Small typo, docstring update, and fixed a bug.

[tcpekin]

size_marker wasn't doing anything in the plot_structure function when I was testing it. It looks like radius was hardcoded with a magic number. I've just slotted in size_marker, added a new magic number, and updated the docstring.

It seems to work for me, but I haven't extensively tested it with a variety of structures, just one.

Also, why is this code written like this, instead of using the commented out ax.scatter method? I was certainly confused.

[sezelt]

This part of the code is for structures with partial occupancy, and it displays the atoms as little pies broken up by the occupancy fraction on each site. For structures with all sites fully occupied, it uses ax.scatter.

[tcpekin]

@sezelt I'm not sure this code ever hits the scatter call to be honest. I am testing with https://legacy.materialsproject.org/materials/mp-81/ (pure gold) and get the following occupancy when I drop a print statement right before the plotting of the atoms.

structure = get_mp_structure('mp-81')
crystal = py4DSTEM.process.diffraction.Crystal.from_pymatgen_structure(
    structure=structure,
)

fig, ax = crystal.plot_structure(
    returnfig=True,
    size_marker=200,
    zone_axis_lattice=[1,1,0],
    perspective_axes=True,
)

So I think the check of if occ is None: will never execute. That's where my confusion stems from. If I remember, I'll submit a small bugfix.

[sezelt]

I think you're right and there's an incorrect assumption, as the Crystal initializer will always set self.occupancy and thus the if occ is None: branches are never taken. Instead we should test for any element of occupancy being non-one.

[tcpekin]

The latest commit should fix it, but I didn't test it to be honest. It may have an extra cast to an array, but that's probably fine.