video-accessibility/backend/app
michael a3b4db104a feat: use Gemini's exact pause point for non-dialogue sections
When the Whisper analysis detects no speech near a Gemini-recommended
pause point, skip the full-gap-overlap algorithm and use the exact
pause point with no overlap (pause_point == resume_from).

This handles cases where Gemini chose a pause point in a silent or
music-only section of the video - there's no dialogue to buffer
around, so we simply pause and resume at the exact same point.

Three outcomes now in snap_pause_point():
1. No speech nearby → exact pause point, no overlap, no warning
2. Speech but no sentence break → warning (existing behavior)
3. Sentence break found → full-gap-overlap (existing behavior)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-28 23:13:12 -06:00
..
__pycache__ fixed middleware silent swallowing of exceptions FINALLY 2025-10-08 18:19:59 -05:00
api/v1 refactor: simplify GCS job deletion to use prefix-based cleanup 2025-12-28 10:07:43 -06:00
core feat: switch Whisper model from large-v3 to medium 2025-12-28 22:35:47 -06:00
lib initial commit 2025-08-24 16:28:33 -05:00
middleware fixed overly aggressive security algorithm 2025-10-08 17:49:20 -05:00
migrations fix: add rendering_video status to MongoDB schema validator 2025-12-27 08:40:23 -06:00
models feat: add rendering_video status for accessible video processing 2025-12-27 06:49:46 -06:00
prompts feat: improve pause point precision with sentence boundary detection 2025-12-28 14:41:12 -06:00
schemas feat: implement full-gap-overlap algorithm for AD pause insertion 2025-12-28 22:51:49 -06:00
services feat: use Gemini's exact pause point for non-dialogue sections 2025-12-28 23:13:12 -06:00
tasks fix: use dedicated whisper worker with FFmpeg dispatch pattern 2025-12-27 08:53:53 -06:00
telemetry initial commit 2025-08-24 16:28:33 -05:00
main.py feat: add video review with timestamped notes to Final Review 2025-12-26 15:30:00 -06:00