Discord OpenAI Bot

This bot uses OpenAI to generate responses to messages in a Discord server. It listens for mentions of the bot and generates responses using OpenAI's GPT chat model. It has some special functionality to track the number of times a user has mentioned the bot recently and limit the number of OpenAI calls.

It has a couple of extra options to do common things. Eg:

@Gepetto create an image of a rocket flying through space
@Gepetto 👀 https://www.example.com/an/article
@Gepetto 👀 <https://www.youtube.com/watch?v=123f830q9>
@Gepetto 👀 <https://www.example.com/an/article> can you give me the main insights on this as bullet points?

The youtube one depends on their being subtitles/transcripts attached to the video. The summarise command is a little limited (currently hard-coded) in scope due to token limits on the text you can send to the cheaper OpenAI models.

Environment Variables

The following environment variables are required:

• DISCORD_BOT_TOKEN: Your Discord bot token
• DISCORD_SERVER_ID: The ID of your Discord server
• OPENAI_API_KEY: Your OpenAI API key
• DEFAULT_MODEL_ENGINE: The engine of the OpenAI model to use (default is 'gpt-3.5-turbo')

Running the Script

To run the bot:

1. Install the required Python dependencies: discord.py and openai. You can install these by running pip install -r requirements.txt.
2. Set your environment variables. These can be set in your shell, or stored in a .env file at the root of your project.
3. Run python main.py in the root of your project to start the bot.

Docker Deployment

A Dockerfile is included in this repository. This allows you to easily build and run your bot inside a Docker container. Please see the run.sh script for an example of building and running the container.