
    shQ                         S SK r S SKrS SKrS SKJr  S SKJrJrJrJ	r	J
r
  S SKJr  SSKJr   " S S5      r " S	 S
5      r " S S5      rg)    N)Fraction)IteratorListMatchOptionalUnion)windowed   )remove_symbols_and_diacriticsc                   t   ^  \ rS rSrSrU 4S jrS\\   S\\   4S jr	S\4S jr
S\4S	 jrS\4S
 jrSrU =r$ )EnglishNumberNormalizer   aZ  
Convert any spelled-out numbers into arabic numbers, while handling:

- remove any commas
- keep the suffixes such as: `1960s`, `274th`, `32nd`, etc.
- spell out currency symbols after the number. e.g. `$20 million` -> `20000000 dollars`
- spell out `one` and `ones`
- interpret successive single-digit numbers as nominal: `one oh one` -> `101`
c                   > [         TU ]  5         1 SkU l        [        / SQSS9 VVs0 s H  u  pX!_M	     snnU l        U R                  R                  5        VVs0 s H  u  p#US:X  a  SOUS-   US4_M     snnU l        SS	S
SSSS.U R                  R                  5        VVs0 s H<  u  p#US:  d  M  US:w  d  M  US:w  d  M  X"R                  S5      (       a  SOS-   US4_M>     snnEU l        0 U R                  EU R                  EU l	        SSSSSSSSS.U l
        U R                  R                  5        VVs0 s H  u  p#UR                  SS5      US4_M     snnU l        U R                  R                  5        VVs0 s H  u  p#UR                  SS 5      US4_M     snnU l        0 U R                  EU R                  EU l        S!S"S#S$S%S&S'S(S)S*S+S,S-.U l        U R                  R                  5        VVs0 s H  u  p#US-   US4_M     snnU l        U R                  R                  5        VVs0 s H  u  p#US-   US4_M     snnU l        0 U R                   EU R"                  EU l        1 U R                  kU R                  kU R                  kU l        S.S.S/S/S0.U l        S1S1S2S2S3S3S4S4S5.U l        [-        [/        U R(                  R1                  5       5      [/        U R*                  R1                  5       5      -   5      U l        S6S70S7S8.U l        1 S9kU l        [-        U R                  U R                  U R                  U R                  U R                  U R                  U R$                  U R(                  U R*                  U R4                  U R6                  4 VVs/ s H  nU  H  nUPM     M     snn5      U l        S:S;1U l        g s  snnf s  snnf s  snnf s  snnf s  snnf s  snnf s  snnf s  snnf )<N>   oohzero)onetwothreefourfivesixseveneightnineteneleventwelvethirteenfourteenfifteensixteen	seventeeneighteennineteenr
   )startr   sixess)r   th)r
   st)   nd)   rd)   r)   )r   r)   )zerothfirstsecondthirdfifthtwelfthr-   r/   r   thr)         (   2   <   F   P   Z   )twentythirtyfortyfiftysixtyseventyeightyninetyyiesiethd     i@B i ʚ;l    J)l     I5 l     NZol     @=7M.cl      B3v^!< l      P ~cegl       73Me'l       (l
F3YHqS )hundredthousandmillionbilliontrillionquadrillionquintillion
sextillion
septillion	octillion	nonillion	decillion-+)minusnegativepluspositive   £u   €$   ¢)poundpoundseuroeurosdollardollarscentcentsrh   %)perpercent>   andpointdoubletripler   ones)super__init__zeros	enumeraterq   itemsones_pluralendswithones_ordinalones_suffixedtensreplacetens_pluraltens_ordinaltens_suffixedmultipliersmultipliers_pluralmultipliers_ordinalmultipliers_suffixeddecimalspreceding_prefixersfollowing_prefixerssetlistvaluesprefixes	suffixersspecialswordsliteral_words)selfinamevaluemappingkey	__class__s         o/Users/tiagomarins/Projetos/claudeai/copy_bank/venv/lib/python3.13/site-packages/whisper/normalizers/english.pyrs    EnglishNumberNormalizer.__init__   sh   (
 %* -
 G
	:  $yy0
0 u}G$*uclB0

