o 8wie*@s,dZddlZddlmZeeZgdZ dZdZdZ dZ d d eDZ ege e e ee e e Z d d ee DZd d ee DZedZedZdd dDZddZddZddZddZddZddZddZd d!Zd"d#Zd$d%Zd&d'Zd(d)Z d*d+Z!d,d-Z"d.d/Z#dS)0z)from https://github.com/keithito/tacotronN) get_logger)TAAAA0AA1AA2AEAE0AE1AE2AHAH0AH1AH2AOAO0AO1AO2AWAW0AW1AW2AYAY0AY1AY2BCHDDHEHEH0EH1EH2ERER0ER1ER2EYEY0EY1EY2FGHHIHIH0IH1IH2IYIY0IY1IY2JHKLMNNGOWOW0OW1OW2OYOY0OY1OY2PRSSHTTHUHUH0UH1UH2UWUW0UW1UW2VWYZZH_z !'(),.:;? -4ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyzcCg|]}d|qS@.0sr]r]_/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/speechbrain/utils/text_to_sequence.py rbcCsi|]\}}||qSr]r]r_ir`r]r]ra rfcCsi|]\}}||qSr]r]rdr]r]rarfrgz(.*?)\{(.+?)\}(.*)z\s+cCs*g|]}td|dtj|dfqS)z\b%s\.r)recompile IGNORECASE)r_xr]r]rarbs))mrsmissus)mrmister)drdoctor)stsaint)cocompany)jrjunior)majmajor)gengeneral)drsdoctors)revreverend)lt lieutenant)hon honorable)sgtsergeant)captcaptain)esqesquire)ltdlimited)colcolonel)ftfortcCs tD] \}}t|||}q|S)z Expand abbreviations pre-defined)_abbreviationsrisub)textregex replacementr]r]raexpand_abbreviationss rcCs|S)zLowercase the text)lowerrr]r]ra lowercasesrcCsttd|S)z&Replaces whitespace by " " in the text )rir_whitespace_rerr]r]racollapse_whitespacercCs|dd}|S)zConverts text to asciiasciiignore)encodedecode)r text_encodedr]r]raconvert_to_asciis rcCst|}t|}|S)zPBasic pipeline that lowercases and collapses whitespace without transliteration.)rrrr]r]rabasic_cleanerssrcCs t|}|S)zLPipeline for German text, that collapses whitespace without transliteration.)rrr]r]ragerman_cleanerssrcCst|}t|}t|}|S)z;Pipeline for non-English text that transliterates to ASCII.)rrrrr]r]ratransliteration_cleanerssrcCs$t|}t|}t|}t|}|S)zGPipeline for English text, including number and abbreviation expansion.)rrrrrr]r]raenglish_cleanerss rcCsrg}t|r7t|}|s|tt||7} |S|tt|d|7}|t|d7}|d}t|s|S)aConverts a string of text to a sequence of IDs corresponding to the symbols in the text. The text can optionally have ARPAbet sequences enclosed in curly braces embedded in it. For example, "Turn left on {HH AW1 S S T AH0 N} Street." Arguments --------- text : str string to convert to a sequence cleaner_names : list names of the cleaner functions to run the text through Returns ------- sequence : list The integers corresponding to the symbols in the text. rh)len _curly_rematch_symbols_to_sequence _clean_textgroup_arpabet_to_sequence)r cleaner_namessequencemr]r]ratext_to_sequences   rcCsZd}|D]"}|tvr&t|}t|dkr"|ddkr"d|dd}||7}q|ddS) z+Converts a sequence of IDs back to a stringrhrr\z{%s}Nz}{r) _id_to_symbolrreplace)rresult symbol_idr`r]r]rasequence_to_texts rcCsV|D]&}|dkr t}|dkrt}|dkrt}|dkrt}|s$td|||}q|S)z.r])symbolsr]r]rar,rrcCstdd|DS)z Prepend "@" to ensure uniquenesscSrZr[r]r^r]r]rarb3rcz(_arpabet_to_sequence..)rsplitrr]r]rar1srcCs|tvo |dko |dkS)z Whether to keep a certain symbolrW~)r)r`r]r]rar6src Csltd|}|D] }|dd}|dd}|dd}|dd}|||}qtd|}z||}WntyItd|tYnwd|d }g}d } z|D]} | d vrn| || d| d 7} qY| | qYWn1t yt d ||D]} | d vr| | } dd| D} | | q| | qYnwd|vr|dd|vs|S)aPDo grapheme to phoneme and keep the punctuations between the words Arguments --------- g2p_model: speechbrain.inference.text.GraphemeToPhoneme Model to apply to the given text while keeping punctuation. text: string the input text. Returns ------- The text string's corresponding phoneme symbols with punctuation symbols. Example ------- >>> from speechbrain.inference.text import GraphemeToPhoneme >>> g2p_model = GraphemeToPhoneme.from_hparams("speechbrain/soundchoice-g2p") # doctest: +SKIP >>> from speechbrain.utils.text_to_sequence import _g2p_keep_punctuations # doctest: +SKIP >>> text = "Hi, how are you?" # doctest: +SKIP >>> _g2p_keep_punctuations(g2p_model, text) # doctest: +SKIP ['HH', 'AY', ',', ' ', 'HH', 'AW', ' ', 'AA', 'R', ' ', 'Y', 'UW', '?'] z\w+[-':\.][-':\.\w]*\w+rXr':.z[\w]+|[-!'(),.:;? ]zerror with text: rrz -!'(),.:;? rhzEDo g2p word by word because of unexpected outputs from g2p for text: cSsg|]}|dkr|qS)rr])r_rer]r]rarbysz*_g2p_keep_punctuations..)rifindallr RuntimeErrorloggerinfoquitjoinrextendappend IndexErrorwarningg2premove) g2p_modelr special_words special_wordrmpall_phonemes word_phonemesphonemes_with_punccountrepp_without_spacer]r]ra_g2p_keep_punctuations;sP                r)$__doc__rispeechbrain.utils.loggerr__name__r valid_symbols_pad _punctuation_special_letters_arpabetlistr enumeraterrrjrrrrrrrrrrrrrrrrrrr]r]r]rasF X