Close moot backlog items; refresh verified commands and suite timing

Co-Authored-By: Claude Opus 5.5 <[email protected]>
This commit is contained in:
Lucas Winther
2026-10-30 14:37:24 +01:00
co-authored by Claude Opus 5.5
parent 978743a36e
commit 80f8004963
3 changed files with 19 additions and 11 deletions
+1 -1
View File
@@ -13,7 +13,7 @@ Install: `bun install --frozen-lockfile`
Dev: `bun dev`
Test one file: `bun test tests/export.test.ts` | One case: `bun test tests/models.test.ts -t "dtype"`
Test (fast, ~6 s): `bun test --path-ignore-patterns='tests/live-real-model.test.ts'`
Test all: `bun test` — ~19 min: `tests/live-real-model.test.ts` downloads and runs real Whisper
Test all: `bun test` — ~22 min: `tests/live-real-model.test.ts` downloads and runs real Whisper
weights whenever huggingface.co is reachable. Run it when touching the live pipeline or decoding.
Build: `bun run build` (rewrites `dist/`)
Opt-in real sidecar test (needs `pip install -r server-py/requirements.txt`): `FW_INTEGRATION=1 bun test tests/transcribe.test.ts`
+7 -4
View File
@@ -6,19 +6,22 @@ Change this page and every tool file in the same commit. (No Cursor/Copilot file
## Verified commands
Run on 2026-10-29 (Bun 1.3.14). Only `worked` rows may appear in a tier, a tool file, or `AGENTS.md` Commands.
Run on 2026-10-29, updated 2026-10-30 (Bun 1.3.14). Only `worked` rows may appear in a tier, a tool file, or `AGENTS.md` Commands.
| Command | Result |
|---|---|
| `bun install --frozen-lockfile` | worked |
| `bun test tests/export.test.ts` | worked — 28 pass |
| `bun test tests/models.test.ts -t "dtype"` | worked — 5 pass, 14 filtered |
| `bun test --path-ignore-patterns='tests/live-real-model.test.ts'` | worked — 239 pass, 1 skip, ~6 s |
| `bun test` | worked — 243 pass, 1 skip (opt-in `FW_INTEGRATION`), ~19 min (after D-039) |
| `bun test --path-ignore-patterns='tests/live-real-model.test.ts'` | worked — 414 pass, 11 skip, ~6 s (2026-10-30) |
| `bun test` | worked — 422 pass, 11 skip (opt-in suites), ~22 min (2026-10-30, after spec 001) |
| `bun run build` | worked |
| `bun run icons` | worked |
| `bun dev` | worked — HTTP 200 on `/` |
| `FW_INTEGRATION=1 bun test tests/transcribe.test.ts` | not run — faster-whisper not installed here |
| `FW_INTEGRATION=1 bun test tests/transcribe.test.ts` | worked — 34 pass with `SIDECAR_PYTHON=/home/user/venvs/faster-whisper/bin/python` (2026-10-30) |
| `MT_INTEGRATION=1 bun test tests/translate-real-model.test.ts` | worked — run by the spec 001 translate workstream (2026-10-30); not re-run in the integration session |
| `PARAKEET_REAL=1 bun test tests/parakeet-real-model.test.ts` | worked — run by the spec 001 Parakeet workstream (2026-10-30); not re-run in the integration session |
| `bun scripts/bench-translate.ts 1 8` | worked (2026-10-30) |
| `bun start` | not run — documented in README |
| `bun run typecheck` | worked — 0 errors (tsc 7.0.2, 57 files); exits 1 on a planted type error |
| `google-chrome --headless=new --no-sandbox --disable-gpu --screenshot=… <url>` | worked (UI screenshots; not in any tier — ad hoc) |
+11 -6
View File
@@ -439,13 +439,16 @@ Found while building transcript translation (D-045), 2026-10-29:
live WebSocket answer 400 via `translateNotSupported` (src/server/transcribe.ts).
Found in spec 001 workstream 6 (D-046), open:
- [ ] Transformers.js 4.2 has no beam search: `num_beams > 1` in a model's
- [x] Transformers.js 4.2 has no beam search: `num_beams > 1` in a model's
`generation_config.json` (edda-v0.2 ships 5) silently decodes greedily
and does a full-vocabulary top-k per step (D-046). Not fixed: it would
mean overriding `num_beams: 1` in the worker, or a patch.
- [ ] On `da_festuge`, neither Danish turbo fine-tune beats the stock turbo
mean overriding `num_beams: 1` in the worker, or a patch. — Moot: edda-v0.2
was not published or added to the catalog (D-046, owner 2026-10-30).
- [x] On `da_festuge`, neither Danish turbo fine-tune beats the stock turbo
(D-046). Judging them needs a longer, harder Danish reference clip than
this repo has.
this repo has. — Closed: the owner kept only whisper-small-danish (D-046);
revisit the turbo conversions (kept in /home/user/hf-publish) only with
a harder Danish clip.
Found while building the long-form pipeline (spec 001 Goal 3, D-049), 2026-10-29:
- [x] "Auto-detect" on the default (chunked) file path does not detect: with no
@@ -453,9 +456,11 @@ Found while building the long-form pipeline (spec 001 Goal 3, D-049), 2026-10-29
English" and decodes Danish audio as an English translation (da_festuge WER
0.94 vs 0.14 with `da`). Long-form ("Skip silence") detects once and pins
the language; the chunked path is unchanged.
- [ ] Spec 001's measurement fallback mentions a hint list of the names in
- [x] Spec 001's measurement fallback mentions a hint list of the names in
`en_alien`'s reference; `tests/meetingFixture.ts` returns audio + reference
only. Add the list when the vocabulary hint (workstream 2) lands.
only. Add the list when the vocabulary hint (workstream 2) lands. — Superseded:
D-048 measured hints directly on en_alien with short and realistic name
lists; the synthetic fixture adds nothing there.
— Fixed for file passes: detect once on the first 30 s and pin (D-053).
- [ ] "Auto-detect" on **live** passes (browser engine) still decodes as English:
each live pass would need its own detection (an extra encoder run per