Repository navigation
S3 — Forecast-file identity invariant (C-25) #54
Description
Epic: #51 · S3 · depends on #61 (S0) · story implementation
Problem (C-25)
The forecast is selected by category="forecast" alone (unfao.py:98, newest-wins) — no loa/name/run filter. A stray upload silently ships the wrong predictions to FAO.
Build shape (per the epic design contract)
delivery/identity.py(representation-free):assert_forecast_identity(selected: dict, expected: dict) -> Noneover plain dicts (name/loa/etc.). No pandas.unfao/extraction.py:file_metadata(downloaded) -> dict.- Manager: constrain the selection to the configured ensemble (name +
loa="pgm"), or validate the selected file's metadata post-download via the invariant; fail loud on mismatch; log the chosen file identity.
Acceptance criteria
-
delivery/identity.pypure + pandas-free; unit-tested on dicts. - A decoy
category="forecast"file with a different name is rejected; the correct one selected. - Mismatch fails loud, naming expected vs found.
Files
views_postprocessing/delivery/identity.py (new), unfao/extraction.py (extend), unfao/managers/unfao.py (_read_forecast_data), tests/.
Activity
[-]S3 — Verify forecast-file identity before enrichment (C-25)[/-][+]S3 — Forecast-file identity invariant (C-25)[/+]feat(delivery): input-integrity invariant S3 — forecast-file identity…
Polichinel commentedon Jun 26, 2026
Delivered and merged to
developmentvia #64 (input-integrity sprint S0–S6). ruff clean, 126 passed / 44 xfailed. Closing — the auto-close keyword did not fire because #64 merged intodevelopment, not the default branch. Live-run validation of the residual contracts (C-25 identity name/loa, C-43 enrichment equivalence) is tracked in the risk register and exercised by the go-global run (views-models#127).