Docs Community Tracker

Overview

This script is designed to aggregate all community-related issues into a single actionable list where we can keep internal notes and not duplicate our efforts. It works by:

• Fetching new issues and discussions from specified hasura GitHub repos
• Fetching new threads from selected channels in Discord
• Taking the body of the issue or discussion and parsing it through an OpenAI model to summarize it succinctly
• Performing an ETL job via a service account to log this into a spreadsheet available to everyone in the org

Usage

Hasura employees can access the file containing the values generated from this script here.

This runs each hour on the hour via a GitHub Action. If you want to run the script locally, follow the steps below.

Development

Step 1. Clone the repo

git clone https://www.github.com/hasura/docs-community-tracker

Step 2. Add a .env

cd docs-community-tracker && cp .env.example .env

Grab the generic values from our 1Password vault by searching for docs-community-tracker env.

Step 3. Create a GitHub PAT with SSO configured

Follow these steps with repo permissions enabled.

Then, use this value to update the GH_TOKEN value in your .env file.

Step 4. Create a development copy of the spreadsheet

We ask that you don't use the "production" spreadsheet for development and testing. Instead, create a copy of the "production" Google Sheet and then choose to share it the same users (service-account@docs-community.iam.gserviceaccount.com with Editor access) as the original.

You can then determine the spreadsheet's id by taking everything after .../d/ and before /edit... in the sheet's URL.

Use this value in your .env file instead of the one provided in the vault.

Step 5. Run the script

Start watch mode:

npm run dev