NEREL.py

import datasets
import json

_NAME = 'NEREL'
_CITATION = '''
@article{loukachevitch2021nerel,
  title={NEREL: A Russian Dataset with Nested Named Entities, Relations and Events},
  author={Loukachevitch, Natalia and Artemova, Ekaterina and Batura, Tatiana and Braslavski, Pavel and Denisov, Ilia and Ivanov, Vladimir and Manandhar, Suresh and Pugachev, Alexander and Tutubalina, Elena},
  journal={arXiv preprint arXiv:2108.13112},
  year={2021}
}
'''.strip()
_DESCRIPTION = 'A Russian Dataset with Nested Named Entities, Relations and Events'
_HOMEPAGE = 'https://doi.org/10.48550/arXiv.2108.13112'
_VERSION = '1.1.0'


class NERELBuilder(datasets.GeneratorBasedBuilder):
    _DATA_URLS = {
        'train': 'data/train.jsonl',
        'test': f'data/test.jsonl',
        'dev': f'data/dev.jsonl',
    }
    _ENT_TYPES_URLS = {
        'ent_types': 'ent_types.jsonl'
    }
    _REL_TYPES_URLS = {
        'rel_types': 'rel_types.jsonl'
    }
    VERSION = datasets.Version(_VERSION)
    BUILDER_CONFIGS = [
        datasets.BuilderConfig('data',
                               version=VERSION,
                               description='Data'),
        datasets.BuilderConfig('ent_types',
                               version=VERSION,
                               description='Entity types list'),
        datasets.BuilderConfig('rel_types',
                               version=VERSION,
                               description='Relation types list')
    ]
    DEFAULT_CONFIG_NAME = 'data'

    def _info(self) -> datasets.DatasetInfo:
        if self.config.name == 'data':
            features = datasets.Features({
                'id': datasets.Value('int32'),
                'text': datasets.Value('string'),
                'entities': datasets.Sequence(datasets.Value('string')),
                'relations': datasets.Sequence(datasets.Value('string')),
                'links': datasets.Sequence(datasets.Value('string'))
            })
        elif self.config.name == 'ent_types':
            features = datasets.Features({
                'type': datasets.Value('string'),
                'link': datasets.Value('string')
            })
        else:
            features = datasets.Features({
                'type': datasets.Value('string'),
                'arg1': datasets.Sequence(datasets.Value('string')),
                'arg2': datasets.Sequence(datasets.Value('string')),
            })
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            citation=_CITATION
        )

    def _split_generators(self, dl_manager: datasets.DownloadManager):
        if self.config.name == 'data':
            files = dl_manager.download(self._DATA_URLS)
            return [
                datasets.SplitGenerator(
                    name=datasets.Split.TRAIN,
                    gen_kwargs={'filepath': files['train']},
                ),
                datasets.SplitGenerator(
                    name=datasets.Split.TEST,
                    gen_kwargs={'filepath': files['test']},
                ),
                datasets.SplitGenerator(
                    name='dev',
                    gen_kwargs={'filepath': files['dev']},
                ),
            ]
        elif self.config.name == 'ent_types':
            files = dl_manager.download(self._ENT_TYPES_URLS)
            return [
                datasets.SplitGenerator(
                    name='ent_types',
                    gen_kwargs={'filepath': files['ent_types']},
                )
            ]
        else:
            files = dl_manager.download(self._REL_TYPES_URLS)
            return [
                datasets.SplitGenerator(
                    name='rel_types',
                    gen_kwargs={'filepath': files['rel_types']},
                )
            ]

    def _generate_examples(self, filepath):
        with open(filepath, encoding='utf-8') as f:
            for i, line in enumerate(f):
                yield i, json.loads(line)