Alias projection plots to particle projection plots

[matthewturk]

yt is able to recognize when you probably want a ParticleProjectionPlot instead of a ProjectionPlot. Maybe we should have it automatically redirect, if you use ProjectionPlot with a particle field.

[chummels]

Isn't this complicated by the fact that particle fields can have SPH smoothing kernels and be turned into true projection plots? Or with the work that @qobilidop did a year or so ago, that even non-SPH particle fields can get smoothing kernels applied to them to make projection plots? Maybe there's still a way to tell what the user wants, but it seems like this might make it a bit more challenging to tell automatically.

[matthewturk]

So the difference in behavior I want to somehow get around is this -- if you do yt.ProjectionPlot(ds, "x", "density") on an SPH field, it does what you probably want. If you do this with any field that isn't SPH, it doesn't, but instead suggests "Why not use a particle plot?" Before, we were able to do deposition fields; now, the viz task requires a call to a different wrapper object.

[chummels]

Oh, that makes sense. I think the only trick here is for it to recognize if a particle based dataset has the appropriate SPH fields, even if it isn't a native SPH dataset, but it may have had those relevant fields added with add_sph_fields() as per PR #2186 . But you probably already considered that.

[neutrinoceros]

Incidentally, I believe that #3489 paves the road to do this in the future.