fix: improve find_segments via numpy diff

[drbh]

This PR updates find_segments to use a numpy array rather than a list. This allow us to a more efficiently find the segments via masking.

when testing the method in isolation this results in a ~3x speedup on 4096 indices

Benchmarking results:
--------------------------------------------------------------------------------
        Size   Original (ms)  Optimized (ms)    Speedup
--------------------------------------------------------------------------------
        4096            0.32            0.08       3.94
        8192            0.59            0.14       4.13
       16384            1.24            0.20       6.06
       32768            2.31            0.32       7.12
       65536            4.57            0.58       7.93