kert e9801f599e feat(opps): recover CY2014-2020 Addendum B + fix ingest finder ...

CMS purged the pre-2021 OPPS addenda in its 2023 site restructure, so
opps.addendum_b only held CY2021-2026. Recover CY2014-2020 from the
Internet Archive Wayback Machine (the `web/{ts}id_/{url}` raw-binary form)
and wire the URLs into download_opps_files.py, with retry/backoff since
Wayback rate-limits.

Key subtlety: the payment data (Relative Weight + Payment Rate per HCPCS)
lives in the real "Addendum B - Final OPPS Payment by HCPCS Code", which
ships inside the COMBINED addenda ZIP — NOT the "Data Addendum B /
2-Times-Rule" file (SI/APC/comment-indicator only, no payment). Point the
older years at the combined ZIPs (CY2016's only surviving snapshot is
2015-vintage but serves the real Addendum B).

Fix the ingest file-finder accordingly:
- match "Addendum B" separator-insensitively (CMS mixes "Addendum B.xlsx"
  and "CMS-1613-FC-Addendum-B.xlsx" across years);
- exclude the "Data Addendum B" file (normalised "dataadd") so it never
  shadows the payment file.

After this, all 13 years ingest and reconcile 100% exact against Addendum B.
Data files remain gitignored; download_opps_files.py + ingest_opps.py
reproduce them.

2026-07-08 14:48:17 -04:00

7.5 KiB

Raw Blame History

View Raw