SuperCutter - "another good size repo actually"

Extract segments from downloaded YouTube videos based on their captions.

Created after I noticed that CGP Estate Agents used phrases such as "a large size room" or "a good size" comically often during their tours. After manually editing a supercut of some occurrences, I endeavored to (somewhat) automate the process.

How to Use

Before starting, update the keywords at the top of cutter.py.

For multiple videos

Download videos using whichever method you prefer. I use yt-dlp since it has proven to be faster than youtube-dl.
Scrape subtitles using download_subtitles.sh. Make sure to modify this file if your video files are named differently.
Run every_vid_in_folder.sh in the folder containing the videos. Make sure to modify the output folder if you'd like. Note that it assumes the subtitle files are named video.mp4.json.

For a single video

Run cutter.py with the arguments [video file], [subtitle file], [clips output directory]

Notes

The subtitles must be in the JSON format as obtained by youtube_transcript_api

Improvements to be Made