157 lines
16 KiB
Plaintext
157 lines
16 KiB
Plaintext
|
Ë
|
|||
|
|
=û gB@ ã óN — d dl Z d dlmZmZmZ ddlmZ ddlmZ G d„ de« Z y)é N)ÚListÚOptionalÚTupleé )Ú
|
|||
|
|
Lemmatizer)ÚTokenc ó ‡ — e Zd ZdZededeee ee f fˆ fd„ « Zde dee fd„Z
|
|||
|
|
dedee de e fd „Z d
|
|||
|
|
edee d ed ee dee f
|
|||
|
|
d
|
|||
|
|
„Z
|
|||
|
|
d
|
|||
|
|
edee d ed ee dee f
|
|||
|
|
d„Zd
|
|||
|
|
edee d ed ee dee f
|
|||
|
|
d„Zd
|
|||
|
|
edee d ed ee dee f
|
|||
|
|
d„Zd
|
|||
|
|
edee d ed ee dee f
|
|||
|
|
d„Zd
|
|||
|
|
edee d e e d ee dee f
|
|||
|
|
d„Zd
|
|||
|
|
edee d e e d ee dee f
|
|||
|
|
d„Zd
|
|||
|
|
edee d e e d ee dee f
|
|||
|
|
d„Zˆ xZS )ÚSpanishLemmatizerzH
|
|||
|
|
Spanish rule-based lemmatizer with morph-based rule selection.
|
|||
|
|
ÚmodeÚreturnc ó< •— |dk( rg d¢}|g fS t ‰| � |« S )NÚrule)Ú lemma_rulesÚlemma_rules_groupsÚ lemma_indexÚ lemma_exc)ÚsuperÚget_lookups_config)Úclsr ÚrequiredÚ __class__s €úYC:\Users\garci\AppData\Roaming\Python\Python312\site-packages\spacy/lang/es/lemmatizer.pyr z$SpanishLemmatizer.get_lookups_config
|
|||
|
|
s) ø€ à �6Š>ÚXˆHØ ˜b�>Ð !ä‘7Ñ-¨dÓ3Ð 3ó Útokenc ód — |j |j t |j « f}|| j v r| j | S |j
|
|||
|
|
}|j j « }t |j « }|dv r|j « gS |dv r%|j r|dk7 r|j « gS |gS |j « }| j j d« j |i « j |« }|� t |« }n†|dk( rd}n|}| j |t |« « } | j j d« j |g « }
|
|||
|
|
t | d|z « ||| |
|
|||
|
|
« }t t j# |« « }|| j |< |S ) N)Ú ÚeolÚspace) ÚadpÚcconjÚintjÚpartÚpropnÚpunctÚsconjÚsymÚxr# r ÚauxÚverbr Ú
|
|||
|
|
lemmatize_)ÚorthÚposÚstrÚmorphÚcacheÚtextÚpos_ÚlowerÚsetÚ
|
|||
|
|
is_sent_startÚlookupsÚ get_tableÚgetÚlistÚ select_ruleÚgetattrÚdictÚfromkeys) Úselfr Ú cache_keyÚstringr, ÚfeaturesÚexcÚlemmasÚrule_posr Úindexs r Úrule_lemmatizez SpanishLemmatizer.rule_lemmatize s… € Ø—Z‘Z §¡¬C°· ± Ó,<Ð=ˆ Ø ˜Ÿ
|
|||
|
|
™
|
|||
|
|
Ñ "Ø—:‘:˜iÑ(Ð (Ø—‘ˆØ�j‰j× Ñ Ó ˆÜ�u—{‘{Ó#ˆØ Ð&Ñ &Ø—L‘L“NÐ#Ð #Ø ð
|
|||
|
|
|
|||
|
|
ñ
|
|||
|
|
|
|||
|
|