  !
 $(99??#4#4KD19 L!&! L05 L}}S11t<udmK#4
 G 0 0FD4E4EF 	
	 GKiiooFW
FW{tDLLe$ucl2FW

  $yy0
0 LLf%t}40
 G 0 0FD4E4EF  $)047;>BF
 9=8H8H8N8N8P#
8PD3J$8P#
 ;?:J:J:P:P:R$
:R;4D4K%&:R$
 %
%%%
&&%
! >$))=dii=$**= 	$
  	$
  ))00234++22456

 C=
 = JJII&&II&&$$--,,,,NNMM  G #C  #  

& $V_Q
6
$

(#
$
HsA   OO&O#6O#>O##O#?!O)!O/ O5:O;*P
r   returnc              #   z  ^^#    S mS mSnS[         4S jnS[        [         [        4   4UU4S jjn[        U5      S:X  a  g [	        S /U-   S /-   S5       GH  u  pVnU(       a  SnM  US L=(       a    [
        R                  " SU5      nUS   U R                  ;   n	U	(       a  US	S  OUn
[
        R                  " SU
5      (       a  U" U
5      nUc   eTbN  [        T[         5      (       a/  TR                  S
5      (       a  [        T5      [        U5      -   mM  U" T5      v   U	(       a  US   OTmUR                  S	:X  a  UR                  mM  U
mM  X`R                  ;  a  Tb
  U" T5      v   U" U5      v   GM'  X`R                  ;   a  [        T=(       d    S5      S-   mGMP  X`R                  ;   a  U R                  U   nTc  UmGMv  [        T[         5      (       d  XPR                  ;   aN  XPR                  ;   a%  US:  a  TS   S:X  d   eTS S [        U5      -   mGM  [        T5      [        U5      -   mGM  US:  a+  TS-  S:X  a  TU-  mGM  [        T5      [        U5      -   mGM  TS-  S:X  a  TU-  mGM*  [        T5      [        U5      -   mGMD  X`R                   ;   GaF  U R                   U   u  pTc  U" [        U5      U-   5      v   GO[        T[         5      (       d  XPR                  ;   a`  XPR                  ;   a.  US:  a(  TS   S:X  d   eU" TS S [        U5      -   U-   5      v   OU" [        T5      [        U5      -   U-   5      v   OUS:  aF  TS-  S:X  a  U" [        TU-   5      U-   5      v   OhU" [        T5      [        U5      -   U-   5      v   OETS-  S:X  a  U" [        TU-   5      U-   5      v   O"U" [        T5      [        U5      -   U-   5      v   S mGM  X`R                  ;   aq  U R                  U   nTc  UmGM  [        T[         5      (       a  [        T5      [        U5      -   mGM  TS-  S:X  a  TU-  mGM   [        T5      [        U5      -   mGM  X`R"                  ;   a  U R"                  U   u  pTc  U" [        U5      U-   5      v   GMV  [        T[         5      (       a%  U" [        T5      [        U5      -   U-   5      v   GM  TS-  S:X  a  U" [        TU-   5      U-   5      v   GM  U" [        T5      [        U5      -   U-   5      v   GM  X`R$                  ;   a  U R$                  U   nTc  UmGM   [        T[         5      (       d  TS:X  aB  U" T5      nUb  X-  OS nUb  UR                  S	:X  a  UR                  mGMN  U" T5      v   UmGM]  TS-  S-  nTS-  nUUU-  -   mGMu  X`R&                  ;   a  U R&                  U   u  pTc  U" [        U5      U-   5      v   O[        T[         5      (       af  U" T5      nUb  X-  OS nUb1  UR                  S	:X  a!  U" [        UR                  5      U-   5      v   OLU" T5      v   U" [        U5      U-   5      v   O+TS-  S-  nTS-  nUUU-  -   mU" [        T5      U-   5      v   S mGMZ  X`R(                  ;   aB  Tb
  U" T5      v   XpR                  ;   d  U(       a  U R(                  U   mGM  U" U5      v   GM  X`R*                  ;   a,  Tb  U R*                  U   mU" T5      v   GM  U" U5      v   GM  X`R,                  ;   a  Tbv  U R,                  U   n[        U[.        5      (       a9  X};   a  U" [        T5      X   -   5      v   SnGM>  U" T5      v   U" U5      v   GMU  U" [        T5      U-   5      v   GMn  U" U5      v   GM{  X`R0                  ;   Ga;  XpR                  ;  a!  U(       d  Tb
  U" T5      v   U" U5      v   GM  US:X  a,  XPR$                  ;  a  Tb
  U" T5      v   U" U5      v   GM  GM  US:X  d  US:X  a  XpR                  ;   d  XpR                  ;   aN  US:X  a  SOSnU R                  R3                  US5      n[        T=(       d    S5      [        U5      U-  -   mSnGMe  Tb
  U" T5      v   U" U5      v   GM  US:X  a3  XpR4                  ;   d  U(       a  [        T=(       d    S5      S
-   mGM  GM  [7        SU 35      e[7        SU 35      e   Tb  U" T5      v   g g 7f)NFr(   c                 :     [        U 5      $ ! [         a     g f = fN)r   
ValueError)r(   s    r   to_fraction:EnglishNumberNormalizer.process_words.<locals>.to_fraction   s#    {" s   
 
