Skill Discovery (Discord bot)

This little project is a "hack" developed as part of MetaFest 2021.

Objective

Create a graphical representation of people skills and interests.

Why?

To facilitate social bonding and skills exchange:

• Find someone that has skill S, which you are looking for
• Find what skills/interests you have in common with person P
• Get a general feel of the interests within a community

Examples

Some interactive visualizations of the data are available here:

Full graph, including all people, all skills/interests and their links:

Word-cloud view of the above:

Project structure

⚠️ HACK ALERT ⚠️

The source code of this project is hack-level quality.

It was stringed together in less than a week. Please calibrate your expectations accordingly.

For example:

• ❌ Most of the code is not covered by tests
• ❌ Code is a hodgepodge of classes, static functions, globals, etc (thanks, Python!)
• ❌ Some calls are not thread-safe (database, graph rendering, ...)
• ❌ The data model is as simple as it gets
• ❌ The rendering is pretty rudimentary
• etc.

That said, it's in pretty good shape for a hack. Future improvements (if anyone is willing) should be possible incrementally.

Code overview

...

Deployment

The bot requires 3 environment variables to run properly. They can be set on the shell, or in a .env file.

DISCORD_TOKEN: Token obtained by creating a new discord application. For examples, see this guide or similar.

GUILD_ID: The unique numeric id of the Discord server (a.k.a guild).

MONITOR_CHANNEL_ID: The unique numeric ID of the Discord channel to monitor.

The easiest way to learn your server and channel IDs is to open Discord in a browser. The URL will look like this:

https://discord.com/channels/<guild id>/<channel id>/

Dependencies

Python 3.9 and it's dependencies listed in requirements.txt. System (apt package) requirements are listed in requirements.system. Each of these can be installed with sudo apt install or equivalent commands.

Using a Python virtual environment is recommended. Alternatively, using nix:

nix-shell -p 'python39.withPackages(ps: with ps; [ graphviz python-dotenv discordpy ])'

The dot program must be installed (on Linux and macOS, this is part of the graphviz package). You can learn more about this format on Wikipedia or on the official website.

Additional fonts may be necessary to properly display emojis on Linux. Consider installing the fonts-noto package. Additional fonts for use with graphviz can also be installed via fontconfig (see https://gist.github.com/jacksonpradolima/840b4a20c617204a65b15da61e124bf6).

A sample installation of dependencies on an Ubuntu distro:

1. sudo xargs apt install <./requirements.system installs system requirements (python, graphviz, etc)
2. python3.9 -m venv ./.venv creates a virtual environment
3. source .venv/bin/activate activates your virtual environment
4. pip install -r ./requirements.txt installs all requisite python packages into your virtual environment

Running locally

Once you have gone through the above steps to creating a new discord app/bot, and have installed necessary depdendencies, you'll want to test the bot locally.

Following the steps below will spin up the bot on your machine, and it'll be listening in the discord server you configured with the DISCORD_TOKEN, GUILD_ID, and MONITOR_CHANNEL_ID. It'll also spin up a SQLite database in the root directory where you're running the bot, to store the skills and users created via the Discord CLI.

The following steps will start the bot on your machine:

1. source ./.venv/bin/activate
2. python3.9 bot.py

That's it!

Permissions

To work properly, the bot needs the following permission in the target channel:

• Read channel messages
• Write messages to channel
• React to messages in channel

License

This project is released under MIT license. See the LICENSE file for details.