Youtube-Word-Finder

Searches a specific said word or phrase within all videos of a YouTube Channel.

Utilizes the auto-generated or written subtitles for that purpose.

Installation

Use the package manager pip to install the requirements.

pip install -r requirements.txt

Usage

Run the script wordFinder.py using the following command line arguments.

python3 wordFinder.py "<WORD|PHRASE>" <YT-CHANNEL-ID> <DEPTH>

Stores search result in the results folder. Result files are JSON formatted.

WORD | PHRASE

The said phrase or the word you want to search for in the subtitles.

Make sure to put quotation marks "" around the word or phrase.

YT CHANNEL ID

The channel ID of the YouTube channel you want to search in.

Can be obtained using sites like: https://www.streamweasels.com/tools/youtube-channel-id-and-user-id-convertor/

DEPTH

Defines how many videos should be scanned. The factor is x100.

This means that with a DEPTH of 3 the 300 newest videos on the channel will be scanned for the WORD or PHRASE.

How it works

1. Gathers all video IDs of the channel until the specified DEPTH and caches them in the id_cache folder for quicker access next time,
2. Divides the acquired video-id list into chunks for multiprocessing
3. Goes through every video and downloads the subtitles using the YouTube-Transcript-API, also caches them into the cache folder.
4. Searches the transcript for the given WORD or PHRASE and returns the timestamp when found
5. Stores results in the results folder in the JSON format for further use.

Credits

• YouTube-Transcript-API - Used for gathering the subtitles of the videos
• YouTube-Search-Python - Used for gathering the video-ids
• NumPy - Used for calculations for handling multiprocessing

License

GPL v3.0