Make get_unique_points_mask more efficient

[uri-granta]

Related issue(s)/PRs: #781

Summary

Make get_unique_points_mask run in parallel. Since this approach detects connected graph components, it can trip up on examples like [1,2,3,4] with tolerance 1, which should return two points not one, even though all the points are connected. However, we can fix this using what will almost certainly be a small number of iterations.

Fully backwards compatible: yes

PR checklist

• The quality checks are all passing
• The bug case / new feature is covered by tests
• Any new features are well-documented (in docstrings or notebooks)

[khurram-ghani]

Looks really good. Glad we could come up with a nice parallel algorithm :-)