U %d@sddlZddlmZddlmZmZmZddlZddlm Z ddl m Z ddl m Z ddlmZdZd Zd Zd Zd d dZddZeeee eeeefdddZGddde ZdS)N)Path)OptionalTupleUnion)Tensor)download_url_to_file)Dataset)extract_archiveZSpeechCommandsspeech_commands_v0.02Z_nohash_Z_background_noise_Z@743935421bb51cccdb6bdd152e04c5c70274e935c82119ad7faeec31780d811dZ@af14739ee7dc311471de98f5f9d2c9191b18aedfe957f4a6ff791c709868ff58)zXhttps://storage.googleapis.com/download.tensorflow.org/data/speech_commands_v0.01.tar.gzzXhttps://storage.googleapis.com/download.tensorflow.org/data/speech_commands_v0.02.tar.gzc sJg}|D]<}tj|}t|}|fdd|D7}W5QRXq|S)Nc s&g|]}tjtj|qS)ospathnormpathjoinstrip).0linerootr F/tmp/pip-unpacked-wheel-lbdmvq91/torchaudio/datasets/speechcommands.py sz_load_list..)r r ropen)r filenamesoutputfilenamefilepathfileobjr rr _load_lists  "r)rr returnc Csptj||}tj|\}}tj|\}}tj|\}}|t\}}t|}t|\} } | | |||fSN) r r relpathsplitsplitext HASH_DIVIDERint torchaudioload) rr r labelrZspeaker_Z speaker_idZutterance_numberZwaveformZ sample_rater r rload_speechcommands_itemsr)c@sjeZdZdZeeddfeeefeee e eddddZ e e ee eee fddd Ze d d d ZdS) SPEECHCOMMANDSa?Create a Dataset for *Speech Commands* [:footcite:`speechcommandsv2`]. Args: root (str or Path): Path to the directory where the dataset is found or downloaded. url (str, optional): The URL to download the dataset from, or the type of the dataset to dowload. Allowed type values are ``"speech_commands_v0.01"`` and ``"speech_commands_v0.02"`` (default: ``"speech_commands_v0.02"``) folder_in_archive (str, optional): The top-level directory of the dataset. (default: ``"SpeechCommands"``) download (bool, optional): Whether to download the dataset if it is not found at root path. (default: ``False``). subset (str or None, optional): Select a subset of the dataset [None, "training", "validation", "testing"]. None means the whole dataset. "validation" and "testing" are defined in "validation_list.txt" and "testing_list.txt", respectively, and "training" is the rest. Details for the files "validation_list.txt" and "testing_list.txt" are explained in the README of the dataset and in the introduction of Section 7 of the original paper and its reference 12. The original paper can be found `here `_. (Default: ``None``) FN)rurlfolder_in_archivedownloadsubsetrc s|dks|dkstd|dkr:d}d}tj|||}t|}tj|}tj||} |ddd}tj||}tj|||_|rtj|jstj | st |d} t || | d t | |jn tj|jstd |jd |d kr t|jd |_n|dkr&t|jd|_n~|dkrvtt|jd dtddt|jdD} fdd| D|_n.tddt|jdD} dd| D|_dS)N)training validationtestingzXWhen `subset` not None, it must take a value from {'training', 'validation', 'testing'}.)zspeech_commands_v0.01r z}sz*SPEECHCOMMANDS.__init__..z*/*.wavcs0g|](}t|krt|krtj|kr|qSr )r# EXCEPT_FOLDERr r rrwZexcludesr rr~s z+SPEECHCOMMANDS.__init__..css|]}t|VqdSrr4r6r r rr8scSs g|]}t|krt|kr|qSr )r#r9r:r r rrs)AssertionErrorr r rfspathbasenamersplit_pathisdirisfile _CHECKSUMSgetrr exists RuntimeErrorr_walkersetsortedrglob) selfrr+r,r-r.base_urlZ ext_archiver?archiveZchecksumwalkerr r<r__init__IsF           zSPEECHCOMMANDS.__init__)nrcCs|j|}t||jS)aLoad the n-th sample from the dataset. Args: n (int): The index of the sample to be loaded Returns: (Tensor, int, str, str, int): ``(waveform, sample_rate, label, speaker_id, utterance_number)`` )rHr)rA)rLrQZfileidr r r __getitem__s zSPEECHCOMMANDS.__getitem__)rcCs t|jSr)lenrH)rLr r r__len__szSPEECHCOMMANDS.__len__)__name__ __module__ __qualname____doc__URLFOLDER_IN_ARCHIVErr5rboolrrPr$rrrRrTr r r rr*3s  > r*)r pathlibrtypingrrrr%ZtorchrZ torch.hubrZtorch.utils.datarZtorchaudio.datasets.utilsr rZrYr#r9rDrr5r$r)r*r r r rs"