from collections import defaultdict import os import json import csv csv.field_size_limit(100000000) import datasets _NAME="commonvoice_benchmark_catalan_accents" _VERSION="1.0.0" _AUDIO_EXTENSIONS=".mp3" _DESCRIPTION = """ A new presentation of the corpus Catalan Common Voice v17.0 - metadata annotated version with the splits redefined to benchmark ASR models with various Catalan accent """ _CITATION = """ @misc{armentanoaccents2024, title={Common Voice Benchmark Catalan Accents}, author={Armentano, Carme}, publisher={Barcelona Supercomputing Center} year={2024}, url={https://huggingface.co/datasets/projecte-aina/commonvoice_benchmark_catalan_accents}, } """ _HOMEPAGE = "https://huggingface.co/datasets/projecte-aina/commonvoice_benchmark_catalan_accents" _LICENSE = "CC-BY-4.0, See https://creativecommons.org/licenses/by/4.0/" _BASE_DATA_DIR = "corpus/" _METADATA_TRAIN = os.path.join(_BASE_DATA_DIR,"files","train.tsv") _METADATA_BALEARIC_FEM = os.path.join(_BASE_DATA_DIR,"files","balearic_female.tsv") _METADATA_BALEARIC_MALE = os.path.join(_BASE_DATA_DIR,"files","balearic_male.tsv") _METADATA_CENTRAL_FEMALE = os.path.join(_BASE_DATA_DIR,"files","central_female.tsv") _METADATA_CENTRAL_MALE = os.path.join(_BASE_DATA_DIR,"files","central_male.tsv") _METADATA_NORTHERN_FEMALE = os.path.join(_BASE_DATA_DIR,"files","northern_female.tsv") _METADATA_NORTHERN_MALE = os.path.join(_BASE_DATA_DIR,"files","northern_male.tsv") _METADATA_NORTHWESTERN_FEMALE = os.path.join(_BASE_DATA_DIR,"files","northwestern_female.tsv") _METADATA_NORTHWESTERN_MALE = os.path.join(_BASE_DATA_DIR,"files","northwestern_male.tsv") _METADATA_VALENCIAN_FEMALE = os.path.join(_BASE_DATA_DIR,"files","valencian_female.tsv") _METADATA_VALENCIAN_MALE = os.path.join(_BASE_DATA_DIR,"files","valencian_male.tsv") _TARS_REPO = os.path.join(_BASE_DATA_DIR,"files","tars_repo.paths") class CommonVoiceBenchmarkCatalanAccentsConfig(datasets.BuilderConfig): """BuilderConfig for The Common Voice Benchmark Catalan Accents""" def __init__(self, name, **kwargs): name=_NAME super().__init__(name=name, **kwargs) class CommonVoiceBenchmarkCatalanAccents(datasets.GeneratorBasedBuilder): """Common Voice Benchmark Catalan Accents""" VERSION = datasets.Version(_VERSION) BUILDER_CONFIGS = [ CommonVoiceBenchmarkCatalanAccentsConfig( name=_NAME, version=datasets.Version(_VERSION), ) ] def _info(self): features = datasets.Features( { "audio": datasets.Audio(sampling_rate=16000), "client_id": datasets.Value("string"), "path": datasets.Value("string"), "sentence": datasets.Value("string"), "up_votes": datasets.Value("int32"), "down_votes": datasets.Value("int32"), "age": datasets.Value("string"), "gender": datasets.Value("string"), "accents": datasets.Value("string"), "variant": datasets.Value("string"), "locale": datasets.Value("string"), "segment": datasets.Value("string"), "mean quality": datasets.Value("string"), "stdev quality": datasets.Value("string"), "annotated_accent": datasets.Value("string"), "annotated_accent_agreement": datasets.Value("string"), "annotated_gender": datasets.Value("string"), "annotated_gender_agreement": datasets.Value("string"), "propagated_gender": datasets.Value("string"), "propagated_accents": datasets.Value("string"), "propagated_accents_norm": datasets.Value("string"), "variant_norm": datasets.Value("string"), "assigned_accent": datasets.Value("string"), "assigned_gender": datasets.Value("string"), "duration": datasets.Value("float32"), } ) return datasets.DatasetInfo( description=_DESCRIPTION, features=features, homepage=_HOMEPAGE, license=_LICENSE, citation=_CITATION, ) def _split_generators(self, dl_manager): metadata_train=dl_manager.download_and_extract(_METADATA_TRAIN) metadata_balearic_fem=dl_manager.download_and_extract(_METADATA_BALEARIC_FEM) metadata_balearic_male=dl_manager.download_and_extract(_METADATA_BALEARIC_MALE) metadata_central_female=dl_manager.download_and_extract(_METADATA_CENTRAL_FEMALE) metadata_central_male=dl_manager.download_and_extract(_METADATA_CENTRAL_MALE) metadata_northern_female=dl_manager.download_and_extract(_METADATA_NORTHERN_FEMALE) metadata_northern_male=dl_manager.download_and_extract(_METADATA_NORTHWESTERN_MALE) metadata_northwestern_female=dl_manager.download_and_extract(_METADATA_NORTHWESTERN_FEMALE) metadata_northwestern_male=dl_manager.download_and_extract(_METADATA_NORTHWESTERN_MALE) metadata_valencian_female=dl_manager.download_and_extract(_METADATA_VALENCIAN_FEMALE) metadata_valencian_male=dl_manager.download_and_extract(_METADATA_VALENCIAN_MALE) tars_repo=dl_manager.download_and_extract(_TARS_REPO) hash_tar_files=defaultdict(dict) with open(tars_repo,'r') as f: hash_tar_files['train']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['balearic_fem']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['balearic_male']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['central_female']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['central_male']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['northern_female']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['northern_male']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['northwestern_female']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['northwestern_male']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['valencian_female']=[path.replace('\n','') for