Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.26%. Comparing base (9e47e70) to head (ef57f6f).
Report is 1 commits behind head on main.

@@            Coverage Diff             @@
##             main     #295      +/-   ##
==========================================
+ Coverage   94.26%   95.26%   +0.99%     
==========================================
  Files           5        5              
  Lines         401      401              
  Branches       90       90              
==========================================
+ Hits          378      382       +4     
  Misses         14       14              
+ Partials        9        5       -4     

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

Re ef57f6f: I'm not sure how I feel about bumping everything at once - does uv lock -P pyarrow suffice to make the tests pass already?

Also: Should we add a version guard requires-python: ">=3.10,<3.14" to pyproject.toml, since this mirrors the versions we explicitly test against?

I decided on a full upgrade in place of trying to upgrade pandas, numpy, polars etc. incrementally, but I can do that too.

I decided on a full upgrade in place of trying to upgrade pandas, numpy, polars etc., but I can do that too.

I guess if it doesn't break anything obvious it doesn't hurt (since we also don't constrain the versions in pyproject.toml).

Also: Should we add a version guard requires-python: ">=3.9,<3.14" to pyproject.toml, since this mirrors the versions we explicitly test against?

Semantically, all current non-development Python versions fulfill this constraint, which makes it an empty condition (until a very brief transition time in October 2025). I've seen these constraints in practice, but I'm not convinced they are all that useful.

Also: Should we add a version guard requires-python: ">=3.9,<3.14" to pyproject.toml, since this mirrors the versions we explicitly test against?

Semantically, all current non-development Python versions fulfill this constraint, which makes it an empty condition (until a very brief transition time in October 2025). I've seen these constraints in practice, but I'm not convinced they are all that useful.

I actually meant >=3.10, but your point still stands somewhat. I guess what it does is give a "proper" notice once Pie-thon 3.14 comes out since you won't be able to install the package without us at least giving it a cursory test beforehand.

I'm happy either way, just wanted to briefly discuss the point while we're thinking about the topic.