Only try to work out the differences between points for multiple #4367
Conversation
There was a problem hiding this comment.
@wjbenfold Awesome! Thanks for picking up this PR 😄
Just a few minor suggestions to service, and perhaps bolster the iris.tests.unit.util.test__coord_regular.Test_points_step class with some minimal test coverage?
Then we're good to go 👍
|
Changes made as requested. Unit tests raised questions of what the correct behaviour with an empty array supplied is. Will rebase then fix whatsnew once @bjlittle is happy with everything else |
|
@wjbenfold if you change “Fixes issue #4250” to “Fixes #4250” in your original post, it should link it so the issue gets closed automatically when this is merged. See https://docs.github.com/en/[email protected]/issues/tracking-your-work-with-issues/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword |
wjbenfold
force-pushed
the
wjbenfold-coord-regular-warn
branch
from
3f97027 to
fe2cb9a
Compare
|
This looks good to go! I'd like to get it merged. @wjbenfold could I ask you to rebase to resolve the conflicting files? |
wjbenfold
force-pushed
the
wjbenfold-coord-regular-warn
branch
from
fe2cb9a to
04b388d
Compare
|
Thanks All for sorting this 😊 |
* main: [pre-commit.ci] pre-commit autoupdate (SciTools#4452) Whatsnew (SciTools#4451) Explicitly define tests so nose can find them (SciTools#4450) Updated environment lockfiles (SciTools#4449) Update CI environment lockfiles (SciTools#4424) Disable GHA workflows on non-SciTools branches (SciTools#4444) Add new contributor to whatsnew (SciTools#4443) Use dask-core instead of dask in ci (SciTools#4434) Mesh recombine (SciTools#4437) Mesh full comparison (SciTools#4439) Only try to work out the differences between points for multiple (SciTools#4367) [pre-commit.ci] pre-commit autoupdate (SciTools#4430) Fix license PyPI classifier (SciTools#4435) Whatsnew for PR SciTools#4367 (SciTools#4440)
* upstream/main: (78 commits) Updated environment lockfiles (SciTools#4458) remove asv package dependency (SciTools#4456) cube.aggregated_by output bounds (SciTools#4315) [pre-commit.ci] pre-commit autoupdate (SciTools#4452) Whatsnew (SciTools#4451) Explicitly define tests so nose can find them (SciTools#4450) Updated environment lockfiles (SciTools#4449) Update CI environment lockfiles (SciTools#4424) Disable GHA workflows on non-SciTools branches (SciTools#4444) Add new contributor to whatsnew (SciTools#4443) Use dask-core instead of dask in ci (SciTools#4434) Mesh recombine (SciTools#4437) Mesh full comparison (SciTools#4439) Only try to work out the differences between points for multiple (SciTools#4367) [pre-commit.ci] pre-commit autoupdate (SciTools#4430) Fix license PyPI classifier (SciTools#4435) Whatsnew for PR SciTools#4367 (SciTools#4440) Suggest type hinting (SciTools#4390) area weight regrid test fixes (SciTools#4432) Update latest.rst (SciTools#4425) ...
🚀 Pull Request
Description
Causes
iris.util.points_stepto check whether there are multiple points to check before trying (thus averting a warning is caused by only one point being provided).Fixes #4250.
Questions for reviewer
coords.from_regularand thus leave this so that it warns when called with a single point (albeit cryptically) so that people calling it are aware that they may not be doing what they meant? I personally think that the np.nan return value of avdiff (which matches the current behaviour) is sufficient.Consult Iris pull request check list