Repository navigation
π Inbox: fact is a status, not a button; machine transcripts become knowledge - #18390
Conversation
β¦nowledge Remove the π Fact button from the Inbox action row. What the item's Knowledge fact holds is now a non-interactive status pill (Pending text, In knowledge (machine text), In knowledge), and opening the fact is a plain link in the metadata line. A finished machine transcription is now written into the item's fact, fenced as untrusted data and marked attr_<x>_content: machine with attr_transcript_provider, keeping source attrs, metadata and highlights. Connector syncs that still find no upstream text leave a machine fact alone; Plaud and meeting refetch keep re-reading machine facts so the source's own text replaces it when it arrives. Signed-off-by: Andy Anderson <andy@clubanderson.com>
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of
|
|
Hive status β updated in place; edits do not notify. Full history is in the Activity Monitor. @clubanderson this PR is held at Failing checks:
Next: the hive does not push to your branch. Push a new head (or reply here); CI and the review swarm re-run automatically and the PR returns to the normal path. |
Mutation testinggremlins found 10 surviving mutant(s) in changed files: the code ran under tests, but changing it did not fail any test.
18 mutant(s) in changed files are not reached by any test (NOT COVERED). A maintainer may apply the
Verdict: |
β¦t the pill Signed-off-by: Andy Anderson <andy@clubanderson.com>
|
[APPROVALNOTIFIER] This PR is APPROVED Approval requirements bypassed by manually added approval. This pull-request has been approved by: The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
What
The "π Fact" control is gone from the Inbox action row. It looked like an action but did one thing:
inboxOpenFact(slug, layer)jumped to Knowledge and opened the item's fact (kbReadFact). It said nothing about what that fact holds. Now:kc-pill,role="status", no hover or pointer) sits next to the "Machine transcript (β¦)" pill. It reads Pending text (metadata only), In knowledge (machine text), or In knowledge (the source's own text). An item that is listed live and has no fact yet shows no pill. It comes from a newfact_statefield on the item.Machine transcripts now reach Knowledge. Before this change, a whisper, Gemini or Deepgram transcript was stored only under
connector-state/transcripts/<source>/<id>.json. Only the Inbox read it, through theContent()fallback. The Knowledge fact stayed in its pending shape (title, date, duration, highlights and the pending note), so knowledge search, nous and agents never saw the words. Now:writeMachineFact(src/pkg/inbox/factwriteback.go) rewrites the item's fact:attr_<x>_content: machine,attr_transcript_provider: <provider>/<model>,attr_transcript_languageandattr_transcribed_at.## Transcript (machine: β¦)section ahead of the highlights. The section is fenced in theuntrusted-inbox-dataenvelope (WrapUntrusted).FactWriter.Writewill not overwrite amachinefact with a page that is stillpending. So the next sync never resets it to pending.machinefacts. Upstream text replaces the machine transcript once it exists (ready).Tests
src/pkg/inbox/factwriteback_test.go:dashboard/test/inbox-item.test.js: the row has no Fact button, the metadata link opens the fact, and the pill shows each state.src/pkg/knowledge/connector/fact_content_test.go:Docs
src/docs/inbox-voice-sources.md: audio-only (pending, with an example fact), after machine transcription, and after upstream transcription.src/docs/inbox-transcription.mdis updated.changelog.d/fixed-inbox-fact-indicator.md.static/index.htmlis regenerated.Coordination
π€ Generated with Claude Code