Update app.py

app.py

@@ -16,16 +16,15 @@ def create_monitor_interface():
         def __init__(self):
             self.client = Groq()
             self.model_name = "llama-3.2-90b-vision-preview"
-            self.max_image_size = (800, 800)  # Increased size for better quality
+            self.max_image_size = (800, 800)
             self.colors = [(0, 0, 255), (255, 0, 0), (0, 255, 0), (255, 255, 0), (255, 0, 255)]
             self.last_analysis_time = 0
-            self.analysis_interval = 2  # Analyze every 2 seconds
-            self.last_observations = []  # Store previous observations
+            self.analysis_interval = 2
+            self.last_observations = []
             
         def resize_image(self, image):
             height, width = image.shape[:2]
             
-            # Only resize if image is too large
             if height > self.max_image_size[1] or width > self.max_image_size[0]:
                 aspect = width / height
                 if width > height:
@@ -50,11 +49,10 @@ def create_monitor_interface():
             frame = self.resize_image(frame)
             frame_pil = PILImage.fromarray(frame)
             
-            # Convert to base64 with better quality
             buffered = io.BytesIO()
             frame_pil.save(buffered, 
                          format="JPEG", 
-                         quality=85,  # Higher quality
+                         quality=85,
                          optimize=True)
             img_base64 = base64.b64encode(buffered.getvalue()).decode('utf-8')
             image_url = f"data:image/jpeg;base64,{img_base64}"
@@ -111,12 +109,10 @@ def create_monitor_interface():
                 'bottom-right': (2*width//3, 2*height//3, width, height)
             }
             
-            # Find the best matching region
             for region_name, coords in regions.items():
                 if region_name in position.lower():
                     return coords
             
-            # Default to center if no match
             return regions['center']
 
         def draw_observations(self, image, observations):
@@ -128,7 +124,6 @@ def create_monitor_interface():
             for idx, obs in enumerate(observations):
                 color = self.colors[idx % len(self.colors)]
                 
-                # Try to extract position from observation
                 parts = obs.split(':')
                 if len(parts) >= 2:
                     position = parts[0]
@@ -137,17 +132,13 @@ def create_monitor_interface():
                     position = 'center'
                     description = obs
                 
-                # Get coordinates based on position
                 x1, y1, x2, y2 = self.get_region_coordinates(position, image.shape)
                 
-                # Draw rectangle
                 cv2.rectangle(image, (x1, y1), (x2, y2), color, 2)
                 
-                # Add label with background
                 label = description[:50] + "..." if len(description) > 50 else description
                 label_size = cv2.getTextSize(label, font, font_scale, thickness)[0]
                 
-                # Ensure label stays within image bounds
                 label_x = max(0, min(x1, width - label_size[0]))
                 label_y = max(20, y1 - 5)
                 
@@ -164,12 +155,10 @@ def create_monitor_interface():
             
             current_time = time.time()
             
-            # Only perform analysis if enough time has passed
             if current_time - self.last_analysis_time >= self.analysis_interval:
                 analysis = self.analyze_frame(frame)
                 self.last_analysis_time = current_time
                 
-                # Parse observations
                 observations = []
                 for line in analysis.split('\n'):
                     line = line.strip()
@@ -183,7 +172,6 @@ def create_monitor_interface():
                 
                 self.last_observations = observations
             
-            # Draw observations on the frame
             display_frame = frame.copy()
             annotated_frame = self.draw_observations(display_frame, self.last_observations)
             
@@ -196,12 +184,12 @@ def create_monitor_interface():
         gr.Markdown("# Safety Analysis System powered by Llama 3.2 90b vision")
         
         with gr.Row():
-            webcam = gr.Image(source="webcam", streaming=True, label="Live Feed")
+            input_image = gr.Image(label="Upload Image")
             output_image = gr.Image(label="Analysis")
         
         analysis_text = gr.Textbox(label="Safety Concerns", lines=5)
             
-        def analyze_stream(image):
+        def analyze_image(image):
             if image is None:
                 return None, "No image provided"
             try:
@@ -211,12 +199,19 @@ def create_monitor_interface():
                 print(f"Processing error: {str(e)}")
                 return None, f"Error processing image: {str(e)}"
             
-        webcam.stream(
-            fn=analyze_stream,
-            outputs=[output_image, analysis_text],
-            show_progress=False
+        input_image.change(
+            fn=analyze_image,
+            inputs=input_image,
+            outputs=[output_image, analysis_text]
         )
 
+        gr.Markdown("""
+        ## Instructions:
+        1. Upload an image to analyze safety concerns
+        2. View annotated results and detailed analysis
+        3. Each box highlights a potential safety issue
+        """)
+
     return demo
 
 demo = create_monitor_interface()