size=0 adjoint montiors for single pixel dims in custom mediums

[tylerflex]

An optimization to reduce data usage for CustomMedium adjoint monitors. size=0 for any dims with just 1 pixel, since we only evaluate at centers.

Note; eventually would be realy nice to just have a point-cloud monitor and do these custom for each structure

[momchil-flex]

So this is probably nice to get in right now but in general these two things should not be coupled? Like I should be able to define my custom medium by a single pixel if I want it extruded in a given dimension, but potentially still integrate fields across that dimension which would be important e.g. for polarization mixing components?

[tylerflex]

Not sure I see the problem with this?

Until we add the option to do full volume integration, I think it makes sense to just restrict the monitor size, like we do with PolySlab.

Agree that the point cloud monitor is kind of a separate issue. In principle all 3 are separate?

[momchil-flex]

Yeah, we can merge this in right away, I was just thinking that we need to make sure this is removed when we incorporate full volume integration.

Have you done any tests that the gradients have not changed? I am not sure if this would be captured by our automatic testing.

[tylerflex]

I tested with a couple different notebooks (mode converter and WDM) both looked the same as before, but were a bit faster