ð ×"Ò" s¨g¢~Ø Ÿ ™ ›Ð'Ð'à �x�à— ‘ “ ˆØ�l‰l×$Ñ$ [Ó1×5Ñ5°c¸2Ó>×BÑBÀ6ÓJˆØ ˆ?ܘ#“Y‰Fà�eŠ|Ø!‘à �Ø×#Ñ# H¬d°8«nÓ=ˆDØ—L‘L×*Ñ*¨=Ó9×=Ñ=¸hÈÓKˆEØ;”W˜T <°(Ñ#:Ó;ؘ $¨óˆFô œ$Ÿ-™-¨Ó/Ó0ˆFà &ˆ�
|
|||
|
|
‰
|
|||
|
|
�9Ñ Øˆ
|
|||
|
|
r r, r@ c óž — | j j d« }||v r.|| D ]& }t |d « j |« sŒ!|d c S y )Nr é r )r5 r6 r3 Úissubset)r= r, r@ ÚgroupsÚgroups r r9 z SpanishLemmatizer.select_ruleB sP € Ø— ‘ ×'Ñ'Ð(<Ó=ˆØ �&‰=Ø œ�Ü�u˜Q‘x“=×)Ñ)¨(Õ3Ø ™8’Oð %ð r Úwordr rD c ó† — g }g }| j j d« j |g « D ]6 \ }}t j |dz ||« } | |k7 sŒ&|j | « Œ8 g }
|
|||
|
|
d|v rƒ|D ]~ } | j
|
|||
|
|
d« s| j
|
|||
|
|
d« sŒ&| j j d« j dg « D ]+ \ }}|
|
|||
|
|
j t j ||| « « Œ- Œ€ |j |
|
|||
|
|
« |D ] } | |v sŒ|j | « Œ t |« dkD r|S t |« dkD r|S |gS )a"
|
|||
|
|
Lemmatize an adjective.
|
|||
|
|
|
|||
|
|
word (str): The word to lemmatize.
|
|||
|
|
features (List[str]): The morphological features as a list of Feat=Val
|
|||
|
|
pairs.
|
|||
|
|
index (List[str]): The POS-specific lookup list.
|
|||
|
|
|
|||
|
|
RETURNS (List[str]): The list of lemmas.
|
|||
|
|
r ú$ú Number=PlurÚnÚsÚaccentsr © r5 r6 r7 ÚreÚsubÚappendÚendswithÚextendÚlen© r= rK r@ r rD Úpossible_lemmasÚselected_lemmasÚoldÚnewÚpossible_lemmaÚadditional_lemmasÚlemmas r Ú
|
|||
|
|
lemmatize_adjz SpanishLemmatizer.lemmatize_adjJ óE € ð ˆØ ˆð Ÿ ™ ×.Ñ.¨}Ó=×AÑAÀ$ÈÖK‰HˆC�Ü ŸV™V C¨#¡I¨s°DÓ 9ˆNØ Ó%Ø ×&Ñ& ~Õ6ð Lð ÐØ ˜HÑ $Û"1�Ø!×*Ñ*¨3Ô/°>×3JÑ3JÈ3Õ3OØ$(§L¡L×$:Ñ$:¸=Ó$I×$MÑ$MØ! 2ö%™˜˜Sð *×0Ñ0´·±¸¸SÀ.Ó1QÕRñ%ð #2ð × Ñ Ð 0Ô1ã$ˆEؘŠ~Ø ×&Ñ& uÕ-ð %ô
|
|||
|
|
ˆÓ !Ò #Ø"Ð "Ü
|
|||
|
|
� Ó
|
|||
|
|
! AÒ
|
|||
|
|
%Ø"Ð "à�6ˆMr c ó‚ — | j j d« j dg « D ] \ }}||k( sŒ |gc S |gS )a
|
|||
|
|
Lemmatize an adverb.
|
|||
|
|
|
|||
|
|
word (str): The word to lemmatize.
|
|||
|
|
features (List[str]): The morphological features as a list of Feat=Val
|
|||
|
|
pairs.
|
|||
|
|
index (List[str]): The POS-specific lookup list.
|
|||
|
|
|
|||
|
|
RETURNS (List[str]): The list of lemmas.