resultc                 6   > [        U 5      n Tb  TU -   n S mS mU $ r   )str)r   prefixr   s    r   output5EnglishNumberNormalizer.process_words.<locals>.output   s*    [F!&EFM    r   r-   z^\d+(\.\d+)?$r
   . 0
   rK   rL   Trm   ro   rp   r+   rn   zUnexpected token: )r   r   intlenr	   rematchr   
isinstancerx   denominator	numeratorr   rt   rq   r{   rz   r   r   r   r   r   r   dictr   getr   r   )r   r   skipr   r   prevcurrentnextnext_is_numeric
has_prefixcurrent_without_prefixfrq   suffixr{   
multiplierpbeforeresidualrepeatsr   r   s                       @@r   process_words%EnglishNumberNormalizer.process_words   sh     $+/	3 		5c? 	 	 u:?#+TFUNdV,CQ#GD4"$.S288<Ld3SO t}}4J4>WQR[G"xx(*@AA 67}$}$!%--%..2E2E #E
S\ 9 $Um+'1v==A%KKE2E

*$ -'Wo%JJ&EKR(3.II%yy)= Es++tyy/@		)dRi$RyC/// %cr
SY 6 #E
SY 6BYrzQ #E
SY 6s{a' #E
SY 6...#11':= TV!344s++tyy/@yy(TBY$RyC///$U3BZ#d)%;f%DEE$SZ#d)%;f%DEEBYrzQ$S%6%?@@$SZ#d)%;f%DEEs{a'$S%6%?@@$SZ#d)%;f%DEEII%yy)= Es++JT2Es{a' #E
SY 6...#11':= TV!344s++ Uc$i!7&!@AAs{a'$S%6%?@@$SZ#d)%;f%DEE,,,!--g6
=&Es++uz#E*A*+-TA}!); !$Um+ *"d]T1F$t|H"X
%::E555%)%>%>w%G"
= Z6!9::s++#E*A*+-TA}!);$S%5%>??$Um+$S_v%=>>"d]T1F$t|H"X
%::E Uf!455444$ -'::%!55g>F /)444$!55g>F -' /)NN*$!^^G4F!&$//>"(Ufl)B"CC#'D"(-/"(/1$SZ&%899 /)MM)zz)/($Um+ /)%#3#33 ,"(-/$Wo- 4 (Gx,?yy(DJJ,>'.(':!#yy}}T15 #EKR 03t9w3F F# ,"(-/$Wo-'}}, #EKR 03 6 1@ %'9'%CDD !#5gY!?@@G $HJ - s   d7d;r(   c                 d   / n[         R                  " SU5      n[        U5       H  u  pE[        UR	                  5       5      S:X  a  M$  U[        U5      S-
  :X  a  UR                  U5        MI  UR                  U5        UR                  SS9S   nX`R                  ;   d  X`R                  ;   a  UR                  S5        M  UR                  S5        M     S	R                  U5      n[         R                  " S
SU5      n[         R                  " SSU5      n[         R                  " SSU5      nU$ )Nz\band\s+a\s+half\br   r
   r+   )maxsplitr   z
point fivez
and a half z([a-z])([0-9])z\1 \2z([0-9])([a-z])z([0-9])\s+(st|nd|rd|th|s)\b\1\2)r   splitru   r   stripappendrsplitr   r   joinsub)r   r(   resultssegmentsr   segment	last_words          r   
