a h-@sddlZddlmZddlmZmZmZddlmZddl m Z ddl m Z ddl mZmZdZd gd gd d d ddddZdZdZedddZeeeeedddZeedddZeedddZGd d!d!e ZGd"d#d#eZGd$d%d%eZdS)&N)Path)ListTupleUnion)Tensor)Dataset)download_url_to_file) _extract_zip_load_waveformi>zvox1_dev_wav.zip)zIhttps://thor.robots.ox.ac.uk/~vgg/data/voxceleb/vox1a/vox1_dev_wav_partaazIhttps://thor.robots.ox.ac.uk/~vgg/data/voxceleb/vox1a/vox1_dev_wav_partabzIhttps://thor.robots.ox.ac.uk/~vgg/data/voxceleb/vox1a/vox1_dev_wav_partaczIhttps://thor.robots.ox.ac.uk/~vgg/data/voxceleb/vox1a/vox1_dev_wav_partad)Z@21ec6ca843659ebc2fdbe04b530baa4f191ad4b0971912672d92c158f32226a0Z@311d21e0c8cbf33573a4fce6c80e5a279d80736274b381c394319fc557159a04Z@92b64465f2b2a3dc0e4196ae8dd6828cbe9ddd1f089419a11e4cbfe2e1750df0Z@00e6190c770b27f27d2a3dd26ee15596b17066b715ac111906861a7d09a211a5) archive_nameurls checksumszvox1_test_wav.zipzGhttps://thor.robots.ox.ac.uk/~vgg/data/voxceleb/vox1a/vox1_test_wav.zipZ@8de57f347fe22b2c24526e9f444f689ecf5096fc2a92018cf420ff6b5b15eaea)r urlchecksumdevtestzBhttps://www.robots.ox.ac.uk/~vgg/data/voxceleb/meta/iden_split.txtzAhttps://www.robots.ox.ac.uk/~vgg/data/voxceleb/meta/veri_test.txt)rootc CsdD]}t|d}tj||}|dkrt|d}t|d}t|d}t||D]d\}}tj|tj|} t|| |dt| d} || WdqZ1s0YqZWdn1s0Yn&t|d }t|d }t|||dt |qdS) Nrr rr r wb)Z hash_prefixrbrr) _ARCHIVE_CONFIGSospathjoinopenzipbasenamerwritereadr ) rarchiver archive_pathr r frr file_pathZf_splitr#K/var/www/auris/lib/python3.9/site-packages/torchaudio/datasets/voxceleb1.py_download_extract_wavs&s       N  r%)rr"subsetreturnc Csg}|dkrd}n|dkr d}nd}t|d<}|D]&}|\}}t||kr4||q4Wdn1sp0Yt|S)Ntrainrr)rsplitintappendsorted) rr"r&f_listindexr!lineidrr#r#r$ _get_flist<s   *r5)rveri_test_pathcCsZg}t|d8}|D]"}|\}}}||||fqWdn1sL0Y|S)Nr,)rr-r/)rr6r1r!r3labelpath1path2r#r#r$_get_paired_flistLs  0r:)r" _ext_audiocCs8|ddd\}}}||d}d|||g}|S)N/-)r-replacer)r"r; speaker_idZ youtube_idZ utterance_idfile_idr#r#r$ _get_file_idUs rCc@sZeZdZdZdZdeeefeddddZ e dd d Z e dd d Z e d ddZ dS) VoxCeleb1a"*VoxCeleb1* :cite:`nagrani2017voxceleb` dataset. Args: root (str or Path): Path to the directory where the dataset is found or downloaded. download (bool, optional): Whether to download the dataset if it is not found at root path. (Default: ``False``). z.wavFN)rdownloadr'cCsJt|}tj|d|_tj|jsF|s>td|jdt|dS)NZwavzDataset not found at z5. Please set `download=True` to download the dataset.)rfspathrr_pathisdir RuntimeErrorr%)selfrrEr#r#r$__init__gs  zVoxCeleb1.__init__)ncCstdSNNotImplementedErrorrJrLr#r#r$ get_metadatarszVoxCeleb1.get_metadatacCstdSrMrNrPr#r#r$ __getitem__uszVoxCeleb1.__getitem__r'cCstdSrMrNrJr#r#r$__len__xszVoxCeleb1.__len__)F)__name__ __module__ __qualname____doc__r;rstrrboolrKr.rQrRrUr#r#r#r$rD\s  rDcseZdZdZdedfeeefeeeddfdd Z e e ee e efdd d Z e e e e e efdd d Ze d ddZZS)VoxCeleb1Identificationu*VoxCeleb1* :cite:`nagrani2017voxceleb` dataset for speaker identification task. Each data sample contains the waveform, sample rate, speaker id, and the file id. Args: root (str or Path): Path to the directory where the dataset is found or downloaded. subset (str, optional): Subset of the dataset to use. Options: ["train", "dev", "test"]. (Default: ``"train"``) meta_url (str, optional): The url of meta file that contains the list of subset labels and file paths. The format of each row is ``subset file_path". For example: ``1 id10006/nLEBBc9oIFs/00003.wav``. ``1``, ``2``, ``3`` mean ``train``, ``dev``, and ``test`` subest, respectively. (Default: ``"https://www.robots.ox.ac.uk/~vgg/data/voxceleb/meta/iden_split.txt"``) download (bool, optional): Whether to download the dataset if it is not found at root path. (Default: ``False``). Note: The file structure of `VoxCeleb1Identification` dataset is as follows: └─ root/ └─ wav/ └─ speaker_id folders Users who pre-downloaded the ``"vox1_dev_wav.zip"`` and ``"vox1_test_wav.zip"`` files need to move the extracted files into the same ``root`` directory. r(FN)rr&meta_urlrEr'cs^t|||dvrtdtj|tj|}tj|sJt||t |j |||_ dS)N)r(rrz0`subset` must be one of ['train', 'dev', 'test']) superrK ValueErrorrrrrexistsrr5rG_flist)rJrr&r]rEmeta_list_path __class__r#r$rKs  z VoxCeleb1Identification.