Composes this distance so that a threshold is applied to point-point distances. If the point-point distance of the underlying measure returns a value less than or equal the given threshold, then instead the value 0L is returned. This allows for quicker searches so that a nearest neighbour becomes an approximate nn within the given threshold (the first arbitrary point encountered with a distance smaller than the threshold will be returned).

Note that the threshold is directly compared to the result of distance, thus if the underlying measure uses a skewed distance, this must be taken into account. For example, if euclideanSq is used, and points within a radius of 4 should be approximated, a threshold of 4 * 4 = 16 must be chosen!

Applies a filter to this measure by constraining distances to objects b which lie within the given IntSquare. That is, if for example distance( a, b ) is called, first it is checked if b is within hyperCube. If so, the underlying measure is calculated, otherwise, Long.MaxValue is returned. This behaviour extends to the minDistance and maxDistance methods.

Calculates the distance between two points.

Calculates the maximum distance between a point and any possible point of a given hyper-cube. In the euclidean case, this is the distance to the hyper-cube b's corner that is furthest to the point a, no matter whether a is contained in b or not.

A value which will never be exceeded by the measure.

Calculates the minimum distance between a point and any possible point of a given hyper-cube. In the euclidean case, this is the distance to the hyper-cube b's corner that is closest to the point a, if a lies outside of b, or zero, if a lies within b.