joaniberian9588 Paper PDF to Markdown
Convert verified paper.pdf bundles produced by paper-pdf-download. Preserve the PDF even when conversion fails.
If the console-script directory is not on PATH, replace paper-pdf in the examples with python -m paper_pdf.
Local-first workflow
Run
paper-pdf doctorand confirmmineru_commandis available.Use one directory batch so MinerU does not restart its model for every paper:
paper-pdf convert ./library --mode local --backend pipelineConfirm each successful bundle contains
paper.md, any referenced files underassets/mineru/, andconversion.status=successinmanifest.json.Report PDF completion and Markdown completion separately. A verified PDF with failed conversion is a partial outcome, not a lost download.
The end-to-end command is:
paper-pdf run --file dois.txt --output ./library --mode local
Cloud API privacy gate
Never select cloud conversion merely because local MinerU is missing or slow. API mode may upload subscription PDFs to a third party and therefore requires the user's explicit permission plus the command-line gate:
paper-pdf convert ./library --mode api --allow-upload
If consent is absent, stop before upload and offer local MinerU installation instead.
Read mineru.md for installation, batch reconciliation, output validation, and troubleshooting.