Update app.py

app.py

@@ -4,9 +4,8 @@ import numpy as np
 from PIL import Image
 from io import BytesIO
 from fastapi.responses import JSONResponse
-from fastapi.middleware.cors import CORSMiddleware  # Add this import
 
-# Load your trained model (make sure it's available in the working directory)
+# Load your trained model
 model = tf.keras.models.load_model('recyclebot.keras')
 
 # Class names for predictions (modify if necessary)
@@ -15,36 +14,29 @@ class_names = ['Glass', 'Metal', 'Paperboard', 'Plastic-Polystyrene', 'Plastic-R
 # Create FastAPI app
 app = FastAPI()
 
-app.add_middleware(
-    CORSMiddleware,
-    allow_origins=["https://recyclesmart.vercel.app/"],  # or use ["*"] for all origins (not recommended for production)
-    allow_credentials=True,
-    allow_methods=["*"],
-    allow_headers=["*"],
-)
-
-# Preprocessing the image
+# Preprocessing the image (resize, reshape without normalization)
 def preprocess_image(image_file):
+    # Load image using PIL (or could use OpenCV, depending on preference)
     image = Image.open(image_file)
-    image = image.resize((240, 240))  # Resize image to match model input
-    img_array = np.array(image)  # Convert to numpy array
-    img_array = img_array.astype(np.float32) / 255.0  # Normalize
-    img_array = np.expand_dims(img_array, axis=0)  # Add batch dimension
-
-    #image = cv2.resize((cv2.imread(image_file)), (240, 240))
-
-    #img_array = np.array(image).reshape(-1, 240, 240, 3)
-               
-    return img_array
+    
+    # Convert image to numpy array
+    image = np.array(image)
+    
+    # Resize to the input shape expected by the model
+    image = cv2.resize(image, (240, 240))  # Resize image to match model input
+    
+    # Reshape the image (similar to your local code)
+    image = image.reshape(-1, 240, 240, 3)  # Add the batch dimension for inference
+    
+    return image
 
 @app.post("/predict")
 async def predict(file: UploadFile = File(...)):
     try:
-        img_array = preprocess_image(file.file)
-        predictions = model.predict(img_array)
-        predicted_class_idx = np.argmax(predictions, axis=1)[0]
-        predicted_class = class_names[predicted_class_idx]
-        #predicted_class = ((class_names[np.argmax(predictions)]))
+        img_array = preprocess_image(file.file)  # Preprocess the image
+        predictions = model.predict(img_array)  # Get predictions
+        predicted_class_idx = np.argmax(predictions, axis=1)[0]  # Get predicted class index
+        predicted_class = class_names[predicted_class_idx]  # Convert to class name
         return JSONResponse(content={"prediction": predicted_class})
     except Exception as e:
         return JSONResponse(content={"error": str(e)}, status_code=400)