tbls-ask-agent-slack

This is a slack-bot version of tbls-ask.

You can call tbls-ask from slack by mentioning the bot.

Usage

To use the tbls-ask-agent-slack bot, mention it in a Slack channel where it's been added. The bot can answer questions about your database schema using natural language.

Example Query

You can ask questions like:

"@tbls-ask name of the users who has the most stars"

The bot will respond with an SQL query that answers your question. For the example above, it would provide a query to find the user with the most stars on their comments.

How It Works

The bot receives your question through a Slack mention.
It interprets your natural language query.
It generates an appropriate SQL query to answer your question.
The bot returns the SQL query in the Slack thread.

Limitations

The bot provides SQL queries but does not execute them directly on your database.
You need to run the provided SQL query on your own MySQL database to get the actual results.
The database document must be prepared using tbls.

How to set up

Environment Variables

OPENAI_API_KEY: API key for OpenAI
SLACK_APP_TOKEN: App token for Slack
SLACK_OAUTH_TOKEN: OAuth token for Slack
GITHUB_TOKEN: Token for GitHub API (optional)

Slack-app settings

Please create an app using manifests.yml and install it to your workspace.

Prepare schema

Copy schemas/config.yml.sample to schemas/config.yml and edit it.

$ cp schemas/config.yml.sample schemas/config.yml

Run locally

go run main.go

It is using socket mode for slack. You don't need to expose the server to the internet.

Deploy to k8s

Build docker image locally

make build-image

Apply manifests

make all

kromiii/tbls-ask-agent-slack