o
    Mi                     @   s   d dl mZ d dlmZ G dd deZG dd deZG dd deZG d	d
 d
eZG dd deZ	G dd deZ
G dd deZG dd deZG dd deZG dd deZG dd deZG dd deZdS )   )BaseStemmer)Amongc                   @   sP  e Zd ZdZh dZh dZh dZh dZh dZdZ	dZ
dZdZd	Zd
d Zdd Zdd Zdd Zdd Zdd Zdd Zdd Zdd Zdd Zdd Zd d! Zd"d# Zd$d% Zd&d' Zd(d) Zd*d+ Zed,d-d.ed/d-d0ed1d-d.ed2d-d.ed3d-d.ed4d-d.ed5d-d.ed6d-d.ed7d-d0ed8d-d0ed9d-d0ed:d-d;ed<d-d=ed>d-d.ed?d-d.ed@d-d.edAd-d.edBd-d.edCd-d.edDd-d.edEd-d.gZ edFd-dGedHd-dIedJd-d0edKd0d.edLd0d=edMd=d;edNd0dOedPd0dQgZ!edRd-dQedSd-d0edTd-d=edUd-d.edVd-d;edWd-dXedYd-dZed[d-dGed\d-dIed]d-dOed^d-d_gZ"ed`d-d;edad-dIedbd-dGedcd-d.eddd-dQeded-dQedfd-dOedgd-dQedhd-d;edid-d=edjd-d;edkd-d0edld-dZedmd-dXgZ#ednd-dXedod-d0edpd-dXedqd-d;edrd-d.edsd-d;edtd-dQedud-d=edvd-dOedwd-dXedxd-d;edyd-dGedzd-dIed{d-dXed|d-d;ed}d-dGgZ$ed~d-d.edd-d.edd-d.gZ%edd-d0edd-d.edd-d;gZ&edd-d.edd-d0edd-d;edd-d=edd-dQedd-dOedd-dIedd-dGedd-dXedd-dZedd-d_edd-dedd-dedd-dedd-dedd-dedd-d=edddedd-dedd-dedd-dedddgZ'edd-d.edd-d0gZ(ed	d-d-eddd.eddd.eddd.edd;d-eddd.gZ)edd-d.edd-d0gZ*edd-d.edd-d0gZ+dS )DutchStemmerz
    This class implements the stemming algorithm defined by a snowball script.
    Generated from dutch.sbl by Snowball 3.0.1 - https://snowballstem.org/
    >   e   è   é   ê   ë>   aiou   à   á   â   ä   ì   í   î   ï   ò   ó   ô   ö   ù   ú   û   ü>   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   yr   r   r   r   r   r   r   r	   r   r   r   r   r   r   r   r   r   r   r   r   >   r
   r   r   r   r   wxr   r   r   r   r   r   r   r   r	   r   r   r   r   r   r   r   r   r   r   r   r   F     c                 C      | j | jkS N)I_p1cursorself r)   Q/home/ubuntu/.local/lib/python3.10/site-packages/snowballstemmer/dutch_stemmer.py__r_R1      zDutchStemmer.__r_R1c                 C   r#   r$   )I_p2r&   r'   r)   r)   r*   __r_R2!   r,   zDutchStemmer.__r_R2c                 C   s   | j | j }z,| j | j }z| tjst t  ty"   Y nw | j | | _| ds1W dS W n	 ty;   Y nw | j | | _dS NijFT)limitr&   in_grouping_br   g_vlab1lab0eq_s_br(   v_1v_2r)   r)   r*   __r_V$   s   
zDutchStemmer.__r_Vc                 C   s   | j | j }| j| jkrdS |  jd8  _z,| j | j }z| tjs&t t  ty1   Y nw | j | | _| ds@W dS W n	 tyJ   Y nw | j | | _dS )NFr   r0   T)	r1   r&   limit_backwardr2   r   r3   r4   r5   r6   r7   r)   r)   r*   __r_VX4   s$   
zDutchStemmer.__r_VXc                 C   sn   | j | j }| j | j }z| dst W dS  ty    Y nw | j | | _| tjs/dS | j | | _dS r/   )r1   r&   r6   r5   out_grouping_br   r3   r7   r)   r)   r*   __r_CG   s   
zDutchStemmer.__r_Cc                 C   sx  | j | j }z| tjst | j| _| tj}|dkr"t | j| _	|dkr| j | j }z-| j | j }z| tj
sAt t  tyL   Y nw | j | | _| j| jkr\t W n	 tyf   Y nw | j | | _|  | _| jdkrzW dS | j}| | j| j| j || _n|dkr| j | j }z-| j | j }z| tj
st t  ty   Y nw | j | | _| j| jkrt W n	 ty   Y nw | j | j }zGz:| j | j }	z| tjst t  ty   Y nw | j |	 | _| tjst | j| jkrt W t  ty   Y t w  ty%   Y nw | j | | _| j | j }
z(| j| jkr=t |  jd8  _| tjsNt | tj
sXt t  tyd   Y nw | j |
 | _| j | | _|  | _| jdkrW dS | j}| | j| j| j || _n|dkr| dsW dS n	| dsW dS W n
 ty   Y nw | j | | _d	S )
Nr!   r   r"   F      u   eëeieeT)r1   r&   r=   r   g_v_WXr5   ketfind_among_ba_0brag_AEIOUlab2r4   r;   slice_toS_chinsertlab4lab3r2   g_AIOUlab7lab6g_Elab5lab8
slice_from)r(   r8   	among_varr9   v_3cv_4v_5v_6v_7v_8r)   r)   r*   __r_lengthen_VU   s   





