Fix the shapes of 2d filters

[mawc2019]

As mentioned here, the 2d filters in filters.py should have kernels with circular boundaries, but these boundaries are rectangular in the current master branch. This PR makes them circular. In addition, this PR adds a round operation preceding each int operation in filters.py to avoid unexpected results, as mentioned here.

[oskooi]

The pre-commit check is failing which likely means you forgot to first install the pre-commit package.

See: https://meep.readthedocs.io/en/latest/Build_From_Source/#meep-for-developers

[codecov-commenter]

Codecov Report

Merging #2390 (383babf) into master (b86b476) will decrease coverage by 0.01%.
The diff coverage is 100.00%.

📣 This organization is not using Codecov’s GitHub App Integration. We recommend you install it so Codecov can continue to function properly for your repositories. Learn more

@@            Coverage Diff             @@
##           master    #2390      +/-   ##
==========================================
- Coverage   72.57%   72.56%   -0.01%     
==========================================
  Files          17       17              
  Lines        5166     5165       -1     
==========================================
- Hits         3749     3748       -1     
  Misses       1417     1417              

Impacted Files	Coverage Δ
python/adjoint/filters.py	60.58% <100.00%> (-0.29%)	⬇️

[mawc2019]

I just installed the pre-commit and committed again, but this online pre-commit check still fails. Does the pre-commit check run automatically when I do git commit on a branch, or should I somehow run it manually?

[mawc2019]

The error in the failed test is as follows.

Traceback (most recent call last):
  File "/home/runner/work/meep/meep/build/meep-1.26.0-beta/_build/sub/python/../../../python/tests/test_adjoint_solver.py", line 685, in test_gradient_backpropagation
    self.assertClose(adj_dd, fnd_dd, epsilon=tol)
  File "/home/runner/work/meep/meep/build/meep-1.26.0-beta/python/tests/utils.py", line 30, in assertClose
    self.assertLessEqual(diff_norm, epsilon * np.maximum(x_norm, y_norm), msg)
AssertionError: 9.003547633008868e-06 not less than or equal to 8.717597578845472e-06 :

This error is related to the code below when single precision is used.

            if nfrq == 1 and frequencies[0] == self.fcen:
                tol = 2e-4 if mp.is_single_precision() else 5e-6

Increasing the tolerance tol from 2e-4 to 2.1e-4 would resolve the issue. Shall we do this?

[stevengj]

Sure, a 5% increase in the tolerance is fine.

[stevengj]

LGTM!

[stevengj]

When you get a chance, try re-running the tutorials so that we can update the images (which will presumably change slightly).