404 solution

app.py

@@ -1,41 +1,45 @@
+import os
 from fastapi import FastAPI, File, UploadFile
 from PIL import Image
-import torch
 import io
-import os
 
+# --- NEW CODE: Import YOLO directly ---
+from ultralytics import YOLO
+
+# Set the cache directory for PyTorch Hub to a writable location
+# (This line is still good practice, so we'll keep it)
 os.environ['TORCH_HOME'] = '/tmp/torch_cache'
-app = FastAPI(title="Car Damage Detection API")
 
-# Load custom YOLOv8 model.
-model = torch.hub.load('ultralytics/yolov8', 'custom', path='best.pt')
+# Initialize the FastAPI app
+app = FastAPI(title="YOLOv8 Car Damage Detection API")
+
+# --- MODIFIED LINE: Load the model directly ---
+# This is the new, recommended way to load a local model
+model = YOLO('best.pt')
 
 
 @app.get("/")
 def read_root():
-    """ A simple endpoint to check if the server is running. """
-    return {"status": "ok", "message": "Car Damage Detection API is running!"}
+    return {"message": "Welcome to the Car Damage Detection API!"}
 
 
 @app.post("/detect")
 async def detect_damage(file: UploadFile = File(...)):
-    """
-    This endpoint receives an image, runs inference, and returns detections.
-    """
-    # Read image content from the uploaded file
+    # ... (the rest of your code remains the same)
     contents = await file.read()
     image = Image.open(io.BytesIO(contents)).convert("RGB")
-
-    # Run the model on the image
     results = model(image)
 
-    # Extract results into a clean format
-    predictions_df = results.pandas().xyxy[0]
-
-    # Filter detections by a confidence threshold (e.g., > 50%)
-    detections = predictions_df[predictions_df['confidence'] > 0.5]
-
-    # Convert the results to a JSON-friendly list of dictionaries
-    output = detections[['name', 'confidence']].to_dict(orient='records')
+    # --- IMPORTANT CHANGE FOR NEW METHOD ---
+    # The output format is slightly different. We need to access the results differently.
+    output = []
+    for result in results:
+        boxes = result.boxes
+        for box in boxes:
+            class_id = int(box.cls[0])
+            class_name = model.names[class_id]
+            confidence = float(box.conf[0])
+            if confidence > 0.5:
+                output.append({'name': class_name, 'confidence': confidence})
 
     return {"detections": output}