__init__rLr'cCs@|j|}t||j}|dd}t|dd}|t||fS)aGet metadata for the n-th sample from the dataset. Returns filepath instead of waveform, but otherwise returns the same fields as :py:func:`__getitem__`. Args: n (int): The index of the sample Returns: Tuple of the following items; str: Path to audio int: Sample rate int: Speaker ID str: File ID r?rr+N)rarCr;r-r. SAMPLE_RATE)rJrLr"rBrAr#r#r$rQs   z$VoxCeleb1Identification.get_metadatacCs2||}t|j|d|d}|f|ddS)adLoad the n-th sample from the dataset. Args: n (int): The index of the sample to be loaded Returns: Tuple of the following items; Tensor: Waveform int: Sample rate int: Speaker ID str: File ID rr)NrQr rG)rJrLmetadataZwaveformr#r#r$rRs z#VoxCeleb1Identification.__getitem__rScCs t|jSrMlenrarTr#r#r$rUszVoxCeleb1Identification.__len__)rVrWrXrY_IDEN_SPLIT_URLrrZrr[rKr.rrQrrRrU __classcell__r#r#rcr$r\|s r\cseZdZdZedfeeefeeddfdd Z e e eee e eefddd Z e e e e e e eefdd d Ze d d dZZS)VoxCeleb1Verificationu*VoxCeleb1* :cite:`nagrani2017voxceleb` dataset for speaker verification task. Each data sample contains a pair of waveforms, sample rate, the label indicating if they are from the same speaker, and the file ids. Args: root (str or Path): Path to the directory where the dataset is found or downloaded. meta_url (str, optional): The url of meta file that contains a list of utterance pairs and the corresponding labels. The format of each row is ``label file_path1 file_path2". For example: ``1 id10270/x6uYqmx31kE/00001.wav id10270/8jEAjG6SegY/00008.wav``. ``1`` means the two utterances are from the same speaker, ``0`` means not. (Default: ``"https://www.robots.ox.ac.uk/~vgg/data/voxceleb/meta/veri_test.txt"``) download (bool, optional): Whether to download the dataset if it is not found at root path. (Default: ``False``). Note: The file structure of `VoxCeleb1Verification` dataset is as follows: └─ root/ └─ wav/ └─ speaker_id folders Users who pre-downloaded the ``"vox1_dev_wav.zip"`` and ``"vox1_test_wav.zip"`` files need to move the extracted files into the same ``root`` directory. FN)rr]rEr'csLt||tj|tj|}tj|s:t||t|j ||_ dSrM) r^rKrrrrr`rr:rGra)rJrr]rErbrcr#r$rKs   zVoxCeleb1Verification.__init__recCs@|j|\}}}t|}t||j}t||j}||t|||fS)agGet metadata for the n-th sample from the dataset. Returns filepaths instead of waveforms, but otherwise returns the same fields as :py:func:`__getitem__`. Args: n (int): The index of the sample Returns: Tuple of the following items; str: Path to audio file of speaker 1 str: Path to audio file of speaker 2 int: Sample rate int: Label str: File ID of speaker 1 str: File ID of speaker 2 )rar.rCr;rf)rJrLr7Zfile_path_spk1Zfile_path_spk2Z file_id_spk1Z file_id_spk2r#r#r$rQs   z"VoxCeleb1Verification.get_metadatacCsJ||}t|j|d|d}t|j|d|d}||f|ddS)aLoad the n-th sample from the dataset. Args: n (int): The index of the sample to be loaded. Returns: Tuple of the following items; Tensor: Waveform of speaker 1 Tensor: Waveform of speaker 2 int: Sample rate int: Label str: File ID of speaker 1 str: File ID of speaker 2 rr*r)Nrg)rJrLrhZ waveform_spk1Z waveform_spk2r#r#r$rRs z!VoxCeleb1Verification.__getitem__rScCs t|jSrMrirTr#r#r$rU4szVoxCeleb1Verification.__len__)rVrWrXrY_VERI_TEST_URLrrZrr[rKr.rrQrrRrUrlr#r#rcr$rms &  rm)rpathlibrtypingrrrZtorchrZtorch.utils.datarZtorchaudio._internalrZtorchaudio.datasets.utilsr r rfrrkrnrZr%r5r:rCrDr\rmr#r#r#r$s2      [