinstein/BOEM/sample_flight/JPG_2024_Jan27/annotated
+  csv_dir: /blue/ewhite/b.weinstein/BOEM/annotations
 
 predict:
   patch_size: 450
@@ -27,28 +30,28 @@ propagate:
 
 detection_model:
   checkpoint: bird
-  checkpoint_dir:
+  checkpoint_dir: /blue/ewhite/b.weinstein/BOEM/detection/checkpoints
   validation_csv_path:
-  train_csv_folder:
-  train_image_dir:
-  crop_image_dir:
+  train_csv_folder: /blue/ewhite/b.weinstein/BOEM/annotations/
+  train_image_dir: /blue/ewhite/b.weinstein/BOEM/sample_flight/JPG_2024_Jan27/annotated
+  crop_image_dir: /blue/ewhite/b.weinstein/BOEM/detection/crops/
   limit_empty_frac: 0
   fast_dev_run: false
   labels:
     - "Bird"
 
 classification_model:
-  checkpoint:
-  checkpoint_dir:
-  validation_csv_path:
-  train_csv_folder:
-  train_image_dir:
-  crop_image_dir:
+  checkpoint: 
+  checkpoint_dir: /blue/ewhite/b.weinstein/BOEM/classification/checkpoints
+  validation_csv_path: 
+  train_csv_folder: /blue/ewhite/b.weinstein/BOEM/annotations/
+  train_image_dir: /blue/ewhite/b.weinstein/BOEM/sample_flight/JPG_2024_Jan27/annotated
+  crop_image_dir: /blue/ewhite/b.weinstein/BOEM/classification/crops/
   under_sample_ratio: 0
-  fast_dev_run: false
+  fast_dev_run: True
 
 pipeline_evaluation:
-  detect_ground_truth_dir:
+  detect_ground_truth_dir: 
   classify_confident_ground_truth_dir:
   classify_uncertain_ground_truth_dir:
   # This is an average mAP threshold for now, but we may want to add a per-iou threshold in the future
@@ -62,23 +65,23 @@ reporting:
   report_dir:
 
 active_learning:
-  image_dir:
-  strategy: 'random'
-  n_images: 100
-  m: 10
+  image_dir: /blue/ewhite/b.weinstein/BOEM/sample_flight/JPG_2024_Jan27/annotated
+  strategy: 'target-labels'
+  n_images: 10
   patch_size: 256
   patch_overlap: 0.5
   min_score: 0.5
   model_checkpoint:
   target_labels:
+    - "Bird"
 
   # Optional parameters:
-  evaluation: null
-  dask_client: null
-  pool_limit: null
+  evaluation: 
+  dask_client: 
+  pool_limit: 100
 
 active_testing:
-  image_dir:
+  image_dir: /blue/ewhite/b.weinstein/BOEM/sample_flight/JPG_2024_Jan27/annotated
   strategy: 'random'
   n_images: 100
   m: 10
@@ -89,4 +92,5 @@ active_testing:
 deepforest:
   train:
     fast_dev_run: True
+    epochs: 1
   workers: 0

conf/server/serenity.yaml

@@ -1,4 +1,4 @@
 # Remote server for image hosting
 user: 'ben'
 host: 'serenity.ifas.ufl.edu'
-key_filename: '/Users/benweinstein/.ssh/id_rsa.pub'
+key_filename: '/home/b.weinstein/.ssh/id_rsa.pub'

main.py

@@ -1,18 +1,21 @@
 import hydra
+import os
 from omegaconf import DictConfig
 from src.pipeline import Pipeline
+from src.label_studio import get_api_key
 
 @hydra.main(config_path="conf", config_name="config")
 def main(cfg: DictConfig):
     """Main entry point for the application"""
+    api_key = get_api_key()
+    os.environ["LABEL_STUDIO_API_KEY"] = api_key
+    if api_key is None:
+        print("Warning: No Label Studio API key found in .comet.config")
+        return None
 
     # Initialize and run pipeline
     pipeline = Pipeline(cfg=cfg)
-    results = pipeline.run(model_path=cfg.model.path)
-
-    # Log results
-    print(f"Images needing review: {len(results['needs_review'])}")
-    print(f"Images not needing review: {len(results['no_review_needed'])}")
+    results = pipeline.run()
 
 if __name__ == "__main__":
     main()

src/label_studio.py

@@ -25,28 +25,24 @@ def upload_to_label_studio(images, sftp_client, label_studio_project, images_to_
     upload_images(sftp_client=sftp_client, images=images, folder_name=folder_name)
     import_image_tasks(label_studio_project=label_studio_project, image_names=images, local_image_dir=images_to_annotate_dir, predictions=preannotations)
 
-def check_for_new_annotations(user, host, key_filename, label_studio_url, label_studio_project_name, train_csv_folder, images_to_annotate_dir, annotated_images_dir, folder_name):
+def check_for_new_annotations(sftp_client, url, project_name, csv_dir, images_to_annotate_dir, annotated_images_dir, folder_name):
     """
     Check for new annotations from Label Studio, move annotated images, and gather new images to annotate.
 