zDutchStemmer.__r_lengthen_Vc                 C   s  | j | _| tj}|dkrdS | j | _|dkr |  sdS dS |dkr`|  s*dS | j| j  }z| 	ds9t
 |  s@t
 W dS  t
yK   Y nw | j| | _ |  sXdS |  s^dS dS |dkrs|  sjdS | dsqdS dS |dkr$z| j| j  }z1| j| j  }| 	d	st |  st |  st | j| | _ |  sW W dS |   t  ty   Y nw | j| | _ z-| j| j  }| 	d
st |  st |  st | j| | _ |  sW W dS t  ty   Y nw | j| | _ |  sW dS |  sW dS | dsW dS W dS  ty#   Y dS w |dkr:|  s0dS | ds8dS dS |dkrW|  sFdS |  sMdS | dsUdS dS |dkrz| j| j  }z"| 	dsnt |  svt | j | _| dsW W dS t  ty   Y nw | j| | _ z| 	dst |  sW W dS t  ty   Y nw | j| | _ z)| 	dst |  st |  st | j | _|  sW W dS t  ty   Y nw | j| | _ zMz.| j| j  }z| 	ds	t t  ty   Y nw | j| | _ | 	ds%t W n
 ty0   Y nw |  s9t |  sBW W dS t  tyN   Y nw | j| | _ |  s]W dS |  seW dS |  smW dS |   W dS  ty~   Y dS w | dsdS dS )Nr!   Fr   r?   tr@   ie   arerr      r      au   hedheidnddr   jT)r&   rC   rD   r   a_1rF   	slice_del_DutchStemmer__r_R1r1   r6   r5   _DutchStemmer__r_CrT   rH   _DutchStemmer__r_lengthen_Vr4   rM   _DutchStemmer__r_VrR   rL   rP   rO   lab10lab9rS   )r(   rU   r8   r9   rV   rX   rY   rZ   r)   r)   r*   
__r_Step_1   sB    
 
{




TR

M


F











zDutchStemmer.__r_Step_1c                 C   s  | j | _| tj}|dkrdS | j | _|dkrzc| j| j  }z| ds*t | j | _| 	 s6W W dS t
  tyA   Y nw | j| | _ z%| dsQt | j | _|  s\t |  sct | 	 skW W dS t
  tyv   Y nw | j| | _ z| dst | j | _| dsW W dS t
  ty   Y nw | j| | _ z%| dst | j | _|  st |  st | 	 sW W dS t
  ty   Y nw | j| | _ z| d	st | j | _| d
sW W dS t
  ty   Y nw | j| | _ z| dst | j | _| dsW W dS t
  ty&   Y nw | j| | _ z!| ds7t | j | _|  sCt | 	 sLW W dS t
  tyX   Y nw | j| | _ | j | _|  skW dS |  ssW dS | 	 s{W dS W dS  t
y   Y dS w |dkr|  sdS | dsdS dS |dkr|  sdS | dsdS dS |dkr|  sdS | dsdS dS |dkr|  sdS |  sdS | 	 sdS dS |dkr|  sdS | dsdS dS |dkr|  s	dS | dsdS dS |dkr)|  sdS | ds'dS dS |dkrK|  s5dS | 	 s<dS | | j | j d |   dS |dkrt|  sWdS |  s^dS | 	 sedS | | j | j d |   dS |  s{dS |  sdS | dsdS dS ) Nr!   Fr   z'tetrntrnr^   inkingmpm'r?   gr@   lijkr`   ischrc   rd   rf   s   r	   l
   eniefT)r&   rC   rD   r   a_2rF   r1   r6   r4   rm   r5   rH   rn   ro   rM   rT   rL   _DutchStemmer__r_VXrR   rP   rO   rK   rp   r(   rU   r8   r)   r)   r*   
__r_Step_2K  sD  












