TCAD adaptive remeshing

[simbilod]

Bisect the mesh according to simulation results, making the simulator usable. For instance:

Initial simulation:

    c = DDComponent(
        component=waveguide,
        xsection_bounds=[(4, -4), (4, 4)],
        full_layerstack=get_layer_stack_generic(),
        physical_layerstack=physical_layerstack,
        doping_info=get_doping_info_generic(),
        contact_info=contact_info,
        resolutions=resolutions,
        mesh_scaling_factor=1e-4,
        background_tag=None,
    )

    c.ddsolver()
    c.save_device("test1.dat")

Now use the simulation results to compute a new mesh size, and generate a new simulation with it:

    meshing_field = c.get_refined_mesh(
        factor=2.0,
        refine_dict={
            "Potential": {
                "func": lambda x0, x1: abs(x0 - x1),
                "threshold": 0.05,
            },
            "Electrons": {
                "func": lambda x0, x1: np.abs(np.log10(x0) - np.log10(x1)),
                "threshold": 1,
            },
        },
        refine_regions=["si"],
    )

   c.delete_device()
    c = DDComponent(
        component=waveguide,
        xsection_bounds=[(4, -4), (4, 4)],
        full_layerstack=get_layer_stack_generic(),
        physical_layerstack=physical_layerstack,
        doping_info=get_doping_info_generic(),
        contact_info=contact_info,
        resolutions=resolutions,
        mesh_scaling_factor=1e-4,
        background_tag=None,
    )

    c.ddsolver(global_meshsize_array=meshing_field)
    c.save_device("test2.dat")

The new mesh is only finer where it needs to be.

TODO:

• Initialize the new simulation with the last results instead of complete reinitialization
• Repeat the remeshing until some criteria (number of bisections, or some tolerance on result)

[sourcery-ai]

Sourcery Code Quality Report

❌  Merging this PR will decrease code quality in the affected files by 13.43%.

Quality metrics	Before	After	Change
Complexity	2.79 ⭐	4.46 ⭐	1.67 👎
Method Length	62.38 🙂	88.23 🙂	25.85 👎
Working memory	7.65 🙂	11.24 😞	3.59 👎
Quality	75.09% ⭐	61.66% 🙂	-13.43% 👎

Other metrics	Before	After	Change
Lines	354	700	346

Changed files	Quality Before	Quality After	Quality Change
gdsfactory/simulation/devsim/get_solver.py	78.57% ⭐	58.38% 🙂	-20.19% 👎
gdsfactory/simulation/gmsh/mesh.py	71.21% 🙂	70.50% 🙂	-0.71% 👎

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

File	Function	Complexity	Length	Working Memory	Quality	Recommendation
gdsfactory/simulation/devsim/get_solver.py	DDComponent.get_refined_mesh	14 🙂	339 ⛔	15 😞	31.02% 😞	Try splitting into smaller methods. Extract out complex expressions
gdsfactory/simulation/devsim/get_solver.py	DDComponent.__init__	4 ⭐	150 😞	28 ⛔	43.74% 😞	Try splitting into smaller methods. Extract out complex expressions
gdsfactory/simulation/gmsh/mesh.py	mesh_from_polygons	4 ⭐	140 😞	16 ⛔	51.03% 🙂	Try splitting into smaller methods. Extract out complex expressions
gdsfactory/simulation/gmsh/mesh.py	tag_interfaces	6 ⭐	84 🙂	10 😞	65.33% 🙂	Extract out complex expressions
gdsfactory/simulation/devsim/get_solver.py	DDComponent.ddsolver	4 ⭐	176 😞	6 ⭐	65.49% 🙂	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 #1074 (05417a0) into main (6bae6aa) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #1074   +/-   ##
=======================================
  Coverage   72.32%   72.32%           
=======================================
  Files         436      436           
  Lines       22569    22569           
  Branches     2993     2993           
=======================================
  Hits        16323    16323           
  Misses       5365     5365           
  Partials      881      881           

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