o -ig@sddlZddlZddlmZddlmZmZmZddlZddl m Z ddl m Z ddl mZmZdZdZd Zgd ZGd d d e Zd ededefddZdS)N)Path)OptionalTupleUnion)Dataset)download_url_to_file) _extract_tar_load_waveformz6https://speech.fit.vutbr.cz/files/quesst14Database.tgzi@@4f869e06bc066bbe9c5dde31dbd3909a0870d70291110ebbb38878dcbc2fc5e4)albanianbasqueczech nnenglishromanianslovakc @seZdZdZ  ddeeefdedeededd f d d Z d e de ee effd dZ d e de e je effddZde fddZd S)QUESST14a*QUESST14* :cite:`Mir2015QUESST2014EQ` dataset. Args: root (str or Path): Root directory where the dataset's top level directory is found subset (str): Subset of the dataset to use. Options: [``"docs"``, ``"dev"``, ``"eval"``]. language (str or None, optional): Language to get dataset for. Options: [``None``, ``albanian``, ``basque``, ``czech``, ``nnenglish``, ``romanian``, ``slovak``]. If ``None``, dataset consists of all languages. (default: ``"nnenglish"``) download (bool, optional): Whether to download the dataset if it is not found at root path. (default: ``False``) rFrootsubsetlanguagedownloadreturnNcCs |dvrtd|dur|tvrtdttt|}tjt}tj||}| ddd}tj|||_ tj |j sZtj |sU|sNt dtt|tdt|||d krht|j |d |_dS|d krvt|j |d |_dS|d krt|j |d|_dSdS)N)docsdevevalz/`subset` must be one of ['docs', 'dev', 'eval']z"`language` must be None or one of .rz9Dataset not found. Please use `download=True` to download) hash_prefixrzlanguage_key_utterances.lstrzlanguage_key_dev.lstrzlanguage_key_eval.lst) ValueError _LANGUAGESstrosfspathpathbasenameURLjoinrsplit_pathisdirisfile RuntimeErrorr _CHECKSUMrfilter_audio_pathsdata)selfrrrrr#archiver0U/home/ubuntu/LTX-2/.venv/lib/python3.10/site-packages/torchaudio/datasets/quesst14.py__init__&s,    zQUESST14.__init__ncCs,|j|}tj||j}|t|djfS)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 to be loaded Returns: Tuple of the following items; str: Path to audio int: Sample rate str: File name )r-r r"relpathr' SAMPLE_RATE with_suffixname)r.r3 audio_pathr5r0r0r1 get_metadataJs zQUESST14.get_metadatacCs2||}t|j|d|d}|f|ddS)a:Load 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 str: File name rN)r:r r')r.r3metadatawaveformr0r0r1 __getitem___s zQUESST14.__getitem__cCs t|jS)N)lenr-)r.r0r0r1__len__ss zQUESST14.__len__)rF)__name__ __module__ __qualname____doc__rrrrboolr2intrr:torchTensorr>r@r0r0r0r1rs$  $rr"rlst_namecCsg}t|}t|d|-}|D]!}|\}}|dur$||kr$qtdd|}|||qWd|S1s>wY|S)z+Extract audio paths for the given language.scoringNz^.*?\/r4)ropenstripsplitresubappend)r"rrI audio_pathsfliner9langr0r0r1r,ws r,)r rNpathlibrtypingrrrrGtorch.utils.datartorchaudio._internalrtorchaudio.datasets.utilsrr r$r6r+rrrr,r0r0r0r1s(    ^