Move to deploy-on-release

neurobagel / annotation_tool

https://annotate.neurobagel.org/

MIT License

3 stars 6 forks source link

Move to deploy-on-release #790

Closed surchs closed 2 weeks ago

surchs commented 2 months ago

Recent events have shown us that deploying on every merge to main is risky, because you can deloy silly things at inopportune times. Let's follow the good examle of our dockerized tools and only deploy to prod on a release.

Specific todos:

[x] Change our deployment process to only run on a release
[x] Find a way to have a running "nightly" version with a persistent name (i.e. not the staging previews, but maybe netlify works)
~~[ ] Setup / carry over the auto-release workflow from other projects~~

surchs commented 1 month ago

Let's not setup the auto-release. This repo is going to get its refactor in the not too distant future. So let's just release by hand for now and save ourselves the setup.

rmanaem commented 3 weeks ago

Find a way to have a running "nightly" version with a persistent name (i.e. not the staging previews, but maybe netlify works)

@surchs Did you have something in mind for this?

We already have netlify on PRs here like we do in the query tool and the openneuro UI: https://github.com/neurobagel/annotation_tool/pull/798#issuecomment-2411349016

surchs commented 3 weeks ago

Yes @rmanaem, in fact the nightly builds (whatever the HEAD of main branch is) are already hosted here: https://neurobagel-annotator.netlify.app/ - we just never use that. Maybe we could link this deployment with a nice clear warning that it is unstable/can break from the repo description or README