beir_hotpotqa.py

"""
""" # TODO
try:
    import ir_datasets
except ImportError as e:
    raise ImportError('ir-datasets package missing; `pip install ir-datasets`')
import datasets

IRDS_ID = 'beir/hotpotqa'
IRDS_ENTITY_TYPES = {'docs': {'doc_id': 'string', 'text': 'string', 'title': 'string', 'url': 'string'}, 'queries': {'query_id': 'string', 'text': 'string'}}

_CITATION = '@inproceedings{Yang2018Hotpotqa,\n    title = "{H}otpot{QA}: A Dataset for Diverse, Explainable Multi-hop Question Answering",\n    author = "Yang, Zhilin  and\n      Qi, Peng  and\n      Zhang, Saizheng  and\n      Bengio, Yoshua  and\n      Cohen, William  and\n      Salakhutdinov, Ruslan  and\n      Manning, Christopher D.",\n    booktitle = "Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing",\n    month = oct # "-" # nov,\n    year = "2018",\n    address = "Brussels, Belgium",\n    publisher = "Association for Computational Linguistics",\n    url = "https://www.aclweb.org/anthology/D18-1259",\n    doi = "10.18653/v1/D18-1259",\n    pages = "2369--2380"\n}\n@article{Thakur2021Beir,\n  title = "BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models",\n  author = "Thakur, Nandan and Reimers, Nils and Rücklé, Andreas and Srivastava, Abhishek and Gurevych, Iryna", \n  journal= "arXiv preprint arXiv:2104.08663",\n  month = "4",\n  year = "2021",\n  url = "https://arxiv.org/abs/2104.08663",\n}'

_DESCRIPTION = "" # TODO

class beir_hotpotqa(datasets.GeneratorBasedBuilder):
    BUILDER_CONFIGS = [datasets.BuilderConfig(name=e) for e in IRDS_ENTITY_TYPES]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features({k: datasets.Value(v) for k, v in IRDS_ENTITY_TYPES[self.config.name].items()}),
            homepage=f"https://ir-datasets.com/beir#beir/hotpotqa",
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        return [datasets.SplitGenerator(name=self.config.name)]

    def _generate_examples(self):
        dataset = ir_datasets.load(IRDS_ID)
        for i, item in enumerate(getattr(dataset, self.config.name)):
            key = i
            if self.config.name == 'docs':
                key = item.doc_id
            elif self.config.name == 'queries':
                key = item.query_id
            yield key, item._asdict()

    def as_dataset(self, split=None, *args, **kwargs):
        split = self.config.name # always return split corresponding with this config to avid returning a redundant DatasetDict layer
        return super().as_dataset(split, *args, **kwargs)