Weird behavior DiffractionSimulation.intensities

[din14970]

Describe the bug
Due to the way intensities is implemented as a property with a setter, some weird behavior can occur due to the direct beam mask.

Similar issues can be created with the coordinates attribute.

To Reproduce

from diffsims.sims.diffraction_simulation import DiffractionSimulation

coords = np.array([[0, 0], [1, 1], [2, 3]])
intensities = np.array([1, 1, 3])
sim = DiffractionSimulation(coordinates = coords, intensities = intensities, with_direct_beam=False)
sim.intensities = sim.intensities

# now throws an error because intensities is of the wrong shape
sim.intensities

Expected behavior
The setter needs to be more restrictive so that not just any array can be used, which creates inconsistencies.
I propose to check the size of the passed arrays and modify behavior depending on whether with_direct_beam is on or not.

[hakonanes]

Sounds good to me. What is the "ground truth" when it comes to shape here, the coordinates? Would be good to check the intensities array shape against the coordinates array shape upon setting these.

Looks like there are some other things in DiffractionSimulation which can be cleaned up, like unnecessary setting of private attributes in __init__ (since they are set upon the call to self.coordinates):

diffsims/diffsims/sims/diffraction_simulation.py

Lines 57 to 65 in 3c2540e

	self._coordinates = None
	self.coordinates = coordinates
	self.indices = indices
	self._intensities = None
	self.intensities = intensities
	self._calibration = (1.0, 1.0)
	self.calibration = calibration
	self.offset = offset
	self.with_direct_beam = with_direct_beam

[hakonanes]

I've marked this as a bug, because sim.intensities = sim.intensities shouldn't result in an error. Thus, this could go in a v0.4.3 patch release, if there is interest. Or, we could wait for a v0.5 release which could be released soon (a week or two?) because there are a few unreleased features in (see changelog).

[din14970]

I'm attempting to create a fix now