Slack chat bot written in Rust that allows the user to interact with a large language model.

A Slack chat bot written in Rust that allows the user to interact with a large language model.

Creating an App on Slack, first steps

• Go to https://api.slack.com/apps and create a new app from scratch.

• Navigate to "OAuth & Permissions" (https://api.slack.com/apps/YOURAPPID/oauth). Go to section "Bot Token Scopes", click on "Add an Oauth Scope", select scopes "app_mentions:read", "chat:write", "files:write", "im:history". Then click on "install to Workspace". You will obtain the "Bot User OAuth Token" on this page then.

• Navigate to "Basic Information" (https://api.slack.com/apps/YOURAPPID/general). Here you will obtain the "Signing secret".

• Now you need to deploy your app following the steps of the next section before continuing with the Slack app configuration.

Configuring, compiling and running

Generate a token from Hugging Face at (https://huggingface.co/settings/tokens).

Create an .env file at the root of the repository (same folder as the Cargo.toml file) with:

    SLACK_OAUTH_TOKEN="bot_user_oauth_token_from_previous_step"
    SLACK_SIGNING_SECRET="signing_secret_from_previous_step"
    LLM_API_TOKEN="api_key_from_huggingface"
    DATABASE_URL="sqlite://db/db.sqlite3"

and run without Docker compose:

docker compose build && docker compose up

or with Docker, but without Docker compose:

docker build . -t rust-slackbot-llm && docker run --rm -it -v $(pwd)/.env:/app/.env rust-slackbot-llm

or with plain cargo:

cargo run

Public IP address

When you run the App, it's exposed via port 51005 and on HTTP only, therefore, it's recommend that install Nginx and add the following configuration (probably at /etc/nginx/sites-available/default) to:

    server {
        server_name your_domain_here.com;
        location / {
            proxy_pass http://127.0.0.1:51005;
            proxy_set_header Host $host;
        }
        listen 80;
    }

And then (after pointing your DNS to your domain) run sudo certbot --nginx.

If you don't own a domain, you might wanna try your luck with https://nip.io/.

You also need to have a reacheable IP address for Slack to deliver the payloads to your bot. Sadly this is not possible with many residential internet ISPs which now use carrier grade NAT, so as a work around you can get a simple machine on a cloud provider (e.g.: AFAIK, Google Cloud has a always free tier, use at your own risk) and run the app there or run at your local machine and redirect the port to cloud machine using SSH ssh your_domain_here.com -R 127.0.0.1:51005:127.0.0.1:51005.

Concluding the deployment of your Slack app

• Navigate to "Event Subscriptions" (https://api.slack.com/apps/YOURAPPID/event-subscriptions). On the requested URL section, fill in https://your_domain_here.com/v1/slack_events and if everything was properly configured in the previous step, you should receive a "verified" status.

• On "Subscribe to bot events", add scopes "app_mention" and "message.im" and click on "Save changes".

• A yellow upper box will show up requesting you the reinstall the App, proceed with that.

• Go to App Home (https://api.slack.com/apps/YOURAPPID/app-home) and check the box "Allow users to send Slash commands and messages from the messages tab".

An extra: plotting

If configured everything correctly, you should have you also have support for plotting out of the box. Just send plot as message to chat bot to get an example: