-
Notifications
You must be signed in to change notification settings - Fork 902
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
CI: numpy version for pandas 1.2 #2687
CI: numpy version for pandas 1.2 #2687
Conversation
On the main branch, this no longer seems necessary? (I just merged a PR, so a recent run https://github.com/geopandas/geopandas/actions/runs/3795324743/jobs/6454340166 shows the docstring failures) Maybe numpy 1.24.1 fixed it in the meantime? (released 2 days ago, https://github.com/numpy/numpy/releases/tag/v1.24.1). numpy/numpy#22849 seems matplotlib related. |
ci/envs/39-pd12-conda-forge.yaml
Outdated
@@ -2,6 +2,7 @@ name: test | |||
channels: | |||
- conda-forge | |||
dependencies: | |||
- numpy <1.24 # compat with matplotlib 3.3.2 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
While it is no longer needed, it might actually make sense to pin numpy here to an older version (eg of around the same age as pandas 1.2) to have some variety in numpy versions as well
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've now pinned this to numpy 1.20 which came out 1 month after pandas 1.2 (on Jan 31 2021). This is dropped from NEP 29 support on Jan 31 2023 - in a month, but I suppose that's fine, if anything I suppose it's a prompt to update the pandas minimum version.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, something to discuss next year
Yep, pretty sure numpy/numpy#22842 is the specific issue we see here and is now fixed (I thought this must have been a behaviour change, didn't check upstream to see if it was a bug). I'll update this with a pinned numpy from pandas 1.2 era |
This is trying to address the crashes in the pandas 1.2 and pandas 3.9 latest environments seen here: https://github.com/geopandas/geopandas/actions/runs/3771037533/jobs/6411224335
Subsequently repurposed to pin numpy versions after this was fixed upstream.