Function text_parse

text_parse/parse.py:24–49

Text document parsing, which is essentially making a copy of the file :param file_path: Original file path :param save_path: Path for the parsed file :return: Path of the parsed file

(file_path,save_path)

    return str(res) if res is not None else ""

def text_parse(file_path,save_path):
    """
    Text document parsing, which is essentially making a copy of the file
    :param file_path: Original file path
    :param save_path: Path for the parsed file
    :return: Path of the parsed file
    """

    # Read file content
    with open(file_path, 'r', encoding='utf-8') as f:
        data = f.readlines()

    print(f"Parsing text file and saving results to {save_path}")

    # Get filename and construct save path
    file_name = file_path.split('/')[-1]  # Extract filename
    base_name = file_name.split('.')[0]  # Extract part without extension
    save_file_path = os.path.join(save_path, base_name + '.txt').replace('\\', '/') # Concatenate full path

    # Ensure the save path directory exists
    os.makedirs(save_path, exist_ok=True)
    # Write to file
    with open(save_file_path, 'w', encoding='utf-8') as f:
        for line in data:
            f.write(line)
    return save_file_path

def pdf_parse(file_path, save_path, progress_callback=None):
    """PDF file parsing, supports progress callback"""