
    h>                        S SK r S r\ R                  R                  S/ SQ5      S 5       r\ R                  R                  SSS/5      S	 5       r\ R                  R                  SS
S/5      S 5       r\ R                  R                  S/ SQ5      S 5       r\ R                  R                  S/ SQ5      S 5       r\ R                  R                  SSS/5      S 5       r	\ R                  R                  S/ SQ5      \ R                  R                  SSS/5      S 5       5       r
\ R                  R                  S/ SQ5      S 5       r\ R                  R                  SSS /5      S! 5       r\ R                  R                  S/ S"Q5      S# 5       rS$ r\ R                  R                  S/ S%Q5      S& 5       r\ R                  R                  S'S(S)S*/4S+S,S-/4S./ S/Q4/5      S0 5       r\ R                  R                  S1S2S3/5      S4 5       rg)5    Nc                     SnU " U5      n[        U5      S:X  d   eUS   R                  S:X  d   eSnU " U5      n[        U5      S:X  d   eUS   R                  S:X  d   eg )	Nzdon't giggle      zn'tzi said don't!      !lentexten_tokenizerr   tokenss      ]/home/james-whalen/.local/lib/python3.13/site-packages/spacy/tests/lang/en/test_exceptions.py+test_en_tokenizer_handles_basic_contractionr      sp    D$Fv;!!9>>U"""D$Fv;!!9>>S       r   )z`ain'tz"isn'tzcan't!c                 6    U " U5      n[        U5      S:X  d   eg )Nr   r
   r   s      r   1test_en_tokenizer_handles_basic_contraction_punctr          $Fv;!r   ztext_poss,text)zRobin'sRobin)zAlexis'sAlexisc                     U " U5      n[        U5      S:X  d   eUS   R                  U:X  d   eUS   R                  S:X  d   eg )N   r   r   z'sr	   )r   	text_possr   r   s       r   *test_en_tokenizer_handles_poss_contractionr      sL     )$Fv;!!9>>T!!!!9>>T!!!r   zschools'zAlexis'c                     U " U5      n[        U5      S:X  d   eUS   R                  UR                  S5      S   :X  d   eUS   R                  S:X  d   eg )Nr   r   'r   r
   r   splitr   s      r   &test_en_tokenizer_splits_trailing_aposr       sW    $Fv;!!9>>TZZ_Q////!9>>S   r   )z'emznothin'zol'c                 `    U " U5      n[        U5      S:X  d   eUS   R                  U:X  d   eg )Nr   r   r	   r   s      r   'test_en_tokenizer_doesnt_split_apos_excr"   '   s4    $Fv;!!9>>T!!!r   )zwe'llzYou'llzthere'llzthis'llzthose'llc                     U " U5      n[        U5      S:X  d   eUS   R                  UR                  S5      S   :X  d   eUS   R                  S:X  d   eg )Nr   r   r   r   'llr   r   s      r   (test_en_tokenizer_handles_ll_contractionr%   .   sW    $Fv;!!9>>TZZ_Q////!9>>U"""r   ztext_lower,text_title)zcan'tzCan't)zain'tzAin'tc                 (   U " U5      nU " U5      nUS   R                   US   R                   R                  5       :X  d   eUS   R                   US   R                   R                  5       :X  d   eUS   R                   US   R                   :X  d   eg Nr   r   )r   titlelower)r   
text_lower
text_titletokens_lowertokens_titles        r   (test_en_tokenizer_handles_capitalizationr.   6   s      
+L
+L?<?#7#7#=#=#?????<?#7#7#=#=#?????<?#7#7777r   pron)IYouHeSheItWeTheycontractionr$   z'dc                 l    U " X-   5      nUS   R                   U:X  d   eUS   R                   U:X  d   eg r'   )r   )r   r/   r7   r   s       r   "test_en_tokenizer_keeps_title_caser9   A   s=     $,-F!9>>T!!!!9>>[(((r   exc)IllillHellhellWellwellc                 6    U " U5      n[        U5      S:X  d   eg Nr   r   )r   r:   r   s      r   $test_en_tokenizer_excludes_ambiguousrC   I   s    #Fv;!r   zwo_punct,w_punct)zWe'vez`We've)zcouldn'tz	couldn't)c                 h    U " U5      n[        U5      S:X  d   eU " U5      n[        U5      S:X  d   eg )Nr   r   r   )r   wo_punctw_punctr   s       r   &test_en_tokenizer_splits_defined_punctrG   O   s<     (#Fv;!'"Fv;!r   )ze.g.zp.m.Jan.zDec.zInc.c                 6    U " U5      n[        U5      S:X  d   eg rB   r   r   s      r   test_en_tokenizer_handles_abbrrJ   Y   r   r   c                 d    SnU " U5      n[        U5      S:X  d   eUS   R                  S:X  d   eg )NzIt's mediocre i.e. bad.   r   zi.e.r	   r   s      r   %test_en_tokenizer_handles_exc_in_textrM   _   s9    $D$Fv;!!9>>V###r   )1amz12a.m.z11p.m.4pmc                 6    U " U5      n[        U5      S:X  d   eg )Nr   r   r   s      r   test_en_tokenizer_handles_timesrQ   f   r   r   z
text,normszI'miamzshan'tshallnotz.Many factors cause cancer 'cause it is complex)manyfactorscausecancerbecauseitiscomplexc                 b    U " U5      nU Vs/ s H  oDR                   PM     snU:X  d   eg s  snf )Nnorm_)r   r   normsr   tokens        r   !test_en_tokenizer_norm_exceptionsrc   l   s0     $F%+,VEKKV,555,s   ,z	text,norm)rH   January)z'cuzrZ   c                 >    U " U5      nUS   R                   U:X  d   eg )Nr   r_   )r   r   normr   s       r   !test_en_lex_attrs_norm_exceptionsrg   |   s"    $F!9??d"""r   )pytestr   markparametrizer   r   r    r"   r%   r.   r9   rC   rG   rJ   rM   rQ   rc   rg    r   r   <module>rl      sr   ! !CD E
 +-CD"" *i!89! :! !:;" <" !WX# Y# 02DE88 !NO6) 7 P)  NO P
 ,.GH !IJ K
$ !CD E
 	d	GU#$<T	

6
6
 ':<O&PQ# R#r   