import json import datasets logger = datasets.logging.get_logger(__name__) _DESCRIPTION = """\ Ukrainian News Dataset This is a dataset of news articles downloaded from various Ukrainian websites. The dataset contains approximately 10 569 428 JSON objects (news) """ _URLS = [ "ukrainian-news1.jsonl", "ukrainian-news2.jsonl", "ukrainian-news3.jsonl", "ukrainian-news4.jsonl", "ukrainian-news5.jsonl" ] class UkrainianNews(datasets.GeneratorBasedBuilder): """Ukrainian News Dataset""" VERSION = datasets.Version("0.0.1") DEFAULT_CONFIG_NAME = "default" BUILDER_CONFIGS = [ datasets.BuilderConfig(name="default", version=VERSION, description=""), ] def _info(self): return datasets.DatasetInfo( description=_DESCRIPTION, features=datasets.Features( { "id": datasets.Value("int32"), "url": datasets.Value("string"), "title": datasets.Value("string"), "text": datasets.Value("string"), "owner": datasets.Value("string"), "datetime": datasets.Value("string"), } ) ) def _split_generators(self, dl_manager): downloaded_files = dl_manager.download(_URLS) return [ datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepaths": downloaded_files}) ] def _generate_examples(self, filepaths): """This function returns the examples in the raw (text) form.""" logger.info("generating examples from = %s", filepaths) for path in filepaths: with open(path, encoding="utf-8") as f: for news_str in f: news = json.loads(news_str) yield news['id'], news