o
    i@/                     @   s  d dl Z d dlmZ d dlmZ ejddd ed dd Zd	d
 Ze j	dddd Z
e j	dddd Ze j	dddd Ze j	dddd Ze j	dddd Ze j	dddd Ze j	dddd Ze j	dddd Ze j	dddd Ze j	ddd d! Ze j	ddd"d# Ze j	ddd$d% Ze j	ddd&d' Ze j	ddd(d) Ze j	ddd*d+ Ze j	ddd,d- Ze j	ddd.d/ Ze j	ddd0d1 Ze j	ddd2d3 Ze j	ddd4d5 Ze j	ddd6d7 Ze j	ddd8d9 Ze j	ddd:d; Z e j	ddd<d= Z!e j	ddd>d? Z"e j	ddd@dA Z#e j	dddBdC Z$e j	dddDdE Z%e j	dddFdG Z&e j	dddHdI Z'e j	dddJdK Z(e j	dddLdM Z)e j	dddNdO Z*e j	dddPdQ Z+e j	dRdS Z,e j	dddTdU Z-e j	dddVdW Z.e j	dddXdY Z/e j	dddZd[ Z0e j	ddd\d] Z1e j	ddd^d_ Z2e j	ddd`da Z3e j	dddbdc Z4e j	ddddde Z5e j	dddfdg Z6e j	dddhdi Z7e j	dddjdk Z8e j	dddldm Z9e j	dddndo Z:e j	dddpdq Z;e j	dddrds Z<e j	dddtdu Z=e j	dddvdw Z>e j	dddxdy Z?e j	dddzd{ Z@e j	ddd|d} ZAe j	ddd~d ZBe j	dddd ZCe j	dddd ZDe j	dddd ZEe j	dddd ZFe j	dddd ZGe j	dddd ZHe j	dddd ZIe j	dddd ZJe j	dddd ZKe j	dddd ZLe j	dddd ZMe j	dddd ZNe j	dddd ZOe j	dddd ZPe j	dddd ZQe j	dddd ZRe j	dddd ZSe j	dddd ZTe j	dddd ZUe j	dddd ZVe j	dddd ZWe j	dddd ZXe j	dddd ZYe j	dddd ZZe j	dddd Z[e j	dddd Z\e j	dddd Z]e j	dddd Z^e j	dddd Z_e j	dddd Z`dS )    N)settings)get_lang_classno_deadlinesi )deadlinec                 C   s<   z| j dddd | j dddd W d S  ty   Y d S w )Nz--slow
store_truezinclude slow tests)actionhelpz--issuestoreztest specific issues)	addoption
ValueError)parser r   H/home/ubuntu/.local/lib/python3.10/site-packages/spacy/tests/conftest.pypytest_addoption   s   r   c                    s   fdd}dD ]}|j v r||std| d q|d}t|tr^dj v rWdd |d	D  d
d jddD }t fdd|D sUtd   d S d S td d S d S )Nc                    s    j d|  dS )Nz--F)config	getoption)opt)itemr   r   getopt   s   z$pytest_runtest_setup.<locals>.getopt)slowzneed --z option to runissuec                 S   s   g | ]}t | qS r   )intstrip).0r   r   r   r   
<listcomp>*   s    z(pytest_runtest_setup.<locals>.<listcomp>,c                 S   s   g | ]}|j d  qS )r   )args)r   markr   r   r   r   ,   s    )namec                    s   g | ]}| v qS r   r   )r   ref)	issue_nosr   r   r   -   s    z"not referencing specified issues: znot referencing any issues)keywordspytestskip
isinstancestrsplititer_markersany)r   r   r   issues
issue_refsr   )r    r   r   pytest_runtest_setup   s   


