Files
stack/dev
kert 1256ae1c04
Some checks failed
CI / skinny-install (aco) (push) Successful in 50s
CI / skinny-install (api) (push) Successful in 35s
CI / lint-test (push) Failing after 1m34s
CI / skinny-install (bcda) (push) Successful in 28s
CI / skinny-install (bib) (push) Successful in 31s
CI / skinny-install (bls) (push) Successful in 24s
CI / skinny-install (pfs) (push) Successful in 31s
CI / skinny-install (rex) (push) Successful in 35s
CI / skinny-install (ccw) (push) Successful in 31s
CI / skinny-install (cli) (push) Successful in 36s
CI / skinny-install (cms) (push) Successful in 27s
CI / skinny-install (conf) (push) Successful in 25s
CI / skinny-install (perf) (push) Successful in 34s
Deploy / build-scan-report (push) Successful in 3m58s
feat: headless PDF retrieval — Unpaywall, PMC, Semantic Scholar, SciHub (refs #257)
Terminal-based PDF retrieval pipeline that does not require Zotero GUI:
- Phase 1: Unpaywall (legal, 100K/day, ~4% coverage for wound care lit)
- Phase 2: PMC ID Converter + OA PDF download (batch 200 DOIs)
- Phase 3: Semantic Scholar batch endpoint (500 DOIs/request)
- Phase 4: SciHub with configurable proxy (--proxy socks5://host:port)

All sources attach PDFs directly to Zotero items in the SQLite DB.

Current blocker: SciHub mirrors blocked from this network (DNS + timeout).
OA coverage is ~4% for this corpus. Need SOCKS proxy or VPN for SciHub.
Script is ready — just needs: SCIHUB_PROXY=socks5://host:port
2026-03-25 18:01:12 -04:00
..