import json import csv import os import datasets logger = datasets.logging.get_logger(__name__) _DESCRIPTION = "HOTPOTQA-PL Dataset" _SPLITS = ["corpus", "queries"] URL = "" _URLs = {subset: URL + f"{subset}.jsonl.gz" for subset in _SPLITS} class BEIR(datasets.GeneratorBasedBuilder): """BEIR-PL BenchmarkDataset.""" BUILDER_CONFIGS = [ datasets.BuilderConfig( name=name, description=f"This is the {name} in the Scifact-PL dataset.", ) for name in _SPLITS ] def _info(self): return datasets.DatasetInfo( description=_DESCRIPTION, features=datasets.Features({ "_id": datasets.Value("string"), "title": datasets.Value("string"), "text": datasets.Value("string"), }), supervised_keys=None, ) def _split_generators(self, dl_manager): """Returns SplitGenerators.""" my_urls = _URLs[self.config.name] data_dir = dl_manager.download_and_extract(my_urls) return [ datasets.SplitGenerator( name=self.config.name, # These kwargs will be passed to _generate_examples gen_kwargs={"filepath": data_dir}, ), ] def _generate_examples(self, filepath): """Yields examples.""" with open(filepath, encoding="utf-8") as f: texts = f.readlines() for i, text in enumerate(texts): text = json.loads(text) if 'metadata' in text: del text['metadata'] if "title" not in text: text["title"] = "" yield i, text