second-brain/tests
Travis Herbranson 5131f3ccf6 web+cli: edit-mode for source metadata + retry/re-run actions
Adds per-record edit mode and retry to the source detail page, plus a
CLI `second-brain retry <id>` for parity.

Edit mode exposes domain (select), title, focus on /sources/{id}.
URL stays read-only — it's the UNIQUE dedupe key and the embeddings
identity tuple. Save goes through a new service helper
`update_source_metadata` so the web route stays a thin wrapper.

Retry exposes two modes, both available on any record:
  - full: status -> PENDING (re-pull + re-process)
  - extract: status -> TRANSCRIBED (re-extract only, requires an
    existing transcript; offered when source_type=video and a
    transcript is present, to avoid the expensive re-download path)

Both modes clear claimed_by/claimed_at and error_message so the queue
dance picks it up cleanly.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-25 16:08:26 -04:00
..
__init__.py project init 2026-05-22 19:08:22 -04:00
test_claim.py deploy/tower + claim race + transcribe smoke tests 2026-05-25 10:20:46 -04:00
test_playlist.py playlists: queue-time YouTube fan-out via yt-dlp extract_flat 2026-05-25 13:59:35 -04:00
test_smoke_embedding.py test: smoke test reads embedding model from config, not hardcoded literal 2026-05-24 22:58:30 -04:00
test_source_edit_retry.py web+cli: edit-mode for source metadata + retry/re-run actions 2026-05-25 16:08:26 -04:00
test_transcribe.py transcripts: capture segment-level output into a new JSONB column 2026-05-25 13:40:49 -04:00