     Args:
-        user (str): The username for the SFTP connection.
-        host (str): The host URL for the SFTP connection.
-        key_filename (str): The path to the SSH key file for the SFTP connection.
-        label_studio_url (str): The URL of the Label Studio server.
-        label_studio_project_name (str): The name of the Label Studio project.
-        train_csv_folder (str): The path to the folder containing training CSV files.
+        sftp_client (paramiko.SFTPClient): The SFTP client for downloading images.
+        url (str): The URL of the Label Studio server.
+        project_name (str): The name of the Label Studio project.
+        csv_dir (str): The path to the folder containing CSV files.
         images_to_annotate_dir (str): The path to the directory of images to annotate.
         annotated_images_dir (str): The path to the directory of annotated images.
         folder_name (str): The name of the folder to upload images to.
-        filter_labels (list, optional): A list of labels to filter images by. Defaults to None.
 
     Returns:
         DataFrame: A DataFrame containing the gathered annotations.
     """
-    sftp_client = create_sftp_client(user=user, host=host, key_filename=key_filename)
-    label_studio_project = connect_to_label_studio(url=label_studio_url, project_name=label_studio_project_name)
-    new_annotations = download_completed_tasks(label_studio_project=label_studio_project, train_csv_folder=train_csv_folder)
+    label_studio_project = connect_to_label_studio(url=url, project_name=project_name)
+    new_annotations = download_completed_tasks(label_studio_project=label_studio_project, csv_dir=csv_dir)
 
    # Move annotated images out of local pool
     if new_annotations is not None:
@@ -63,7 +59,7 @@ def check_for_new_annotations(user, host, key_filename, label_studio_url, label_
         return None
 
     # Choose new images to annotate
-    label_studio_annotations = gather_data(train_csv_folder)
+    label_studio_annotations = gather_data(csv_dir)
 
     return label_studio_annotations
 
@@ -266,7 +262,7 @@ def import_image_tasks(label_studio_project, image_names, local_image_dir, predi
         tasks.append(upload_dict)
     label_studio_project.import_tasks(tasks)
 
-def download_completed_tasks(label_studio_project, train_csv_folder):
+def download_completed_tasks(label_studio_project, csv_dir):
     labeled_tasks = label_studio_project.get_labeled_tasks()
     if not labeled_tasks:
         print("No new annotations")
@@ -280,11 +276,11 @@ def download_completed_tasks(label_studio_project, train_csv_folder):
         if len(label_json) == 0:
             result = {
                     "image_path": image_path,
-                    "xmin": None,
-                    "ymin": None,
-                    "xmax": None,
-                    "ymax": None,
-                    "label": None,
+                    "xmin": 0,
+                    "ymin": 0,
+                    "xmax": 0,
+                    "ymax": 0,
+                    "label": 0,
                     "annotator":labeled_task["annotations"][0]["created_username"]
                 }
             result = pd.DataFrame(result, index=[0])
@@ -302,7 +298,7 @@ def download_completed_tasks(label_studio_project, train_csv_folder):
 
     # Save csv in dir with timestamp
     timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
-    train_path = os.path.join(train_csv_folder, "train_{}.csv".format(timestamp))
+    train_path = os.path.join(csv_dir, "train_{}.csv".format(timestamp))
     annotations.to_csv(train_path, index=False)
 
     return annotations

src/pipeline.py

@@ -36,16 +36,23 @@ def run(self):
         # Check for new annotations if the check_annotations flag is set
         if self.config.check_annotations:
             new_annotations = label_studio.check_for_new_annotations(
-                **self.config.label_studio)
+                sftp_client=self.sftp_client,
+                url=self.config.label_studio.url,
+                csv_dir=self.config.label_studio.csv_dir,
+                project_name=self.config.label_studio.project_name,
+                folder_name=self.config.label_studio.folder_name,
+                images_to_annotate_dir=self.config.label_studio.images_to_annotate_dir,
+                annotated_images_dir=self.config.label_studio.annotated_images_dir,
+            )
             if new_annotations is None:
                 print("No new annotations, exiting")
                 if self.config.force_upload:
-                    image_paths = glob.glob(os.path.join(self.config.active_learning.image_dir, "*.jpg"))
+                    image_paths = glob.glob(os.path.join(self.config.label_studio.images_to_annotate_dir, "*.jpg"))
                     image_paths = random.sample(image_paths, 10)
                     label_studio.upload_to_label_studio(images=image_paths, 
                                                     sftp_client=self.sftp_client, 
                                                     label_studio_project=self.label_studio_project, 
-                                                    images_to_annotate_dir=self.config.active_learning.image_dir, 
+                                                    images_to_annotate_dir=self.config.label_studio.images_to_annotate_dir,
                                                     folder_name=self.config.label_studio.folder_name, 
                                                         preannotations=None
                                                            )

tests/conftest.py

@@ -39,10 +39,10 @@ def config(tmpdir_factory):
     # Create sample bounding box annotations
     train_data = {
         'image_path': ['empty.jpg', 'birds.jpg', "birds.jpg"],
-        'xmin': [0, 200, 150],
-        'ymin': [0, 300, 250],
-        'xmax': [20, 300, 250],
-        'ymax': [20, 400, 350],
+        'xmin': [None, 200, 150],
+        'ymin': [None, 300, 250],
+        'xmax': [None, 300, 250],
+        'ymax': [None, 400, 350],
         'label': ['Bird', 'Bird1', 'Bird2'],
         'annotator': ['test_user', 'test_user', 'test_user']
     }