# coding=utf-8 # Copyright 2021 The HuggingFace Datasets Authors and the current dataset script contributor. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ OpenSLR Dataset""" from __future__ import absolute_import, division, print_function import os import re from pathlib import Path import datasets from datasets.tasks import AutomaticSpeechRecognition _DATA_URL = "https://openslr.org/resources/{}" _CITATION = """\ SLR70, SLR71: @inproceedings{guevara-rukoz-etal-2020-crowdsourcing, title = {{Crowdsourcing Latin American Spanish for Low-Resource Text-to-Speech}}, author = {Guevara-Rukoz, Adriana and Demirsahin, Isin and He, Fei and Chu, Shan-Hui Cathy and Sarin, Supheakmungkol and Pipatsrisawat, Knot and Gutkin, Alexander and Butryna, Alena and Kjartansson, Oddur}, booktitle = {Proceedings of The 12th Language Resources and Evaluation Conference (LREC)}, year = {2020}, month = may, address = {Marseille, France}, publisher = {European Language Resources Association (ELRA)}, url = {https://www.aclweb.org/anthology/2020.lrec-1.801}, pages = {6504--6513}, ISBN = {979-10-95546-34-4}, } """ _DESCRIPTION = """\ OpenSLR is a site devoted to hosting speech and language resources, such as training corpora for speech recognition, and software related to speech recognition. We intend to be a convenient place for anyone to put resources that they have created, so that they can be downloaded publicly. """ _HOMEPAGE = "https://openslr.org/" _LICENSE = "" _RESOURCES = { "SLR70": { "Language": "Nigerian English", "LongName": "Crowdsourced high-quality Nigerian English speech data set", "Category": "Speech", "Summary": "Data set which contains recordings of Nigerian English", "Files": ["en_ng_female.zip", "en_ng_male.zip"], "IndexFiles": ["line_index.tsv", "line_index.tsv"], "DataDirs": ["", ""], }, "SLR71": { "Language": "Chilean Spanish", "LongName": "Crowdsourced high-quality Chilean Spanish speech data set", "Category": "Speech", "Summary": "Data set which contains recordings of Chilean Spanish", "Files": ["es_cl_female.zip", "es_cl_male.zip"], "IndexFiles": ["line_index.tsv", "line_index.tsv"], "DataDirs": ["", ""], }, } class OpenSlrConfig(datasets.BuilderConfig): """BuilderConfig for OpenSlr.""" def __init__(self, name, **kwargs): """ Args: data_dir: `string`, the path to the folder containing the files in the downloaded .tar citation: `string`, citation for the data set url: `string`, url for information about the data set **kwargs: keyword arguments forwarded to super. """ self.language = kwargs.pop("language", None) self.long_name = kwargs.pop("long_name", None) self.category = kwargs.pop("category", None) self.summary = kwargs.pop("summary", None) self.files = kwargs.pop("files", None) self.index_files = kwargs.pop("index_files", None) self.data_dirs = kwargs.pop("data_dirs", None) description = ( f"Open Speech and Language Resources dataset in {self.language}. Name: {self.name}, " f"Summary: {self.summary}." ) super(OpenSlrConfig, self).__init__(name=name, description=description, **kwargs) class OpenSlr(datasets.GeneratorBasedBuilder): DEFAULT_WRITER_BATCH_SIZE = 32 BUILDER_CONFIGS = [ OpenSlrConfig( name=resource_id, language=_RESOURCES[resource_id]["Language"], long_name=_RESOURCES[resource_id]["LongName"], category=_RESOURCES[resource_id]["Category"], summary=_RESOURCES[resource_id]["Summary"], files=_RESOURCES[resource_id]["Files"], index_files=_RESOURCES[resource_id]["IndexFiles"], data_dirs=_RESOURCES[resource_id]["DataDirs"], ) for resource_id in _RESOURCES.keys() ] def _info(self): features = datasets.Features( { "path": datasets.Value("string"), "audio": datasets.Audio(sampling_rate=48_000), "sentence": datasets.Value("string"), } ) return datasets.DatasetInfo( description=_DESCRIPTION, features=features, supervised_keys=None, homepage=_HOMEPAGE, license=_LICENSE, citation=_CITATION, task_templates=[AutomaticSpeechRecognition(audio_column="audio", transcription_column="sentence")], ) def _split_generators(self, dl_manager): """Returns SplitGenerators.""" resource_number = self.config.name.replace("SLR", "") urls = [f"{_DATA_URL.format(resource_number)}/{file}" for file in self.config.files] if urls[0].endswith(".zip"): dl_paths = dl_manager.download_and_extract(urls) path_to_indexs = [os.path.join(path, f"{self.config.index_files[i]}") for i, path in enumerate(dl_paths)] path_to_datas = [os.path.join(path, f"{self.config.data_dirs[i]}") for i, path in enumerate(dl_paths)] archives = None else: archives = dl_manager.download(urls) path_to_indexs = dl_manager.download(self.config.index_files) path_to_datas = self.config.data_dirs return [ datasets.SplitGenerator( name=datasets.Split.TRAIN, gen_kwargs={ "path_to_indexs": path_to_indexs, "path_to_datas": path_to_datas, "archive_files": [dl_manager.iter_archive(archive) for archive in archives] if archives else None, }, ), ] def _generate_examples(self, path_to_indexs, path_to_datas, archive_files): """Yields examples.""" counter = -1 for i, path_to_index in enumerate(path_to_indexs): with open(path_to_index, encoding="utf-8") as f: lines = f.readlines() for id_, line in enumerate(lines): # Following regexs are needed to normalise the lines, since the datasets # are not always consistent and have bugs: line = re.sub(r"\t[^\t]*\t", "\t", line.strip()) field_values = re.split(r"\t\t?", line) if len(field_values) != 2: continue filename, sentence = field_values # set absolute path for audio file path = os.path.join(path_to_datas[i], f"{filename}.wav") counter += 1 yield counter, {"path": path, "audio": path, "sentence": sentence}