C@

;

6

1



*

$












zDutchStemmer.__r_Step_2c                 C   sn  | j | _| tj}|dkrdS | j | _|dkr'|  sdS | ds%dS dS |dkr=|  s1dS |  s7dS | 	  dS |dkrO|  sGdS |  sMdS dS |dkr\| dsZdS dS |d	krz?| j
| j  }z| d
spt | dsyW W dS t  ty   Y nw | j
| | _ |  sW dS |  sW dS | 	  W dS  ty   Y dS w |dkr|  sdS |  sdS | dsdS dS |dkr|  sdS |  sdS | | j | j d | 	  dS |dkr|  sdS |  sdS | | j | j d | 	  dS |dkr|  sdS |  sdS | dsdS dS |  s&dS |  s-dS | ds5dS dS )Nr!   Fr   eerr?   r@   r`   r   rc   ildrb   rd   aarrf   fr   r}   r   r^   rj   T)r&   rC   rD   r   a_3rF   rn   rT   rm   rp   r1   r6   r4   r5   ro   _DutchStemmer__r_R2rK   r   r)   r)   r*   
__r_Step_3  s   
HA<
8


'&






zDutchStemmer.__r_Step_3c                 C   s  zw| j | j }z| j| _| tj}|dkrt | j| _|dkr5|  s)t | 	ds2W W dS t |dkrL|  s@t | 	dsIW W dS t |dkrb|  sWt | 
 s_W W dS t |dkr|  smt |  stt | 	d	s}W W dS t |d
kr|  st |  st | 	dsW W dS t |dkr|  st |  st | 	dsW W dS t |dkr|  st | 	dsW W dS t |dkr|  st | 	dsW W dS t |  st |  st | 
 sW W dS |   t  ty   Y nw | j | | _| j| _| tjdkr(W dS | j| _|  s4W dS | j | j }z| dsDt | j| jkrNt W W dS  ty[   Y nw | j | | _|  sjW dS | 
 srW dS |   W dS  ty   Y dS w )Nr!   r   r_   Fr?   r   r@   r`   nrc   r   rd   r   rf   teerr   r~   innT)r1   r&   rC   rD   r   a_4r4   rF   rn   rT   rm   rq   ro   rp   r5   a_5r6   rH   r;   )r(   r8   rU   r9   r)   r)   r*   
__r_Step_4?  s   
4
/*
"









zDutchStemmer.__r_Step_4c                 C   sn   | j | _| tj}|dkrdS | j | _|dkr!| dsdS dS |dkr.| ds,dS dS | ds5dS dS )	Nr!   Fr   kr?   r   pT)r&   rC   rD   r   a_6rF   rT   )r(   rU   r)   r)   r*   
__r_Step_7  s    


zDutchStemmer.__r_Step_7c                 C   s  | j | _| tj}|dkrdS | j | _|dkr!| dsdS d+S |dkr.| ds,dS d+S |dkr;| ds9dS d+S |d	krH| d
sFdS d+S |dkrU| dsSdS d+S |dkrb| ds`dS d+S |dkro| dsmdS d+S |dkr|| dszdS d+S |dkr| dsdS d+S |dkr| dsdS d+S |dkr| j| j  }z| dst	 | j | j
krt	 W dS  t	y   Y nw | j| | _ | dsdS d+S |dkr| dsdS d+S |dkr| dsdS d+S |dkr| dsdS d+S |d kr| d!sdS d+S |d"kr| d#sdS d+S |d$kr!| d%sdS d+S |d&kr0| d's.dS d+S |d(kr?| d)s=dS d+S | d*sGdS d+S ),Nr!   Fr   br?   rW   r@   rj   r`   r   rc   r}   rd   hrf   rk   r   r   r   r   r   r{      r   r      r      q   r      r      r^      v   r      r    zT)r&   rC   rD   r   a_7rF   rT   r1   r6   r5   r;   r   r)   r)   r*   
__r_Step_6  s   
C
@
=
:
7
4
1
.
+
(










