""" NOTE: This file implements translation tasks using datasets from WMT conferences, provided by sacrebleu. Traditionally they are evaluated with BLEU scores. TER and CHRF are other options. We defer citations and descriptions of the many translations tasks used here to the SacreBLEU repo from which we've obtained the datasets: https://github.com/mjpost/sacrebleu/blob/master/sacrebleu/dataset.py Homepage: https://github.com/mjpost/sacrebleu/blob/master/sacrebleu/dataset.py """ from sacrebleu import sacrebleu import datasets import os import json _CITATION = """ @inproceedings{post-2018-call, title = "A Call for Clarity in Reporting {BLEU} Scores", author = "Post, Matt", booktitle = "Proceedings of the Third Conference on Machine Translation: Research Papers", month = oct, year = "2018", address = "Belgium, Brussels", publisher = "Association for Computational Linguistics", url = "https://www.aclweb.org/anthology/W18-6319", pages = "186--191", } """ class SacrebleuManual(datasets.GeneratorBasedBuilder): VERSION = datasets.Version("1.0.0") names = [name for name in list(sacrebleu.get_available_testsets()) if name not in ["wmt23", "wmt24"] and "wmt21" not in name] print(names) BUILDER_CONFIGS = [ datasets.BuilderConfig(name=f"{name.replace('/', '_')}_{langpair}", version=datasets.Version("1.0.0"), description="") for name in names for langpair in sacrebleu.get_langpairs_for_testset(name) ] def _info(self): features = datasets.Features( { "translation": datasets.Value("string"), } ) return datasets.DatasetInfo( description=f"Sacrebleu\n{self.config.description}", features=features, homepage="", license="", citation=_CITATION, ) def _split_generators(self, dl_manager): downloaded_files = dl_manager.download(f"{os.path.join(*self.config.name.split('_'))}.jsonl") print(downloaded_files) return [ datasets.SplitGenerator( name=datasets.Split.TEST, gen_kwargs={"path": downloaded_files}, ) ] # method parameters are unpacked from `gen_kwargs` as given in `_split_generators` def _generate_examples(self, path): with open(path, encoding="utf-8") as f: for key, row in enumerate(f): yield key, json.loads(row)