[ImgBot] Optimize images - #1
Open
imgbot[bot] wants to merge 1 commit into
Open
Conversation
*Total -- 7,655.77kb -> 6,967.33kb (8.99%) /app/assets/trayIconTemplate@2x.png -- 2.85kb -> 1.74kb (38.98%) /app/assets/trayIconRecordingTemplate@2x.png -- 2.89kb -> 1.79kb (38.21%) /website/public/readme.png -- 755.56kb -> 492.84kb (34.77%) /app/assets/trayIconTemplate.png -- 1.28kb -> 0.87kb (32.42%) /app/assets/trayIconRecordingTemplate.png -- 1.33kb -> 0.90kb (32.18%) /.github/assets/transcription_model_selector.png -- 35.44kb -> 26.64kb (24.83%) /docs/logo/dragonfly.png -- 358.07kb -> 278.79kb (22.14%) /website/public/dragonfly-logo-512.png -- 358.07kb -> 278.79kb (22.14%) /app/renderer/src/assets/app-icon.png -- 358.07kb -> 278.79kb (22.14%) /.github/assets/keep_recordings.png -- 23.72kb -> 19.66kb (17.11%) /website/public/stenoai-logo.svg -- 0.78kb -> 0.71kb (8.84%) /website/public/demo.gif -- 5,733.45kb -> 5,561.61kb (3%) /website/public/logos/aws.svg -- 0.27kb -> 0.26kb (2.19%) /website/public/logos/rutgers.svg -- 0.27kb -> 0.27kb (2.14%) /website/public/logos/european-union.svg -- 0.28kb -> 0.27kb (2.09%) /website/public/logos/hashicorp.svg -- 0.35kb -> 0.35kb (0.83%) /website/public/logos/deliveroo.svg -- 0.59kb -> 0.59kb (0.49%) /docs/logo/wordmark-dark.svg -- 6.75kb -> 6.74kb (0.13%) /docs/logo/wordmark.svg -- 6.75kb -> 6.74kb (0.13%) /website/public/logos/tesco.svg -- 2.42kb -> 2.42kb (0.12%) /docs/logo/mark.svg -- 6.56kb -> 6.56kb (0.07%) Signed-off-by: ImgBotApp <ImgBotHelp@gmail.com>
Optic00
added a commit
that referenced
this pull request
Jun 30, 2026
…tenolabs#207) (stenolabs#247) * fix: rescue live transcript when batch transcription returns empty (stenolabs#207) When the batch transcription of a recording returns empty or fails, fall back to the live transcript that was accumulated during the recording session rather than saving "No speech detected". Gap 1 — live transcript fallback: - Electron snapshots live segments at stop time, writes to a temp file, and passes it to process-streaming via --live-transcript - Python uses the live transcript when batch is empty (sentinel match) or transcription_failed; marks the meeting as is_live_transcript=true and keeps the audio file for future reprocessing - Per-instance drain promise (on close, not exit) with 8 s safety-net ensures the final utterance is captured before the snapshot - Quick stop→start guard: stale sidecar stdout handler exits early when a newer process has taken over; kill watchdog bound to the concrete proc instance Gap 2 — empty result on energetic channel is a failure: - transcribe_audio() sets transcription_empty=True before substituting the silence sentinel, so transcribe_diarised() can distinguish a truly empty result from short speech - Only fails the meeting when no channel produced usable text; one channel with text + one noisy empty channel still succeeds * fix: use UTF-8 for transcript writes and rescue short live transcripts - _write_transcript_file: open() now always uses encoding='utf-8' so multilingual transcripts are safe on Windows (cubic P2 #1) - live-transcript fallback: drop the 100-char minimum-length gate; any non-whitespace live content is worth rescuing when batch fails or returns only silence — a brief session should never be lost (cubic P2 #2) * test(e2e): model-free T2 spec for the stenolabs#207 live-transcript rescue Drives the real bundled backend's process-streaming with a silent stereo WAV (batch transcription returns the silence sentinel with no model loaded) plus --live-transcript, and asserts the rescue lands on disk: is_live_transcript: true in the note, the live text written to the transcript file (not the sentinel), and the audio preserved for a later retry. Extends makeWav with channels/silence so the energy gate skips both channels (the model-free trigger). Summarisation goes to the mock Ollama, so it runs in the fast model-free t2 lane, not @pipeline.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Beep boop. Your images are optimized!
Your image file size has been reduced by 9% 🎉
Details
📝 docs |
repo | 🙋🏾 issues | 🏪 marketplace
~Imgbot - Part of Optimole family