Fix guess_bounds precision for circular coords at numpy v2

[rcomer]

🚀 Pull Request

Description

Fixes #6738, or at least it fixes my use case 😉

The coordinate's modulus is a python float. It seems that at numpy v1 doing arithmetic between a np.float32 and a float gave a np.float64. At numpy v2 it quite reasonably gives a np.float32, but that means we are now calculating at lower precision than we were. Making the modulus into a np.float64 seems like the simplest way to reinstate the previous behaviour.

---

Consult Iris pull request check list

---

Add any of the below labels to trigger actions on this PR:

• benchmark_this Request that this pull request be benchmarked to check if it introduces performance shifts

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.39%. Comparing base (20e4d88) to head (13f9d4f).
⚠️ Report is 2 commits behind head on v3.14.x.

Additional details and impacted files

@@           Coverage Diff            @@
##           v3.14.x    #6793   +/-   ##
========================================
  Coverage    90.39%   90.39%           
========================================
  Files           91       91           
  Lines        24761    24763    +2     
  Branches      4639     4639           
========================================
+ Hits         22382    22384    +2     
  Misses        1608     1608           
  Partials       771      771           

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[trexfeathers]

Thanks for tackling this @rcomer! I have one suggestion, for honouring the original dtype instead of always using float64.

[trexfeathers]

Great, thanks! Once we've got v3.14.0 out tomorrow I will get this merged into v3.14.x.

In case it isn't clear: I can't merge it now because that would mean cutting a new release candidate and waiting another two weeks.

[rcomer]

Thanks @trexfeathers - I didn't know about promote_types, so I've learned something along the way 😎

[rcomer]

Do I need to rebase against v3.14.x? Also will the whatnew need to move?

[rcomer]

I have attempted to follow the pattern of recent patches for the whatsnew.

[rcomer]

Thank you for being so accommodating @trexfeathers!