Update app.py

app.py

@@ -1,28 +1,20 @@
 import cv2
 import random
-import time
 from ultralytics import YOLO
 from gtts import gTTS
-import pygame
-import threading
 from datetime import datetime, timedelta
 import gradio as gr
 
-# Initialize pygame mixer
-pygame.mixer.quit()
-pygame.mixer.init()
-
 # Load YOLOv8 model
 yolo = YOLO("yolov8n.pt")
 
 # Audio alert settings
-audio_temp_dir = "audio_temp_files"
 alert_categories = {"person", "cat", "dog", "knife", "fire", "gun"}
 last_alert_time = {}
 alert_cooldown = timedelta(seconds=10)
 
-# Create audio alert
-def play_audio_alert(label, position):
+# Create audio alert as downloadable file
+def generate_audio_alert(label, position):
     phrases = [
         f"Be careful, there's a {label} on your {position}.",
         f"Watch out! {label} detected on your {position}.",
@@ -30,22 +22,11 @@ def play_audio_alert(label, position):
     ]
     caution_note = random.choice(phrases)
 
-    temp_file_path = f"{audio_temp_dir}/temp_{datetime.now().strftime('%Y%m%d_%H%M%S_%f')}.mp3"
+    # Save audio alert as an MP3 file
+    temp_file_path = f"audio_alert_{datetime.now().strftime('%Y%m%d_%H%M%S')}.mp3"
     tts = gTTS(caution_note)
     tts.save(temp_file_path)
-
-    try:
-        pygame.mixer.music.load(temp_file_path)
-        pygame.mixer.music.play()
-
-        def cleanup_audio_file():
-            while pygame.mixer.music.get_busy():
-                time.sleep(0.1)
-            pygame.mixer.music.stop()
-
-        threading.Thread(target=cleanup_audio_file, daemon=True).start()
-    except Exception as e:
-        print(f"Error playing audio alert: {e}")
+    return temp_file_path
 
 # Process a single frame
 def process_frame(image, enable_audio):
@@ -53,6 +34,8 @@ def process_frame(image, enable_audio):
     result = results[0]
 
     detected_objects = {}
+    audio_files = []
+
     for box in result.boxes:
         x1, y1, x2, y2 = map(int, box.xyxy[0])
         label = result.names[int(box.cls[0])]
@@ -69,14 +52,15 @@ def process_frame(image, enable_audio):
                 label not in last_alert_time
                 or current_time - last_alert_time[label] > alert_cooldown
             ):
-                play_audio_alert(label, position)
+                audio_file = generate_audio_alert(label, position)
+                audio_files.append(audio_file)
                 last_alert_time[label] = current_time
 
         # Draw bounding boxes
         cv2.rectangle(image, (x1, y1), (x2, y2), (0, 255, 0), 2)
         cv2.putText(image, label, (x1, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)
 
-    return image
+    return image, audio_files
 
 # Gradio interface function
 def object_detection_webcam(enable_audio):
@@ -90,17 +74,21 @@ def object_detection_webcam(enable_audio):
             return "Error: Unable to read from camera."
 
         frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
-        processed_frame = process_frame(frame, enable_audio)
-        yield processed_frame
+        processed_frame, audio_files = process_frame(frame, enable_audio)
+
+        yield {"image": processed_frame, "audio": audio_files}
 
     cap.release()
 
-# Create Gradio interface
+# Gradio UI
 def gradio_app():
     return gr.Interface(
         fn=object_detection_webcam,
         inputs=[gr.Checkbox(label="Enable Audio Alerts", value=False)],
-        outputs="image",
+        outputs=[
+            gr.Image(label="Processed Frame"),
+            gr.File(label="Audio Alerts"),
+        ],
         live=True,
     )