S5 — Structured delivery provenance invariant (C-15, supersedes #12)

[Polichinel]

Epic: #51 · S5 · depends on #61 (S0); uses #53 counts · supersedes #12 · story implementation

Problem (C-15)

Upload metadata lacks structured provenance (lookup version is only in a free-text description; no region, cell/unmapped counts, fill count). FAO can't audit what produced their data.

Build shape (per the epic design contract)

• delivery/provenance.py (representation-free): build_provenance(...) -> dict from primitives — lookup_version, region, expected_cell_count, actual_cell_count, unmapped_count (+ fill_count if S2 provides it). No pandas.
• Manager: attach the dict to both dsm.upload_data(...) metadata payloads (_save).

Acceptance criteria

• delivery/provenance.py pure + pandas-free; unit-tested on the metadata dict.
• Both historical + forecast uploads carry the structured fields, sourced from the enricher + S1 coverage (not hardcoded).
• Add shapefile version and error count to upload metadata (C-15) #12 closed as superseded (already done).

Files

views_postprocessing/delivery/provenance.py (new), unfao/managers/unfao.py (_save), tests/.

[Polichinel]

Delivered and merged to development via #64 (input-integrity sprint S0–S6). ruff clean, 126 passed / 44 xfailed. Closing — the auto-close keyword did not fire because #64 merged into development, 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).