A simple app for logging twitch's chat to a PostgreSQL database.
Python Shell
Switch branches/tags
Nothing to show
Clone or download
Latest commit d11d623 Feb 10, 2017

README.rst

twitch-chat-logger

A simple python app for logging twitch's chat to a PostgreSQL database. It logs an arbitrary ammount of channels (default is a 100) ordered by the numbers of viewers or an specific list of channels. Twitch seems to not like a single bot joining a large amount of channels, so each bot is limitted to 20 channels. This app automatically scales the number of bots according to how many channels are to be logged (e.g. logging a 100 channels will result in 5 bots being created). The list of most popular channels is updated every 60 seconds and the bots join and leave channels as needed.

Logging 100 channels for 24 hours seems to amount to ~4 million chat lines (~400 MB).

Setup

Install this repo using git.

git clone https://github.com/bernardopires/twitch-chat-logger.git

A twitch account is required to connect and log the chat channels. Create a file named settings.py (an example is provided with the name settings.py.example) and update the IRC settings dictionary with your account credentials. Hint: You can get your oauth password from the Twitch Chat OAuth Password Generator.

IRC = {
    'SERVER': 'irc.twitch.tv',
    'NICK': 'twitch_username',
    'PASSWORD': 'your_oauth_password',
    'PORT': 6667,
}

The project makes requests to pull the most popular channels from the Twitch API. To do this you need to get your Client-Id which you can get going to Twitch > Account Settings > Conections` and register a new aplication (at the bottom of the page). Then add it to the settings file in the API dictionary.

API = {
    'CLIENTID': 'YOUR ID GOES HERE'
}

If you're using docker and fig you're all set.

fig up

Otherwise, install the PostgreSQL database if you haven't yet and create a database named twitch. Update the DATABASE dictionary inside settings.py with your credentials.

DATABASE = {
    'NAME': 'twitch',
    'USER': 'database_username',
    'PASSWORD': 'database_password',
    'HOST': 'localhost',
}

Create the needed tables by running create_tables.sql.

psql twitch -f create_tables.sql -U your_db_username -h localhost -W

Install the python library dependencies with pip.

pip install -r requirements.txt

Finally, you're ready! If you encounter any errors installing psycopg2, you may have to execute apt-get install libpq-dev python-dev.

python main.py

The command above will start 5 bots logging the 100 most popular twitch channels. To log a different amount use the parameter n, to log a specific list of channels use the parameter c and to save the output to a file use the parameter f. For example, use the command below to log the 50 most popular channels with the output being saved to a file named log.txt

python main.py -n 50 -f log.txt

To log a specific list of channels, separate the channels name by whitespace.

python main.py -c channel1 channel2 channel3