o j@sddlZddlmZddlmZmZmZdefddZddedee fd d Z defd d Z ddedeefddZ defddZ ddedefddZdS)N)Optional)download_kaggle_dataset download_urlextract_archivepathcCJtj|ddd}t||tj|}tj||}tdt|dS)zDownload and extract LJSpeech dataset Args: path (str): path to the directory where the dataset will be stored. Texist_okz:https://data.keithito.com/data/speech/LJSpeech-1.1.tar.bz2 > Extracting archive file...NosmakedirsrrbasenamejoinprintrrurlrarchiverH/home/kuhnn/.local/lib/python3.10/site-packages/TTS/utils/downloaders.pydownload_ljspeech   rF use_kagglecCs^|r tdd|dStj|ddd}t||tj|}tj||}tdt|dS)zDownload and extract VCTK dataset. Args: path (str): path to the directory where the dataset will be stored. use_kaggle (bool, optional): Downloads vctk dataset from kaggle. Is generally faster. Defaults to False. z5mfekadu/english-multispeaker-corpus-for-voice-cloningVCTKTrzKhttps://datashare.ed.ac.uk/bitstream/handle/10283/3443/VCTK-Corpus-0.92.zipr N) rr r rrrrrr)rrrrrrrr download_vctks   rcCstdd|dS)z}Download and extract Tweb dataset Args: path (str): Path to the directory where the dataset will be stored. z0bryanpark/the-world-english-bible-speech-datasetTWEBN)r)rrrr download_tweb*srallsubsetcCsdddddddd}tj|d d |d krF|D]&\}}td |d t||tj|}tj||}tdt|qtddS||}t||tj|}tj||}tdt|dS)aDownload and extract libri tts dataset. Args: path (str): Path to the directory where the dataset will be stored. subset (str, optional): Name of the subset to download. If you only want to download a certain portion specify it here. Defaults to 'all'. z:http://www.openslr.org/resources/60/train-clean-100.tar.gzz:http://www.openslr.org/resources/60/train-clean-360.tar.gzz:http://www.openslr.org/resources/60/train-other-500.tar.gzz4http://www.openslr.org/resources/60/dev-clean.tar.gzz4http://www.openslr.org/resources/60/dev-other.tar.gzz5http://www.openslr.org/resources/60/test-clean.tar.gzz5http://www.openslr.org/resources/60/test-other.tar.gz)zlibri-tts-clean-100zlibri-tts-clean-360zlibri-tts-other-500zlibri-tts-dev-cleanzlibri-tts-dev-otherzlibri-tts-test-cleanzlibri-tts-test-otherTrrz > Downloading z...r z > All subsets downloadedN) r r itemsrrrrrr)rr subset_dictsubvalrrrrrrdownload_libri_tts3s0         r#cCr)zDownload and extract Thorsten german male voice dataset. Args: path (str): Path to the directory where the dataset will be stored. Trz8https://www.openslr.org/resources/95/thorsten-de_v02.tgzr Nr rrrrdownload_thorsten_deZrr$englishlanguagecCs^dddddd}tj|dd||}t||tj|}tj||}td t|d S) zDownload and extract Mailabs dataset. Args: path (str): Path to the directory where the dataset will be stored. language (str): Language subset to download. Defaults to english. z5https://data.solak.de/data/Training/stt_tts/en_US.tgzz5https://data.solak.de/data/Training/stt_tts/de_DE.tgzz5https://data.solak.de/data/Training/stt_tts/fr_FR.tgzz5https://data.solak.de/data/Training/stt_tts/it_IT.tgzz5https://data.solak.de/data/Training/stt_tts/es_ES.tgz)r%germanfrenchitalianspanishTrr Nr )rr& language_dictrrrrrrdownload_mailabsis    r,)F)r)r%)r typingrTTS.utils.downloadrrrstrrboolrrr#r$r,rrrrs  '