Update app.py

app.py

@@ -9,6 +9,7 @@ import cv2
 import constants as const
 from get_drive_model import ensure_model_download
 import atexit  # Import atexit module
+
 # Define the path for the directory
 FRAMES_FOLDER = 'static/frames'
 
@@ -26,11 +27,6 @@ second_model = tf.keras.models.load_model('anomaly_classification_model.h5', cus
 UPLOAD_FOLDER = 'uploads'
 FRAMES_FOLDER = 'static/frames'
 
-# Preloaded video paths for different types of detection
-VIOLENT_DETECTION_VIDEO = 'uploads/violent_detection_video.mp4'
-EXPLOSION_DETECTION_VIDEO = 'uploads/explosion_detection_video.mp4'
-NORMAL_DETECTION_VIDEO = 'uploads/normal_detection_video.mp4'
-
 def process_video(filepath):
     print(f"Processing video: {filepath}")
 
@@ -58,12 +54,6 @@ def process_video(filepath):
     frame_paths = save_frames_to_filesystem(frames_for_display)
     return prediction_label, frame_paths
 
-def read_video(filepath):
-    frames_for_prediction, frames_for_display = preprocess(filepath)
-    print(frames_for_prediction.shape)
-    print(frames_for_display.shape)
-    return frames_for_prediction, frames_for_display
-
 def save_frames_to_filesystem(frames):
     frame_paths = []
     for i, frame in enumerate(frames):
@@ -75,18 +65,9 @@ def save_frames_to_filesystem(frames):
     return frame_paths
 
 def cleanup_uploads_folder():
-    # List of files to keep (make exceptions for these files)
-    exceptions = {
-        'violent_detection_video.mp4',
-        'explosion_detection_video.mp4',
-        'normal_detection_video.mp4'
-    }
-
     # Clean up the UPLOAD_FOLDER
     if os.path.exists(UPLOAD_FOLDER):
         for filename in os.listdir(UPLOAD_FOLDER):
-            if filename in exceptions:
-                continue  # Skip the files that are in the exceptions list
             file_path = os.path.join(UPLOAD_FOLDER, filename)
             try:
                 if os.path.isfile(file_path) or os.path.islink(file_path):
@@ -114,26 +95,16 @@ def cleanup_uploads_folder():
 # Register the cleanup function
 atexit.register(cleanup_uploads_folder)
 
-# Define functions to trigger each type of detection
-def detect_violent():
-    return process_video(VIOLENT_DETECTION_VIDEO)
-
-def detect_explosion():
-    return process_video(EXPLOSION_DETECTION_VIDEO)
-
-def detect_normal():
-    return process_video(NORMAL_DETECTION_VIDEO)
-
-# Create Gradio Interface with buttons for specific detections
+# Create Gradio Interface
 iface = gr.Interface(
     fn=process_video,
-    inputs=gr.File(type="filepath"),
+    inputs=gr.File(type="filepath"),  # Add file upload for all detections
     outputs=[
         gr.Textbox(label="Prediction", elem_id="prediction-box"),
         gr.Gallery(label="Video Frames", elem_id="frame-gallery", columns=5, rows=10)
     ],
     title="Anomaly Detection in Videos",
-    description="Upload a video file or use the preloaded one to detect anomalies.",
+    description="Upload a video file and detect anomalies, violent activity, or explosions.",
     theme="default",
     css="""
         body {
@@ -176,10 +147,11 @@ iface = gr.Interface(
     """
 )
 
-# Creating additional interfaces for button-based triggers
+# Additional interfaces for individual triggers (with file upload option)
+
 violent_iface = gr.Interface(
-    fn=detect_violent,
-    inputs=None,
+    fn=process_video,  # Use the same process_video for Violent Detection
+    inputs=gr.File(type="filepath"),  # File upload for violent detection
     outputs=[
         gr.Textbox(label="Prediction"),
         gr.Gallery(label="Video Frames", columns=5, rows=10)
@@ -188,8 +160,8 @@ violent_iface = gr.Interface(
 )
 
 explosion_iface = gr.Interface(
-    fn=detect_explosion,
-    inputs=None,
+    fn=process_video,  # Use the same process_video for Explosion Detection
+    inputs=gr.File(type="filepath"),  # File upload for explosion detection
     outputs=[
         gr.Textbox(label="Prediction"),
         gr.Gallery(label="Video Frames", columns=5, rows=10)
@@ -198,8 +170,8 @@ explosion_iface = gr.Interface(
 )
 
 normal_iface = gr.Interface(
-    fn=detect_normal,
-    inputs=None,
+    fn=process_video,  # Use the same process_video for Normal Detection
+    inputs=gr.File(type="filepath"),  # File upload for normal detection
     outputs=[
         gr.Textbox(label="Prediction"),
         gr.Gallery(label="Video Frames", columns=5, rows=10)
@@ -213,3 +185,5 @@ combined_iface = gr.TabbedInterface([iface, violent_iface, explosion_iface, norm
 
 if __name__ == "__main__":
     combined_iface.launch()
+
+