import gradio as gr
import os
import torch

#from demos.foodvision_mini.model import create_effnetb2_model
from model import create_effnetb2_model
from timeit import default_timer as timer
from typing import Tuple, Dict

#setup classnames

class_names = ['pizza', 'steak', 'sushi']

# model and trandorms preparations
effnetb2, effnetb2_transforms = create_effnetb2_model(num_classes=3)

#load save weights

effnetb2.load_state_dict(
    torch.load(f = "09_preptrained_effnetb2_20_percent (2).pth",
               map_location = torch.device('cpu'))

)

# make predictions

def predict(img) -> Tuple[Dict,float] :
  start_time = timer()
  # this returns the prediction, and then, time
  #start a timers
  # transform the input image for use with effnetb2
  #put model into eval mode
  # create a prediction label and prediction probability dictionary
  img =  effnetb2_transforms(img).unsqueeze(0)
  effnetb2.eval()

  with torch.inference_mode():
    pred_probs = torch.softmax(effnetb2(img), dim = 1)

  pred_labels_and_probs = {class_names[i]:float(pred_probs[0][i]) for i in range(len(class_names))}

  end_time = timer()
  pred_time = round(end_time - start_time, 4)

  return pred_labels_and_probs, pred_time

import os
example_list = [["examples/" + example] for example in os.listdir("examples")]

title = "FoodVision Mini"
Description = "An EfficientNetB2 feature computer vision model to classify images as pizza, steak or sushi"
article = "Cretated at......"

demo = gr.Interface(fn=predict,inputs=gr.Image(type='pil'),
                    outputs =[gr.Label(num_top_classes=3, label = "Predictions"),
                              gr.Number(label="Prediction time (s)")],
                    examples= example_list,
                    title = title,
                    description=Description,
                    article=article)

demo.launch(debug=False,share = True) # print errors locally, generate a publically shareable URL