2D slice plotting

[hervasa2]

A long awaited feature has finally arrived! 2D slice plotting is now available via seriestype = :slice

plot(sim.electric_potential, φ = 0)
plot!(sim.detector, show_passives = false, st = :slice, φ = 0, legend = :topleft)

It is based on the sample surface method. Only a subset of these points are selected and projected on a 2D plane which can be set with the x,y,z,φ attributes. Default is set to y = 0. Sampling density can be set as usual with n_samples. Even though the default density is set 5 times higher than for :samplesurface plots can be generated without lag. Example:

plot(
    plot(sim.detector, st = :slice, x = 2u"mm"),
    plot(sim.detector, st = :slice, z = 0u"mm"),
    plot(sim.detector, st = :slice, φ = 90u"°"),
    layout = (1,3), size = (1000,300)
    )

Sampling density was not set correctly when multiple CSGs where plotted at once (eg. :samplesurface in the plot recipe for SolidStateDetectors). This was corrected and carried over to :slice.

Default values for n_samples are 40 and 100 or 200 for samplesurface and :slice respectively. :slice checks if any surface is coplanar with the plotting plane. In such cases n_samples = 100, else n_samples = 200.

Looping over values for the plotting planes gives:

Scan along y	Scan along z	Scan along φ

[fhagemann]

At first glance, this looks very promising to me. Are you planning to add some text for the plotting documentation? https://juliaphysics.github.io/SolidStateDetectors.jl/stable/man/plotting/

[hervasa2]

Yes. But first I would like to consider if this is the most intuitive front end slice_plane selection. What do you think of it now versus having x,y,... as keyword arguments?

[fhagemann]

I think that having x,y,... as keyword argument might indeed be more intuitive, especially keeping in mind that most users would use this together with potential slice plot.

[fhagemann]

I tried out the code, works very smoothly. Great job!!

Some questions I would have:

• [x] The plot recipe is based on sampling points and plotting them as scatter seriestype. What would be the recommended syntax to draw bigger/thinner lines?
• [x] Regarding the point above: Would it make sense to use the linewidth series attribute to also be able to modify the thickness of the contacts in the slice plot? The slice plot looks basically like a line and might have people thinking that it is actually a path plot, where the line thickness could be adjusted via linewidth/lw.
• [x] When making slices in ϕ, is there any convenient way to let the method also sample points in "negative" ϕ direction, similar to the full_det keyword in the electric potential plotting?
• [x] Can we add simple examples explaining how to use the slice seriestype to the documentation?

[hervasa2]

All the requested changes have been pushed. linewidth/lw is reinterpreted as markersize in the internal plot recipes for st = :slice. full_det is now also suported. Additionally when a z slice is chosen the projection/proj = :polar keyword is supported.

  plot(sim.electric_potential, z = 30u"mm")
  plot!(sim.detector, st = :slice, z = 30u"mm", proj = :polar, lw = 2)