import gradio as gr
import os
import torch

from model import create_effnetb2_model
from typing import Tuple, Dict 
from timeit import default_timer as timer

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

effnetb2, effnetb2_transform = create_effnetb2_model(num_classes = len(class_names))

effnetb2.load_state_dict(torch.load(f="09_pretrained_effnetb2_feature_extractor_pizza_steak_sushi_20_percent.pth",
                                    map_location = torch.device("cpu")))

def predict(img) -> Tuple[Dict, float]:

  start = timer()

  effnetb2.eval()
  with torch.inference_mode():

    img = effnetb2_transform(img)
    img = img.unsqueeze(dim=0)

    pred = effnetb2(img)
    pred_probs = torch.softmax(pred, dim=1)
    class_label = torch.argmax(pred_probs, dim=1)

  pred_dict = {class_names[i]: float(pred_probs[0][i].cpu().item()) for i in range(len(class_names))}
  end = timer()
  pred_time = round(end-start, 5)

  return pred_dict, pred_time

title = "FoodVision Mini 🍕🥩🍣"
description = "An EfficientNetB2 feature extractor computer vision model to classify images of food as pizza, steak or sushi"

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

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="")

demo.launch(debug=False,
            share=True)