mswc.py

# coding=utf-8

"""MSWC keyword spotting classification dataset."""


import os
import textwrap
import datasets
import itertools
import typing as tp
from pathlib import Path

from ._mswc import (
    TRAIN_ENG, VALIDATION_ENG, TEST_ENG, 
    TRAIN_SPA, VALIDATION_SPA, TEST_SPA, 
    TRAIN_IND, VALIDATION_IND, TEST_IND, 
)

FOLDER_IN_ARCHIVE = "genres"
SAMPLE_RATE = 16_000

_ENG_FILENAME = 'eng-kw-archive.tar.gz'
_SPA_FILENAME = 'spa-kw-archive.tar.gz'
_IND_FILENAME = 'ind-kw-archive.tar.gz'

CLASS_ENG = list(set([fileid.split('_')[0] for fileid in TRAIN_ENG]))
CLASS_SPA = list(set([fileid.split('_')[0] for fileid in TRAIN_SPA]))
CLASS_IND = list(set([fileid.split('_')[0] for fileid in TRAIN_IND]))


class MswcConfig(datasets.BuilderConfig):
    """BuilderConfig for MSWC."""
    
    def __init__(self, features, **kwargs):
        super(MswcConfig, self).__init__(version=datasets.Version("0.0.1", ""), **kwargs)
        self.features = features


class MSWC(datasets.GeneratorBasedBuilder):

    BUILDER_CONFIGS = [
        MswcConfig(
            features=datasets.Features(
                {
                    "file": datasets.Value("string"),
                    "audio": datasets.Audio(sampling_rate=SAMPLE_RATE),
                    "keyword": datasets.Value("string"),
                    "label": datasets.ClassLabel(names=CLASS_ENG), 
                }
            ),
            name="english", 
            description=textwrap.dedent(
                """\
                Keyword spotting classifies each audio for its keywords as a multi-class
                classification, where keywords are in the same pre-defined set for both training and testing. 
                The evaluation metric is accuracy (ACC).
                """
            ), 
        ), 
        MswcConfig(
            features=datasets.Features(
                {
                    "file": datasets.Value("string"),
                    "audio": datasets.Audio(sampling_rate=SAMPLE_RATE),
                    "keyword": datasets.Value("string"),
                    "label": datasets.ClassLabel(names=CLASS_SPA), 
                }
            ),
            name="spanish", 
            description=textwrap.dedent(
                """\
                Keyword spotting classifies each audio for its keywords as a multi-class
                classification, where keywords are in the same pre-defined set for both training and testing. 
                The evaluation metric is accuracy (ACC).
                """
            ), 
        ), 
        MswcConfig(
            features=datasets.Features(
                {
                    "file": datasets.Value("string"),
                    "audio": datasets.Audio(sampling_rate=SAMPLE_RATE),
                    "keyword": datasets.Value("string"),
                    "label": datasets.ClassLabel(names=CLASS_IND), 
                }
            ),
            name="indian", 
            description=textwrap.dedent(
                """\
                Keyword spotting classifies each audio for its keywords as a multi-class
                classification, where keywords are in the same pre-defined set for both training and testing. 
                The evaluation metric is accuracy (ACC).
                """
            ), 
        ), 
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description="",
            features=self.config.features,
            supervised_keys=None,
            homepage="",
            citation="",
            task_templates=None,
        )

    def _split_generators(self, dl_manager):
        """Returns SplitGenerators."""

        if self.config.name == "english":
            archive_path = dl_manager.extract(_ENG_FILENAME)
        elif self.config.name == "spanish":
            archive_path = dl_manager.extract(_SPA_FILENAME)
        elif self.config.name == "indian":
            archive_path = dl_manager.extract(_IND_FILENAME)
            
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN, gen_kwargs={"archive_path": archive_path, "split": "train"}
            ), 
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION, gen_kwargs={"archive_path": archive_path, "split": "validation"}
            ), 
            datasets.SplitGenerator(
                name=datasets.Split.TEST, gen_kwargs={"archive_path": archive_path, "split": "test"}
            ), 
        ]

    def _generate_examples(self, archive_path, split=None):
        
        if self.config.name == 'english':
            extensions = ['.wav']
            _, _walker = fast_scandir(archive_path, extensions, recursive=True)
            if split == 'train':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in TRAIN_ENG]
            elif split == 'validation':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in VALIDATION_ENG]
            elif split == 'test':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in TEST_ENG]
                
        elif self.config.name == 'spanish':
            extensions = ['.wav']
            _, _walker = fast_scandir(archive_path, extensions, recursive=True)
            if split == 'train':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in TRAIN_SPA]
            elif split == 'validation':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in VALIDATION_SPA]
            elif split == 'test':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in TEST_SPA]
                
        elif self.config.name == 'indian':
            extensions = ['.wav']
            _, _walker = fast_scandir(archive_path, extensions, recursive=True)
            if split == 'train':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in TRAIN_IND]
            elif split == 'validation':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in VALIDATION_IND]
            elif split == 'test':
                _walker = [fileid for fileid in _walker if Path(fileid).stem in TEST_IND]

        for guid, audio_path in enumerate(_walker):
            yield guid, {
                "id": str(guid),
                "file": audio_path, 
                "audio": audio_path, 
                "keyword": Path(audio_path).stem.split('_')[0], 
                "label": Path(audio_path).stem.split('_')[0], 
            }


def fast_scandir(path: str, exts: tp.List[str], recursive: bool = False):
    # Scan files recursively faster than glob
    # From github.com/drscotthawley/aeiou/blob/main/aeiou/core.py
    subfolders, files = [], []

    try:  # hope to avoid 'permission denied' by this try
        for f in os.scandir(path):
            try:  # 'hope to avoid too many levels of symbolic links' error
                if f.is_dir():
                    subfolders.append(f.path)
                elif f.is_file():
                    if os.path.splitext(f.name)[1].lower() in exts:
                        files.append(f.path)
            except Exception:
                pass
    except Exception:
        pass

    if recursive:
        for path in list(subfolders):
            sf, f = fast_scandir(path, exts, recursive=recursive)
            subfolders.extend(sf)
            files.extend(f)  # type: ignore

    return subfolders, files