A library that implements RPLSH. The idea is that given a vector space and a query point (vector) we try to find the closest points (neighbours) as fast as possible.
How do we do this with RPLSH?
Points that are closest to each other are more likely to share a subspace, more on this point further. We split the vector space with n hyperplanes and create subspaces. Once several subspaces are created, for a given query point we search the from the subspace it belongs to, effectively reducing the number of points (neighbours) to search from.