multidim.PointCloud.dists

PointCloud.dists(indices0, indices1)

Compute distances points indices0 and indices1. indices0 and indices1 must be 1-dimensional np arrays, so use not “5” but “np.array([5])”

The return is a np array with shape == (len(indices0), len(indices1))

This uses the distance cache, depending on self.cache_type. You can query the size of the cache with cache_usage().