Unit tests for MeshIndexSet.

[trexfeathers]

• Check raw code coverage
• Better use of fixtures for setting up tests
• Compare with Tests for indexing meshes #7218 for other coverage misses

[trexfeathers]

This is best understood by comparing to other metadata tests such as test_MeshMetadata.py ¹

Footnotes

1. the only differences should be resulting from the different memberships, and possibly some slight ordering differences. ↩

[trexfeathers]

• Should have a more thorough parameterisation of the three locations
• Should test whether the original MeshXY is all lazy connectivities/coords versus all NumPy versions

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.26%. Comparing base (a77af27) to head (1c2ed20).

Additional details and impacted files

@@                  Coverage Diff                  @@
##           FEATURE_index_set    #7219      +/-   ##
=====================================================
+ Coverage              89.27%   90.26%   +0.99%     
=====================================================
  Files                     92       92              
  Lines                  25555    25570      +15     
  Branches                4762     4767       +5     
=====================================================
+ Hits                   22813    23080     +267     
+ Misses                  1945     1703     -242     
+ Partials                 797      787      -10     

☔ View full report in Codecov by Harness.
📢 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.

[stephenworsley]

I think it's worth passing in a more complex case here to properly exercise the maths that's going on under the hood here. I'd say it's worth coming up with a mesh, indices pair with the following properties:

• Multiple cells are indexed.
• The indices are not all consecutive (e.g. [1,2,4])
• Some of the faces being indexed have masked connectivities (e.g. there is a mix of triangles and rectangles).
• Some of the faces are neighbouring (i.e. they share nodes).
• Some of the faces are disconnected from each of the other faces (i.e. they share no nodes).

It's also worth having such a case for edge/node calculations. Once you have such examples, it's worth introducing them to other tests where there are complex calculations happening under the hood.

[stephenworsley]

Probably worth making MeshIndexSetMetadata private since _MeshIndexSet is private.

[stephenworsley]

I think it's worth getting a bit more specific here about what the resulting mesh ought to actually look like. It's worth maybe reusing the more complex example mesh I mentioned before.