Skip to content

feat: jump to a speaker's turns from the Speakers panel (#145) - #147

Open
raphaelm22 wants to merge 2 commits into
ObscureAintSecure:masterfrom
raphaelm22:feat/speaker-jump
Open

raphaelm22 wants to merge 2 commits into
ObscureAintSecure:masterfrom
raphaelm22:feat/speaker-jump

Conversation

@raphaelm22

Copy link
Copy Markdown

Closes #145

What

Each speaker ID in the Speakers panel is now a link to that speaker's turns in the transcript.

  • Click: scroll to the speaker's first turn (a run of consecutive segments), with the line before it for context, and highlight it for 1.5 s.
  • Click again: next turn, counter "2/12" next to the ID; wraps after the last.
  • Click another speaker: starts at their first turn (the hard one to find when someone barely spoke).
  • Tooltip: number of turns and talk time.

speaker_turns / next_turn / _speaker_summary are plain functions in speaker_name_panel.py; TranscriptViewer._jump_to_speaker does the scrolling. The flash never overrides the playback highlight and is dropped when the transcript is rebuilt.

Tests

13 new tests in tests/test_speaker_name_panel.py, including _jump_to_speaker against a stub viewer. Full suite: 421 passed.

Checked with an offscreen TranscriptViewer on a real 40-minute transcript (890 segments): the quietest speaker's first turn is segment 518 (20:32), reached in one click, "1/12".

🤖 Generated with Claude Code

Someone who spoke for a few seconds in a long meeting is hard to find
in the transcript, which is exactly when you need to hear them to name
them. Each speaker ID in the Speakers panel is now a link: clicking it
scrolls to that speaker's first turn, and clicking again walks through
the following ones ("3/12"), wrapping at the end. The turn is briefly
highlighted, and the tooltip shows the speaker's turns and talk time.
style.qss gives every QPushButton a filled background, which made the
speaker IDs look like buttons.
@raphaelm22

Copy link
Copy Markdown
Author

Ready for review. Checked in an offscreen viewer on a real 40 min transcript (890 segments); not yet in daily use.

@raphaelm22
raphaelm22 marked this pull request as ready for review October 22, 2026 13:28
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Jump to a speaker's lines from the Speakers panel

1 participant