r+   module)scopec                   C      t d jS Nxxr   	tokenizerr   r   r   r   r2   6      r2   sessionc                   C   r.   )Nafr1   r   r   r   r   af_tokenizer;   r3   r6   c                   C   r.   )Namr1   r   r   r   r   am_tokenizer@   r3   r8   c                   C   r.   )Narr1   r   r   r   r   ar_tokenizerE   r3   r:   c                   C   r.   )Nbgr1   r   r   r   r   bg_tokenizerJ   r3   r<   c                   C   r.   )Nbnr1   r   r   r   r   bn_tokenizerO   r3   r>   c                   C   r.   )Nbor1   r   r   r   r   bo_tokenizerT   r3   r@   c                   C   r.   )Ncar1   r   r   r   r   ca_tokenizerY   r3   rB   c                   C   r.   )Ncsr1   r   r   r   r   cs_tokenizer^   r3   rD   c                   C   r.   )Ndar1   r   r   r   r   da_tokenizerc   r3   rF   c                   C   r.   Nder1   r   r   r   r   de_tokenizerh   r3   rI   c                   C   r.   rG   r   vocabr   r   r   r   de_vocabm   r3   rL   c                   C   r.   )Ndsbr1   r   r   r   r   dsb_tokenizerr   r3   rN   c                   C   r.   )Nelr1   r   r   r   r   el_tokenizerw   r3   rP   c                   C   r.   Nenr1   r   r   r   r   en_tokenizer|   r3   rS   c                   C   r.   rQ   rJ   r   r   r   r   en_vocab   r3   rT   c                 C   s   t d| }|dS )NrR   r   )r   create_pipe)rT   nlpr   r   r   	en_parser   s   
rW   c                   C   r.   Nesr1   r   r   r   r   es_tokenizer   r3   rZ   c                   C   r.   rX   rJ   r   r   r   r   es_vocab   r3   r[   c                   C   r.   )Netr1   r   r   r   r   et_tokenizer   r3   r]   c                   C   r.   )Neur1   r   r   r   r   eu_tokenizer   r3   r_   c                   C   r.   )Nfar1   r   r   r   r   fa_tokenizer   r3   ra   c                   C   r.   )Nfir1   r   r   r   r   fi_tokenizer   r3   rc   c                   C   r.   )Nfor1   r   r   r   r   fo_tokenizer   r3   re   c                   C   r.   Nfrr1   r   r   r   r   fr_tokenizer   r3   rh   c                   C   r.   rf   rJ   r   r   r   r   fr_vocab   r3   ri   c                   C   r.   )Ngar1   r   r   r   r   ga_tokenizer   r3   rk   c                   C   r.   )Ngrcr1   r   r   r   r   grc_tokenizer   r3   rm   c                   C   r.   )Ngur1   r   r   r   r   gu_tokenizer   r3   ro   c                   C   r.   )Nher1   r   r   r   r   he_tokenizer   r3   rq   c                   C   r.   )Nhir1   r   r   r   r   hi_tokenizer   r3   rs   c                   C   r.   )Nhrr1   r   r   r   r   hr_tokenizer   r3   ru   c                   C   r.   Nhtr1   r   r   r   r   ht_tokenizer   r3   rx   c                   C   r.   rv   rJ   r   r   r   r   ht_vocab   r3   ry   c                   C   r.   )Nhur1   r   r   r   r   hu_tokenizer   r3   r{   c                   C   r.   )Nidr1   r   r   r   r   id_tokenizer   r3   r}   c                   C   r.   )Nisr1   r   r   r   r   is_tokenizer   r3   r   c                   C   r.   Nitr1   r   r   r   r   it_tokenizer   r3   r   c                   C   r.   r   rJ   r   r   r   r   it_vocab   r3   r   c                   C      t d td jS )N	sudachipyjar"   importorskipr   r2   r   r   r   r   ja_tokenizer      
r   c                   C   r.   )Nhsbr1   r   r   r   r   hsb_tokenizer   r3   r   c                   C   r   )Nnattokor   r   r   r   r   ko_tokenizer  r   r   c                  C   s$   ddddiii} t d| }|jS )NrV   r2   @tokenizerszspacy.Tokenizer.v1r   )r   from_configr2   r   rV   r   r   r   ko_tokenizer_tokenizer  s   r   c                   C   r.   )Nlar1   r   r   r   r   la_tokenizer  r3   r   c                   C   r.   )Nlbr1   r   r   r   r   lb_tokenizer  r3   r   c                   C   r.   )Nlgr1   r   r   r   r   lg_tokenizer"  r3   r   c                   C   r.   )Nltr1   r   r   r   r   lt_tokenizer'  r3   r   c                   C   r.   )Nlvr1   r   r   r   r   lv_tokenizer,  r3   r   c                   C   r.   )Nmkr1   r   r   r   r   mk_tokenizer1  r3   r   c                   C   r.   )Nmlr1   r   r   r   r   ml_tokenizer6  r3   r   c                   C   r.   )Nmsr1   r   r   r   r   ms_tokenizer;  r3   r   c                   C   r.   )Nnbr1   r   r   r   r   nb_tokenizer@  r3   r   c                   C   r.   )Nner1   r   r   r   r   ne_tokenizerE  r3   r   c                   C   r.   NnlrJ   r   r   r   r   nl_vocabJ  r3   r   c                   C   r.   r   r1   r   r   r   r   nl_tokenizerO  r3   r   c                   C   r.   )Nnnr1   r   r   r   r   nn_tokenizerT  r3   r   c                   C   r.   )Nplr1   r   r   r   r   pl_tokenizerY  r3   r   c                   C   r.   Nptr1   r   r   r   r   pt_tokenizer^  r3   r   c                   C   r.   r   rJ   r   r   r   r   pt_vocabc  r3   r   c                   C   r.   )Nror1   r   r   r   r   ro_tokenizerh  r3   r   c                   C   r   )N	pymorphy3rur   r   r   r   r   ru_tokenizerm  r   r   c                   C   s   t d td dS )Nr   r   
lemmatizerr"   r   r   add_piper   r   r   r   ru_lemmatizers  s   
r   c                   C   s"   t d td jdddidS )Nr   r   r   modepymorphy3_lookupr   r   r   r   r   r   ru_lookup_lemmatizery  s   

