Automate UMLS data download and artifact build pipeline #61
Description
Summary
Script to authenticate with UMLS API, download release files, filter to relevant subsets, and rebuild automata artifacts.
Details
- Replaces manual download process
- Complements/supersedes feat: Python backup script to download SNOMED CT without UMLS API key #35 (Python SNOMED download script)
- UMLS API supports programmatic access with API key authentication
- Key files to download: MRCONSO.RRF, MRREL.RRF, MRSTY.RRF
- Filter to relevant SABs (SNOMEDCT_US, RXNORM, NCI, NDFRT, ICD11)
- Rebuild umls_automata.bin.zst artifact after download
- Should be runnable in CI for periodic updates
Acceptance Criteria
- Script authenticates with UMLS API using stored credentials
- Downloads and extracts required RRF files
- Filters to relevant vocabularies
- Rebuilds automata artifact
- Can run headless in CI environment
- Documented in README
See also: #35
Priority: P3