Module blue_data_selector

source code

This takes anticorrelated samples from a sequence of data. In other words, it tends to choose data that have not been seen frequently before. With blueness>=1, it never shows a sample n times until all samples have been shown n-1 times.

(Note that it is possible for an item to appear twice in sucession, if it appears as the Nth item in a set of N, then appears as the first item in the second set of N.)

Imports: heapq, random, math