r   c                   C   r.   )Nsar1   r   r   r   r   sa_tokenizer  r3   r   c                   C   r.   )Nskr1   r   r   r   r   sk_tokenizer  r3   r   c                   C   r.   )Nslr1   r   r   r   r   sl_tokenizer  r3   r   c                   C   r.   )Nsrr1   r   r   r   r   sr_tokenizer  r3   r   c                   C   r.   )Nsqr1   r   r   r   r   sq_tokenizer  r3   r   c                   C   r.   )Nsvr1   r   r   r   r   sv_tokenizer  r3   r   c                   C   r.   )Ntar1   r   r   r   r   ta_tokenizer  r3   r   c                   C   r   )N	pythainlpthr   r   r   r   r   th_tokenizer  r   r   c                   C   r.   )Ntir1   r   r   r   r   ti_tokenizer  r3   r   c                   C   r.   )Ntlr1   r   r   r   r   tl_tokenizer  r3   r   c                   C   r.   )Ntrr1   r   r   r   r   tr_tokenizer  r3   r   c                   C   r.   )Nttr1   r   r   r   r   tt_tokenizer  r3   r   c                   C   r.   )Nkyr1   r   r   r   r   ky_tokenizer  r3   r   c                   C   r   )Nr   ukr   r   r   r   r   uk_tokenizer  r   r   c                   C   s$   t d t d td dS )Nr   pymorphy3_dicts_ukr   r   r   r   r   r   r   uk_lemmatizer  s   

r   c                   C   s,   t d t d td jdddidS )Nr   r   r   r   r   r   r   r   r   r   r   r   uk_lookup_lemmatizer  s
   


r   c                   C   r.   )Nurr1   r   r   r   r   ur_tokenizer  r3   r   c                   C   r   )Npyvivir   r   r   r   r   vi_tokenizer  r   r   c                   C   r.   r/   r1   r   r   r   r   xx_tokenizer  r3   r   c                   C   r.   )Nyor1   r   r   r   r   yo_tokenizer  r3   r   c                  C   s   t d } | jS )Nzhr1   )rV   r   r   r   zh_tokenizer_char  s   
r   c                  C   s0   t d dddddii} td| }|jS )NjiebarV   r2   spacy.zh.ChineseTokenizerr   	segmenterr   )r"   r   r   r   r2   r   r   r   r   zh_tokenizer_jieba  s   
r   c                  C   sB   t d ddddidddiid} td	| }|  |jS )
Nspacy_pkusegr2   r   pkusegr   pkuseg_modelweb)rV   
initializer   )r"   r   r   r   r   r2   r   r   r   r   zh_tokenizer_pkuseg  s   

	r   c                   C   r.   )Nhyr1   r   r   r   r   hy_tokenizer  r3   r   )ar"   
hypothesisr   
spacy.utilr   register_profileload_profiler   r+   fixturer2   r6   r8   r:   r<   r>   r@   rB   rD   rF   rI   rL   rN   rP   rS   rT   rW   rZ   r[   r]   r_   ra   rc   re   rh   ri   rk   rm   ro   rq   rs   ru   rx   ry   r{   r}   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   <module>   sj   


 










































































































































































