rh"TdZddlZejdk\rddlZnddlZGddZy)z"English Normalizer class for CLVP.N) ceZdZdZdedefdZdedefdZdedefdZdedefd Z dedefd Z dedefd Z dedefd Z dedefd Z dedefdZdedefdZdZy)EnglishNormalizercdDcgc]1}tjd|dztj|df3c}|_gd|_gd|_gd|_ycc}w)N))mrsmisess)mrmister)drdoctor)stsaint)cocompany)jrjunior)majmajor)gengeneral)drsdoctors)revreverend)lt lieutenant)hon honorable)sgtsergeant)captcaptain)esqesquire)ltdlimited)colcolonel)ftfortz\b%s\.r) onetwothreefourfivesixseveneightnine) teneleventwelvethirteenfourteenfifteensixteen seventeeneighteennineteen) r-r-twentythirtyfortyfiftysixtyseventyeightyninety)recompile IGNORECASE_abbreviationsonesteenstens)selfxs }/var/www/html/ai-insurance-compliance-backend/venv/lib/python3.12/site-packages/transformers/models/clvp/number_normalizer.py__init__zEnglishNormalizer.__init__sc ZZ QqT)2== 91Q4 @ 0a    k K s6Anumreturnc|dk(ry|dkrd|jt|zS|dkr|j|S|dkr|j|dz S|dkr6|j|dz|dzdk7rd|j|dzzzSdzS|d kr9|j|dzd z|dzdk7rd |j|dzzzSdzS|d kr;|j|d zd z|d zdk7rd|j|d zzzSdzS|dkr;|j|d zdz|d zdk7rd|j|d zzzSdzS|dkr;|j|dzdz|dzdk7rd|j|dzzzSdzS|dkr;|j|dzdz|dzdk7rd|j|dzzzSdzS|dkr;|j|dzdz|dzdk7rd|j|dzzzSdzSy)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 thousand, iʚ;z millionlJ)z billionlI5z trillionlNZoz quadrillionznumber out of range)number_to_wordsabsrMrNrO)rPrTs rRr`z!EnglishNormalizer.number_to_wordsFs  !8 1Wd223s8<< < 2X99S> ! 2X::cBh' ' 3Y99SBY'SVY[S[_`S`31E1EcBh1O+Oi ifhi i 4Z #*% 2_beh_hlm_mcDADjAo4$..sTz::W TVW  = $$SI%56CF?VWCW4$..sY??a ^`a  $ $$$SM%9:GJ]GZ^_G_4$..s]/BCCi fhi  ( ($$S,=%=>KNQbKbfgKg4$..s5F/FGGq npq  , ,$$S,A%AB !22a74//6K0KLL    )textcD|jddjdS)z+ Converts unicode to ascii asciiignorezutf-8)encodedecoderPrcs rRconvert_to_asciiz"EnglishNormalizer.convert_to_asciis {{7H-44W==rbmcl|jd}|jd}t|dkDr|dzS|drt|dnd}t|dkDr|drt|dnd}|r!|r|dk(rdnd}|dk(rdnd }|d |d |d |S|r|dk(rdnd}|d |S|r|dk(rdnd }|d |Sy ) zZ This method is used to expand numerical dollar values into spoken words. r,.z dollarsrdollardollarscentcentsr^r_z zero dollars)groupsplitlenint)rPrkmatchpartsrprr dollar_unit cent_units rR_expand_dollarsz!EnglishNormalizer._expand_dollarss  C  u:>:% %#(8#eAh-!$UaE!HE!H ! u&-l( K"'1*'I%,k5)L L &-l( K%{3 3 "'1*'I#Y/ /!rbcD|jdjddS)zF This method is used to remove commas from sentences. r,,r-rsreplacerPrks rR_remove_commasz EnglishNormalizer._remove_commasswwqz!!#r**rbcD|jdjddS)zO This method is used to expand '.' into spoken word ' point '. r,rmz point r~rs rR_expand_decimal_pointz'EnglishNormalizer._expand_decimal_pointswwqz!!#y11rbcdddd}t|jddd}d|d zkr |d zd krd }n|j|dzd }|j||zS) z` This method is used to expand ordinals such as '1st', '2nd' into spoken words. rndrd)r,rnrrNrXrZrYth)rvrsgetr`)rPrTordinal_suffixessuffixs rR_expand_ordinalz!EnglishNormalizer._expand_ordinalsp $6#))A,s#$ s?sSyBF%))#(D9F##C(611rbct|jd}|dkDr\|dkrW|dk(ry|dkDr|dkrd|j|dzzS|dzdk(r|j|dzd zS|j|S|j|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 rZr])rvrsr`)rPrkrTs rR_expand_numberz EnglishNormalizer._expand_numbers !''!*o :#*d{%td &)=)=cCi)HHHsa++C3J7*DD++C00'', ,rbc~tjd|j|}tjdd|}tjd|j|}tjd|j|}tjd|j |}tjd|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]+)rIsubrr{rrrris rRnormalize_numbersz#EnglishNormalizer.normalize_numberss vv,d.A.A4Hvv)<>vv*D,@,@$Gvv*D,F,FMvv-t/C/CTJvvi!4!4d; rbc\|jD]\}}tj|||}|S)z/ Expands the abbreviate words. )rLrIr)rPrcregex replacements rRexpand_abbreviationsz&EnglishNormalizer.expand_abbreviationss5#'"5"5 4 E;66%d3D 4 rbcVtjtjdd|S)z. Removes multiple whitespaces z\s+r^)rIrrJris rRcollapse_whitespacez%EnglishNormalizer.collapse_whitespaces vvbjj(#t44rbc|j|}|j}|j|}|j|}|j |}|j dd}|S)z Converts text to ascii, numbers / number-like quantities to their spelt-out counterparts and expands abbreviations "r-)rjlowerrrrrris rR__call__zEnglishNormalizer.__call__sd $$T*zz|%%d+((.''-||C$ rbN)__name__ __module__ __qualname__rSrvstrr`rjr{rrrrrrrrrbrRrrs'kR9)39)39)v>S>S> """0+++ 2s2s2 23 23 2---( c c 555 rbr)__doc__sys version_inforIrrrrbrRrs/ ) w XXrb