feat(datasets) Add VerticalSizePartitioner

[adam-narozniak]

Proposal

Enable users to specify the vertical split in terms of an absolute number of columns per partition or fraction of the total columns.

from flwr_datasets import FederatedDataset
from flwr_datasets.partitioner import VerticalSizePartitioner

partitioner = VerticalSizePartitioner(
    partition_sizes=[8, 4, 2],
    active_party_column="income",
    active_party_columns_mode="create_as_last"
)
fds = FederatedDataset(
    dataset="scikit-learn/adult-census-income",
    partitioners={"train": partitioner}
)
partitions = [fds.load_partition(i) for i in range(fds.partitioners["train"].num_partitions)]
print([partition.column_names for partition in partitions])
# [['occupation', 'relationship', 'sex', 'education', 'age', 'capital.loss', 'native.country', 'capital.gain'], ['marital.status', 'fnlwgt', 'education.num', 'hours.per.week'], ['workclass', 'race'], ['income']]

easier to read output format (print with pprint)

[['occupation',
  'relationship',
  'sex',
  'education',
  'age',
  'capital.loss',
  'native.country',
  'capital.gain'],
 ['marital.status', 'fnlwgt', 'education.num', 'hours.per.week'],
 ['workclass', 'race'],
 ['income']]