|
|||
|
|
r Úadverbs)r5 r6 r7 )r= rK r@ r rD r\ r] s r Ú
|
|||
|
|
lemmatize_advz SpanishLemmatizer.lemmatize_advz sF € ð Ÿ ™ ×.Ñ.¨}Ó=×AÑAÀ)ÈRÖP‰HˆC�Ø�s‹{Ø�u’ ð Qð
|
|||
|
|
ˆvˆ
|
|||
|
|
r c óz — g }g }| j j d« j dg « D ] \ }}||k( sŒ |gc S | j j d« j dg « D ] \ }}||k( sŒ |gc S | j j d« j dg « D ]0 \ }}t j |dz ||« } |j | « Œ2 |j |« t
|
|||
|
|
|« dk( r|S t
|
|||
|
|
|« dkD r/|D ] }
|
|||
|
|
|
|
|||
|
|
|v sŒ|j |
|
|||
|
|
« Œ t
|
|||
|
|
|« dk\ r|S |S g S )a"
|
|||
|
|
Lemmatize a determiner.
|
|||
|
|
|
|||
|
|
word (str): The word to lemmatize.
|
|||
|
|
features (List[str]): The morphological features as a list of Feat=Val
|
|||
|
|
pairs.
|
|||
|
|
index (List[str]): The POS-specific lookup list.
|
|||
|
|
|
|||
|
|
RETURNS (List[str]): The list of lemmas.
|
|||
|
|
r ÚdetÚdet_and_pron_fixedÚdet_and_pron_generalrM rG ©r5 r6 r7 rS rT rU rX © r= rK r@ r rD rZ r[ r\ r] r^ r` s r Ú
|
|||
|
|
lemmatize_detz SpanishLemmatizer.lemmatize_det� sL € ð ˆØ ˆð Ÿ ™ ×.Ñ.¨}Ó=×AÑAÀ%ÈÖL‰HˆC�Ø�s‹{Ø�u’ ð Mð Ÿ ™ ×.Ñ.¨}Ó=×AÑAØ "ö
|
|||
|
|
‰HˆC�ð �s‹{Ø�u’ ð
|
|||
|
|
ð Ÿ ™ ×.Ñ.¨}Ó=×AÑAØ " Bö
|
|||
|
|
‰HˆC�ô ŸV™V C¨#¡I¨s°DÓ 9ˆNØ × "Ñ " >Õ 2ð
|
|||
|
|
ð
|
|||
|
|
× Ñ ˜tÔ$ä ˆÓ 1Ò $Ø"Ð "Ü
|
|||
|
|
� Ó
|
|||
|
|
! AÒ
|
|||
|
|
%Û(�ؘE’>Ø#×*Ñ*¨5Õ1ð )ô �?Ó# qÒ(Ø&Ð&à&Ð&àˆIr c ó† — g }g }| j j d« j |g « D ]6 \ }}t j |dz ||« } | |k7 sŒ&|j | « Œ8 g }
|
|||
|
|
d|v rƒ|D ]~ } | j
|
|||
|
|
d« s| j
|
|||
|
|
d« sŒ&| j j d« j dg « D ]+ \ }}|
|
|||
|
|
j t j ||| « « Œ- Œ€ |j |
|
|||
|
|
« |D ] } | |v sŒ|j | « Œ t |« dkD r|S t |« dkD r|S |gS )a
|
|||
|
|
Lemmatize a noun.
|
|||
|
|
|
|||
|
|
word (str): The word to lemmatize.
|
|||
|
|
features (List[str]): The morphological features as a list of Feat=Val
|
|||
|
|
pairs.
|
|||
|
|
index (List[str]): The POS-specific lookup list.
|
|||
|
|
|
|||
|
|
RETURNS (List[str]): The list of lemmas.
|
|||
|
|
r rM rN rO rP rQ r rR rY s r Úlemmatize_nounz SpanishLemmatizer.lemmatize_nounÅ rb r c ó2 — | j j d« j dg « D ] \ }}||k( sŒ |gc S |j d« }t j
|
|||
|
|
d|d « rt j dd|« }t j dd|« }|gS ) a
|
|||
|
|
Lemmatize a numeral.
