Marker PDF-to-Markdown Converter

Convert PDFs to Markdown while preserving LaTeX formulas and document structure. Uses the marker_single CLI from the marker-pdf package.

Dependencies

• marker_single on PATH (pip install marker-pdf if missing)
• Python 3.10+ (available in the task image)

Quick Start

from scripts.marker_to_markdown import pdf_to_markdown

markdown_text = pdf_to_markdown("paper.pdf")
print(markdown_text)

Python API

• pdf_to_markdown(pdf_path, *, timeout=600, cleanup=True) -> str
  • Runs marker_single --output_format markdown --disable_image_extraction
  • cleanup=True: use a temp directory and delete after reading the Markdown
  • cleanup=False: keep outputs in <pdf_stem>_marker/ next to the PDF
  • Exceptions: FileNotFoundError if the PDF is missing, RuntimeError for marker failures, TimeoutError if it exceeds the timeout
• Tips: bump timeout for large PDFs; set cleanup=False to inspect intermediate files

Command-Line Usage

# Basic conversion (prints markdown to stdout)
python scripts/marker_to_markdown.py paper.pdf

# Keep temporary files
python scripts/marker_to_markdown.py paper.pdf --keep-temp

# Custom timeout
python scripts/marker_to_markdown.py paper.pdf --timeout 600

Output Locations

• cleanup=True: outputs stored in a temporary directory and removed automatically
• cleanup=False: outputs saved to <pdf_stem>_marker/; markdown lives at <pdf_stem>_marker/<pdf_stem>/<pdf_stem>.md when present (otherwise the first .md file is used)

Troubleshooting

• marker_single not found: install marker-pdf or ensure the CLI is on PATH
• No Markdown output: re-run with --keep-temp/cleanup=False and check stdout/stderr saved in the output folder