zDutchStemmer.__r_Step_6c                 C   s  | j | _| tj}|dkrdS | j | _|  sdS |  s dS |dkr| j| j  }z| 	ds3t
 |  s:t
 W dS  t
yE   Y nw | j| | _ z=| j| j  }z| 	ds\t | j | jkret | dsnW W dS t  tyy   Y nw | j| | _ |  sW dS W dS  ty   Y dS w | j| j  }z| 	dst |  st W dS  ty   Y nw | j| | _ | j| j  }z| 	dst | j | jkrt W dS  ty   Y nw | j| | _ |  sdS dS )	Nr!   Fr   r   inr   r   T)r&   rC   rD   r   a_8rF   rn   ro   r1   r6   r5   rH   r;   rT   r4   rm   rM   rL   )r(   rU   r8   r9   rV   rX   r)   r)   r*   __r_Step_1c  sv   




zDutchStemmer.__r_Step_1cc           
      C   sx  | j | _| dsdS | j | _| j }| j d }|| jkrdS || _ || _ | j }zV	 | j }z4z(| j }z| ds:t t  tyE   Y nw || _ | tj	sRt
 W t  ty`   Y t w  t
yi   Y nw || _ | j | jkrvW dS |  j d7  _ q) ty   Y nw 	 | j }z2z&| j }z| dst t  ty   Y nw || _ | tj	st W n	 ty   Y nw W q ty   Y nw || _ 	 z| j | jk rt W dS  ty   Y nw || _ | tj}|dkrdS d| _|  sdS | j }	z0| j | _| tj}|dkrt | j | _|dkr"| ds!W dS n	| d	s+W dS W n
 ty6   Y nw |	| _ dS )
NgeFr@   Tr0   r   r!   r   r   )r&   rF   eq_srC   r1   rM   rH   in_groupingr   r3   r4   r5   rP   rR   rL   rO   
find_amonga_9B_GE_removedrm   a_10rS   rT   )
r(   r8   rW   r9   rV   rX   rY   rZ   rU   r\   r)   r)   r*   __r_Lose_prefix5  s   






zDutchStemmer.__r_Lose_prefixc           
      C   s  | j | jkrdS |  j d7  _ z/	 z| j | _| dst | j | _t  ty-   Y nw | j | jkr7W dS |  j d7  _ q tyG   Y nw | j }| j d }|| jkrWdS || _ || _ | j }zV	 | j }z4z(| j }z| dsst t  ty~   Y nw || _ | 	t
jst W t  ty   Y t w  ty   Y nw || _ | j | jkrW dS |  j d7  _ qb ty   Y nw 	 | j }z2z&| j }z| dst t  ty   Y nw || _ | 	t
jst W n	 ty   Y nw W q ty   Y nw || _ 	 z| j | jk rt W dS  ty   Y nw || _ d| _|  s)dS | j }z0| j | _| t
j}	|	dkr?t | j | _|	dkrR| dsQW dS n	| d	s[W dS W n
 tyf   Y nw || _ dS )
NFr   Tr   r@   r0   r!   r   r   )r&   r1   rF   r   r4   rC   r5   rR   rL   r   r   r3   rM   rH   rS   rO   rP   rs   r   rm   r   a_11rr   rT   )
r(   r9   rW   rV   rX   rY   rZ   r[   v_9rU   r)   r)   r*   __r_Lose_infix  s   







zDutchStemmer.__r_Lose_infixc                 C   s  | j | _| j | _| j}z	 z| tjst W q ty!   Y nw 	 d}	 | j}z6z&| j}z| ds7t	 t
  t	yB   Y nw || _| tjsOt W n	 t
yY   Y nw |d8 }W q% tyh   Y nw || _	 |dkrtt | tjs}t | j| _	 z| tjst W q ty   Y nw 	 d}	 | j}z6z&| j}z| dst t  ty   Y nw || _| tjst W n	 ty   Y nw |d8 }W q ty   Y nw || _	 |dkrt | tjst | j| _W n
 ty   Y nw || _dS )NTr   r0   r!   )r1   r%   r-   r&   out_groupingr   r3   r4   r   rL   rM   r   rH   r5   rR   rS   rO   rP   )r(   r8   rV   rX   rY   r[   r\   r   r)   r)   r*   __r_measure  s   

