doc_generator/src/textdocs/utils.py

# textdocs/utils.py
import re
from django.conf import settings

def convert_media_urls_to_file_paths(html_string):
    """
    Replaces relative media URLs in an HTML string with absolute
    file:// paths so that WeasyPrint can read images directly
    from the shared filesystem, bypassing HTTP entirely.
    
    Example:
        src="/media/tinymce_uploads/image.png"
        becomes
        src="file:///app/media/tinymce_uploads/image.png"
    """
    media_url = settings.MEDIA_URL  # e.g., '/media/'
    media_root = str(settings.MEDIA_ROOT)  # e.g., '/app/media'

    # This regex finds src attributes that start with MEDIA_URL
    # It handles both single and double quotes
    pattern = r'(src=["\'])' + re.escape(media_url) + r'(.*?["\'])'
    
    replacement = r'\1file://' + media_root + r'/\2'
    
    converted_html = re.sub(pattern, replacement, html_string)
    
    return converted_html