### imports and class names setup ###
import gradio as gr
import os
import torch

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

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

### model and transforms prepration ###
effnetb2, effnetb2_transforms = create_effnetb2_model(num_classes=3,
                                                      seed=29)
# loading the saved weights
effnetb2.load_state_dict(
  torch.load(
    f="pretrained_effnetb2_feature_extractor.pth",
    map_location=torch.device("cpu") # loading the model to cpu
  )
)

### predict function ###
def predict(img) -> Tuple[Dict, float]:
  # start a timer
  start_time = timer()

  # transforming the input image
  img = effnetb2_transforms(img).unsqueeze(0)

  # putting the model into eval mode & making prediction
  effnetb2.eval()
  with torch.inference_mode():
    # passing transformed img through the model and turn pred logits into probs
    pred_probs = torch.softmax(effnetb2(img), dim=1)

  # creating a prediction label & prediction probability dictionary
  pred_labels_and_probs = {class_names[i]: float(pred_probs[0][i]) for i in range(len(class_names))}

  # calculate pred time
  end_time = timer()
  pred_time = round(end_time-start_time, 4)

  # return pred dict and pred time
  return pred_labels_and_probs, pred_time

### gradio app ###
# creating title, description and article
title = "FoodVision Mini"
description = "An EfficientNetB2 feature extractor computer vision model to classify images as pizza, steak or sushi."

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

# creating the gradio demo
demo = gr.Interface(fn=predict, # maps inputs to outputs
                    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)

# launching the demo
demo.launch(debug=False) # don't need `share=True` in HuggingFace spaces