zDutchStemmer.__r_measurec                 C   s\  d| _ |   | j| _| j| _| j| j }z|  st d| _ W n	 ty*   Y nw | j| | _| j| j }z|  s?t d| _ W n	 tyL   Y nw | j| | _| j| j }z| 	 sat
 d| _ W n	 t
yn   Y nw | j| | _| j| j }z|  st d| _ W n	 ty   Y nw | j| | _| j| _d| _| j}z| j}|  st || _|   W n	 ty   Y nw || _| j| _| j| _| j| j }z| jst d| _ |  st W n	 ty   Y nw | j| | _| j| _d| _| j}z| j}	|  st |	| _|   W n
 ty   Y nw || _| j| _| j| _| j| j }
z| js3t d| _ |  s>t W n
 tyI   Y nw | j|
 | _| j| _| j| _| j| _| j| j }z|  skt d| _ W n
 tyy   Y nw | j| | _| j| j }z| j st |  st W n
 ty   Y nw | j| | _| j| _dS )NFT)	B_stemmed_DutchStemmer__r_measurer&   r;   r1   _DutchStemmer__r_Step_1r5   _DutchStemmer__r_Step_2r4   _DutchStemmer__r_Step_3rH   _DutchStemmer__r_Step_4rM   r   _DutchStemmer__r_Lose_prefixrL   rR   _DutchStemmer__r_Step_1c_DutchStemmer__r_Lose_infixrP   rO   _DutchStemmer__r_Step_7rS   rs   _DutchStemmer__r_Step_6)r(   r8   r9   rV   rX   rY   rZ   r[   r\   r   v_10v_11v_12r)   r)   r*   _stem+  s   








zDutchStemmer._stemr
   r   r   r?   r   r   r   r   r   r   r   r   r   u   eër@   u   iër`   r   r   r   r   r   r   r   r   nder   r   rf   r   z'sesiesausrd   u   ésrc   der   ischejelijkeler   ener   reseteiever   rh   fiegieatieismery   arijerijselrdersteriteitdsttstendatieferigachtigioneelbaarlaarnaarraarerigerachtigerlijkertanterigstachtigstlijkstigigerigstftktptbbccddffgghhjjkkllmmnnppr   qqr   rrr   ssr   ttr   r   vvr   wwr   xxr   r   zz   rj   r^   eftvaavalvalivarer	   r   N),__name__
__module____qualname____doc__rQ   rN   rG   r3   rB   r   r   r-   r%   rJ   rn   r   rq   r   ro   rp   r   r   r   r   r   r   r   r   r   r   r   r   rE   rl   r   r   r   r   r   r   r   r   r   r   r)   r)   r)   r*   r      s2   Z  !T\M<R[I
c













































































































r   c                   @      e Zd ZdS )r5   Nr  r  r  r)   r)   r)   r*   r5          r5   c                   @   r  )r4   Nr  r)   r)   r)   r*   r4   #  r  r4   c                   @   r  )rH   Nr  r)   r)   r)   r*   rH   &  r  rH   c                   @   r  )rM   Nr  r)   r)   r)   r*   rM   )  r  rM   c                   @   r  )rL   Nr  r)   r)   r)   r*   rL   ,  r  rL   c                   @   r  )rR   Nr  r)   r)   r)   r*   rR   /  r  rR   c                   @   r  )rP   Nr  r)   r)   r)   r*   rP   2  r  rP   c                   @   r  )rO   Nr  r)   r)   r)   r*   rO   5  r  rO   c                   @   r  )rS   Nr  r)   r)   r)   r*   rS   8  r  rS   c                   @   r  )rs   Nr  r)   r)   r)   r*   rs   ;  r  rs   c                   @   r  )rr   Nr  r)   r)   r)   r*   rr   >  r  rr   N)basestemmerr   amongr   r   BaseExceptionr5   r4   rH   rM   rL   rR   rP   rO   rS   rs   rr   r)   r)   r)   r*   <module>   s0             "