RocketScrape

A command line tool to process Discord message streams.

Setup

To get started, install the rocketscrape Python package from within the directory using:

pip install .

RocketScrape requires a Discord authentication token for API requests, which should be stored in a DISCORD_USER_TOKEN environment variable. For more information on how to obtain a token for your account see https://discordpy-self.readthedocs.io/en/latest/authenticating.html.

export DISCORD_USER_TOKEN=<token>

Usage

The tool primarily expects a source (flag) and analysis type (positional argument) to be specified. Below are examples of how to use RocketScrape:

rocketscrape -c CHANNEL1 [-c CHANNEL2, ...] (--include-threads | --exclude-threads) ANALYSIS

Replace CHANNEL1, CHANNEL2, etc., with either the Discord channel / thread ID or one of the predefined channel names from the Channel enum in rocketscrape.py. You can choose to include or exclude threads within specified channels using the required --include-threads or --exclude-threads flag. Similarly, you can run rocketscrape on an entire server using:

rocketscrape --server SERVER (--include-threads | --exclude-threads) ANALYSIS

SERVER can either be a server ID or a name defined in the Server enum. There are optional -s / --start and -e / --end arguments that accept an ISO format datetime string to restrict the date range. For instance, to get the top contributors for the support channel over the last month, you can use:

rocketscrape -c support -s $(date -d "-1 month" +"%Y-%m-%d") --include-threads contributors

For a complete list of global options and analysis types, run the help command:

rocketscrape -h

Additionally, you can access help commands for specific analysis types, e.g.:

rocketscrape contributor-history -h

Adding a custom analysis class

Each analysis type in RocketScrape is implemented as a subclass of MessageAnalysis. These classes are required to implement _require_reactions(self), _prepare(self), _on_message(self, message), _finalize(self), _display_result(self, result, client, max_results) and subcommand(). Optionally, custom_args(cls) can be overridden to add analysis-specific command line arguments. Examples can be found in analysis.py.