feat(translation): tunable acceptance floor (0.90) + per-post sticky override (#155) #210

Merged
bvandeusen merged 1 commits from dev into main 2026-07-10 22:47:45 -04:00
Owner

Makes the translation gate strict-but-tunable and adds a per-post manual override, so the operator can roll back the confidently-mis-flagged translations the old fixed 0.80 floor couldn't catch.

Why: a probe (fresh == cached, verified) showed Interpreter still confidently mis-detects short ASCII English like "… WIP Part 1" as German at 0.86 — above the old floor — so it was accepted and re-translate reproduced it. Confidence alone can't separate the 0.86 collision (genuine German lands there too), and single-word mis-flags sit at a confident 1.0 no floor catches.

Two levers:

  • Acceptance floor is now a live Settings value (ImportSettings.translation_min_confidence, default 0.90, in the Translation card) — tunable with no redeploy. The gate reads it per-run.
  • Per-post sticky override (Post.translation_override: auto / force / original). force stores a translation even below the floor; original keeps the original and clears any stored translation. The sweep honors it every run and _reset_translations skips original, so it survives a Re-translate-all. POST /api/posts/<id>/translation-override applies it immediately (translate now if the service is up, else queue). UI: PostTranslationControl on the posts-feed card.

Rollback recipe this enables: raise the floor → "Re-translate all" → the 0.86 mis-flags are rejected and restored to the original; force / keep-original handle the residual.

Migration 0084 — adds both columns (+ a CHECK on the override). Feed + provenance serializers expose translation_override.

Commit aea2701. CI green on dev tip (run 2122: lint + unit + integration w/ migration 0084 + frontend build). +581/−54 across 16 files.

Post-merge: curator :latest redeploy runs migration 0084 automatically.

🤖 Generated with Claude Code

https://claude.ai/code/session_01CgZP9v2otxVJymiYsnVuMy

Makes the translation gate strict-but-tunable and adds a per-post manual override, so the operator can roll back the confidently-mis-flagged translations the old fixed 0.80 floor couldn't catch. **Why:** a probe (fresh == cached, verified) showed Interpreter still confidently mis-detects short ASCII English like "… WIP Part 1" as German at 0.86 — above the old floor — so it was accepted and re-translate reproduced it. Confidence alone can't separate the 0.86 collision (genuine German lands there too), and single-word mis-flags sit at a confident 1.0 no floor catches. **Two levers:** - **Acceptance floor is now a live Settings value** (`ImportSettings.translation_min_confidence`, default **0.90**, in the Translation card) — tunable with no redeploy. The gate reads it per-run. - **Per-post sticky override** (`Post.translation_override`: auto / force / original). `force` stores a translation even below the floor; `original` keeps the original and clears any stored translation. The sweep honors it every run and `_reset_translations` skips `original`, so it survives a Re-translate-all. `POST /api/posts/<id>/translation-override` applies it immediately (translate now if the service is up, else queue). UI: `PostTranslationControl` on the posts-feed card. **Rollback recipe this enables:** raise the floor → "Re-translate all" → the 0.86 mis-flags are rejected and restored to the original; force / keep-original handle the residual. **Migration 0084** — adds both columns (+ a CHECK on the override). Feed + provenance serializers expose `translation_override`. Commit `aea2701`. CI green on dev tip (run 2122: lint + unit + integration w/ migration 0084 + frontend build). +581/−54 across 16 files. **Post-merge:** curator `:latest` redeploy runs migration 0084 automatically. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01CgZP9v2otxVJymiYsnVuMy
bvandeusen added 1 commit 2026-07-10 22:47:39 -04:00
feat(translation): tunable acceptance floor (0.90) + per-post sticky override (#155)
CI / lint (push) Successful in 3s
CI / frontend-build (push) Successful in 18s
CI / backend-lint-and-test (push) Successful in 27s
CI / integration (push) Successful in 3m52s
aea2701c28
The gate at a fixed 0.80 couldn't catch the real pain: Interpreter (fresh ==
cached, verified by probe) confidently mis-detects short ASCII English like
"... WIP Part 1" as German at 0.86 — above the floor — so it was accepted and a
re-translate reproduced it. Confidence alone can't separate the 0.86 collision
(genuine German lands there too), and single-word mis-flags sit at a confident
1.0 no floor catches.

Two operator-approved levers:

- Acceptance floor is now a live Settings value (ImportSettings.
  translation_min_confidence, default 0.90; surfaced in the Translation card), so
  it's tunable without a redeploy. _accept takes the threshold as a parameter.

- Per-post sticky override (Post.translation_override: auto/force/original).
  'force' stores a translation even below the floor (rescue a skipped
  legit-foreign title); 'original' keeps the original and clears any stored
  translation (kill a confident mis-flag no floor catches). The sweep honors it
  on every run and _reset_translations skips 'original', so the choice survives a
  Re-translate-all. POST /api/posts/<id>/translation-override applies it
  immediately (translate now when the service is up, else queue for the sweep).
  UI: PostTranslationControl on the posts-feed card.

Migration 0084 (both columns + a CHECK on the override). The feed + provenance
serializers expose translation_override.

With a stricter floor the rollback finally works: raise it -> Re-translate all ->
the 0.86 mis-flags are rejected and restored to the original; force /
keep-original handle the residual either way.

Tests: gate thresholds against the param (0.86 rejected at 0.90, explicit-floor
cases); sweep force/original + re-translate-skips-original; override endpoint
(validation, original clears, force queues when disabled, feed exposes it);
settings min_confidence default/save/validate.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CgZP9v2otxVJymiYsnVuMy
bvandeusen merged commit 77bfc32b02 into main 2026-07-10 22:47:45 -04:00
Sign in to join this conversation.
No Reviewers
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: bvandeusen/FabledCurator#210