/stfu

Primary LanguagePythonMIT LicenseMIT

CDawgVA and Friends Saying STFU for 20 Minutes

Resulting video: https://www.youtube.com/watch?v=wlj69LSS94k

Intall dependencies

Ensure you have the following tools installed:

  • python 3
  • pipenv
  • ffmpeg
pipenv install

Usage

You need up to 1 TB of storage available to store the clips and the various stages of processing.

Download subtitles

./fetch_subtitles.sh

Download rough cut of all stfu occurances

pipenv run extract_stfu_occurances --delete-clips-directory true --rough True --clips-directory ${PATH_TO_STORE_ROUGH_CLIPS?}

Improve the timestamps using Whisper AI

./improve_timestamps.sh ${PATH_TO_STORE_ROUGH_CLIPS?}

Cut the clips exactly using the original timestamps and improved timestamps

pipenv run improve_stfu_occurances --clips-directory ${PATH_TO_STORE_ROUGH_CLIPS?} --improved-clips-directory ${PATH_TO_STORE_IMPROVED_CLIPS?}

Concatenate the clips into a single video, normalize audio and generate description

./concat_video.sh

The final video can be found at ./data/video_concatenated.mkv and the description at ./data/description.txt.