Update app.py

app.py

@@ -27,8 +27,8 @@ SAVE_DIR = Path("/tmp/emotion_results")
 SAVE_DIR.mkdir(exist_ok=True)
 
 # Create directories
-(SAVE_DIR / "faces").mkdir(exist_ok=True)  # For raw face crops
-(SAVE_DIR / "annotated").mkdir(exist_ok=True)  # For annotated images
+(SAVE_DIR / "faces").mkdir(exist_ok=True)
+(SAVE_DIR / "annotated").mkdir(exist_ok=True)
 for emotion in EMOTION_MAP.keys():
     (SAVE_DIR / "faces" / emotion).mkdir(exist_ok=True, parents=True)
     (SAVE_DIR / "annotated" / emotion).mkdir(exist_ok=True, parents=True)
@@ -56,7 +56,7 @@ def predict_emotion(batch_no: str, image):
     try:
         # Convert PIL Image to OpenCV format
         frame = np.array(image)
-        if frame.ndim == 3:  # Color image
+        if frame.ndim == 3:
             frame = cv2.cvtColor(frame, cv2.COLOR_RGB2BGR)
         
         # Try different backends for face detection
@@ -86,14 +86,14 @@ def predict_emotion(batch_no: str, image):
         # Extract face coordinates
         x, y, w, h = region['x'], region['y'], region['w'], region['h']
         
-        # 1. Save raw face crop (for training)
+        # 1. Save raw face crop
         face_crop = frame[y:y+h, x:x+w]
         timestamp = int(time.time())
         face_dir = SAVE_DIR / "faces" / emotion
         face_path = face_dir / f"{batch_no}_{timestamp}.jpg"
         cv2.imwrite(str(face_path), face_crop)
         
-        # 2. Create and save annotated image (for display)
+        # 2. Create and save annotated image
         annotated_frame = frame.copy()
         cv2.rectangle(annotated_frame, (x, y), (x+w, y+h), (0, 255, 0), 2)
         cv2.putText(annotated_frame, f"{emotion} {EMOTION_MAP[emotion]} {confidence:.1f}%", 
@@ -173,11 +173,7 @@ def update_gallery(emotion, image_type):
     for emotion, images in image_dict.items():
         for img_path in images:
             gallery.append((img_path, f"{emotion}: {Path(img_path).name}"))
-    return gr.Gallery(value=gallery, label="Image Gallery"), gr.CheckboxGroup(
-        choices=[img[0] for img in gallery], 
-        label="Selected Images",
-        value=[]
-    )
+    return gallery, [img[0] for img in gallery]
 
 def get_logs():
     if LOG_FILE.exists():
@@ -264,12 +260,10 @@ with gr.Blocks(title="Emotion Capture", css="""
     )
 
 # Data Management Interface
-with gr.Blocks(title="Data Management", css="""
-    .gradio-container { max-width: 1200px !important }
-    .data-section { border: 1px solid #ccc; padding: 20px; border-radius: 5px; margin-top: 20px; }
-    .gallery { grid-template-columns: repeat(auto-fill, minmax(200px, 1fr)); }
-    .action-buttons { margin-top: 20px; }
-""") as data_interface:
+with gr.Blocks(title="Data Management") as data_interface:
+    
+    # Initialize gallery data
+    initial_gallery, initial_selection = update_gallery("All Emotions", "faces")
     
     gr.Markdown("# Data Management Interface")
     
@@ -290,12 +284,13 @@ with gr.Blocks(title="Data Management", css="""
                 refresh_btn = gr.Button("Refresh Gallery")
             
             gallery = gr.Gallery(
+                value=initial_gallery,
                 label="Image Gallery",
-                elem_classes="gallery",
                 columns=4
             )
             selected_images = gr.CheckboxGroup(
                 label="Selected Images",
+                choices=initial_selection,
                 interactive=True
             )
             
@@ -339,7 +334,7 @@ with gr.Blocks(title="Data Management", css="""
             )
             
             download_all_btn.click(
-                lambda e, t: create_custom_zip([img[0] for img in get_image_gallery(e, t).items() for img in img[1]]),
+                lambda e, t: create_custom_zip([img for img_list in get_image_gallery(e, t).values() for img in img_list]),
                 inputs=[emotion_selector, image_type_selector],
                 outputs=download_output
             )
@@ -392,26 +387,11 @@ with gr.Blocks(title="Data Management", css="""
                 lambda: update_gallery("All Emotions", "faces"),
                 outputs=[gallery, selected_images]
             )
-            
-            # Initial load of logs
-            data_interface.load(
-                view_logs,
-                outputs=logs_display
-            )
 
 # Combine interfaces
 demo = gr.TabbedInterface(
     [capture_interface, data_interface],
-    ["Emotion Capture", "Data Management"],
-    css="""
-    .tab { padding: 20px; }
-    """
-)
-
-# Initialize gallery and logs
-demo.load(
-    lambda: update_gallery("All Emotions", "faces"),
-    outputs=[gallery, selected_images]
+    ["Emotion Capture", "Data Management"]
 )
 
 if __name__ == "__main__":