|
|||
|
|
|
|||
|
|
word (str): The word to lemmatize.
|
|||
|
|
features (List[str]): The morphological features as a list of Feat=Val
|
|||
|
|
pairs.
|
|||
|
|
index (List[str]): The POS-specific lookup list.
|
|||
|
|
|
|||
|
|
RETURNS (List[str]): The list of lemmas.
|
|||
|
|
r Únumú,z(\.)([0-9]{3})$r z\.r ú.)r5 r6 r7 ÚsplitrS ÚsearchrT )r= rK r@ r rD r\ r] Ú
|
|||
|
|
splitted_words r Ú
|
|||
|
|
lemmatize_numz SpanishLemmatizer.lemmatize_numõ s‹ € ð Ÿ ™ ×.Ñ.¨}Ó=×AÑAÀ%ÈÖL‰HˆC�Ø�s‹{Ø�u’ ð Mð
|
|||
|
|
Ÿ
|
|||
|
|
™
|
|||
|
|
3›ˆ
|
|||
|
|
Ü
|
|||
|
|
�9‰9Ð'¨ °qÑ)9Ô :Ü—6‘6˜% dÓ+ˆDÜ�v‰v�d˜D $Ó'ˆØˆvˆ
|
|||
|
|
r c ó† — g }g }| j j d« j dg « D ] \ }}||k( sŒ |gc S | j j d« j dg « D ] \ }}||k( sŒ |gc S | j j d« j dg « D ]6 \ }}t j |dz ||« } | |k7 sŒ&|j | « Œ8 |j |« t
|
|||
|
|
|« dk( r|S t
|
|||
|
|
|« dkD r/|D ] }
|
|||
|
|
|
|
|||
|
|
|v sŒ|j |
|
|||
|
|
« Œ t
|
|||
|
|
|« dk\ r|S |S g S )a
|
|||
|
|
Lemmatize a pronoun.
|
|||
|
|
|
|||
|
|
word (str): The word to lemmatize.
|
|||
|
|
features (List[str]): The morphological features as a list of Feat=Val
|
|||
|
|
pairs.
|
|||
|
|
index (List[str]): The POS-specific lookup list.
|
|||
|
|
|
|||
|
|
RETURNS (List[str]): The list of lemmas.
|
|||
|
|
r Úpronrh ri rM rG rj rk s r Úlemmatize_pronz SpanishLemmatizer.lemmatize_pron sT € ð ˆØ ˆð Ÿ ™ ×.Ñ.¨}Ó=×AÑAÀ&È"ÖM‰HˆC�Ø�s‹{Ø�u’ ð Nð Ÿ ™ ×.Ñ.¨}Ó=×AÑAØ "ö
|
|||
|
|
‰HˆC�ð �s‹{Ø�u’ ð
|
|||
|
|
ð Ÿ ™ ×.Ñ.¨}Ó=×AÑAØ " Bö
|
|||
|
|
‰HˆC�ô ŸV™V C¨#¡I¨s°DÓ 9ˆNØ Ó%Ø ×&Ñ& ~Õ6ð
|
|||
|
|
ð × Ñ ˜tÔ$ä ˆÓ 1Ò $Ø"Ð "Ü
|
|||
|
|
� Ó
|
|||
|
|
! AÒ
|
|||
|
|
%Û(�ؘE’>Ø#×*Ñ*¨5Õ1ð )ô �?Ó# qÒ(Ø&Ð&à&Ð&àˆIr c ób — d|v r| j ||||« S g }g }t |xs d« }| j j d« j |g « D ]6 \ }}t j |dz ||« } | |k7 sŒ&|j | « Œ8 |D ] }
|
|||
|
|
|
|
|||
|
|
|v sŒ|j |
|
|||
|
|
« Œ t |« dk( rà|D ]Û }
|
|||
|
|
| j j d« j dg « D ]J \ }}||
|
|||
|
|
v sΠt |
|
|||
|
|
« D ]2 \ } } | |k( sŒ |
|
|||
|
|
d| |z |
|
|||
|
|
| dz d z }
|
|||
|
|
|
|