import os from pathlib import Path from typing import Dict, List, Tuple import datasets from seacrowd.utils.configs import SEACrowdConfig from seacrowd.utils.constants import Tasks from seacrowd.utils import schemas import json _CITATION = """\ @inproceedings{koto2020liputan6, title={Liputan6: A Large-scale Indonesian Dataset for Text Summarization}, author={Koto, Fajri and Lau, Jey Han and Baldwin, Timothy}, booktitle={Proceedings of the 1st Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 10th International Joint Conference on Natural Language Processing}, pages={598--608}, year={2020} } """ _LOCAL = False _LANGUAGES = ["ind"] # We follow ISO639-3 language code (https://iso639-3.sil.org/code_tables/639/data) _DATASETNAME = "liputan6" _DESCRIPTION = """ A large-scale Indonesian summarization dataset consisting of harvested articles from Liputan6.com, an online news portal, resulting in 215,827 document-summary pairs. """ _HOMEPAGE = "https://github.com/fajri91/sum_liputan6" _LICENSE = "CC-BY-SA 4.0" _URLS = { _DATASETNAME: "https://storage.googleapis.com/babert-pretraining/IndoNLG_finals/downstream_task/downstream_task_datasets.zip", } _SUPPORTED_TASKS = [Tasks.SUMMARIZATION] _SOURCE_VERSION = "1.0.0" _SEACROWD_VERSION = "2024.06.20" class Liputan6(datasets.GeneratorBasedBuilder): """A large-scale Indonesian summarization dataset consisting of harvested articles from Liputan6.com, an online news portal, resulting in 215,827 document-summary pairs.""" SOURCE_VERSION = datasets.Version(_SOURCE_VERSION) SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION) TYPE_LIST = ['canonical', 'xtreme'] BUILDER_CONFIGS = ( [ SEACrowdConfig( name="liputan6_{fold_name}_source".format(fold_name=i), version=_SOURCE_VERSION, description="liputan6 source schema", schema="source", subset_id="liputan6_{fold_name}".format(fold_name=i), ) for i in TYPE_LIST ] + [ SEACrowdConfig( name="liputan6_{fold_name}_seacrowd_t2t".format(fold_name=i), version=_SEACROWD_VERSION, description="liputan6 Nusantara schema", schema="seacrowd_t2t", subset_id="liputan6_{fold_name}".format(fold_name=i), ) for i in TYPE_LIST ] ) DEFAULT_CONFIG_NAME = "liputan6_canonical_source" def _info(self) -> datasets.DatasetInfo: if self.config.schema == "source": features = datasets.Features( { "document": datasets.Value("string"), "id": datasets.Value("string"), "summary": datasets.Value("string") } ) elif self.config.schema == "seacrowd_t2t": features = schemas.text2text_features return datasets.DatasetInfo( description=_DESCRIPTION, features=features, homepage=_HOMEPAGE, license=_LICENSE, citation=_CITATION, ) def _get_fold_name(self): subset_id = self.config.subset_id idx_fold = subset_id.index("_") file_id = subset_id[(idx_fold + 1):] return file_id def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]: fold_name = self._get_fold_name() urls = _URLS[_DATASETNAME] data_dir = Path(dl_manager.download_and_extract(urls)) location = { "train": "IndoNLG_downstream_tasks/liputan6/{fold_name}_train.json", "test": "IndoNLG_downstream_tasks/liputan6/{fold_name}_test.json", "dev": "IndoNLG_downstream_tasks/liputan6/{fold_name}_dev.json" } return [ datasets.SplitGenerator( name=datasets.Split.TRAIN, gen_kwargs={ "filepath": os.path.join(data_dir, location["train"].format(fold_name=fold_name)), "split": "train", }, ), datasets.SplitGenerator( name=datasets.Split.TEST, gen_kwargs={ "filepath": os.path.join(data_dir, location["test"].format(fold_name=fold_name)), "split": "test", }, ), datasets.SplitGenerator( name=datasets.Split.VALIDATION, gen_kwargs={ "filepath": os.path.join(data_dir, location["dev"].format(fold_name=fold_name)), "split": "dev", }, ), ] def _generate_examples(self, filepath: Path, split: str) -> Tuple[int, Dict]: if self.config.schema == "source": if "xtreme_train.json" in filepath: with open(filepath) as f: lines = f.read().split("{") LEN = len(lines) for i, line in enumerate(lines): if 0 < i < LEN-1: idx = line.index("}") line = "{"+line[:idx+1] each_data = json.loads(line) ex = { "id": each_data["id"], "document": each_data['text'], "summary": each_data['label'] } yield each_data["id"], ex else: with open(filepath) as f: data = json.load(f) for i, each_data in enumerate(data): ex = { "id": each_data["id"], "document": each_data['text'], "summary": each_data['label'] } yield each_data["id"], ex elif self.config.schema == "seacrowd_t2t": if "xtreme_train.json" in filepath: with open(filepath) as f: lines = f.read().split("{") LEN = len(lines) for i, line in enumerate(lines): if 0 < i < LEN-1: idx = line.index("}") line = "{"+line[:idx+1] each_data = json.loads(line) ex = { "id": each_data["id"], "text_1": each_data['text'], "text_2": each_data['label'], "text_1_name": "document", "text_2_name": "summary" } yield each_data["id"], ex else: with open(filepath) as f: data = json.load(f) for i, each_data in enumerate(data): ex = { "id": each_data["id"], "text_1": each_data['text'], "text_2": each_data['label'], "text_1_name": "document", "text_2_name": "summary" } yield each_data["id"], ex