|
from collections import defaultdict |
|
import os |
|
import json |
|
import csv |
|
|
|
import datasets |
|
|
|
_NAME="dummy_corpus_asr_es" |
|
_VERSION="1.0.0" |
|
|
|
_DESCRIPTION = """ |
|
An extremely small corpus of 40 audio files taken from Common Voice (es) with the objective of testing how to share datasets in Hugging Face. |
|
""" |
|
|
|
_CITATION = """ |
|
@misc{dummy-corpus-asr-es, |
|
title={Dummy Corpus for ASR in Spanish.}, |
|
author={Hernandez Mena, Carlos Daniel}, |
|
year={2022}, |
|
url={https://huggingface.co/datasets/carlosdanielhernandezmena/dummy_corpus_asr_es}, |
|
} |
|
""" |
|
|
|
_HOMEPAGE = "https://huggingface.co/datasets/carlosdanielhernandezmena/dummy_corpus_asr_es" |
|
|
|
_LICENSE = "CC-BY-4.0, See https://creativecommons.org/licenses/by/4.0/" |
|
|
|
_BASE_DATA_DIR = "data/" |
|
_METADATA_TRAIN = _BASE_DATA_DIR + "train.tsv" |
|
_METADATA_TEST = _BASE_DATA_DIR + "test.tsv" |
|
_METADATA_DEV = _BASE_DATA_DIR + "dev.tsv" |
|
|
|
class DummyCorpusAsrEsConfig(datasets.BuilderConfig): |
|
"""BuilderConfig for Dummy Corpus ASR ES.""" |
|
|
|
def __init__(self, name, **kwargs): |
|
name=_NAME |
|
super().__init__(name=name, **kwargs) |
|
|
|
class DummyCorpusAsrEs(datasets.GeneratorBasedBuilder): |
|
"""The Dummy Corpus ASR ES dataset.""" |
|
|
|
VERSION = datasets.Version(_VERSION) |
|
BUILDER_CONFIGS = [ |
|
DummyCorpusAsrEsConfig( |
|
name=_NAME, |
|
version=datasets.Version(_VERSION), |
|
) |
|
] |
|
|
|
def _info(self): |
|
features = datasets.Features( |
|
{ |
|
"audio_id": datasets.Value("string"), |
|
"audio": datasets.Audio(sampling_rate=16000), |
|
"split": datasets.Value("string"), |
|
"gender": datasets.Value("string"), |
|
"normalized_text": datasets.Value("string"), |
|
"relative_path": datasets.Value("string"), |
|
} |
|
) |
|
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_test=dl_manager.download_and_extract(_METADATA_TEST) |
|
metadata_dev=dl_manager.download_and_extract(_METADATA_DEV) |
|
|
|
meta_paths={"train":metadata_train,"test":metadata_test,"dev":metadata_dev} |
|
|
|
with open(metadata_train) as f: |
|
hash_meta_train = {x["audio_id"]: x for x in csv.DictReader(f, delimiter="\t")} |
|
|
|
with open(metadata_test) as f: |
|
hash_meta_test = {x["audio_id"]: x for x in csv.DictReader(f, delimiter="\t")} |
|
|
|
with open(metadata_dev) as f: |
|
hash_meta_dev = {x["audio_id"]: x for x in csv.DictReader(f, delimiter="\t")} |
|
|
|
hash_audios=defaultdict(dict) |
|
hash_audios["train"]=[] |
|
for audio_in in hash_meta_train: |
|
hash_audios["train"].append(hash_meta_train[audio_in]["relative_path"]) |
|
|
|
hash_audios["test"]=[] |
|
for audio_in in hash_meta_test: |
|
hash_audios["test"].append(hash_meta_test[audio_in]["relative_path"]) |
|
|
|
hash_audios["dev"]=[] |
|
for audio_in in hash_meta_dev: |
|
hash_audios["dev"].append(hash_meta_dev[audio_in]["relative_path"]) |
|
|
|
relative_paths=hash_audios |
|
|
|
audio_paths = dl_manager.download(hash_audios) |
|
|
|
local_extracted_audio_paths = dl_manager.download_and_extract(audio_paths) |
|
|
|
return [ |
|
datasets.SplitGenerator( |
|
name=datasets.Split.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": meta_paths["train"], |
|
"relative_paths":relative_paths["train"], |
|
} |
|
), |
|
datasets.SplitGenerator( |
|
name=datasets.Split.VALIDATION, |
|
gen_kwargs={ |
|
"audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["dev"]], |
|
"local_extracted_archives_paths": local_extracted_audio_paths["dev"], |
|
"metadata_paths": meta_paths["dev"], |
|
"relative_paths":relative_paths["dev"], |
|
} |
|
), |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TEST, |
|
gen_kwargs={ |
|
"audio_archives": [dl_manager.iter_archive(archive) for archive in audio_paths["test"]], |
|
"local_extracted_archives_paths": local_extracted_audio_paths["test"], |
|
"metadata_paths": meta_paths["test"], |
|
"relative_paths":relative_paths["test"], |
|
} |
|
), |
|
] |
|
|
|
def _generate_examples(self, audio_archives, local_extracted_archives_paths, metadata_paths,relative_paths): |
|
|
|
features = ["normalized_text","gender","split","relative_path"] |
|
|
|
meta_path = metadata_paths |
|
|
|
with open(meta_path) as f: |
|
metadata = {x["audio_id"]: x for x in csv.DictReader(f, delimiter="\t")} |
|
|
|
for audio_archive,local_path,rel_path in zip(audio_archives,local_extracted_archives_paths,relative_paths): |
|
|
|
audio_id =os.path.splitext(os.path.basename(rel_path))[0] |
|
path = local_path |
|
|
|
yield audio_id, { |
|
"audio_id": audio_id, |
|
**{feature: metadata[audio_id][feature] for feature in features}, |
|
"audio": {"path": path}, |
|
} |
|
|
|
|
