import gradio as gr
import tensorflow as tf
import numpy as np
from PIL import Image
model_path = "Xception.keras"
model = tf.keras.models.load_model(model_path)
# Define the core prediction function
def predict_pokemon(image):
    # Preprocess image
    print(type(image))
    image = Image.fromarray(image.astype('uint8'))  # Convert numpy array to PIL image
    image = image.resize((150, 150)) #resize the image to 28x28 and converts it to gray scale
    image = np.array(image)
    image = np.expand_dims(image, axis=0) # same as image[None, ...]
    
    # Predict
    prediction = model.predict(image)
    
    # Because the output layer was dense(0) without an activation function, we need to apply sigmoid to get the probability
    # we could also change the output layer to dense(1, activation='sigmoid')
    prediction = np.round(prediction, 2)
        # Separate the probabilities for each class
    P_aloevera = prediction[0][0]  # Probability for class 'abra'
    P_curcuma = prediction[0][1]   # Probability for class 'beedrill'
    p_guava = prediction[0][2]    # Probability for class 'sandshrew'
    return {'aloevera': P_aloevera, 'curcuma':  P_curcuma, 'guava': p_guava}
# Create the Gradio interface
input_image = gr.Image()
iface = gr.Interface(
    fn=predict_pokemon,
    inputs=input_image, 
    outputs=gr.Label(),
    examples=["images/aloevera0.jpg", "images/curcuma51.jpg", "images/guava10.jpg"],   
    description="A simple mlp classification model for image classification using the mnist dataset.")
iface.launch(share=True)