Source code for pykeen.datasets.hetionet

# -*- coding: utf-8 -*-

"""The Hetionet dataset.

Get a summary with ``python -m pykeen.datasets.hetionet``

import click
from docdata import parse_docdata
from more_click import verbose_option

from .base import SingleTabbedDataset
from ..typing import TorchRandomHint

__all__ = [

URL = ""

[docs]@parse_docdata class Hetionet(SingleTabbedDataset): """The Hetionet dataset from [himmelstein2017]_. In its publication [himmelstein2017]_, it is demonstrated to be useful for link prediction in drug repositioning and made publicly available through its `GitHub repository <>`_ in several formats. The link prediction algorithm showcased does not rely on embeddings, which leaves room for interesting comparison. One such comparison was made during the master's thesis of Lingling Xu [xu2019]_. --- name: Hetionet citation: author: Himmelstein year: 2017 link: github: hetio/hetionet single: true statistics: entities: 45158 relations: 24 triples: 2250197 training: 1800157 testing: 225020 validation: 225020 """ def __init__( self, random_state: TorchRandomHint = 0, **kwargs, ): """Initialize the `Hetionet <>`_ dataset from [himmelstein2017]_. :param random_state: The random seed to use in splitting the dataset. Defaults to 0. :param kwargs: keyword arguments passed to :class:`pykeen.datasets.base.SingleTabbedDataset`. """ super().__init__( url=URL, random_state=random_state, **kwargs, )
@click.command() @verbose_option def _main(): from pykeen.datasets import get_dataset ds = get_dataset(dataset=Hetionet) ds.summarize() if __name__ == "__main__": _main()