defeasible-nli.py

import datasets
import json
import os

citation='''
@inproceedings{rudinger-etal-2020-thinking,
    title = "Thinking Like a Skeptic: Defeasible Inference in Natural Language",
    author = "Rudinger, Rachel  and
      Shwartz, Vered  and
      Hwang, Jena D.  and
      Bhagavatula, Chandra  and
      Forbes, Maxwell  and
      Le Bras, Ronan  and
      Smith, Noah A.  and
      Choi, Yejin",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2020",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2020.findings-emnlp.418",
    doi = "10.18653/v1/2020.findings-emnlp.418",
    pages = "4661--4675"
}
'''

class DefeasibleNLIConfig(datasets.BuilderConfig):
    citation=citation

configs = ['atomic','snli','social']                                                              
splits=['train', 'test', 'dev']
_URLs = {(f,s):f"https://huggingface.co/datasets/metaeval/defeasible-nli/resolve/main/{f}_{s}.jsonl" for f in configs for s in splits}

class DefeasibleNLI(datasets.GeneratorBasedBuilder):
    
    BUILDER_CONFIGS = [
            DefeasibleNLIConfig(
                name=n,
                data_dir=n
            ) for n in configs
    ]

    def _split_generators(self, dl_manager: datasets.DownloadManager):
        path = lambda split: dl_manager.download(_URLs[self.config.name,split])
        return [ datasets.SplitGenerator(name=name, gen_kwargs={'path':path(split),'split':split})
                for name,split in zip([datasets.Split.TRAIN,datasets.Split.VALIDATION,datasets.Split.TEST],
                                       ['train','dev','test'])]

    def _info(self):
        return datasets.DatasetInfo()
    
    def _generate_examples(self,path,split):
        """Yields examples."""
        with open(path, "r", encoding="utf-8") as f:
            for id_, line in enumerate(f):
                line_dict = json.loads(line)
                if not line_dict['UpdateTypeImpossible']:
                    fields = ["Premise","Hypothesis","Update","UpdateType"]#,"UpdateTypeImpossible","UpdateTypeImpossibleReason"]
                    line_dict = {k:v for k,v in line_dict.items() if k in fields}
                    yield id_, line_dict