import tensorflow as tf
from keras.models import load_model
import gradio as gr
from matplotlib import pyplot as plt
import cv2
import numpy as np
model = load_model('eee4.keras')

def image_mod(image):
    raw_image = image
    temp_image = "temporary_image.jpg"
    cv2.imwrite(temp_image, raw_image)
    
    img = cv2.imread("temporary_image.jpg")
    resize = tf.image.resize(img, (256, 256))
    plt.imshow(resize.numpy().astype(int))
    yhat = model.predict(np.expand_dims(resize,0))
    display_old = np.argmax(yhat)
    display = str(display_old)
    #display = str(yhat)
    #display = str(display)
    if display == "0":
        message = "Cloudy" # 
    if display == "1":
        message = "Snowy" # Morning_fog_-_Flickr_-_tmoravec.jpg
    if display == "2":
        message = "Foggy" # Stuyvesant_Fish_House_25_E78_St_cloudy_jeh.jpg
    if display == "3":
        message = "Rainy" #Snow_on_Branches,_Beechview,_2020-12-17,_01.jpg
    if display == "4":
        message = "Sunny" # Daedalus_000355_171913_516869_4578_(36155269413).jpg
    return message

gr.Interface(fn=image_mod,
             inputs=gr.Image(shape=(256, 256)),
             outputs=gr.Label(num_top_classes=3),
             examples=["640px-Hopetoun_house_sunny_day.jpg","Panoramic_of_water_reflection_of_the_mountains_of_Vang_Vieng_with_cloudy_sky_in_paddy_fields.jpg","Snow_Scene_at_Shipka_Pass_1.JPG","Jida,_Zhuhai,_rainy_day.jpg","The_lift_bridge_on_Cherry_Street_over_the_ship_channel_to_the_Turning_Basin_on_a_foggy_day,_2012-03-17_-a.jpg"]).launch()