import gradio as gr
import tensorflow as tf
import numpy as np
from PIL import Image

model_path = "Cat_transfer_learning_MobileNetV2.keras"
model = tf.keras.models.load_model(model_path)

# Define the core prediction function
def predict_cat(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)
    
     # No need to apply sigmoid, as the output layer already uses softmax
    # Convert the probabilities to rounded values
    prediction = np.round(prediction, 3)

    # Separate the probabilities for each class
    p_americanshorthair = prediction[0][0]  # Probability for class 'articuno'
    p_bengal = prediction[0][1]   # Probability for class 'moltres'
    p_mainecoon = prediction[0][2]    # Probability for class 'zapdos'
    p_ragdoll = prediction[0][3]    # Probability for class 'zapdos'
    p_scottishfold = prediction[0][4]
    p_sphinx = prediction[0][5]

 
    return {'americanshorthair':  p_americanshorthair, 'bengal': p_bengal, 'mainecoon': p_mainecoon, 'ragdoll': p_ragdoll, 'scottishfold': p_scottishfold, 'sphinx': p_sphinx }

# Create the Gradio interface
input_image = gr.Image()
iface = gr.Interface(
    fn=predict_cat,
    inputs=input_image,
    outputs=gr.Label(),
    examples=["images/americanshorthair_1.jpg", "images/americanshorthair_2.jpg", "images/americanshorthair_3.jpg", "images/bengal_1.jpg", "images/bengal_2.jpeg", "images/bengal_3.jpg", "images/mainecoon_1.jpg", "images/mainecoon_2.jpeg", "images/mainecoon_3.jpg", "images/ragdoll_1.jpg", "images/ragdoll_2.jpg", "images/ragdoll_3.jpeg", "images/scottishfold_1.jpeg", "images/scottishfold_2.jpg", "images/scottishfold_3.jpg", "images/sphinx_1.jpg", "images/sphinx_2.jpg", "images/sphinx_3.jpg"],  
    description="Let's predict some cats!")

iface.launch()