o Zh"@sdZddlZGdddZdS)z"English Normalizer class for CLVP.Nc@seZdZddZdedefddZdedefdd Zd edefd d Zd edefd dZ d edefddZ dedefddZ d edefddZ dedefddZ dedefddZdedefddZddZdS)EnglishNormalizercCs2dddD|_gd|_gd|_gd|_dS)NcSs*g|]}td|dtj|dfqS)z\b%s\.r)recompile IGNORECASE).0xr Y/var/www/auris/lib/python3.10/site-packages/transformers/models/clvp/number_normalizer.py sz.EnglishNormalizer.__init__..))ZmrsZmisess)mrZmister)ZdrZdoctor)stZsaint)coZcompany)ZjrZjunior)Zmajmajor)genZgeneral)ZdrsZdoctors)revZreverend)ltZ lieutenant)ZhonZ honorable)ZsgtZsergeant)ZcaptZcaptain)ZesqZesquire)ZltdZlimited)colZcolonel)ftZfort) onetwothreefourfivesixseveneightnine) tenZelevenZtwelveZthirteenZfourteenZfifteenZsixteenZ seventeenZeighteenZnineteen) rrZtwentyZthirtyZfortyZfiftyZsixtyZseventyZeightyZninety)_abbreviationsonesteenstens)selfr r r __init__s    zEnglishNormalizer.__init__numreturncCs|dkrdS|dkrd|t|S|dkr|j|S|dkr'|j|dS|dkrD|j|d|ddkrAd||dSdS|d krc|j|dd |ddkr`d ||dSdS|d kr||d d |d dkrd||d SdS|dkr||d d|d dkrd||d SdS|dkr||dd|ddkrd||dSdS|dkr||dd|ddkrd||dSdS|dkr||dd|ddkrd||dSdSdS)ax Converts numbers(`int`) to words(`str`). Please note that it only supports upto - "'nine hundred ninety-nine quadrillion, nine hundred ninety-nine trillion, nine hundred ninety-nine billion, nine hundred ninety-nine million, nine hundred ninety-nine thousand, nine hundred ninety-nine'" or `number_to_words(999_999_999_999_999_999)`. rzerozminus d-r hundred i@Bz thousandz, iʚ;z millionlJ)z billionlI5z trillionlNZoz quadrillionznumber out of range)number_to_wordsabsr!r"r#)r$r&r r r r0?s 2.       z!EnglishNormalizer.number_to_wordstextcCs|dddS)z+ Converts unicode to ascii asciiignorezutf-8)encodedecoder$r2r r r convert_to_asciizz"EnglishNormalizer.convert_to_asciimcCs|d}|d}t|dkr|dS|drt|dnd}t|dkr0|dr0t|dnd}|rN|rN|dkrr@Z dollar_unitZ cent_unitr r r _expand_dollarss"   $  z!EnglishNormalizer._expand_dollarscC|dddS)zF This method is used to remove commas from sentences. r,rrAreplacer$r:r r r _remove_commasr9z EnglishNormalizer._remove_commascCrH)zO This method is used to expand '.' into spoken word ' point '. rr;z point rJrLr r r _expand_decimal_pointr9z'EnglishNormalizer._expand_decimal_pointcCs^dddd}t|ddd}d|d kr |d d kr d }n||dd }|||S) z` This method is used to expand ordinals such as '1st', '2nd' into spoken words. r ndrd)rr<rNr)r+r*th)rDrAgetr0)r$r&Zordinal_suffixessuffixr r r _expand_ordinals z!EnglishNormalizer._expand_ordinalcCs~t|d}|dkr:|dkr:|dkrdS|dkr&|dkr&d||dS|ddkr5||dd S||S||S) a  This method acts as a preprocessing step for numbers between 1000 and 3000 (same as the original repository, link : https://github.com/neonbjb/tortoise-tts/blob/4003544b6ff4b68c09856e04d3eff9da26d023c2/tortoise/utils/tokenizer.py#L86) rr-i iz two thousandiz two thousand r+r.)rDrAr0)r$r:r&r r r _expand_numbers   z EnglishNormalizer._expand_numbercCsttd|j|}ttdd|}ttd|j|}ttd|j|}ttd|j|}ttd|j|}|S)z This method is used to normalize numbers within a text such as converting the numbers to words, removing commas, etc. z([0-9][0-9\,]+[0-9])u£([0-9\,]*[0-9]+)z \1 poundsz\$([0-9\.\,]*[0-9]+)z([0-9]+\.[0-9]+)z[0-9]+(st|nd|rd|th)z[0-9]+)rsubrrMrGrNrVrWr7r r r normalize_numberssz#EnglishNormalizer.normalize_numberscCs"|jD] \}}t|||}q|S)z/ Expands the abbreviate words. )r rrX)r$r2regex replacementr r r expand_abbreviationssz&EnglishNormalizer.expand_abbreviationscCsttdd|S)z. Removes multiple whitespaces z\s+r/)rrXrr7r r r collapse_whitespacesz%EnglishNormalizer.collapse_whitespacecCs@||}|}||}||}||}|dd}|S)z Converts text to ascii, numbers / number-like quantities to their spelt-out counterparts and expands abbreviations "r)r8lowerrYr\r]rKr7r r r __call__s     zEnglishNormalizer.__call__N)__name__ __module__ __qualname__r%rDstrr0r8rGrMrNrVrWrYr\r]r`r r r r rs);   r)__doc__rrr r r r s