Spaces:
Sleeping
Sleeping
feat: deploy Phase 1 architecture fixes and Phase 2 dataset collection & evaluation infrastructure
632bdc4 verified MEXAR Medical Corpus — Real Data Directory
This directory is populated by running:
python backend/scripts/fetch_pubmed.py --email your_email@example.com
Contents (after fetch):
PMC<ID>.txt— one plain-text file per PubMed Central OA articlemanifest.json— PMCID → title → subdomain → source URL mapping
Target: ~220 documents across 3 subdomains (cardiology, oncology, internal medicine)
See MEXAR_Phase2_Dataset_Collection.md for full instructions.