splitter

K-Fold Data Splitting with persistence

installation:

no pypi package planned so far.

git clone https://github.com/p-sodmann/splitter/
cd splitter
pip install .

usage:

from splitter.splitter import Splitter

splitter = Splitter(k_splits)

example_data = ["a", "b", "c", "d", "e", "f"]

for example in example_data:
    splitter.add(example)
    
splitter.splits
>>> defaultdict(list, {8: ['a'], 4: ['b'], 2: ['c', 'f'], 0: ['d'], 1: ['e']})

# to generate a [train, valid, tests] split with the first 8 splits as training data:
splitter.get_split([[0,1,2,3,4,5,6,7],[8],[9]])
>>> [['a', 'c', 'f', 'b', 'd'], ['e'], []]

# note that the tests data is empty.