Create app.py

app.py

@@ -0,0 +1,47 @@
+from fastapi import FastAPI, UploadFile, File
+from keras.models import load_model
+from PIL import Image
+import numpy as np
+import cv2
+
+app = FastAPI()
+
+# Load the model
+model = load_model('model1.h5')
+
+# Define the target size for the images
+target_size = (256, 256)
+
+# Define a function to preprocess the uploaded image
+def preprocess_image(image):
+    # Convert the image to RGB
+    image = image.convert('RGB')
+    # Resize the image
+    img = np.array(image)
+    resized_img = cv2.resize(img, target_size)
+    resized_img = resized_img / 255.0
+    return np.expand_dims(resized_img, axis=0)
+
+# Define the class names (adjust based on your model's classes)
+class_names = ['Blight', 'Healthy', 'Gray Leaf Spot', 'Common Rust']
+
+@app.post("/predict/")
+async def predict(file: UploadFile = File(...)):
+    # Read the image
+    image = Image.open(file.file)
+    
+    # Preprocess the image
+    processed_image = preprocess_image(image)
+    
+    # Make predictions
+    predictions = model.predict(processed_image)[0]
+    
+    # Interpret the predictions
+    predicted_class = class_names[np.argmax(predictions)]
+    confidence = np.max(predictions)
+    
+    return {
+        "prediction": predicted_class,
+        "confidence": confidence,
+        "raw_predictions": {class_name: float(predictions[i]) for i, class_name in enumerate(class_names)}
+    }