chore: Relax dependencies to numpy 2.0 compatibility

[effigies]

25bdb2b had very steep minimum version upgrades, which can be awkward on downstream tools.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.74%. Comparing base (5c52e2b) to head (3df8641).
⚠️ Report is 9 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #284      +/-   ##
==========================================
+ Coverage   95.68%   96.74%   +1.05%     
==========================================
  Files          16       16              
  Lines        1994     1994              
  Branches      267      267              
==========================================
+ Hits         1908     1929      +21     
+ Misses         64       41      -23     
- Partials       22       24       +2     

Flag	Coverage Δ
unittests	96.72% <ø> (+1.05%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.

[effigies]

Damn. Looks like we depend on something in numpy 2.0. We're running slightly faster than SPEC0...

Anyway, #271 broke the spatial grid API by transposing ndindex and ndcoords. @oesteban I suggest that after this we release a 25.1.0 and yank 25.0.1 from PyPI.

[oesteban]

Looks like we depend on something in numpy 2.0. We're running slightly faster than SPEC0

Python 3.9 is supported only with numpy<2.10 May be the 4D bspline interpolation from scipy that requires numpy 2.0

[effigies]

@oesteban Not sure why CircleCI failed to build the docker image, but I slimmed that process down a lot. It shaves 20min off CircleCI runs. Possibly more once we get caching.