# Import dependencies
from keras.models import load_model
from PIL import Image, ImageOps
import numpy as np
import gradio as gr


# Definition of the main function for predictions
def predict_nevus(image):
    # Load the model
    model = load_model('keras_model.h5')

    # Create the array of the right shape to feed into the keras model
    # The 'length' or number of images you can put into the array is
    # determined by the first position in the shape tuple, in this case 1.
    data = np.ndarray(shape=(1, 224, 224, 3), dtype=np.float32)

    #turn the image into a numpy array
    image_array = np.asarray(image)
    
    # Normalize the image
    normalized_image_array = (image_array.astype(np.float32) / 127.0) - 1
    
    # Load the image into the array
    data[0] = normalized_image_array

    # run the inference
    prediction = model.predict(data)
    return { 
        'Melanoma': float(prediction[0][0]), 
        'Lunar': float(prediction[0][1])
    }

# Deploy with Gradio
examples=[
  ['2.jpg'],
  ['37.jpg'],
  ['186.jpg']
]

iface = gr.Interface(
    fn=predict_nevus, 
    inputs=gr.inputs.Image(shape=(224, 224)), 
    outputs="label",
    title="Detector de melanomas",
    description="Herramienta online que utiliza inteligencia artificial para detectar posibles melanomas en fotografías de lunares.",
    examples=examples,
    allow_flagging='never',
    theme="peach"
)

iface.launch(debug=True)