Add fft grid and raz grid to test against master

[unalmis]

• Adds FFT grids and field aligned grids to test_compute_everything. Should be merged before stuff that changes grids New Grid API #2053 or stuff that changes how the dependencies are ordered/selected for computation Switch compute from recursive to looped #1557 to ensure they do not cause bugs in the computations.
• Deletes a bunch of files that are now not necessary and removes bounce objective docstring boilerplate
• uses nondiff_argnums instead of nondiff_argnames so that the code works on old jax versions
• Adds comments to explain per bounce_comments PR
• Resolves nufft Error #2162 .

[github-actions]

Memory benchmark result

|               Test Name                |      %Δ      |    Master (MB)     |      PR (MB)       |    Δ (MB)    |    Time PR (s)     |  Time Master (s)   |
| -------------------------------------- | ------------ | ------------------ | ------------------ | ------------ | ------------------ | ------------------ |
  test_objective_jac_w7x                 |    1.69 %    |     3.937e+03      |     4.003e+03      |    66.48     |       30.63        |       28.41        |
  test_proximal_jac_w7x_with_eq_update   |   -1.35 %    |     6.539e+03      |     6.450e+03      |    -88.55    |       141.31       |       142.22       |
  test_proximal_freeb_jac                |   -0.01 %    |     1.337e+04      |     1.337e+04      |    -1.28     |       85.52        |       82.79        |
  test_proximal_freeb_jac_blocked        |    0.06 %    |     7.681e+03      |     7.686e+03      |     4.64     |       73.47        |       72.87        |
  test_proximal_freeb_jac_batched        |   -0.39 %    |     7.673e+03      |     7.643e+03      |    -29.79    |       72.16        |       72.47        |
  test_proximal_jac_ripple               |   -1.89 %    |     3.593e+03      |     3.525e+03      |    -67.75    |       49.82        |       51.35        |
  test_proximal_jac_ripple_bounce1d      |    0.70 %    |     3.777e+03      |     3.803e+03      |    26.48     |       64.90        |       64.90        |
  test_eq_solve                          |    0.35 %    |     2.081e+03      |     2.088e+03      |     7.23     |       77.48        |       77.06        |

For the memory plots, go to the summary of Memory Benchmarks workflow and download the artifact.

[codecov]

Codecov Report

❌ Patch coverage is 92.47312% with 7 lines in your changes missing coverage. Please review.
✅ Project coverage is 94.36%. Comparing base (6b3978f) to head (9b9d666).

Files with missing lines	Patch %	Lines
desc/integrals/_interp_utils.py	44.44%	5 Missing ⚠️
desc/integrals/bounce_integral.py	96.07%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2157      +/-   ##
==========================================
- Coverage   94.37%   94.36%   -0.02%     
==========================================
  Files         101      101              
  Lines       28763    28785      +22     
==========================================
+ Hits        27146    27162      +16     
- Misses       1617     1623       +6     

Files with missing lines	Coverage Δ
desc/compute/_fast_ion.py	100.00% <ø> (ø)
desc/compute/_neoclassical.py	100.00% <ø> (ø)
desc/equilibrium/equilibrium.py	96.12% <ø> (ø)
desc/integrals/__init__.py	100.00% <ø> (ø)
desc/integrals/_bounce_utils.py	94.26% <100.00%> (+0.12%)	⬆️
desc/objectives/_fast_ion.py	95.18% <100.00%> (ø)
desc/objectives/_neoclassical.py	95.00% <100.00%> (ø)
desc/objectives/objective_funs.py	94.99% <100.00%> (+<0.01%)	⬆️
desc/integrals/bounce_integral.py	98.05% <96.07%> (+0.07%)	⬆️
desc/integrals/_interp_utils.py	87.05% <44.44%> (-1.36%)	⬇️

... and 2 files with indirect coverage changes

🚀 New features to boost your workflow:

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

[unalmis]

someone else will need to cllick merge when ready

[unalmis]

@unalmis Please do not mark comments as resolved unless they are addressed in the code or the discussion has concluded. Responding is not the same as resolving. Comments are helpful for everyone to read, but when they are resolved it makes them less visible for other reviewers to see and join the conversation.

I can show you how to enable GitHub notifications to see when there are new comments.

You had reopened two which is outside scope of this PR as master already has them. Such discussion should be done in github issues. I think the effort it takes for a reviewer to click see comment button is much less than the increased effort for me, the maintainer, to track which issues are actually resolvable and which are just unactionable due to jax or cosmetic personal preference that should be resolved elsewhere as they don't need my input. Github notifications don't work selectively on a PR. i get notifs for all of desc or none of it.

[unalmis]

@ddudt consider https://github.com/pullpo-io/conventional-comments for comprimise

[dpanici]

Looks good to me, thanks for taking some of the comments from that PR I had made. I would also like @ddudt to re-review since he had made a lot of comments his first pass around.

If possible, let's merge this one before the sparse pullback PR (#2170 , have not reviewed it yet) gets merged into here, so we can isolate the changes a bit and make the reviews of each more manageable

[dpanici]

Non-Blocking, just curiosity: I'm fine with skipping the axis if it is sensitive, but out of curiosity what OS are you using?

I find I get minor diffs that still trip the test if I say re-saved quantities I computed on my mac, which then will have slight differences compared to the quantities computed on the CI's linux (I believe) runners. Usually I would just make sure I saved the quantities on some linux (non-wsl though I think I may have had issues there too) os like one of the university clusters.

[unalmis]

#2199

[unalmis]

i don't have codecov override

[ddudt]

I want to echo this comment from Rory about having smaller PRs, each with a narrow scope. This was a big PR that covered a lot more than it originally advertised. More digestible PRs are easier and faster to review, and will make you a better software developer.

Also I really like this Chrome extension that Kaya recommended and I would recommend it to others!

[unalmis]

i don't have codecov override

need codecov override