Crate rand_split

This crate achieves the functionality of sklearn’s train_test_split to generate splits of the data (in this case, a slice), generalized for an arbitrary number of splits. It both provides functions (see split_parts) that work on slices and iterator traits (see PartsSplit)to work with streams of data.

Check out the examples in the repository for more information.

Traits

PartsSplit

An iterator trait to generate splits from a data stream of unknown lenght.

TTVSplit

An iterator type to generate train/test/validaton splits from a data stream of unknown lenght.

Functions

split_parts

Split the elements of a container in randomized sets which contain a a part (in splits) of the input.

train_test_split

Generate train-test splits. Wrapper around split_parts

ttv_split

Generate train-test-validation splits. Wrapper around split_parts