Method _preprocess

ms_agent/utils/push_to_hub.py:366–423

(folder_path: str,
                    path_in_repo_url: str,
                    add_powered_by: bool = True)

    @staticmethod
    def _preprocess(folder_path: str,
                    path_in_repo_url: str,
                    add_powered_by: bool = True) -> 'Tuple[str, str]':

        report_filename = 'report.md'
        file_path = os.path.join(folder_path, report_filename)
        file_path_hash: str = text_hash(text=file_path, keep_n_chars=8)
        new_report_filename: str = f'report_{file_path_hash}.md'
        current_cache_path: str = os.path.join(folder_path, '.cache')
        os.makedirs(current_cache_path, exist_ok=True)
        new_file_path = os.path.join(current_cache_path, new_report_filename)

        if not os.path.exists(file_path):
            logger.warning(
                f'The report file: {file_path} does not exist. Skipping preprocessing.'
            )
            return '', ''

        try:
            shutil.copy(file_path, new_file_path)
        except Exception as e:
            logger.error(f'Error copying the report file: {e}')
            return '', ''

        try:
            with open(file_path, 'r', encoding='utf-8') as f:
                content = f.read()
                if add_powered_by and not content.lstrip().startswith(
                        '<span style='):
                    content = """<span style="color: darkgreen; font-weight: bold; font-family: monospace;
                    ">Powered by [MS-Agent](https://github.com/modelscope/ms-agent) |
                    [DocResearch](https://github.com/modelscope/ms-agent/blob/main/projects/doc_research/README.md)
                    </span>""" + '\n\n' + content

            pattern = r'!\[(.*?)\]\((resources/.*?)\)'
            replacement = rf'![\1]({path_in_repo_url}\2)'
            new_content, count = re.subn(pattern, replacement, content)

            if count > 0:
                with open(file_path, 'w', encoding='utf-8') as f:
                    f.write(new_content)
                logger.info(
                    f"Preprocessed {count} 'resources/' links in {file_path}.")
            else:
                logger.info(
                    f'No "resources/" links found in {file_path}. No changes made.'
                )

        except IOError as e:
            logger.error(f'Error reading or writing the report file: {e}')
            return '', ''
        except Exception as e:
            logger.error(
                f'Unexpected error during preprocessing of PushToModelScope: {e}'
            )
            return '', ''

        return file_path, new_file_path