preprocess"EnglishNumberNormalizer.preprocess  s    88115#H-JA7==?#q(CMA%%w'w'#NNAN6r:	->N>N1NNN<0NN<0 . HHW FF$h2FF$h2 FF17A>r   c                     S[         4S jnS[         4S jn[        R                  " SX!5      n[        R                  " SX15      n[        R                  " SSU5      nU$ )Nmc                      U R                  S5      nU R                  S5      n[        U R                  S5      5      nU U SUS 3$ ! [         a    U R                  s $ f = f)Nr
   r+   r-   r   02d)groupr   r   string)r   currencyintegerri   s       r   combine_cents:EnglishNumberNormalizer.postprocess.<locals>.combine_cents  sa     771:''!*AGGAJ"G9AeC[99  xx s   AA A! A!c                 v     S[        U R                  S5      5       3$ ! [         a    U R                  s $ f = f)Nra   r
   )r   r   r   r   )r   s    r   extract_cents:EnglishNumberNormalizer.postprocess.<locals>.extract_cents  s9     C
O,--  xx s    88u,   ([€£$])([0-9]+) (?:and )?¢([0-9]{1,2})\bu   [€£$]0.([0-9]{1,2})\bz	\b1(s?)\bzone\1)r   r   r   )r   r(   r   r   s       r   postprocess#EnglishNumberNormalizer.postprocess  sW    	 U 	 	 U 	  FFBMUFF.A FF<1-r   c                     U R                  U5      nSR                  S U R                  UR                  5       5       5       5      nU R	                  U5      nU$ )Nr   c              3   .   #    U  H  oc  M  Uv   M     g 7fr    ).0words     r   	<genexpr>3EnglishNumberNormalizer.__call__.<locals>.<genexpr>  s     X&CdTT&Cs   	)r   r   r   r   r   r   r(   s     r   __call__ EnglishNumberNormalizer.__call__  sJ    OOAHHXd&8&8&CXXQr   )r   r   r   r   r   r   r   rq   ry   rw   rz   r   r   r   r   r{   r~   r}   r   r   rt   )__name__
__module____qualname____firstlineno____doc__rs   r   r   r   r   r   r   r   __static_attributes____classcell__)r   s   @r   r   r      sW    L-\] 49 ] # ] ~C :S 2#  r   r   c                   ,    \ rS rSrSrS rS\4S jrSrg)EnglishSpellingNormalizeri  zr
Applies British-American spelling mappings as listed in [1].

