Sweep Fabrication

[simbilod]

some cleaned up work I had lying around to treat morphological transformations as parameters to the model builder, e.g.

Initial geometry:

Dilation and erosion (ports are shrunk and grown too):

Offsets between layers (ports are also translated):

Corner rounding (ports are padded to maintain continuity):

This can be swept just like any other Component or LayerStack parameters in the model builder:

Next step is to include better distributed processing in the generic model builder. I think with Ray we can scale from local, to cluster, to cloud seamlessly, so just need to format the remote decorator properly

[sourcery-ai]

Sourcery Code Quality Report

❌  Merging this PR will decrease code quality in the affected files by 10.04%.

Quality metrics	Before	After	Change
Complexity	2.20 ⭐	6.27 ⭐	4.07 👎
Method Length	56.69 ⭐	77.81 🙂	21.12 👎
Working memory	7.45 🙂	8.44 🙂	0.99 👎
Quality	74.97% 🙂	64.93% 🙂	-10.04% 👎

Other metrics	Before	After	Change
Lines	420	706	286

Changed files	Quality Before	Quality After	Quality Change
gdsfactory/simulation/sax/build_model.py	71.61% 🙂	71.63% 🙂	0.02% 👍
gdsfactory/simulation/sax/femwell_waveguide_model.py	75.52% ⭐	76.19% ⭐	0.67% 👍
gdsfactory/simulation/sax/meep_FDTD_model.py	79.57% ⭐	79.74% ⭐	0.17% 👍
gdsfactory/simulation/sax/parameter.py	80.67% ⭐	57.19% 🙂	-23.48% 👎

Here are some functions in these files that still need a tune-up:

File	Function	Complexity	Length	Working Memory	Quality	Recommendation
gdsfactory/simulation/sax/parameter.py	LithoParameter.layer_dilation_erosion	26 😞	202 😞	12 😞	32.87% 😞	Refactor to reduce nesting. Try splitting into smaller methods. Extract out complex expressions
gdsfactory/simulation/sax/parameter.py	LithoParameter.layer_round_corners	17 🙂	225 ⛔	14 😞	34.41% 😞	Try splitting into smaller methods. Extract out complex expressions
gdsfactory/simulation/sax/parameter.py	LithoParameter.layer_x_offset	10 🙂	134 😞	10 😞	54.99% 🙂	Try splitting into smaller methods. Extract out complex expressions
gdsfactory/simulation/sax/parameter.py	LithoParameter.layer_y_offset	10 🙂	134 😞	10 😞	54.99% 🙂	Try splitting into smaller methods. Extract out complex expressions
gdsfactory/simulation/sax/parameter.py	LayerStackThickness.__init__	4 ⭐	142 😞	9 🙂	61.87% 🙂	Try splitting into smaller methods

Legend and Explanation

The emojis denote the absolute quality of the code:

• ⭐ excellent
• 🙂 good
• 😞 poor
• ⛔ very poor

The 👍 and 👎 indicate whether the quality has improved or gotten worse with this pull request.

---

Please see our documentation here for details on how these metrics are calculated.

We are actively working on this report - lots more documentation and extra metrics to come!

Help us improve this quality report!

[codecov]

Codecov Report

Merging #1309 (a99fdbe) into main (7c5bb0e) will decrease coverage by 0.18%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main    #1309      +/-   ##
==========================================
- Coverage   70.14%   69.97%   -0.18%     
==========================================
  Files         363      364       +1     
  Lines       21468    21569     +101     
  Branches     3074     3086      +12     
==========================================
+ Hits        15058    15092      +34     
- Misses       5524     5583      +59     
- Partials      886      894       +8     

