devrajsinh2012's picture
feat: deploy Phase 1 architecture fixes and Phase 2 dataset collection & evaluation infrastructure
632bdc4 verified
|
Raw
History Blame Contribute Delete
500 Bytes

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 article
  • manifest.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.