fix bug in plot_boundaries of plot2D

[oskooi]

Fixes #2022.

Currently, the plot_boundaries subroutine of plot2D produces the incorrect output when the output_plane parameter passed to plot2D defines an area which does not include the PML boundaries. This can be demonstrated using the examples/oblique-planewave.py. The output region defined by output_plane does not include the PMLs but they incorrectly appear in the wrong location in the output.

Also includes some general formatting tweaks to visualization.py to improve readability.

cell_size = mp.Vector3(14,10,0)

pml_layers = [mp.PML(thickness=2,direction=mp.X)]

nonpml_vol = mp.Volume(center=mp.Vector3(), size=mp.Vector3(10,10,0))

sim.plot2D(fields=mp.Ez,
           output_plane=nonpml_vol)

With this PR, the output does not include the PML boundaries in the x direction.

cc @hammy4815

[stevengj]

Does it work correctly if the output_plane partially intersects the PML?

[codecov-commenter]

Codecov Report

Merging #2079 (2928e64) into master (5275f43) will decrease coverage by 0.14%.
The diff coverage is 68.75%.

@@            Coverage Diff             @@
##           master    #2079      +/-   ##
==========================================
- Coverage   73.25%   73.11%   -0.15%     
==========================================
  Files          17       17              
  Lines        4902     4895       -7     
==========================================
- Hits         3591     3579      -12     
- Misses       1311     1316       +5     

Impacted Files	Coverage Δ
python/visualization.py	43.47% <68.75%> (-1.61%)	⬇️

[oskooi]

Does it work correctly if the output_plane partially intersects the PML?

Yes, it works correctly in this case as well. Here is a demonstration based on the example above with a PML thickness of 2.0. The output_plane only intersects the PML over a distance of 1.0 on each side in the x direction.

nonpml_vol = mp.Volume(center=mp.Vector3(),
                       size=mp.Vector3(12,10,0))
sim.plot2D(fields=mp.Ez,
           output_plane=nonpml_vol)

For comparison, here is the output using the master branch (5275f43). The PML thickness incorrectly appears as 2.0.