MushroomObserver / mushroom-observer / 28201916460

98%

main: 98%

Pull #4601

github

Split residual into copy (historic) vs remote_manual (iNat-side) (#4565)

The materializer treated the whole residual bucket as `copy`, but that conflated
the defunct iNat-team copy service with manual links created on the iNat side
(field 5005 set there, no MO-side trace) — including recent obs that can't be
copies. The copy service ran in multi-user machine bursts through 2021; 2022+
"copies" are individual users bulk-linking their own obs (e.g. johnplischke 1393
in one day).

- Add `remote_manual` (6) enum value for iNat-side manual links.
- classify reads the iNat obs created_at and only assigns `copy` to residual
  rows predating COPY_CUTOFF (2022-01-01); newer residual + import-extras ->
  remote_manual. `manual` (0) is now MO-side only (notes URL).

Dry run after the fix: copy 73314 (was 77936, the historic set), remote_manual
5160 (new), manual 1872, mirror 26598 — same 106944 total, correctly relabeled.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

10 of 10 new or added lines in 2 files covered. (100.0%)

49659 of 50473 relevant lines covered (98.39%)

726.47 hits per line