YoutubeWordCatcher

Extract every clip of a Youtube channel's videos where a specific word is pronounced and build a big video out of it.

Install

python3 -m venv venv
source ./venv/bin/activate
pip install -r requirements.txt

Configure

You can edit the configuration by creating a config.ini file.

The available sections are:

• all : will be loaded in every configuration
• command : will be loaded only for the specified command (example: catch)

You can find all the possible values for the configuration in the config.py files located in the commands folder.

Note that whenever there are dictionaries or lists, you can represent the data on a multiline but beware of the indentation.

Run

In order to run properly, you need to get a Youtube API key, which you can obtain by following the instructions here.

python ywc.py catch --help

⚠️ If you are using Windows

You must set the environment variables FFMPEG_BINARY and IMAGEMAGICK_BINARY or you will get the error "FileNotFoundError: [WinError 2] The system cannot find the file specified".

Chart

To visualize the data generated by this program, I wrote a command to generate charts.

To use it, make sure you generated the data by running the main program, then execute:

python ymc.py chart --help

You can look at the code to edit some variables if necessary, or add your own data visualization.

Example

A complete example is available in EXAMPLE.md.

Todo

• Manual clip sorter to remove false positives
• Improve timestamps with SpeechRecognition plugin

Authors

• Marc Villain marc.villain@epita.fr