Create CI to build & push PyBaMM's Docker Images

[arjxn-py]

Description

This PR aims to create a CI that builds all PyBaMM's docker images & then pushes them to DockerHub.

Type of change

Please add a line in the relevant section of CHANGELOG.md to document the change (include PR #) - note reverse order of PR #s. If necessary, also add to the list of breaking changes.

• New feature (non-breaking change which adds functionality)
• Optimization (back-end change that speeds up the code)
• Bug fix (non-breaking change which fixes an issue)

Key checklist:

• No style issues: $ pre-commit run (or $ nox -s pre-commit) (see CONTRIBUTING.md for how to set this up to run automatically when committing locally, in just two lines of code)
• All tests pass: $ python run-tests.py --all (or $ nox -s tests)
• The documentation builds: $ python run-tests.py --doctest (or $ nox -s doctests)

You can run integration tests, unit tests, and doctests together at once, using $ python run-tests.py --quick (or $ nox -s quick).

Further checks:

• Code is commented, particularly in hard-to-understand areas
• Tests added that prove fix is effective or that feature works

[arjxn-py]

Related to #3312
cc: @agriyakhetarpal @brosaplanella @Saransh-cpp

[arjxn-py]

I'd also need help with setting up the docker secrets for this one.

[agriyakhetarpal]

Would building and pushing images in their separate jobs be a better way, rather than in a series of steps?

[codecov]

Codecov Report

All modified lines are covered by tests ✅

Comparison is base (d02a78b) 99.58% compared to head (b059ebe) 99.58%.

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #3316   +/-   ##
========================================
  Coverage    99.58%   99.58%           
========================================
  Files          256      256           
  Lines        19998    19998           
========================================
  Hits         19915    19915           
  Misses          83       83           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[arjxn-py]

Would building and pushing images in their separate jobs be a better way, rather than in a series of steps?

Can they run parallelly in case of separate jobs? I'd be happy to switch to separate jobs either way though 🙂
Thanks!

[agriyakhetarpal]

Can they run parallelly in case of separate jobs?

Yes, they will run and upload images in parallel. We would need to check out the repository in each job, however. Since it would be developers who would be using the image, could you remove the sparse checkout and let it fetch the entire commit depth of the repository?

[agriyakhetarpal]

Actually, multiple jobs might not be needed in this case, a matrix for the solver arguments should suffice after the check out step

[Saransh-cpp]

I'd also need help with setting up the docker secrets for this one.

Do you want us to add them in the GitHub repo or do you want us to generate it from Dockerhub?

[arjxn-py]

Do you want us to add them in the GitHub repo or do you want us to generate it from Dockerhub?

I'd need help with adding them to github repo, i can generate & send secrets to you on slack.

[agriyakhetarpal]

Note: to push to Docker Hub we should use https://github.com/docker/build-push-action.

I noticed that they have some examples about setting up QEMU, we should be able to build both AMD64 and ARM64 images with that (I may take a look in another PR).

[Saransh-cpp]

Thanks, @arjxn-py! Looks great! Should be ready to merge once the comment below is resolved -

[agriyakhetarpal]

Also, taking into consideration that we would further be pushing images on release candidates, releases, and possibly nightly releases too (not in this PR): I was wondering about storage limits on Docker Hub. @arjxn-py do you know if there are any? If so, we could push just the "All solvers" image because the other images would then be redundant, and document this change in the Docker installation page. If someone wants to build an image with a specific solver locally, we already have instructions on how to do that

[arjxn-py]

storage limits on Docker Hub

As of my knowledge & research there's no storage limit for images on DockerHub (i.e. we can have as many tags we want) however there's an individual image size limit which we don't have to bother about as our image size is smaller.

[agriyakhetarpal]

That's great, thanks for the info! Pushing all images should be fine then

[agriyakhetarpal]

I think we did not tag the images properly. On https://hub.docker.com/r/pybamm/pybamm/tags I can see that this image was pushed with the latest tag, and

[agriyakhetarpal]

this one also has the latest tag

[agriyakhetarpal]

the earlier one (with all solvers) should have had the all tag. I do have a local branch which fixes things up a bit for this workflow (including logging in and uploading these images in parallel jobs) and will write a PR soon, I might look into adding the multi-platform images there itself.