Repository navigation
v0.19.0 — TUI parity for the screenrecording iteration loop
0.19.0 — TUI parity for the screenrecording iteration loop
No migration.
The TUI can now originate the full video loop from 0.18.0 — import a
screen recording, upload it with the iteration-plan template, and browse
the resulting cue frames — plus small API/CLI surface completions.
Added
- Video import in the TUI record screen. The Import picker (ctrl+u)
accepts.mp4/.movalongside audio: the flat recordings list
includes video (marked 🎬), the browse-mode directory tree filters to
the full accepted set, and the picker hints/docs name every extension. - Iteration-plan toggle. New sticky "Iter. plan" toggle in the
record screen's toggles row (default ON, persisted). Video uploads
sendsummary_template=iteration-plan; audio uploads are unaffected
(a plan keyed to cue frames makes no sense for a meeting). vezir scribe --template—run_scribeaccepts and forwards
summary_template, mirroringsummary_preset; every upload path
(CLI, scribe, TUI, API) now has template parity.- Retry-summary with template from the TUI. The detail screen's
preset picker gains a template field (prefilled with the session's own
template, oriteration-planfor video sessions that lack one);
VezirClient.retry_summary(template=)sends it (server >= 0.18.0). - Video + template visibility.
Sessionkeeps the server'svideo
andsummary_templatefields (newis_video): video sessions show a
· 🎬badge in the sessions list and🎬 video/template:lines in
the detail meta block. - Cue-frame grouping.
cue_*.pngattachments (up to 45 on a video
session) collapse into a single "N cue frames" row in the artifacts
table; Enter opens a frame-list modal, and each frame opens through
the existing attachment viewer (OS image viewer). All other
attachments keep their own rows. Inline image preview stays deferred.
Tests
- 1121 passing (was 1110 at 0.18.0). New: video scan/label/picker tests,
_template_forrules, record-body upload kwargs, preset-picker
3-tuple + template roundtrip, retry-summary template body,
Session.from_dictvideo/template fields, detail-screen badges +
frame grouping incl. the FrameListScreen → ArtifactScreen chain.