import numpy as np
import torch
import torch.nn.functional as F
from torchvision.transforms.functional import normalize
from huggingface_hub import hf_hub_download
import gradio as gr
from PIL import Image

def resize_image(image):
    image = image.convert('RGB')
    model_input_size = (1024, 1024)
    image = image.resize(model_input_size, Image.BILINEAR)
    return image

def process(image):
    # 이미지가 numpy 배열인 경우에만 PIL.Image 객체로 변환
    if isinstance(image, np.ndarray):
        orig_image = Image.fromarray(image)
    else:
        # 이미 PIL.Image.Image 객체인 경우, 변환 없이 사용
        orig_image = image

    w, h = orig_im_size = orig_image.size
    image = resize_image(orig_image)
    im_np = np.array(image)
    im_tensor = torch.tensor(im_np, dtype=torch.float32).permute(2, 0, 1)
    im_tensor = torch.unsqueeze(im_tensor, 0)
    im_tensor = torch.divide(im_tensor, 255.0)
    im_tensor = normalize(im_tensor, [0.5, 0.5, 0.5], [1.0, 1.0, 1.0])
    if torch.cuda.is_available():
        im_tensor = im_tensor.cuda()

    # 모델 로딩 및 예측
    net = BriaRMBG()
    model_path = hf_hub_download("briaai/RMBG-1.4", 'model.pth')
    if torch.cuda.is_available():
        net.load_state_dict(torch.load(model_path))
        net = net.cuda()
    else:
        net.load_state_dict(torch.load(model_path, map_location="cpu"))
    net.eval()

    result = net(im_tensor)
    # post process
    result = torch.squeeze(F.interpolate(result[0][0], size=(h, w), mode='bilinear'), 0)
    ma = torch.max(result)
    mi = torch.min(result)
    result = (result - mi) / (ma - mi)
    # image to pil
    im_array = (result * 255).cpu().data.numpy().astype(np.uint8)
    pil_im = Image.fromarray(np.squeeze(im_array))
    # paste the mask on the original image
    new_im = Image.new("RGBA", pil_im.size, (0, 0, 0, 0))
    new_im.paste(orig_image, mask=pil_im)

    return new_im

with gr.Blocks() as demo:
    with gr.Column():
        gr.Markdown("누끼따기의 왕 '누킹'(Nuking)")
        input_image = gr.Image(type="pil")
        output_image = gr.Image()
        process_button = gr.Button("Remove Background")
        process_button.click(fn=process, inputs=input_image, outputs=output_image)

demo.launch()