dev-document_builder

Run the convert_md_to_docx.py script with your Markdown file. It preprocesses the content by removing image alt text and horizontal rules, resolves relative image paths via Pandoc's resource-path option, then inserts a formatted cover page and Word table of contents.

Use the pdf_section_split.py script, which automatically reads the PDF's built-in TOC bookmarks and splits it into per-section files organized by chapter. For PDFs without bookmarks, it falls back to scanning running page headers.

Use pdfplumber for table extraction because it detects cell boundaries and table structure accurately. For plain text extraction, PyMuPDF (fitz) is faster, and pypdf handles merging and splitting operations.

Pandoc renders Markdown image alt text as visible captions in Word. Remove the alt text by using empty brackets like  instead of ; the conversion script does this automatically during preprocessing.

Standard parsers cannot read scanned pages because they contain only images. Use pytesseract with pdf2image to OCR each page, or use the pdf_to_image_md.py script which integrates Pix2Text for text and formula recognition.

Use the convert_docx_mammoth.py script, which converts DOCX to Markdown via the mammoth library and extracts embedded images to an images folder. For batch conversion of a whole directory, use batch_convert.py with Pandoc.