Update app.py

app.py

@@ -112,66 +112,54 @@ def create_monitor_interface():
                 """Get coordinates based on position description."""
                 # Basic regions
                 regions = {
+                    'center': (width//3, height//3, 2*width//3, 2*height//3),
+                    'background': (0, 0, width, height),
                     'top-left': (0, 0, width//3, height//3),
                     'top': (width//3, 0, 2*width//3, height//3),
                     'top-right': (2*width//3, 0, width, height//3),
-                    'center-left': (0, height//3, width//3, 2*height//3),
-                    'center': (width//3, height//3, 2*width//3, 2*height//3),
-                    'center-right': (2*width//3, height//3, width, 2*height//3),
+                    'left': (0, height//3, width//3, 2*height//3),
+                    'right': (2*width//3, height//3, width, 2*height//3),
                     'bottom-left': (0, 2*height//3, width//3, height),
                     'bottom': (width//3, 2*height//3, 2*width//3, height),
                     'bottom-right': (2*width//3, 2*height//3, width, height),
-                    'left': (0, height//4, width//3, 3*height//4),
-                    'right': (2*width//3, height//4, width, 3*height//4)
+                    'ground': (0, 2*height//3, width, height)
                 }
                 
                 # Find best matching region
-                best_match = 'center'
-                max_words = 0
-                pos_lower = position.lower()
+                position = position.lower()
+                for key in regions.keys():
+                    if key in position:
+                        return regions[key]
                 
-                for region in regions.keys():
-                    words = region.split('-')
-                    matches = sum(1 for word in words if word in pos_lower)
-                    if matches > max_words:
-                        max_words = matches
-                        best_match = region
-                
-                return regions[best_match]
+                return regions['center']  # Default to center if no match
             
             for idx, obs in enumerate(observations):
                 color = self.colors[idx % len(self.colors)]
                 
-                # Parse location and description
-                parts = obs.split(':')
-                if len(parts) >= 2:
-                    position = parts[0]
-                    description = ':'.join(parts[1:])
-                    
-                    # Get region coordinates
-                    x1, y1, x2, y2 = get_region_coordinates(position)
-                    
-                    # 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)
-                    
-                    # Position text above the box
-                    text_x = max(0, x1)
-                    text_y = max(label_size[1] + padding, y1 - padding)
-                    
-                    # Draw text background
-                    cv2.rectangle(image, 
-                                 (text_x, text_y - label_size[1] - padding),
-                                 (text_x + label_size[0] + padding, text_y),
-                                 color, -1)
-                    
-                    # Draw text
-                    cv2.putText(image, label,
-                                (text_x + padding//2, text_y - padding//2),
-                                font, font_scale, (255, 255, 255), thickness)
+                # Get coordinates for this observation
+                x1, y1, x2, y2 = get_region_coordinates(obs['location'])
+                
+                # Draw rectangle
+                cv2.rectangle(image, (x1, y1), (x2, y2), color, 2)
+                
+                # Add label with background
+                label = obs['description'][:50] + "..." if len(obs['description']) > 50 else obs['description']
+                label_size, _ = cv2.getTextSize(label, font, font_scale, thickness)
+                
+                # Position text above the box
+                text_x = max(0, x1)
+                text_y = max(label_size[1] + padding, y1 - padding)
+                
+                # Draw text background
+                cv2.rectangle(image, 
+                             (text_x, text_y - label_size[1] - padding),
+                             (text_x + label_size[0] + padding, text_y),
+                             color, -1)
+                
+                # Draw text
+                cv2.putText(image, label,
+                            (text_x + padding//2, text_y - padding//2),
+                            font, font_scale, (255, 255, 255), thickness)
             
             return image
 
@@ -180,27 +168,34 @@ def create_monitor_interface():
                 return None, "No image provided"
                 
             analysis = self.analyze_frame(frame)
-            display_frame = self.resize_image(frame.copy())
+            display_frame = frame.copy()
             
-            # Parse observations from the analysis
+            # Parse observations from the formatted response
             observations = []
-            for line in analysis.split('\n'):
-                line = line.strip()
-                if line.startswith('-'):
-                    # Extract text between <location> tags if present
-                    if '<location>' in line and '</location>' in line:
-                        start = line.find('<location>') + len('<location>')
-                        end = line.find('</location>')
-                        observation = line[end + len('</location>'):].strip()
-                    else:
-                        observation = line[1:].strip()  # Remove the dash
-                    if observation:
-                        observations.append(observation)
+            lines = analysis.split('\n')
+            for line in lines:
+                # Look for location tags in the line
+                if '<location>' in line and '</location>' in line:
+                    start = line.find('<location>') + len('<location>')
+                    end = line.find('</location>')
+                    location = line[start:end].strip()
+                    
+                    # Get the description that follows the location tag
+                    desc_start = line.find('</location>') + len('</location>:')
+                    description = line[desc_start:].strip()
+                    
+                    if location and description:
+                        observations.append({
+                            'location': location,
+                            'description': description
+                        })
             
-            # Draw observations on the image
-            annotated_frame = self.draw_observations(display_frame, observations)
+            # Draw observations if we found any
+            if observations:
+                annotated_frame = self.draw_observations(display_frame, observations)
+                return annotated_frame, analysis
             
-            return annotated_frame, analysis
+            return display_frame, analysis
 
     # Create the main interface
     monitor = SafetyMonitor()