returns table in chinese, when I explicitly download english weights

[maifeeulasad]

请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem

• 系统环境/System Environment：Ubuntu 22.04 LTS
• 版本号/Version：Paddle：2.5.1 PaddleOCR：2.7.0.2 问题相关组件/Related components：-
• 运行指令/Command Code：

First I'm downloading all the necessery weights:

# Download the PP-OCRv3 text detection model and unzip it
wget https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_infer.tar && tar xf en_PP-OCRv3_det_infer.tar
# Download the PP-OCRv3 text recognition model and unzip it
wget https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_infer.tar && tar xf en_PP-OCRv3_rec_infer.tar
# Download the PP-StructureV2 form recognition model and unzip it
wget https://paddleocr.bj.bcebos.com/ppstructure/models/slanet/en_ppstructure_mobile_v2.0_SLANet_infer.tar && tar xf en_ppstructure_mobile_v2.0_SLANet_infer.tar

And also using the recommended version:

pip install "paddleocr>=2.0.1"

I've modified the predict_table.py file based upon my need (rest is available @ https://github.com/PaddlePaddle/PaddleOCR/blob/b318d204fa4b984a76edee167a5cb867e5b85838/ppstructure/table/predict_table.py):

def process_image_and_extract_table(det_model_dir, rec_model_dir, table_model_dir,rec_char_dict_path, table_char_dict_path, image_path, output_dir):
    args = parse_args()
    args.det_model_dir = det_model_dir
    args.rec_model_dir = rec_model_dir
    args.table_model_dir = table_model_dir
    args.rec_char_dict_path = rec_char_dict_path
    args.table_char_dict_path = table_char_dict_path
    args.image_dir = image_path
    args.output = output_dir
    
    os.makedirs(args.output, exist_ok=True)
    
    table_sys = TableSystem(args)
    output_path = table_sys.extract_tables(image_path)
    return output_path

And calling this function like this:

import os
from PIL import Image
from paddleocr import PaddleOCR,draw_ocr
# import table.meaw as meaw
# from table import meaw
# from table.predict_table import process_image_and_extract_table
from predict_table import process_image_and_extract_table

c_dir = "files"

# print(os.listdir(c_dir))

files = [os.path.join(c_dir,x) for x in os.listdir(c_dir) if x.endswith(".PNG")]

print(files)

ocr = PaddleOCR(use_angle_cls=True, lang='en')

for file in files:
    img_file = file + ".tem.png"
    print(file)
    result = ocr.ocr(file, cls=True)
    # result = ocr.ocr(file, cls=True, type='structure')
    print("---------------------------------------")
    print(result)
    print(len(result))
    print(len(result[0]))
    print("***************************************")

    result = result[0]
    image = Image.open(file).convert('RGB')
    boxes = [line[0] for line in result]
    txts = [line[1][0] for line in result]
    scores = [line[1][1] for line in result]
    im_show = draw_ocr(image, boxes, txts, scores, font_path='./Roboto-Thin.ttf')
    im_show = Image.fromarray(im_show)
    im_show.save(img_file)

    output_path = process_image_and_extract_table(
        det_model_dir='./en_PP-OCRv3_det_infer',
        rec_model_dir='./en_PP-OCRv3_rec_infer',
        table_model_dir='./en_ppstructure_mobile_v2.0_SLANet_infer',
        rec_char_dict_path='./ppocr_keys_v1.txt',
        table_char_dict_path='./table_structure_dict.txt',
        image_path=file,
        output_dir='./out'
    )

    print("Table extracted and saved to:", output_path)

    # im_show.show("demo")

    # break

I've downloaded other necessery files from here: ppocr_keys_v1.txt, table_structure_dict.txt

• 完整报错/Complete Error Message：
  No specific error messages, it's simply returning output in Chinese.

I just wanted to generate HTML table out of a document.

[Shreejan-git]

Having a same issue

[UserWangZz]

ppocr_keys_v1.txt
this dict is a Chinese character dict.
you should use en dict which can download from this link.
https://github.com/PaddlePaddle/PaddleOCR/blob/main/ppocr/utils/en_dict.txt