numpy 2.0 compatibility

[xylar]

There has been a change in numpy as of 2.0 that affects the behavior of np.unique:
https://numpy.org/doc/stable/release/2.0.0-notes.html#np-unique-return-inverse-shape-for-multi-dimensional-inputs

This seems to affect:

jigsaw-python/jigsawpy/bisect.py

Lines 205 to 227 in d9d70e6

	if (edge.size != 0):
	__, efwd, erev = np.unique(
	np.sort(edge, axis=+1),
	return_index=True,
	return_inverse=True, axis=0)
	edge = edge[efwd, :]
	if (quad.size != 0):
	__, qfwd, qrev = np.unique(
	np.sort(quad, axis=+1),
	return_index=True,
	return_inverse=True, axis=0)
	quad = quad[qfwd, :]
	if (hexa.size != 0):
	__, hfwd, hrev = np.unique(
	np.sort(hexa, axis=+1),
	return_index=True,
	return_inverse=True, axis=0)
	hexa = hexa[hfwd, :]

I will submit a PR to hopefully address this shortly.