Update app.py

app.py

@@ -27,6 +27,8 @@ os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
 import tensorflow as tf
 tf.get_logger().setLevel('ERROR')
 
+@spaces.GPU(duration=500)
+
 # Initialize models and other global variables
 device = 'cuda' if torch.cuda.is_available() else 'cpu'
 
@@ -92,6 +94,12 @@ def extract_frames(video_path, output_folder, fps):
         '-vf', f'fps={fps}',
         f'{output_folder}/frame_%04d.jpg'
     ]
+    
+    print(f"FFmpeg command: {command}")
+    
+    if any(arg is None for arg in command):
+        raise ValueError(f"None value found in FFmpeg command: {command}")
+    
     try:
         result = subprocess.run(command, check=True, capture_output=True, text=True)
         print(f"FFmpeg stdout: {result.stdout}")
@@ -294,7 +302,7 @@ def lstm_anomaly_detection(X, feature_columns, num_anomalies=10, epochs=100, bat
     else:
         mse_comp = mse_all  # If no components, use all features
 
-    top_indices_comp = mse_comp.argsort()[-num_anomalies:][::-1]
+top_indices_comp = mse_comp.argsort()[-num_anomalies:][::-1]
     anomalies_comp = np.zeros(len(mse_comp), dtype=bool)
     anomalies_comp[top_indices_comp] = True
 
@@ -332,8 +340,6 @@ def plot_emotion(df, emotion, num_anomalies, color):
     plt.tight_layout()
     return fig
 
-import base64
-
 def get_random_face_sample(organized_faces_folder, largest_cluster, output_folder):
     person_folder = os.path.join(organized_faces_folder, f"person_{largest_cluster}")
     face_files = [f for f in os.listdir(person_folder) if f.endswith('.jpg')]
@@ -344,15 +350,22 @@ def get_random_face_sample(organized_faces_folder, largest_cluster, output_folde
         
         # Read the image and resize it to be smaller
         face_img = cv2.imread(face_path)
-        small_face = cv2.resize(face_img, (100, 100))  # Resize to NxN pixels
+        small_face = cv2.resize(face_img, (100, 100))  # Resize to 100x100 pixels
         cv2.imwrite(output_path, small_face)
         
         return output_path
     return None
 
-@spaces.GPU(duration=300)  # Set to 5 minutes (300 seconds)
-
 def process_video(video_path, num_anomalies, num_components, desired_fps, batch_size, progress=gr.Progress()):
+    print(f"Video path: {video_path}")
+    print(f"Num anomalies: {num_anomalies}")
+    print(f"Num components: {num_components}")
+    print(f"Desired FPS: {desired_fps}")
+    print(f"Batch size: {batch_size}")
+    
+    if not os.path.exists(video_path):
+        raise FileNotFoundError(f"Video file not found: {video_path}")
+    
     output_folder = "output"
     os.makedirs(output_folder, exist_ok=True)
     
@@ -393,7 +406,7 @@ def process_video(video_path, num_anomalies, num_components, desired_fps, batch_
         try:
             anomalies_all, anomaly_scores_all, top_indices_all, anomalies_comp, anomaly_scores_comp, top_indices_comp, _ = lstm_anomaly_detection(X, feature_columns, num_anomalies=num_anomalies, batch_size=batch_size)
         except Exception as e:
-            print(f"Error details: {str(e)}")
+            print(f"Error in anomaly detection: {str(e)}")
             print(f"X shape: {X.shape}")
             print(f"X dtype: {X.dtype}")
             return f"Error in anomaly detection: {str(e)}", None, None, None, None, None, None
@@ -413,7 +426,6 @@ def process_video(video_path, num_anomalies, num_components, desired_fps, batch_
         except Exception as e:
             return f"Error generating plots: {str(e)}", None, None, None, None, None, None, None, None, None
 
-
         # Get a random face sample
         face_sample = get_random_face_sample(organized_faces_folder, largest_cluster, output_folder)
         
@@ -431,7 +443,7 @@ def process_video(video_path, num_anomalies, num_components, desired_fps, batch_
             results += "\n".join([f"{df[emotion].iloc[i]:.4f} at {df['Timecode'].iloc[i]}" for i in top_indices])
     
         return results, face_sample, anomaly_plot_all, anomaly_plot_comp, *emotion_plots
-        
+
 # Gradio interface
 iface = gr.Interface(
     fn=process_video,