Conversation
* Pin pandas <3.0 in requirements and pyproject * Split NumPy constraints by Python version for Apple M4 error * Make xr.concat data_vars explicit to silence FutureWarning * Expand zarr error regex to accept "contains data" * Ignore .hypothesis in pytest collection
jeromekelleher
left a comment
There was a problem hiding this comment.
This all LGTM - thanks @Billyzhang1229!
923fefb to
27d6a19
Compare
|
We could also consider dropping Mergify - the built in GitHub support has been working well over in tskit-dev. |
|
The only failing job is the Cubed workflow. The Cubed workflow explicitly installs dev builds from GitHub, overriding the pinned dependencies in requirements.txt. In .github/workflows/cubed.yml it runs: python -m pip install -U git+https://github.com/cubed-dev/cubed.git -U git+https://github.com/cubed-dev/cubed-xarray.git -U git+https://github.com/pydata/xarray.git The Cubed failure is due to an upstream FutureWarning / behavior change related to tuple-based chunk specification in xarray 2026.1.1.dev18. |
It would be OK to skip this until we can fix it. |
|
Cool, I’ll go ahead and squash the commits. |
|
These commits all seem like they do something well specified, so no need to squash from my perspective |
6ba76c3 to
021e4ca
Compare
|
I’ve tidied things up a bit now. removed the dev commits. |
#1338