Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Extend the CI jobs in the "GMT Tests" workflow #3615

Closed
seisman opened this issue Nov 13, 2024 · 3 comments · Fixed by #3618
Closed

Extend the CI jobs in the "GMT Tests" workflow #3615

seisman opened this issue Nov 13, 2024 · 3 comments · Fixed by #3618
Labels
discussions Need more discussion before taking further actions
Milestone

Comments

@seisman
Copy link
Member

seisman commented Nov 13, 2024

Currently, in the "GMT Tests" workflow, we have 7 CI jobs:

  • Python 3.10 + core packages (pinned to minimum versions) [Linux/macOS/Windows]
  • Python 3.12 + core packages (pinned to maximum versions) + optional packages [Linux/macOS/Windows]
  • Python 3.11 + core packages (pinned to specific versions) + a few optional packages [Linux only, mainly for testing old versions of optional packages]

As you can see, the CI jobs can't guarantee that PyGMT works with Python 3.10 + optional packages.

What about extending the number of the CI jobs to 9:

@seisman seisman added the discussions Need more discussion before taking further actions label Nov 13, 2024
@weiji14
Copy link
Member

weiji14 commented Nov 13, 2024

  • Python 3.10 + core packages (pinned to minimum versions) + optional packages (pinned to the minimum supported version if any, related to

I've been wanting to create a matrix build using something like uv's --resolution lowest flag that will install the lowest possible version for all dependencies, both direct and indirect (transitive), or --resolution lowest-direct that will use the lowest compatible versions for all direct dependencies, while using the latest compatible versions for all other dependencies.

The main complication is that we have non-PyPI dependencies - GMT (and also GDAL). We could set up a build such that GMT is pulled from conda-forge, and the rest are installed via uv, but that might need to be in a separate file to ci_tests.yaml.

@seisman
Copy link
Member Author

seisman commented Nov 13, 2024

Maybe we can try pixi, which can install packages from both conda-forge and PyPI (using uv).

Some related posts:

@weiji14
Copy link
Member

weiji14 commented Nov 13, 2024

Yes, I thought about pixi, but it doesn't support a mode similar to uv's --lowest resolution yet. I've tried searching the issues but couldn't find it documented anywhere, though there is some work on integrating uv as the PyPI resolver in pixi at prefix-dev/pixi#1295.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
discussions Need more discussion before taking further actions
Projects
None yet
2 participants