baselines/UIED-3.3/run_batch.py

import multiprocessing
import glob
import time
import json
from tqdm import tqdm
import argparse
import os
from os.path import join as pjoin, exists
import cv2

import detect_compo.ip_region_proposal as ip

def resize_height_by_longest_edge(img_path, resize_length=800):
    org = cv2.imread(img_path)
    height, width = org.shape[:2]
    if height > width:
        return resize_length
    else:
        return int(resize_length * (height / width))


def main(args):
    # initialization
    # input_img_root = "E:/Mulong/Datasets/rico/combined"
    # output_root = "E:/Mulong/Result/rico/rico_uied/rico_new_uied_v3"
    # data = json.load(open('E:/Mulong/Datasets/rico/instances_test.json', 'r'))
    input_img_root = args.input_img_root
    output_root = args.output_root
    with open(args.ann, 'r') as f:
        data = json.load(f)

    os.makedirs(output_root, exist_ok=True)
    # input_imgs = [pjoin(input_img_root, img['file_name'].split('/')[-1]) for img in data['images']]
    # input_imgs = sorted(input_imgs, key=lambda x: int(x.split('/')[-1][:-4]))  # sorted by index
    input_imgs = [pjoin(input_img_root, img['file_name']) for img in data['images']]

    # key_params = {'min-grad': 10, 'ffl-block': 5, 'min-ele-area': 50, 'merge-contained-ele': True,
    #               'max-word-inline-gap': 10, 'max-line-ingraph-gap': 4, 'remove-bar': True}
    key_params = {'min-grad': 10, 'ffl-block': 5, 'min-ele-area': 50, 'merge-contained-ele': True,
                  'max-word-inline-gap': 10, 'max-line-ingraph-gap': 4, 'remove-bar': False}
    

    is_ip = True
    is_clf = False
    is_ocr = True
    is_merge = True

    # Load deep learning models in advance
    compo_classifier = None
    if is_ip and is_clf:
        compo_classifier = {}
        from cnn.CNN import CNN
        # compo_classifier['Image'] = CNN('Image')
        compo_classifier['Elements'] = CNN('Elements')
        # compo_classifier['Noise'] = CNN('Noise')
    ocr_model = None
    if is_ocr:
        import detect_text.text_detection as text

    # set the range of target inputs' indices
    num = 0
    # start_index = 30800  # 61728
    # end_index = 100000
    start_time = time.time()
    for input_img in tqdm(input_imgs, total=len(input_imgs)):
        name = input_img.split('/')[-1].split('.')[-2]
        resized_height = resize_height_by_longest_edge(input_img)
        # index = input_img.split('/')[-1][:-4]
        # if int(index) < start_index:
        #     continue
        # if int(index) > end_index:
        #     break

        if is_ocr:
            text.text_detection(input_img, output_root, show=False, method='paddle')

        if is_ip:
            ip.compo_detection(input_img, output_root, key_params, classifier=compo_classifier, resize_by_height=resized_height, show=False)

        if is_merge:
            import detect_merge.merge as merge
            # compo_path = pjoin(output_root, 'ip', str(index) + '.json')
            # ocr_path = pjoin(output_root, 'ocr', str(index) + '.json')
            compo_path = pjoin(output_root, 'ip', name + '.json')
            ocr_path = pjoin(output_root, 'ocr', name + '.json')
            merge.merge(input_img, compo_path, ocr_path, output_root, is_remove_bar=key_params['remove-bar'], show=False)

        num += 1
    
    end_time = time.time()
    print(f'Total inference time: {end_time - start_time} s')
    os.makedirs(pjoin(output_root, 'time'), exist_ok=True)
    with open(pjoin(output_root, 'time', 'inference time.txt'), 'w') as f:
        f.write(f'First stage total inference time: {end_time - start_time} s')


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('--input_img_root', type=str, required=True)
    parser.add_argument('--output_root', type=str, required=True)
    parser.add_argument('--ann', type=str, required=True)
    args = parser.parse_args()
    main(args)