Skip to content

Latest commit

 

History

History
160 lines (118 loc) · 5.79 KB

README.md

File metadata and controls

160 lines (118 loc) · 5.79 KB
Raw

Clojurians Log v2

YAY! This project has been funded by ClojuristsTogether: https://www.clojuriststogether.org/news/q3-2021-funding-announcement/

This is a v2 of clojurians-log-app

Why v2?

I & the Gaiwan team help in maintaining the original clojurians-app instance, and feel that it's quite possible to fix all the existing issues in clojurians-log and keep maintaining it.

But the decision to make an independent v2 makes a lot of sense to me personally to get a chance to completely re-architect the solution from all the lessons we've learnt over the years. It also sounds much more exciting 🙈🙈

We will still be able to carry some of the independent modules over and improvise on them, so it's not all waste :)

Some of the issues I want to solve/implement with v2:

  • mobile responsive & usability (this is something I really really want personally)
  • integrating real time search for entire logs (super excited for this)
  • using newer slack api's (the current ones we used are deprecated)
  • real-time logging (no more waiting till the end of the day for archive to update), permalinks to threads/conversations
  • slack imports from the beginning of time
  • idempotent incremental archive imports
  • ability to backfill data in case of server downtime
  • avoid server downtimes, better monitoring
  • improve performance & reduce resource consumption so we can run this on a much smaller instance
  • better statistics reporting (automated stats email to admins/staff)
  • open analytics
  • support for images and attachments
  • fix all of the SEO issues (internal links showing up on google, better indexing, etc)
  • more documentation to make it even more easier for folks to contribute
  • possibly automated email digests tracking specific keywords
  • slack-bot custom commands
    • delete / anonymise a user
    • automated book & resource recommendations
    • thread saving / voting using emojis
    • many more custom use cases could be supported using emojis & custom commands
  • many more ideas!

Why not a static site?

I really really love static site generation, but I think we could discuss a bit more about this:

Some pros of SSG:

  • Static site won't go down
  • Fast page load times

Some cons of SSG:

  • regenerating pages on delete events (these can go back upto a few weeks)
  • regenerating pages on new reactions added
  • incremental generation is hard
  • generating links (which date preceed and succeed #channel-foo on 2021-05-30 which have existing messages)
  • changing rendering code / fixing a frontend bug means we need to re-render everything, "selective rendering" of only last month would lead to divergence in pages and subtle bugs
  • we still need to maintain a server to actually fetch the slack events real-time, so it's not "ops-free"
  • the amount of archived data will only keep on increasing, which will keep increasing the magnitude of all of the above issues
  • On top of this, we've got a request by a member if they can opt-out from having their messages archived. This wasn't a strict request from their end, but it would be a nice to have feature to at least anonymize the messages if a user says so (by using a slack command like /clojurians-log anonymize-me) I don't want to dismiss the idea of static generation (especially as an add-on), but it might be more work than expected.

A lot of the above features we want to implement is simply not possible with the static site. This is why I am trying to create clojurians-log v2 as a dynamic server rendered site and keeping in mind of all of the issues we're facing in the way we log currently.

Screenshots

image

image

Development

Use launchpad to run the database + app, and launch a REPL connected to your Emacs server:

bin/launchpad dev --emacs --go

The app will run at http://localhost:8000.

To start css watchers, run make dev.

To use psql on the docker postgres:

docker compose exec -it db psql -U myuser clojurians_log

To use pgcli instead:

pip install -U pgcli
make pgcli
# if it asks for a pass enter "mypass"

For populating the db from a slack archive, check bulk_import.clj file.

Migration files

We use migratus library to manage migrations.

To create new migration files, from the REPL run the following

user> (go)
Launching with profile  :dev
:initiated
user> (def mc (migrations-config))
#'user/mc
user> (require '[clojurians-log.db.migrations :as m])
nil
;; now create first-time migrations or run unapplied migrations
user> (m/migrate mc)
;; create new migration sql files
user> (m/create mc "add-name-to-user-table")

This will create two files, but just delete the down files as we don't want to rollback.

Secrets

Secrets are stored in resources/config/secrets.edn

Deployment

GitHub actions builds a docker image from the main branch. Right now they aren't tagged.

You can use the docker image ghcr.io/clojurians-log-v2:main to run an instance.

Read more in ./docs/database.md to see how to init the DB and populate it with slack data.

You also need to mount a secrets.edn file for the docker container to launch properly

A mount bind like this should work:

/home/user/clojurians-log/secrets.edn:/src/resources/config/secrets.edn

Contribute

This project is currently undergoing major changes. Before you contribute please create an issue to discuss your idea as things may change drastically and rapidly over the next few weeks.

Please feel free to create issues to discuss if you have any cool ideas which we could implement in the future, or let me know if you just want to have a chat about this :)