paper-metadata

---
name: paper-metadata
description: Resolve a paper from a PDF path, exact title, DOI, or vague description, then render its OpenAlex metadata (authors, affiliations, journal, year, citation count, DOI, topics, concepts, abstract) as a clean one-page PDF. Saves alongside the source PDF or in CWD, opens in evince. Use when the user wants a quick metadata card for a paper.
---

# paper-metadata

One-page PDF metadata card from OpenAlex.

## Arguments

`<query>` — one of:
- **PDF path** — e.g. `/path/to/paper.pdf` (extracts title via pdfinfo + filename)
- **Exact title** — e.g. `"Neural Inverse Operators for Solving PDE Inverse Problems"`
- **DOI** — e.g. `10.1016/j.apm.2024.05.028` or `https://doi.org/...`
- **Vague description** — e.g. `"the wei explicit inverse paper"` or `"Mei GANs elastography 2023"` (resolved by fuzzy search)

Optional flags:
- `--out <path>` — output PDF path (default: alongside source PDF if path given, else slug-based name in CWD)
- `--no-open` — skip opening in viewer

## Pipeline

### Step 1 — Detect input mode
The helper auto-detects:
- DOI pattern → direct lookup at `https://api.openalex.org/works/doi:<DOI>`
- Existing path / `.pdf` ending → extract title via `pdfinfo`, fall back to filename slug
- Otherwise → treat as title or vague description, fuzzy-search OpenAlex

### Step 2 — Run the helper
```bash
python3 ~/.claude/skills/paper-metadata/helpers/build_metadata_pdf.py "<query>" [--out <path>] [--no-open]
```

The helper:
1. Resolves the query to an OpenAlex work record
2. Reconstructs the abstract from OpenAlex's inverted index
3. Renders a one-page PDF with reportlab: title, journal+year+citation+OA banner, authors with affiliations, identifiers (DOI / OpenAlex ID / type / language / reference count), topics, concepts, abstract, and up to 3 alternative candidates as a footer disclaimer
4. Saves the PDF and opens it in the default viewer

### Step 3 — Report
Tell the user where the PDF was saved and the resolution mode used (doi / search). If alternative candidates were listed, highlight that.

## Notes
- Uses OpenAlex (no auth, free). Sets a User-Agent. Citation counts and topic tags come straight from OpenAlex.
- Abstract reconstruction handles OpenAlex's inverted-index format.
- For ambiguous fuzzy queries, picks the top-scoring match and lists up to 3 alternatives in the PDF footer so you can see what else came up.

## Do-not-touch rules
- Read-only on the source PDF.
- Never try to scrape full text from publisher sites — only OpenAlex's structured metadata.