Impacted Files	Coverage Δ
gdsfactory/generic_tech/__init__.py	70.37% <0.00%> (-29.63%)	⬇️
gdsfactory/export/to_stl.py	25.92% <0.00%> (-6.08%)	⬇️
gdsfactory/components/edge_coupler_array.py	82.22% <0.00%> (-3.83%)	⬇️
gdsfactory/widgets/layout_viewer.py	14.01% <0.00%> (-3.32%)	⬇️
gdsfactory/show.py	47.05% <0.00%> (-2.95%)	⬇️
gdsfactory/pdk.py	56.88% <0.00%> (-1.87%)	⬇️
gdsfactory/component.py	67.24% <0.00%> (-1.06%)	⬇️
gdsfactory/port.py	66.08% <0.00%> (-0.17%)	⬇️
gdsfactory/__init__.py	94.00% <0.00%> (-0.12%)	⬇️
gdsfactory/config.py	66.00% <0.00%> (ø)
... and 7 more

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[joamatab]

Looks great Simon,

I think you should merge the polygons before rounding them

[joamatab]

The docs also have some issues, can merge once it's fixed

https://github.com/gdsfactory/gdsfactory/actions/runs/4218770431/jobs/7323514671

[simbilod]

Thanks @joamatab , I was already fusing the polygons for the erosion/dilation so that was an easy addition :)

@HelgeGehring the notebook now crashes because femwell tries to convert something complex to real. The plugin test fails for the same reason. Do you know the easiest fix?

Traceback (most recent call last):
  File "/home/simbil/Github/gdsfactory/gdsfactory/simulation/fem/test_mode_solver.py", line 71, in <module>
    test_compute_cross_section_mode_cache()
  File "/home/simbil/Github/gdsfactory/gdsfactory/simulation/fem/test_mode_solver.py", line 56, in test_compute_cross_section_mode_cache
    lams, basis, xs = compute_modes(with_cache=True, overwrite=False)
  File "/home/simbil/Github/gdsfactory/gdsfactory/simulation/fem/test_mode_solver.py", line 35, in compute_modes
    lams, basis, xs = compute_cross_section_modes(
  File "/home/simbil/Github/gdsfactory/gdsfactory/simulation/fem/mode_solver.py", line 63, in compute_cross_section_modes
    return compute_component_slice_modes(
  File "/home/simbil/Github/gdsfactory/gdsfactory/simulation/fem/mode_solver.py", line 191, in compute_component_slice_modes
    lams, basis, xs = mode_solver.compute_modes(
  File "/home/simbil/miniconda3/envs/gdsfactory/lib/python3.10/site-packages/femwell/mode_solver.py", line 83, in compute_modes
    lams, xs = solve(
  File "/home/simbil/miniconda3/envs/gdsfactory/lib/python3.10/site-packages/skfem/utils.py", line 262, in solve
    out = solve_eigen(A, b, x, I, solver, **kwargs)  # type: ignore
  File "/home/simbil/miniconda3/envs/gdsfactory/lib/python3.10/site-packages/skfem/utils.py", line 215, in solve_eigen
    return solver(A, M, **kwargs)
  File "/home/simbil/miniconda3/envs/gdsfactory/lib/python3.10/site-packages/femwell/solver.py", line 114, in solver
    K_ = PETSc.Mat().createAIJ(size=K.shape, csr=(K.indptr, K.indices, K.data))
  File "PETSc/Mat.pyx", line 357, in petsc4py.PETSc.Mat.createAIJ
  File "PETSc/petscmat.pxi", line 823, in petsc4py.PETSc.Mat_AllocAIJ
  File "PETSc/petscmat.pxi", line 794, in petsc4py.PETSc.Mat_AllocAIJ_CSR
  File "include/arraynpy.pxi", line 155, in petsc4py.PETSc.iarray_s
  File "include/arraynpy.pxi", line 130, in petsc4py.PETSc.iarray
TypeError: Cannot cast array data from dtype('complex128') to dtype('float64') according to the rule 'safe'

[HelgeGehring]

Might it be, that the "real" conversion of slepc is installed instead of the "complex" version?
Also, you can try to not specify the solver (to use scipy)

[simbilod]

strange, the gdsfactory makefile installs the complex version

Anyway, I am now making it use whatever is default in femwell and now it works!

[simbilod]

seems like scipy works locally but not on the build server for docs

[joamatab]

Happy to merge it when the tests pass

[simbilod]

Changed it to slepc again. Worked locally on my VM, and seems to work here too