dagshub-docs/create_single_doc.py

import os
import re
import sys


def adjust_image_paths(markdown_content, original_md_file_path, root_folder):
    """
    Adjusts image paths in the markdown content to be relative to the root folder.
    This includes markdown image syntax and HTML <img> tags.

    :param markdown_content: The content of the markdown file.
    :param original_md_file_path: The file path of the original markdown file.
    :param root_folder: The root folder that contains all the markdown files.
    :return: Updated markdown content with adjusted image paths.
    """
    def adjust_path(match):
        original_path = match.group(1)

        # Construct the absolute path to the image
        absolute_image_path = os.path.abspath(os.path.join(os.path.dirname(original_md_file_path), original_path))

        # Make the path relative to the root folder
        relative_image_path = os.path.relpath(absolute_image_path, root_folder)

        # Ensure the path starts with './'
        if not relative_image_path.startswith('.'):
            relative_image_path = './' + relative_image_path

        return match.group(0).replace(original_path, relative_image_path)

    # Adjust paths in markdown image syntax
    adjusted_content = re.sub(r'!\[.*?\]\((.*?)\)', adjust_path, markdown_content)

    # Adjust paths in HTML <img> tags
    adjusted_content = re.sub(r'<img\s+[^>]*?src=["\'](.*?)["\'][^>]*>', adjust_path, adjusted_content)

    return adjusted_content


def concatenate_markdown_files(root_folder, output_file):
    """
    Searches for markdown (.md) files starting from the root folder, adjusts their image paths,
    concatenates their contents, and writes them to a specified output markdown file.

    :param root_folder: The root directory to search for markdown files.
    :param output_file: The file path for the output markdown file.
    """
    concatenated_content = ""

    for subdir, dirs, files in os.walk(root_folder):
        for file in files:
            if file.endswith('.md'):
                file_path = os.path.join(subdir, file)
                with open(file_path, 'r') as md_file:
                    file_content = md_file.read()
                    # Adjust image paths
                    file_content = adjust_image_paths(file_content, file_path, root_folder)
                    concatenated_content += file_content + '\n\n'

    with open(output_file, 'w') as output_md_file:
        output_md_file.write(concatenated_content)


if __name__ == "__main__":
    if len(sys.argv) != 3:
        print("Usage: python script.py <root_folder> <output_file>")
        sys.exit(1)

    root_folder = sys.argv[1]
    output_file = sys.argv[2]

    concatenate_markdown_files(root_folder, output_file)
    print(f"All markdown files from {root_folder} have been concatenated into {output_file}.")