unit tests for large values of l and m for basis functions

[junchaoxia]

Current the unit tests for basis functions using small values of l and m. We should add tests for large values so that the overflow problems will be identified more obvious when switching dtype from float64 to float32, such as shown in the line,
https://github.com/ComputationalCryoEM/ASPIRE-Python/blob/develop/src/aspire/basis/basis_utils.py#L129
https://github.com/ComputationalCryoEM/ASPIRE-Python/blob/develop/src/aspire/basis/basis_utils.py#L138

[janden]

Well, also for large images, not just large ell and m.

[garrettwrong]

I think we can add these as parameterized cases that are marked expensive now.

[j-c-c]

I've added parametrized cases for larger images (res = 64) that are marked expensive to the tests for FBbasis2d and FFBbasis2d. The tests for the 3D bases are hardcoded so we'll have to do more than just adding parametrized cases. Any suggestions on how to proceed in the 3D case? Do we want to build out more logical based tests there?

[garrettwrong]

Hrmm, 2D is good to have, but the issue was actually more intended for the 3D case. The function noted by Junchao test is only used for 3D.

The draft PR doesn't really add any new test coverage because 64 is not a large image (we regularly use that elsewhere already).

Yes, if you would like to use parameterization directly as was suggested, that would involve parameterizing the test, or using some class based logic; whichever gets the job done with the least intrusive changes.

[janden]

3D tests will be moved from hardcoded to something more flexible in #605, but I need to sit down and figure out what the formulas should look like. Think we'll have to put larger tests for 3D on hold until that one's merged.