Implement efficient approximate search structures

[tavianator]

See https://github.com/erikbern/ann-benchmarks

The state of the art is currently HNSW.

[tavianator]

Or apparently https://github.com/google-research/google-research/tree/master/scann