Module abd_clam::knn

Algorithms for K Nearest Neighbor search.

The stable algorithms are Linear, RepeatedRnn, GreedySieve, Sieve, and SieveSepCenter. The default algorithm is GreedySieve, as it was the best overall performer in our scaling experiments.

We will experiment with other algorithms in the future, and they will be added to this enum as they are being implemented. They should not be considered stable until they are documented as such.

Structs

• OrdNumber
  Field by which we rank elements in priority queue of hits.
• RevNumber
  Field by which we reverse-rank elements in priority queue of hits.

Enums

• Algorithm
  The algorithm to use for K-Nearest Neighbor search.