condensce code

src/active_learning.py

@@ -166,3 +166,23 @@ def update_sys_path():
         raise ValueError("Invalid strategy. Must be one of 'random', 'most-detections', or 'target-labels'.")
 
     return chosen_images
+
+def predict_and_divide(trained_detection_model, trained_classification_model, image_paths, patch_size, patch_overlap, confident_threshold):
+    predictions = detection.predict(
+        model=trained_detection_model,
+        crop_model=trained_classification_model,
+        image_paths=image_paths,
+        patch_size=patch_size,
+        patch_overlap=patch_overlap,
+    )
+    combined_predictions = pd.concat(predictions)
+
+    # Split predictions into confident and uncertain
+    uncertain_predictions = combined_predictions[
+        combined_predictions["score"] <= confident_threshold]
+
+    confident_predictions = combined_predictions[
+        ~combined_predictions["image_path"].isin(
+            uncertain_predictions["image_path"])]
+
+    return confident_predictions, uncertain_predictions

src/pipeline.py

@@ -4,12 +4,12 @@
 import pandas as pd
 from omegaconf import DictConfig
 
-from src.active_learning import choose_train_images, choose_test_images
+from src.active_learning import choose_train_images, choose_test_images, predict_and_divide
 from src import propagate
 from src import label_studio
 from src.classification import preprocess_and_train_classification
 from src.data_processing import density_cropping
-from src.detection import preprocess_and_train, predict
+from src.detection import preprocess_and_train
 from src.pipeline_evaluation import PipelineEvaluation
 from src.reporting import Reporting
 
@@ -77,32 +77,35 @@ def run(self):
             return None
         else:
             train_images_to_annotate = choose_train_images(
-                performance, trained_detection_model,
-                **self.config.active_learning)
-            test_images_to_annotate = choose_test_images(
-                performance, **self.config.active_testing)
-
-            predictions = predict(
-                m=trained_detection_model,
-                crop_model=trained_classification_model,
-                image_paths=train_images_to_annotate,
+                evaluation=performance,
+                image_dir=self.config.active_learning.image_dir,
+                model=trained_detection_model,
+                strategy=self.config.active_learning.strategy,
+                n=self.config.active_learning.n_images,
                 patch_size=self.config.active_learning.patch_size,
                 patch_overlap=self.config.active_learning.patch_overlap,
+                min_score=self.config.active_learning.min_score
             )
-            combined_predictions = pd.concat(predictions)
-
-            # Split predictions into confident and uncertain
-            uncertain_predictions = combined_predictions[
-                combined_predictions["score"] <=
-                self.config.active_learning.confident_threshold]
 
-            confident_predictions = combined_predictions[
-                ~combined_predictions["image_path"].isin(
-                    uncertain_predictions["image_path"])]
+            test_images_to_annotate = choose_test_images(
+                image_dir=self.config.active_testing.image_dir,
+                model=trained_detection_model,
+                strategy=self.config.active_testing.strategy,
+                n=self.config.active_testing.n_images,
+                patch_size=self.config.active_testing.patch_size,
+                patch_overlap=self.config.active_testing.patch_overlap,
+                min_score=self.config.active_testing.min_score)
+
+
+            confident_predictions, uncertain_predictions = predict_and_divide(
+                trained_detection_model, trained_classification_model,
+                train_images_to_annotate, self.config.active_learning.patch_size,
+                self.config.active_learning.patch_overlap,
+                self.config.active_learning.confident_threshold)
 
             reporter.confident_predictions = confident_predictions
             reporter.uncertain_predictions = uncertain_predictions
-            
+
             print(f"Images requiring human review: {len(confident_predictions)}")
             print(f"Images auto-annotated: {len(uncertain_predictions)}")
 

tests/test_active_learning.py

@@ -36,5 +36,5 @@ def test_choose_test_images(detection_model, config):
         n=config.active_testing.n_images,
         patch_size=config.active_testing.patch_size,
         patch_overlap=config.active_testing.patch_overlap,
-        min_score=config.active_testing.min_score        )
+        min_score=config.active_testing.min_score)
     assert len(test_images_to_annotate) > 0