twat-video 2.1.7

Video processing helpers for the twat plugin ecosystem

twat-video

twat-video is the video domain package for the twat plugin ecosystem. It builds reusable, testable ffmpeg command wrappers and delegates AI video generation to twat_genai when such provider engines are available.

Font-organizer preservation

This repository used to package font_organizer. The canonical copy now exists in plugins/repos/twat_font as twat_font, with the same organizer/core modules and CLI surface. The old source tree is left in this repo only as historical material; the build now packages src/twat_video.

Requirements

Install ffmpeg and ffprobe for real media processing. Tests exercise command construction with dry_run=True and do not require media binaries.

Python API

from twat_video import crop_scale, import_audio, split_segment

crop_scale("input.mp4", "square.mp4", crop="1080:1080:420:0", scale="720:720")
split_segment("input.mp4", "clip.mp4", start=12.5, duration=4.0)
import_audio("silent.mp4", "voice.wav", "dubbed.mp4")

Implemented deterministic helpers:

• ffmpeg boundary: run_command, run_ffmpeg, probe_video.
• video geometry/timing: crop_scale, change_fps, split_segment, reverse_video.
• composition: merge_by_gap, ken_burns.
• audio/subtitle operations: import_audio, extract_subtitles, repair_srt_text, add_reverb.
• visual effect: add_grain.

CLI

python -m twat_video --help
python -m twat_video crop-scale in.mp4 out.mp4 --crop 1080:1080:420:0 --scale 720:720 --dry-run
python -m twat_video fps in.mp4 out.mp4 24 --dry-run
python -m twat_video split in.mp4 clip.mp4 --start 1.5 --duration 3 --dry-run
python -m twat_video import-audio silent.mp4 voice.wav dubbed.mp4 --dry-run
python -m twat_video ken-burns still.png out.mp4 --duration 5 --dry-run

AI video boundary

generate_video() is a narrow adapter. It imports twat_genai and calls a future twat_genai.generate_video provider API if present; it does not embed SkyReels, WAN, Chutes, or other provider clients.

Reference scripts

Reusable patterns from vidcropscale.py, vidfps, vidframedrop.py, vidsplit.py, vidreverse, vidmergebygap.py, vidkenburns.py, vidgrainer.py, vidreverb.py, vidimportaudio, mkv2srt.py, and srtmultifix.py are represented as package APIs.

Watermark removal, menu-bar automation, JSON presets, batch shell wrappers, and local renaming/upload workflows remain in reference/bin-img-vid/ because they are workstation-specific one-offs or policy-sensitive scripts rather than reusable package behavior.