path in f] with open(tars_repo,'r') as f: hash_tar_files['valencian_male']=[path.replace('\n','') for path in f] hash_meta_paths={"train":metadata_train, "balearic_fem":metadata_balearic_fem, "balearic_male":metadata_balearic_male, "central_female":metadata_central_female, "central_male":metadata_central_male, "northern_female":metadata_northern_female, "northern_male":metadata_northern_male, "northwestern_female":metadata_northwestern_female, "northwestern_male":metadata_northwestern_male, "valencian_female":metadata_valencian_female, "valencian_male":metadata_valencian_male} audio_paths = dl_manager.download(hash_tar_files) splits=["train","balearic_fem","balearic_male","central_female","central_male","northern_female", "northern_male","northwestern_female","northwestern_male","valencian_female","valencian_male"] local_extracted_audio_paths = ( dl_manager.extract(audio_paths) if not dl_manager.is_streaming else { split:[None] * len(audio_paths[split]) for split in splits } ) return [ datasets.SplitGenerator( name="train", gen_kwargs={ "audio_archives":[dl_manager.iter_archive(archive) for archive in audio_paths["train"]], "local_extracted_archives_paths": local_extracted_audio_paths["train"], "metadata_paths": hash_meta_paths["train"], } ), datasets.SplitGenerator( name="balearic_fem", gen_kwargs={ "audio_archives":[dl_manager.iter_archive(archive) for archive in audio_paths["balearic_fem"]], "local_extracted_archives_paths": local_extracted_audio_paths["balearic_fem"], "metadata_paths": hash_meta_paths["balearic_fem"], } ), datasets.SplitGenerator( name="balearic_male", gen_kwargs={ "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["balearic_male"]], "local_extracted_archives_paths": local_extracted_audio_paths["balearic_male"], "metadata_paths": hash_meta_paths["balearic_male"], } ), datasets.SplitGenerator( name="central_female", gen_kwargs={ "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["central_female"]], "local_extracted_archives_paths": local_extracted_audio_paths["central_female"], "metadata_paths": hash_meta_paths["central_female"], } ), datasets.SplitGenerator( name="central_male", gen_kwargs={ "audio_archives":[dl_manager.iter_archive(archive) for archive in audio_paths["central_male"]], "local_extracted_archives_paths": local_extracted_audio_paths["central_male"], "metadata_paths": hash_meta_paths["central_male"], } ), datasets.SplitGenerator( name="northern_female", gen_kwargs={ "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["northern_female"]], "local_extracted_archives_paths": local_extracted_audio_paths["northern_female"], "metadata_paths": hash_meta_paths["northern_female"], } ), datasets.SplitGenerator( name="northern_male", gen_kwargs={ "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["northern_male"]], "local_extracted_archives_paths": local_extracted_audio_paths["northern_male"], "metadata_paths": hash_meta_paths["northern_male"], } ), datasets.SplitGenerator( name="northwestern_female", gen_kwargs={ "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["northwestern_female"]], "local_extracted_archives_paths": local_extracted_audio_paths["northwestern_female"], "metadata_paths": hash_meta_paths["northwestern_female"], } ), datasets.SplitGenerator( name="northwestern_male", gen_kwargs={ "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["northwestern_male"]], "local_extracted_archives_paths": local_extracted_audio_paths["northwestern_male"], "metadata_paths": hash_meta_paths["northwestern_male"], } ), datasets.SplitGenerator( name="valencian_female", gen_kwargs={ "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["valencian_female"]], "local_extracted_archives_paths": local_extracted_audio_paths["valencian_female"], "metadata_paths": hash_meta_paths["valencian_female"], } ), datasets.SplitGenerator( name="valencian_male", gen_kwargs={ "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["valencian_male"]], "local_extracted_archives_paths": local_extracted_audio_paths["valencian_male"], "metadata_paths": hash_meta_paths["valencian_male"], } ), ] def _generate_examples(self, audio_archives, local_extracted_archives_paths, metadata_paths): features = ["client_id","sentence","up_votes","down_votes","age","gender", "accents","variant","locale","segment","mean quality","stdev quality", "annotated_accent","annotated_accent_agreement","annotated_gender", "annotated_gender_agreement","propagated_gender","propagated_accents", "propagated_accents_norm","variant_norm","assigned_accent","assigned_gender", "duration","path"] with open(metadata_paths) as f: metadata = {x["path"]: x for x in csv.DictReader(f, delimiter="\t")} for audio_archive, local_extracted_archive_path in zip(audio_archives, local_extracted_archives_paths): for audio_filename, audio_file in audio_archive: audio_id =os.path.splitext(os.path.basename(audio_filename))[0] audio_id=audio_id+_AUDIO_EXTENSIONS path = os.path.join(local_extracted_archive_path, audio_filename) if local_extracted_archive_path else audio_filename try: yield audio_id, { "path": audio_id, **{feature: metadata[audio_id][feature] for feature in features}, "audio": {"path": path, "bytes": audio_file.read()}, } except: continue