Your friendly Slack-to-Discourse archivist.
This is a Slack bot that can ship threads off to a Discourse instance as forum posts on request.
It was inspired by Dgraph's Wisemonk, and built for the Zeebe.io community Slack.
One of the issues with hosting a community on Slack is the loss of history. Valuable discussions and answers to questions quickly scroll over the 10,000 message horizon. A number of technical communities have moved to Discord to deal with this.
We looked at doing that, but then saw an opportunity to build a searchable knowledge base on our forum by sending valuable threads there as posts. There they can be curated by editors, indexed by Google, and discovered by other members of our community searching for answers.
Dgraph had a similar idea a few years ago, and built Wisemonk in Go. It hasn't been updated for a few years, and I couldn't get it to work - so I coded this up in TypeScript, using Slack's Web API and Events API.
In a thread (this is important - your conversations need to be threaded), @ the bot with what you want as the title. The bot will then roll up the thread and turn it into a forum post.
Example:
Here is the post generated from this thread.
Obviously you will need a Slack where you can add bot users, and a Discourse instance where you can get an API key. You will also need to run the bot with a resolvable DNS address, or at least an external IP, as it needs to listen to Push notifications from Slack's Event API.
Here are the instructions about creating a bot user on Slack. You only need to do steps 1 - 3.
The bot will need the following scopes in its OAuth settings:
- app_mentions:read
- channels:history
- channels:join
- channels:read
- chat:write
- im:history
- im:read
- im:write
- incoming-webhook
- links:read
- links:write
- reactions:write
- users.profile:read
- users:read
- users:write
You will need to set up the Event Subscriptions for the bot, like this:
The Slack Archivist bot uses a PouchDB to track the threads that it has archived. The database is created locally using the leveldown adapter in the slack-archivist-db directory, so you should mount that into the docker container to ensure its persistence across container lifecycles.
You can optionally provide a COUCHDB_URL via the environment to sync the database with a remote CouchDB instance. This is recommended to ensure the persistence of data. You can start a CouchDB instance easily in Google Cloud using Bitnami, or use IBM Cloudant.
Read the docker-compose.yml file and set up either the environment variables or a .env file. Then run:
docker-compose up -d
If you want to run this using https, you can do it easily using the Nginx / LetsEncrypt Docker sidecar.
To install, clone the repository, then run:
npm i
You configure Slack Archivist via environment variables:
# Required
DISCOURSE_TOKEN
DISCOURSE_USER
DISCOURSE_CATEGORY
DISCOURSE_URL
SLACK_TOKEN
SLACK_SIGNING_SECRET
SLACK_BOTNAME
# Optional
SLACK_PORT # Default: 3000
COUCHDB_URL # For syncing
LOG_LEVEL # winston log level. Default: info
You can set these through your environment, or put them into a .env file.
Rename env to .env, and fill in your Slack bot and Discord details.
The message templates for the Slack Archivist messages can be found in the messages directory.
You can run the bot using ts-node:
npm i -g ts-node
ts-node src/main.ts
Or by transpiling to JS:
npm run build
npm run start
You can also deploy the bot using docker, following the instructions and template in the deploy directory.
The bot behaviour is described in the Behaviour.bpmn file.
I livestreamed a lot of the coding: