app.py

import gradio as gr
import re
import os
from pdf2image import convert_from_path
from transformers import DonutProcessor, VisionEncoderDecoderModel
import torch
from PIL import Image
from pathlib import Path

from models.experimental import attempt_load
from utils.datasets import LoadImage
from utils.general import check_img_size, non_max_suppression, scale_coords, set_logging
from utils.torch_utils import select_device
import cv2
#sudo apt-get install poppler-utils Necesario

key = str(os.environ.get('key'))

def check_image(image):
    try:
        images = convert_from_path(Path(image.name), fmt="jpeg", size=(960,1280))
        return images
    except:
        return [Image.open(image)]

def crop(files = '', #files
           weights = 'yolov7.pt', #model.pt path(s)
           classes = None, #filter by class: --class 0, or --class 0 2 3
           imgsz = 640, #inference size (pixels)
           device = '', #cuda device, i.e. 0 or 0,1,2,3 or cpu
           conf_thres = 0.25, #object confidence threshold
           iou_thres = 0.45, #IOU threshold for NMS
           augment = False, #augmented inference
           agnostic_nms = False): #class-agnostic NMS 

    # Initialize
    set_logging()
    device = select_device(device)
    half = device.type != 'cpu'  # half precision only supported on CUDA

    # Load model
    model = attempt_load(weights, map_location=device)  # load FP32 model
    stride = int(model.stride.max())  # model stride
    imgsz = check_img_size(imgsz, s=stride)  # check img_size

    if half:
        model.half()  # to FP16

    # Set Dataloader
    dataset = LoadImage(files = files, img_size=imgsz, stride=stride)

    # Get names and colors
    names = model.module.names if hasattr(model, 'module') else model.names
    
    # Run inference
    if device.type != 'cpu':
        model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters())))  # run once
    old_img_w = old_img_h = imgsz
    old_img_b = 1
    list_cropobj = []
    for img, img0s in dataset:
        img = torch.from_numpy(img).to(device)
        img = img.half() if half else img.float()  # uint8 to fp16/32
        img /= 255.0  # 0 - 255 to 0.0 - 1.0
        if img.ndimension() == 3:
            img = img.unsqueeze(0)

        # Inference
        with torch.no_grad():   # Calculating gradients would cause a GPU memory leak
            pred = model(img, augment=augment)[0]

        # Apply NMS
        pred = non_max_suppression(pred, conf_thres, iou_thres, classes=classes, agnostic=agnostic_nms)

        # Process detections
        for i, det in enumerate(pred):  # detections per image
            if len(det):
                # Rescale boxes from img_size to img0s size
                det[:, :4] = scale_coords(img.shape[2:], det[:, :4], img0s.shape).round()

                # Write results
                for *xyxy, conf, cls in reversed(det):
                    #crop an image based on coordinates
                    object_coordinates = [int(xyxy[0]),int(xyxy[1]),int(xyxy[2]),int(xyxy[3])]
                    cropobj_bgr = img0s[int(xyxy[1]):int(xyxy[3]),int(xyxy[0]):int(xyxy[2])]
                    cropobj_rgb = cv2.cvtColor(cropobj_bgr, cv2.COLOR_BGR2RGB)
                    clase = names[int(cls)]
                    list_cropobj.append([Image.fromarray(cropobj_rgb),int(cls)])
                    
    return list_cropobj


def get_attributes(input_img):
    #access_token = str(os.environ.get('key'))
    access_token = key
    processor = DonutProcessor.from_pretrained("AA-supply/donut-finetuned-cedula-crop", use_auth_token=access_token)
    model = VisionEncoderDecoderModel.from_pretrained("AA-supply/donut-finetuned-cedula-crop", use_auth_token=access_token)

    device = "cuda" if torch.cuda.is_available() else "cpu"

    model.eval()
    model.to(device)
    
    images = check_image(input_img)
    images = crop(weights="best.pt", files= images)
    image_cedula = [img[0] for img in images if img[1]==0][0] #0 en 'img[1]==0' es el label de cedula, si se reemplaza por 1
                                                              #entrega licencias
    
    pixel_values = processor(image_cedula, return_tensors="pt").pixel_values
    pixel_values = pixel_values.to(device)
    print(pixel_values.size())
    # prepare decoder inputs
    task_prompt = "<s_cord-v2>"
    decoder_input_ids = processor.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
    decoder_input_ids = decoder_input_ids.to(device)

    # autoregressively generate sequence
    outputs = model.generate(
        pixel_values,
        decoder_input_ids=decoder_input_ids,
        max_length=model.decoder.config.max_position_embeddings,
        early_stopping=True,
        pad_token_id=processor.tokenizer.pad_token_id,
        eos_token_id=processor.tokenizer.eos_token_id,
        use_cache=True,
        num_beams=1,
        bad_words_ids=[[processor.tokenizer.unk_token_id]],
        return_dict_in_generate=True,
    )

    # turn into JSON
    seq = processor.batch_decode(outputs.sequences)[0]
    seq = seq.replace(processor.tokenizer.eos_token, "").replace(processor.tokenizer.pad_token, "")
    seq = re.sub(r"<.*?>", "", seq, count=1).strip()  # remove first task start token
    seq = processor.token2json(seq)

    return str(seq)

demo = gr.Interface(get_attributes, "file", "label")
demo.launch()