inference.py

import tensorflow as tf
import numpy as np
from PIL import Image
import os

# ========================
# Custom layer (مطلوبة)
# ========================
from tensorflow.keras.layers import Layer

class CustomScaleLayer(Layer):
    def __init__(self, scale=1.0, **kwargs):
        super(CustomScaleLayer, self).__init__(**kwargs)
        self.scale = scale

    def call(self, inputs):
        if isinstance(inputs, (list, tuple)):
            x = tf.add_n(inputs)
        else:
            x = inputs
        return x * self.scale

    def get_config(self):
        config = super().get_config()
        config.update({"scale": self.scale})
        return config

# ========================
# Load the model
# ========================
def load_model(model_path):
    try:
        model = tf.keras.models.load_model(model_path, custom_objects={'CustomScaleLayer': CustomScaleLayer})
        print("Model loaded successfully!")
        return model
    except Exception as e:
        print(f"Error loading model: {e}")
        return None

# ========================
# Image preprocessing
# ========================
def preprocess_image(image_path, target_size=(299, 299), normalize=True):
    try:
        img = Image.open(image_path)
        if img.mode != 'RGB':
            img = img.convert('RGB')
        img = img.resize(target_size)
        img_array = np.array(img)
        if normalize:
            img_array = img_array / 255.0
        img_array = np.expand_dims(img_array, axis=0)
        return img_array
    except Exception as e:
        print(f"Error preprocessing image: {e}")
        return None

# ========================
# Prediction
# ========================
def predict_disease(model, image_array):
    try:
        prediction = model.predict(image_array)
        predicted_class = np.argmax(prediction[0])
        confidence = prediction[0][predicted_class]
        return predicted_class, confidence, prediction[0]
    except Exception as e:
        print(f"Error making prediction: {e}")
        return None, None, None

# ========================
# Class names
# ========================
def get_class_name(class_index):
    classes = [
        "Apple___Apple_scab", "Apple___Black_rot", "Apple___Cedar_apple_rust", "Apple___healthy",
        "Blueberry___healthy", "Cherry_(including_sour)___Powdery_mildew", "Cherry_(including_sour)___healthy",
        "Corn_(maize)___Cercospora_leaf_spot Gray_leaf_spot", "Corn_(maize)___Common_rust_",
        "Corn_(maize)___Northern_Leaf_Blight", "Corn_(maize)___healthy", "Grape___Black_rot",
        "Grape___Esca_(Black_Measles)", "Grape___Leaf_blight_(Isariopsis_Leaf_Spot)", "Grape___healthy",
        "Orange___Haunglongbing_(Citrus_greening)", "Peach___Bacterial_spot", "Peach___healthy",
        "Pepper,_bell___Bacterial_spot", "Pepper,_bell___healthy", "Potato___Early_blight",
        "Potato___Late_blight", "Potato___healthy", "Raspberry___healthy", "Soybean___healthy",
        "Squash___Powdery_mildew", "Strawberry___Leaf_scorch", "Strawberry___healthy",
        "Tomato___Bacterial_spot", "Tomato___Early_blight", "Tomato___Late_blight", "Tomato___Leaf_Mold",
        "Tomato___Septoria_leaf_spot", "Tomato___Spider_mites Two-spotted_spider_mite",
        "Tomato___Target_Spot", "Tomato___Tomato_Yellow_Leaf_Curl_Virus", "Tomato___Tomato_mosaic_virus",
        "Tomato___healthy"
    ]
    if 0 <= class_index < len(classes):
        return classes[class_index]
    else:
        return "Unknown"

# ========================
# Main function
# ========================
def main():
    model_path = "Pretrained_model.h5"
    sample_image_path = "sample_image.jpg"  # 👈 ضع هنا اسم الصورة

    if not os.path.exists(model_path):
        print(f"Model file not found at: {model_path}")
        return

    if not os.path.exists(sample_image_path):
        print(f"Image file not found at: {sample_image_path}")
        return

    model = load_model(model_path)
    if model is None:
        return

    # تحقق هل الموديل فيه طبقة Rescaling
    has_rescaling = any(isinstance(layer, tf.keras.layers.Rescaling) for layer in model.layers)
    image_array = preprocess_image(sample_image_path, target_size=(299, 299), normalize=not has_rescaling)
    if image_array is None:
        return

    predicted_class, confidence, all_predictions = predict_disease(model, image_array)
    if predicted_class is not None:
        class_name = get_class_name(predicted_class)
        print(f"\nPrediction Results:")
        print(f"Predicted Class: {class_name}")
        print(f"Confidence: {confidence:.2%}")
        print(f"Class Index: {predicted_class}")

        # Show top 3 predictions
        top_3_indices = np.argsort(all_predictions)[-3:][::-1]
        print(f"\nTop 3 Predictions:")
        for i, idx in enumerate(top_3_indices):
            class_name = get_class_name(idx)
            confidence = all_predictions[idx]
            print(f"{i+1}. {class_name}: {confidence:.2%}")

if __name__ == "__main__":
    main()