basque_parliament_1.py

# Copyright 2023 GTTS (http://gtts.ehu.eus)
#
# 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.

# Lint as: python3
"""Basque Parliament dataset"""


import datasets
from datasets.utils.py_utils import size_str
import os
import csv
from tqdm import tqdm

from .languages import LANGUAGES
from .release_stats import STATS


_CITATION = """\
"""

_HOMEPAGE = "https://huggingface.co/datasets/gttsehu/basque_parliament_1"

_LICENSE = "https://creativecommons.org/publicdomain/zero/1.0/"

_DESCRIPTION = (
    f"Basque Parliament dataset blah blah blah..."
    f"blah blah blah..."
    f"blah blah blah..."
)

_BASE_URL = "https://huggingface.co/datasets/gttsehu/basque_parliament_1/resolve/main/"

_AUDIO_URL = _BASE_URL + "audio/{split}_{shard_idx}.tar"
_METADATA_URL =  _BASE_URL + "metadata/{split}.tsv"

class BasqueParliamentConfig(datasets.BuilderConfig):
    """BuilderConfig for BasqueParliament."""

    def __init__(self, name, version, **kwargs):
        self.language = kwargs.pop("language", None)
        self.release_date = kwargs.pop("release_date", None)
        self.num_clips = kwargs.pop("num_clips", None)
        self.num_speakers = kwargs.pop("num_speakers", None)
        self.validated_hr = kwargs.pop("validated_hr", None)
        self.total_hr = kwargs.pop("total_hr", None)
        self.size_bytes = kwargs.pop("size_bytes", None)
        self.size_human = size_str(self.size_bytes)
        description = _DESCRIPTION

        super(BasqueParliamentConfig, self).__init__(
            name = name,
            version = datasets.Version(version),
            description = _DESCRIPTION,
            **kwargs,
        )


class BasqueParliament(datasets.GeneratorBasedBuilder):
    """Basque Parliament is a free Basque-Spanish speech corpus."""

    DEFAULT_CONFIG_NAME = "all"
    
    BUILDER_CONFIGS = [
        BasqueParliamentConfig(
            name=lang,
            version=STATS["version"],
            language=LANGUAGES[lang],
            release_date=STATS["date"],
            num_clips=lang_stats["clips"],
            num_speakers=lang_stats["users"],
            total_hr=float(lang_stats["totalHrs"]) if lang_stats["totalHrs"] else None,
            size_bytes=int(lang_stats["size"]) if lang_stats["size"] else None,
        )
        for lang, lang_stats in STATS["locales"].items()
    ]

    def _info(self):
        description = (
            f"Basque Parliament dataset blah blah blah..."
            f"blah blah blah..."
            f"blah blah blah..."
        )
        features = datasets.Features(
            {
                "path": datasets.Value("string"),
                "audio": datasets.features.Audio(sampling_rate=16_000),
                "sentence": datasets.Value("string"),
                "speaker_id": datasets.Value("string"),
                "language": datasets.Value("string"),
                "PRR": datasets.Value("float32"),
                "length": datasets.Value("float32"),
            }
        )

        return datasets.DatasetInfo(
            description = _DESCRIPTION,
            features = features,
            supervised_keys = None,
            homepage = _HOMEPAGE,
            license = _LICENSE,
            citation = _CITATION,
            version = self.config.version,
        )


    def _split_generators(self, dl_manager):
        lang = self.config.name
        
        audio_urls = {}
        splits = ("train", "train_clean", "dev", "test")
        for split in splits:
            if split == "train_clean": continue
            audio_urls[split] = [
                _AUDIO_URL.format(split=split, shard_idx=i) for i in range(STATS["n_shards"][split])
            ]
        audio_urls["train_clean"]=audio_urls["train"]
        archive_paths = dl_manager.download(audio_urls)
        local_extracted_archive_paths = dl_manager.extract(archive_paths) if not dl_manager.is_streaming else {}

        metadata_urls = {split: _METADATA_URL.format(lang=lang, split=split) for split in splits}
        metadata_paths = dl_manager.download_and_extract(metadata_urls)

        split_generators = []
        split_names = {
            "train": datasets.Split.TRAIN,
            "dev": datasets.Split.VALIDATION,
            "test": datasets.Split.TEST,
        }
        for split in splits:
            split_generators.append(
                datasets.SplitGenerator(
                    name=split_names.get(split, split),
                    gen_kwargs={
                        "local_extracted_archive_paths": local_extracted_archive_paths.get(split),
                        "archives": [dl_manager.iter_archive(path) for path in archive_paths.get(split)],
                        "metadata_path": metadata_paths[split],
                    },
                ),
            )

        return split_generators

    def _generate_examples(self, local_extracted_archive_paths, archives, metadata_path):
        lang = self.config.name
        data_fields = list(self._info().features.keys())
        metadata = {}
        with open(metadata_path, encoding="utf-8") as f:
            reader = csv.DictReader(f, delimiter="\t", quoting=csv.QUOTE_NONE)
            metadata = { row["path"]:row for row in tqdm(reader, desc="Reading metadata...") }

        excluded = 0
        for i, audio_archive in enumerate(archives):
            for path, file in audio_archive:
                if path not in metadata :
                    excluded += 1
                    continue
                result = dict(metadata[path])
                if lang == "all" or lang == result["language"] :
                    # set the audio feature and the path to the extracted file
                    path = os.path.join(local_extracted_archive_paths[i], path) if local_extracted_archive_paths else path
                    result["audio"] = {"path": path, "bytes": file.read()}
                    result["path"] = path
                    yield path, result
        print(excluded,'audio files not found in metadata')