Skip to content

Tiktak#147

Open
dam2452 wants to merge 3 commits intomainfrom
tiktak
Open

Tiktak#147
dam2452 wants to merge 3 commits intomainfrom
tiktak

Conversation

@dam2452
Copy link
Copy Markdown
Owner

@dam2452 dam2452 commented Apr 5, 2026

Description

Testing

Remember to run the tests when your changes are ready. Workflows can be started from here. Use your branch.

Also, deploy and tests will automatically begin upon PR approval.

dam2452 added 3 commits April 4, 2026 16:51
Introduce TikTak feature: add TikTakHandler command (/tiktak, /tt) and responses, register the handler in sending_videos exports and subscribed permission factory, and add ClipType.TIKTAK enum. Implement TikTakProcessor to probe videos, read per-episode detection JSONs, compute person-centered 9:16 crop trajectories, and run ffmpeg (NVENC then libx264 fallback) to produce vertical clips. Persist TikTak outputs as the last clip and log success. Also add TIKTAK_DETECTION_DIR setting to configure detection file location.
Enrich EPISODE semantic search results with clip start/end times so UI can show a short preview. Added SemanticSegmentsFinder.fetch_first_dialogue_times to query Elasticsearch for the earliest start_time per episode (using a per-episode agg + top_hits). SemanticSearchHandler now calls that helper for EPISODE mode, sets SegmentKeys.START_TIME and END_TIME (END_TIME = START_TIME + SENSODCINEK_PREVIEW_DURATION_SEC), and then persists the enriched segments. The episode response formatter was updated to display a "klip od M:SS" snippet when start time is available. Also added SENSODCINEK_PREVIEW_DURATION_SEC (default 30s) to settings and adjusted imports accordingly.
Replace dual-path NVENC/software encoding logic with a single, configurable FFmpeg pipeline. Introduces codec-related constants (_CODEC, _PRESET, _CRF, _PROFILE, _LEVEL, _PIX_FMT), adjusts the crop filter (removed eval=frame), and builds one ffmpeg command with explicit video profile/level/pix_fmt options. On failure the process now raises an FFMpegException with the ffmpeg stderr snippet instead of attempting fallbacks or per-codec logging. This simplifies processing, reduces dependency on NVENC, and produces a consistent output format.
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.

1 participant