Archilyzer · Source

archilyzer

Archilyzer
git clone https://archilyzer.pages.dev/source/archilyzer.git
Log | Files | Refs | README | LICENSE

commit 626a0d055f5c6120fa278525b8cb6940cec15ff8
parent cea8c991af9aac51f5a4b57b4043c42d9ea1c1be
Author: I Mean I'm Just Saying <imeanimjustsaying@kiwifarms.st>
Date:   Fri, 28 Aug 2026 00:27:52 -0400

backfill: review touch-ups after the hand-off

Three things review caught, none of them behavioural.

transcriptSourceTarget had been inserted directly beneath the comment block
belonging to attributionTargetFor, so that comment ("The identity, minus the
per-video half…") read as if it described the new helper and
attributionTargetFor was left with none. The helper moves below
attributionTargetFor, carrying only its own comment.

The `status: "fetched"` return's ctx object literal was indented eight spaces
where the two `"failed"` call sites above it use six.

And STATE.md's "Last updated" line named commit 3 in words, because its sha did
not exist when the line was written. It says `cea8c99` now.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Diffstat:
Mcommon/controller/backfillReacquire.ts | 8++++----
Mcommon/lib/operations.ts | 14+++++++-------
Mplans/STATE.md | 2+-
3 files changed, 12 insertions(+), 12 deletions(-)

diff --git a/common/controller/backfillReacquire.ts b/common/controller/backfillReacquire.ts @@ -277,10 +277,10 @@ export async function reacquireMediaFor(opts: { return { status: "fetched", cleanup: buildCleanup(opts.videoDir, before, opts.videoId, log, { - paths: opts.paths, - channelSlug: opts.channelSlug, - platform: detectPlatform(config.url), - }), + paths: opts.paths, + channelSlug: opts.channelSlug, + platform: detectPlatform(config.url), + }), }; } diff --git a/common/lib/operations.ts b/common/lib/operations.ts @@ -823,6 +823,13 @@ function attributionApplies(files: VideoFiles): boolean { // includes the generatedAt of the diarization.json it names clusters from, and // that is a disk read — so it is added inside the one state() branch that has // already paid for the read. See AttributionProvenance.diarizationGeneratedAt. +function attributionTargetFor( + settings: SiteSettings, + method: AttributionMethod, +): AttributionFreshnessTarget { + return resolveAttributionTarget(method, settings.attribution).target; +} + // The transcript half of the per-video identity, for both lanes: which text the // names would be made from now. ZERO I/O — pickIndexTranscript reads the // already-loaded VideoFiles, and this runs per video per job start, where the @@ -836,13 +843,6 @@ function transcriptSourceTarget( return source ? { transcriptSource: source } : {}; } -function attributionTargetFor( - settings: SiteSettings, - method: AttributionMethod, -): AttributionFreshnessTarget { - return resolveAttributionTarget(method, settings.attribution).target; -} - const attributionText: Operation = { id: "attribution-text", label: "Speaker names (from the transcript)", diff --git a/plans/STATE.md b/plans/STATE.md @@ -4,7 +4,7 @@ The working memory for the local-AI derived-corpus work. Rewritten at the end of session, before context is cleared. See [`README.md`](README.md) for the protocol. **Last updated:** 2026-08-28 — **the re-acquire hand-off shipped** (`e450c2c`, `f661677`, -and the e2e/docs commit after them): re-acquired audio on a subtitle channel is no longer +`bec4775`): re-acquired audio on a subtitle channel is no longer deleted under a transcription that the auto-queue policy would have started on it. It is handed over when `autoQueue.transcription` would draw the video from `downloadedAutoSubsOnly`, kept while a transcription task is running on it, and removed