Implement ClassBreakOp quantile/jenks variants based on histogram instead of single values

Description

Using histograms allows to process larger rasters without using too much heap, thus avoiding OOMs, and in the particular case of Jenks, which has a complexity of N^2 (N being the samples used), also significantly reduce calculation time. Of course, this comes with a (small) loss of precision compared to using the full set of values.

Environment

None

Assignee

Andrea Aime

Reporter

Andrea Aime

Triage

None

Components

Fix versions

Priority

Medium
Configure