avoid duplicate computation during classification

CONTRIBUTING.md

@@ -42,7 +42,7 @@ git push origin feature/my-feature
 
 Before submitting a contribution:
 
-1. **Follow the code style guidelines** - JABS uses Ruff for linting and formatting
+1. **Follow the code style guidelines** - JABS uses Ruff for linting and formatting, and follows the Google docstring style.
 2. **Add tests** for new functionality
 3. **Update documentation** as needed (docstrings, user guide, developer guide)
 4. **Run the test suite** to ensure nothing is broken: `pytest`

src/jabs/scripts/classify.py

@@ -137,13 +137,14 @@ def classify_pose(
             data = Classifier.combine_data(per_frame_features, window_features)
 
             if data.shape[0] > 0:
-                pred = classifier.predict(data, features["frame_indexes"])
+                # Get probabilities for all classes
                 pred_prob = classifier.predict_proba(data, features["frame_indexes"])
 
+                # Derive predictions by taking argmax (class with highest probability)
+                # This is equivalent to predict() but avoids duplicate computation
+                pred = np.argmax(pred_prob, axis=1).astype(np.int8)
+
                 # Keep the probability for the predicted class only.
-                # The following code uses some
-                # numpy magic to use the pred array as column indexes
-                # for each row of the pred_prob array we just computed.
                 pred_prob = pred_prob[np.arange(len(pred_prob)), pred]
 
                 # Copy results into results matrix

src/jabs/ui/classification_thread.py

@@ -133,19 +133,18 @@ def check_termination_requested() -> None:
 
                     check_termination_requested()
                     if data.shape[0] > 0:
-                        # make predictions
-                        predictions[identity] = self._classifier.predict(
-                            data, feature_values["frame_indexes"]
-                        )
-
-                        # also get the probabilities
+                        # Get probabilities for all classes
                         prob = self._classifier.predict_proba(
                             data, feature_values["frame_indexes"]
                         )
-                        # Save the probability for the predicted class only.
-                        # The following code uses some
-                        # numpy magic to use the _predictions array as column indexes
-                        # for each row of the 'prob' array we just computed.
+
+                        # Derive predictions by taking argmax (class with highest probability)
+                        # This is equivalent to predict() but avoids duplicate computation
+                        # Consider using np.where with a threshold here if we want to be more conservative
+                        # about predictions
+                        predictions[identity] = np.argmax(prob, axis=1).astype(np.int8)
+
+                        # Use predictions as column indexes for each row of prob
                         probabilities[identity] = prob[np.arange(len(prob)), predictions[identity]]
                     else:
                         predictions[identity] = np.array(0)