re-instate undocumented get_array_slice_dimensions in Python interface

[oskooi]

#1456 removed get_array_slice_dimensions from the Python interface which unfortunately was discovered afterwards to be used by the adjoint solver in a couple of places:

meep/python/adjoint/optimization_problem.py

Lines 191 to 195 in 37117fe

	# store design region voxel parameters
	self.design_grids = []
	for drm in self.design_region_monitors:
	s = [self.sim.get_array_slice_dimensions(c,vol=drm.where)[0] for c in [mp.Ex,mp.Ey,mp.Ez]]
	self.design_grids += [YeeDims(*s)]

meep/python/adjoint/optimization_problem.py

Lines 408 to 415 in 37117fe

	# get monitor dft output array dimensions
	dims = []
	for m in mon:
	ci_dims = []
	for ci in c:
	comp = ci if yee_grid else mp.Dielectric
	ci_dims += [simob.get_array_slice_dimensions(comp,vol=m.where)[0]]
	dims.append(ci_dims)

Rather than modify the above lines to use the low-level _get_array_slice_dimensions, this PR simply re-instates get_array_slice_dimensions (in its original form). Note that this is an internal function used only by the adjoint solver that is not documented in the user manual.

We will need to do a version 1.17.1 release after this is merged since 1.17.0 broke the adjoint solver.

Also, it would be good to add some unit tests for the adjoint solver to prevent these types of fixes in the future.

[smartalecH]

Looks good to me. I ran locally just to make sure we weren't missing anything else:

The bend tutorial gradients look fine.

I agree though -- a simple adjoint unit test would really help (it's on my list of things to do).