### 1. Imports and class names setup ### 
import gradio as gr
import os
import torch

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

class_names = ['appaloosa_leopard', 'dutch_warmblood', 'thoroughbred_chestnut']

model_ft, model_transforms = create_model(num_classes=len(class_names))

model_ft.load_state_dict(
    torch.load("horses_swin_t.pt", map_location=torch.device("cpu"))
)

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

    img = model_transforms(img).unsqueeze(0)

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

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

    pred_time = round(timer() - start_time, 5)

    return pred_labels_and_probs, pred_time

title = "HorseVision Mini 🐎"
description = "A feature extractor computer vision model to classify images of horses."
#article = "Created at [09. PyTorch Model Deployment](https://www.learnpytorch.io/09_pytorch_model_deployment/)."

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=article
                    )

demo.launch()