Update app.py

app.py

@@ -1,4 +1,3 @@
-# Line 1
 import os
 import base64
 from io import BytesIO
@@ -68,14 +67,24 @@ def crop_image_to_food(img, yolo_results, confidence_threshold=0.2):
         detected_objects = []
         seen_labels = set()
         for result in yolo_results:
+            # Log YOLO output shapes
+            logger.info(f"YOLO output - Boxes: {len(result.boxes.xyxy)}, Classes: {len(result.boxes.cls)}, Confidences: {len(result.boxes.conf)}")
+            # Validate array lengths
+            if not (len(result.boxes.xyxy) == len(result.boxes.cls) == len(result.boxes.conf)):
+                logger.error(f"Mismatched YOLO output lengths: Boxes={len(result.boxes.xyxy)}, Classes={len(result.boxes.cls)}, Confidences={len(result.boxes.conf)}")
+                continue
             for box, cls, conf in zip(result.boxes.xyxy, result.boxes.cls, result.boxes.conf):
+                cls_idx = int(cls)
+                if cls_idx >= len(result.names):
+                    logger.error(f"Invalid class index {cls_idx}, max allowed: {len(result.names)-1}")
+                    continue
                 if conf >= confidence_threshold:
-                    label = result.names[int(cls)]
+                    label = result.names[cls_idx]
                     if label not in seen_labels:  # Avoid duplicates
                         detected_objects.append({
                             "label": label,
                             "box": box,
-                            "confidence": conf
+                            "confidence": float(conf)
                         })
                         seen_labels.add(label)
 
@@ -286,21 +295,31 @@ def calculate_nutrients(food_items, portion_size, yolo_detections):
 
     return nutrients, micronutrients, 0
 
-# FastAPI endpoint for food analysis (around line 271)
+# FastAPI endpoint for food analysis
 @app.post("/analyze_food")
 async def analyze_food(request: ImageRequest):
     try:
         img = decode_base64_image(request.image)
         yolo_results = yolo_model(img)
+        
+        # Log YOLO model details
+        logger.info(f"YOLO model classes: {len(yolo_model.names)}, Names: {yolo_model.names}")
+        
         cropped_img, was_cropped = crop_image_to_food(img, yolo_results)
         
         # Store YOLO detections for fallback
-        yolo_detections = [
-            {"label": r.names[int(cls)], "confidence": float(conf)}
-            for r in yolo_results
-            for cls, conf in zip(r.boxes.cls, r.boxes.conf)
-            if float(conf) >= 0.2
-        ]
+        yolo_detections = []
+        for r in yolo_results:
+            if len(r.boxes.cls) != len(r.boxes.conf):
+                logger.error(f"Mismatched YOLO cls and conf lengths: Classes={len(r.boxes.cls)}, Confidences={len(r.boxes.conf)}")
+                continue
+            for cls, conf in zip(r.boxes.cls, r.boxes.conf):
+                cls_idx = int(cls)
+                if cls_idx >= len(r.names):
+                    logger.error(f"Invalid class index {cls_idx}, max allowed: {len(r.names)-1}")
+                    continue
+                if float(conf) >= 0.2:
+                    yolo_detections.append({"label": r.names[cls_idx], "confidence": float(conf)})
 
         # Combine LogMeal and Clarifai results with weighted scoring
         logmeal_results = logmeal_classifier(cropped_img)
@@ -321,10 +340,15 @@ async def analyze_food(request: ImageRequest):
         food_items = {k: v for k, v in food_items.items() if v >= 0.15}
         logger.info(f"Combined food items detected: {food_items}")
 
-        non_food_items = [
-            r.names[int(cls)] for r in yolo_results
-            for cls in r.boxes.cls if r.names[int(cls)] not in food_items and float(r.boxes.conf[int(cls)]) >= 0.2
-        ]
+        non_food_items = []
+        for r in yolo_results:
+            for cls, conf in zip(r.boxes.cls, r.boxes.conf):
+                cls_idx = int(cls)
+                if cls_idx >= len(r.names):
+                    logger.error(f"Invalid class index {cls_idx} in non_food_items, max allowed: {len(r.names)-1}")
+                    continue
+                if r.names[cls_idx] not in food_items and float(conf) >= 0.2:
+                    non_food_items.append(r.names[cls_idx])
         is_non_food = len(food_items) == 0 or (len(non_food_items) > len(food_items) and max(food_items.values(), default=0) < 0.3)
         logger.info(f"Non-food items: {non_food_items}, is_non_food: {is_non_food}")
 
@@ -338,7 +362,6 @@ async def analyze_food(request: ImageRequest):
             for food, prob in food_items.items()
         ]
 
-        # Line ~275: Ensured proper dictionary syntax
         return {
             "is_non_food": is_non_food,
             "non_food_items": non_food_items,
@@ -368,6 +391,7 @@ async def gradio_analyze(image):
         request = ImageRequest(image=base64_image, portion_size=100.0)
         return await analyze_food(request)
     except Exception as e:
+        logger.error(f"Gradio analysis failed: {str(e)}")
         return {"error": str(e)}
 
 # Gradio interface