[1] https://www.tysto.com/uk-us-spelling-list.html
c                     [         R                  R                  [         R                  R                  [        5      S5      n[
        R                  " [        U5      5      U l        g )Nzenglish.json)	ospathr   dirname__file__jsonloadopenr   )r   mapping_paths     r   rs   "EnglishSpellingNormalizer.__init__  s8    ww||BGGOOH$=~Nyyl!34r   r(   c                 V   ^  SR                  U 4S jUR                  5        5       5      $ )Nr   c              3   Z   >#    U  H   nTR                   R                  X5      v   M"     g 7fr   )r   r   )r   r   r   s     r   r   5EnglishSpellingNormalizer.__call__.<locals>.<genexpr>  s#     K((44s   (+)r   r   r   s   ` r   r   "EnglishSpellingNormalizer.__call__  s    xxKKKKr   )r   N)	r   r   r   r   r   rs   r   r   r   r   r   r   r   r     s    5L# Lr   r   c                   (    \ rS rSrS rS\4S jrSrg)EnglishTextNormalizeri  c                 P   SU l         0 SS_SS_SS_SS	_S
S_SS_SS_SS_SS_SS_SS_SS_SS_SS_SS_SS _S!S"_0 S#S$_S%S&_S'S(_S)S*_S+S,_S-S._S/S0_S1S2_S3S4_S5S6_S7S8_S9S:_S;S<_S=S>_S?S@_SASB_SCSD_ESESFSGSHSISJSKSLSMSNSOSPSLSQSRSS.EU l        [        5       U l        [	        5       U l        g )TNz\b(hmm|mm|mhm|mmm|uh|um)\bz	\bwon't\bzwill notz	\bcan't\bzcan notz	\blet's\bzlet usz	\bain't\baintz	\by'all\bzyou allz	\bwanna\bzwant toz	\bgotta\bzgot toz	\bgonna\bzgoing toz\bi'ma\bzi am going toz\bimma\bz
\bwoulda\bz
would havez
\bcoulda\bz
could havez\bshoulda\bzshould havez	\bma'am\bmadamz\bmr\bzmister z\bmrs\bzmissus z\bst\bzsaint z\bdr\bzdoctor z\bprof\bz
professor z\bcapt\bzcaptain z\bgov\bz	governor z\bald\bz	alderman z\bgen\bzgeneral z\bsen\bzsenator z\brep\bzrepresentative z\bpres\bz
president z\brev\bz	reverend z\bhon\bz
honorable z\basst\bz
assistant z	\bassoc\bz
associate z\blt\bzlieutenant z\bcol\bzcolonel z\bjr\bzjunior z\bsr\bzsenior zesquire z	 had beenz	 has beenz	 had gonez	 has gonez	 had donez has gotz notz arez isz wouldz willz havez am)z\besq\bz	'd been\bz	's been\bz	'd gone\bz	's gone\bz	'd done\bz's got\bzn't\bz're\bz's\bz'd\bz'll\bz't\bz've\bz'm\b)ignore_patterns	replacersr   standardize_numbersr   standardize_spellings)r   s    r   rs   EnglishTextNormalizer.__init__  s   <6
*6
 )6
 (	6

 &6
 )6
 )6
 (6
 *6
 6
 6
 <6
 <6
 M6
 '6
" y#6
$ 	%6
& x'6
( y)6
* +6
, -6
. /6
0 16
2 
36
4 
56
6 )76
8 96
: ;6
< =6
> ?6
@ ,A6
B }C6
D 
E6
F yG6
H yI6
J #%%%%%#k6
n $;#< %>%@"r   r(   c                    UR                  5       n[        R                  " SSU5      n[        R                  " SSU5      n[        R                  " U R                  SU5      n[        R                  " SSU5      nU R                  R                  5        H  u  p#[        R                  " X#U5      nM     [        R                  " SSU5      n[        R                  " SS	U5      n[        US
S9nU R                  U5      nU R                  U5      n[        R                  " SS	U5      n[        R                  " SSU5      n[        R                  " SSU5      nU$ )Nz[<\[][^>\]]*[>\]]r   z\(([^)]+?)\)z\s+''z	(\d),(\d)r   z\.([^0-9]|$)z \1u
   .%$¢€£)keepu   [.$¢€£]([^0-9])z	([^0-9])%z\1 z\s+r   )	lowerr   r   r	  r
  rv   r   r  r  )r   r(   patternreplacements       r   r   EnglishTextNormalizer.__call__  s   GGIFF'Q/FF?B*FF4''Q/FF7C#$(NN$8$8$: GwQ/A %; FF<!,FF?FA.)!,?$$Q'&&q) FF)615FF<+FF63"r   )r	  r
  r  r  N)r   r   r   r   rs   r   r   r   r   r   r   r  r    s    :Ax# r   r  )r   r   r   	fractionsr   typingr   r   r   r   r   more_itertoolsr	   basicr   r   r   r  r   r   r   <module>r     sB     	 	  9 9 # 0s slL LU Ur   