|
|
|
from collections import defaultdict |
|
import os |
|
import csv |
|
import datasets |
|
|
|
_NAME="RUV TV unknown speakers" |
|
_VERSION="1.0.0" |
|
_DESCRIPTION = """ |
|
RUV TV unknown speakers |
|
""" |
|
|
|
class RuvConfig(datasets.BuilderConfig): |
|
"""BuilderConfig for The RUV TV unknown speakers Corpus""" |
|
|
|
def __init__(self, name, **kwargs): |
|
name=_NAME |
|
super().__init__(name=name, **kwargs) |
|
|
|
class RuvAsr(datasets.GeneratorBasedBuilder): |
|
"""RUV TV unknown speakers 1.0""" |
|
|
|
VERSION = datasets.Version(_VERSION) |
|
BUILDER_CONFIGS = [ |
|
RuvConfig( |
|
name=_NAME, |
|
version=datasets.Version(_VERSION), |
|
) |
|
] |
|
|
|
def _info(self): |
|
features = datasets.Features( |
|
{ |
|
"audio_id": datasets.Value("string"), |
|
"audio": datasets.Audio(sampling_rate=16000), |
|
"show_name": datasets.Value("string"), |
|
"episode_id": datasets.Value("string"), |
|
"text": datasets.Value("string"), |
|
} |
|
) |
|
return datasets.DatasetInfo( |
|
description=_DESCRIPTION, |
|
features=features, |
|
) |
|
|
|
def _split_generators(self, dl_manager: datasets.DownloadManager): |
|
metadata_path = dl_manager.download_and_extract("metadata.tsv") |
|
audio_dir = "audio" |
|
audio_paths = {} |
|
audio_archives = {} |
|
texts_path = dl_manager.download_and_extract("text") |
|
|
|
with open(metadata_path) as f: |
|
metadata = csv.DictReader(f, fieldnames=["audio_id", "episode_id", "show_name", "duration"], delimiter="\t") |
|
for line in metadata: |
|
audio_path = os.path.join(audio_dir, line["show_name"], line["episode_id"], line["audio_id"] + ".flac") |
|
audio_paths[line["audio_id"]] = audio_path |
|
try: |
|
audio_archives[line["audio_id"]] = dl_manager.download_and_extract(audio_path) |
|
except FileNotFoundError as e: |
|
print(e) |
|
print("Failed to download", audio_path, "continuing...") |
|
continue |
|
|
|
return [ |
|
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"metadata_path": metadata_path, "audio_archives": audio_archives, "texts_path": texts_path}), |
|
] |
|
|
|
def _generate_examples(self, metadata_path, audio_archives, texts_path): |
|
with open(metadata_path) as f, open(texts_path) as g: |
|
metadata = csv.DictReader(f, fieldnames=["audio_id", "episode_id", "show_name", "duration"], delimiter="\t") |
|
for line, text_line in zip(metadata, g.readlines()): |
|
audio_id = line["audio_id"] |
|
audio_file = audio_archives[audio_id] |
|
|
|
yield audio_id, { |
|
"audio_id": audio_id, |
|
"audio": {"path": audio_file, "bytes": open(audio_file, "rb").read()}, |
|
"show_name": line["show_name"], |
|
"episode_id": line["episode_id"], |
|
"text": text_line.split(" ", maxsplit=1)[1].rstrip(), |
|
} |
|
|
