import-papers

Import Papers Skill

Process all PDFs in research/dump/, create bibtex entries, and move them to research/papers/.

Instructions

For each PDF in research/dump/:

1. Extract metadata using pdftotext (first 30 lines only to avoid context overload):

   pdftotext "filename.pdf" - 2>&1 | head -30
   

2. Identify:

   • Title (full, not truncated)
   • Authors (first few primary authors)
   • Year (infer from content or use the current year for recent papers)
   • Venue (if clearly stated: conference/journal)

3. Generate citation key using format: firstauthorYEARkeyword

   • Use lowercase only
   • First author's last name
   • Year
   • One descriptive keyword from title
   • Examples: ren2015faster, choi2024neuromamba, park2024processing

4. Create bibtex entry matching existing format in research/papers.bib:

   @article{citationkey,
    author = {Author One and Author Two and Author Three},
    ejournal = {arXiv},
    title = {Full Paper Title},
    url = {https://arxiv.org/abs/XXXX.XXXXX},
    year = {YYYY}
   }
   

   Use @inproceedings if it's a conference paper (like NIPS, ICML, etc.) with:

   @inproceedings{citationkey,
    author = {Authors},
    booktitle = {Conference Name},
    title = {Title},
    year = {YYYY}
   }
   

5. Append to papers.bib:

   cat >> research/papers.bib << 'EOF'
   
   @article{citationkey,
    ...
   }
   EOF
   

6. Move and rename PDF to research/papers/:

   mv "research/dump/original.pdf" "research/papers/citationkey.pdf"
   

7. Handle duplicates: If two PDFs are identical (same title), keep only one.

8. Verify after processing all files:

   • List moved files: ls -1 research/papers/ | tail -10
   • Check dump is empty: ls -1 research/dump/*.pdf 2>/dev/null || echo "empty"

9. Run bibify.py to update the changelog and sync with Google Drive:

   ./bin/bibify.py
   

   If the output contains Missing notes for, run bibify a second time — newly uploaded papers need a second pass to generate their note stubs.