import gradio as gr
import tensorflow as tf
import numpy as np
from PIL import Image
 
model_path = "fungi-model_transferlearning_xception.keras"
model = tf.keras.models.load_model(model_path)
 

def predict_fungi(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, ...]

    prediction = model.predict(image)
 

    # Convert the probabilities to rounded values
    prediction = np.round(prediction, 2)
 
    # Separate the probabilities for each class
    p_ediblemushroomsporocarp = prediction[0][0]  
    p_ediblesporocarp = prediction[0][1]   
    p_poisonousmushroomsporocarp = prediction[0][2]    
    p_poisonoussporocarp = prediction[0][2]  
 
    return {'Edible Mushroom Sporocarp':  p_ediblemushroomsporocarp, 'Edible Sporocarp': p_ediblesporocarp, 'Poisonous Mushroom Sporocarp': p_poisonousmushroomsporocarp, 'Poisonous Sporocarp': p_poisonoussporocarp}
 
 
input_image = gr.Image()
iface = gr.Interface(
    fn=predict_fungi,
    inputs=input_image,
    outputs=gr.Label(),
    examples=["images/ce (1).jpg", "images/ce (2).jpg", "images/ncvc (1).png", "images/ncvc (5).png", "images/nncv (2).png", "images/nncv (4).png", "images/cv (2).png", "images/cv (4).png"],
    title="Is it edible?",
    description="'sporocarp' is a broad term encompassing any spore-producing structure in various organisms, while 'mushroom sporocarp' specifically refers to the spore-producing structures of fungi that are typically visible and recognizable as mushrooms.")
 
iface.launch()