Function parse_pdf

demo/Paper2Poster/utils/src/model_utils.py:98–141

Parse a PDF file and extract text and images. Args: pdf_path (str): The path to the PDF file. output_path (str): The directory to save the extracted content. model_lst (list): A list of models for processing the PDF. Returns: str: The full text extracte

(
    pdf_path: str,
    output_path: str = None,
    model_lst: list = None,
    save_file: bool = True,
)

def parse_pdf(
    pdf_path: str,
    output_path: str = None,
    model_lst: list = None,
    save_file: bool = True,
) -> str:
    """
    Parse a PDF file and extract text and images.

    Args:
        pdf_path (str): The path to the PDF file.
        output_path (str): The directory to save the extracted content.
        model_lst (list): A list of models for processing the PDF.

    Returns:
        str: The full text extracted from the PDF.
    """
    if save_file:
        os.makedirs(output_path, exist_ok=True)
    config_parser = ConfigParser(
        {
            "output_format": "markdown",
        }
    )
    converter = PdfConverter(
        config=config_parser.generate_config_dict(),
        artifact_dict=model_lst,
        processor_list=config_parser.get_processors(),
        renderer=config_parser.get_renderer(),
    )
    rendered = converter(pdf_path)
    full_text, _, images = text_from_rendered(rendered)
    if save_file:
        with open(pjoin(output_path, "source.md"), "w+", encoding="utf-8") as f:
            f.write(full_text)
        for filename, image in images.items():
            image_filepath = os.path.join(output_path, filename)
            image.save(image_filepath, "JPEG")
        with open(pjoin(output_path, "meta.json"), "w+") as f:
            f.write(json.dumps(rendered.metadata, indent=4))

    if not save_file:
        return full_text, rendered
    return full_text


# def get_text_embedding(