import datasets """ domain in {'alarm', 'calling', 'event', 'messaging', 'music', 'news', 'people', 'recipes', 'reminder', 'timer', 'weather'} """ _URL = "https://fb.me/mtop_dataset" _CITATION = """@article{li2020mtop, title={MTOP: A comprehensive multilingual task-oriented semantic parsing benchmark}, author={Li, Haoran and Arora, Abhinav and Chen, Shuohui and Gupta, Anchit and Gupta, Sonal and Mehdad, Yashar}, journal={arXiv preprint arXiv:2008.09335}, year={2020} }""" _DESCRIPTION = """ """ class MtopConfig(datasets.BuilderConfig): """BuilderConfig for Mtop.""" def __init__(self, **kwargs): """BuilderConfig for Mtop. Args: **kwargs: keyword arguments forwarded to super. """ super(MtopConfig, self).__init__(**kwargs) class Mtop(datasets.GeneratorBasedBuilder): BUILDER_CONFIGS = [ MtopConfig( name="mtop", version=datasets.Version("1.0.0", ""), description="Plain text", ), ] def _info(self): return datasets.DatasetInfo( description=_DESCRIPTION, features=datasets.Features( { "idx": datasets.Value("string"), "intent": datasets.Value("string"), "spans": datasets.Value("string"), "question": datasets.Value("string"), "domain": datasets.Value("string"), "lang": datasets.Value("string"), "logical_form": datasets.Value("string"), "tokenized_question": datasets.Value("string"), } ), # No default supervised_keys (as we have to pass both question # and context as input). supervised_keys=None, homepage="", citation=_CITATION, ) def _split_generators(self, dl_manager): filepath = dl_manager.download_and_extract(_URL) return [ datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": filepath,"split":"train"}), datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": filepath,"split":"eval"}), datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": filepath,"split":"test"}), ] def _generate_examples(self, filepath, split): """This function returns the examples in the raw (text) form.""" key = 0 for lang in "de en es fr hi th".split(): with open(f"{filepath}/mtop/{lang}/{split}.txt", encoding="utf-8") as f: for example in f: example = example.split("\t") dict_example = dict(idx=example[0], intent=example[1], spans=example[2], question=example[3], domain=example[4], lang=example[5], logical_form=example[6], tokenized_question=example[7]) yield key, dict_example key += 1