/usr/local/lib/python3.6/site-packages/sacremoses/__pycache__
Edit: /usr/local/lib/python3.6/site-packages/sacremoses/__pycache__/tokenize.cpython-36.pyc (16593B)
3
Egz @ s d dl Z d dlmZ d dlmZ d dlmZ d dlmZ d dlm Z m
Z
e Ze ZG dd de
ZG d d
d
e
Zdd
gZdS ) N) text_type)Perluniprops)NonbreakingPrefixes)is_cjk)VIRAMASNUKTASc 2 s e Zd ZdZedjejdZedjejddje dje
ZedjejdZedjejdZ
edjejddje dje
ZedjejdZdZdZdZdZdZdjedfZdjeddfZdZdZdZdjedfZdjedfZdjedfZdZdZdZ dZ!dZ"dZ#dZ$dZ%dZ&dZ'd0jed1d2fZ(d3jed1d2fZ)d4jed1d2fZ*d5jee
dfZ+d6jedd7fZ,dZ-dZ.dZ/dZ0dZ1dZ2dZ3dZ4dZ5dZ6dZ7dZ8dZ9dZ:dZ;dZ
dZ?dZ@dZAdZBdZCdZDdZEdZFdZGdZHdZIdZJdZKdZLdZMdZNdZOdZPdZQdZRdZSdZTdZUdZVdZWdZXdZYdjeddfZZdjeeddfZ[djeddfZ\djeddfZ]djeddfZ^eZe[e\e]e^gZ_djeddfZ`djeddfZadjeddfZbdjeddfZce`eaebecgZddZedZfdZgdZhdZidZjdZkeeeee e!e"e#e$e%e&e(e)e*e+e,e-e.e/e0e1e2e3e4e5e6e7e8e9e:e;ee?e@eAeBeCeDeEeFeGeHeIeJeKeLeMeNg2Zle'eOePeQeReSeTeUeVeWg
ZmeReSeTeUeVeWeXeYgZnegeheiejekgZoddddgZpd fdd Zqdd Zrdd Zsdd Ztdd Zudd Zvdd Zwdd ZxdddZydddZz Z{S )MosesTokenizerz
This is a Python port of the Moses Tokenizer from
https://github.com/moses-smt/mosesdecoder/blob/master/scripts/tokenizer/tokenizer.perl
IsNIsAlnumIsScIsSoIsAlphaIsLower\s+ [\000-\037] +^ $z([^{}\s\.'\`\,\-]) \1 z ([{alphanum}])\-(?=[{alphanum}]))alphanumz\1 @-@ \.([\.]+)
DOTMULTIDOTMULTI\.([^\.])
DOTDOTMULTI
DOTMULTI\.DOTDOTMULTIz
([^{}])[,]z\1 , z
[,]([^{}])z , \1z
([{}])[,]$^```` ^"^`([^`])` \1^'`
([ ([{<])"\1 `` ([ ([{<])``([ ([{<])`([^`])\1 ` \2
([ ([{<])'\1 ` \.\.\. _ELLIPSIS_
_ELLIPSIS_z([^{numbers}])[,]([^{numbers}]))Znumbersz\1 , \2z([{numbers}])[,]([^{numbers}])z([^{numbers}])[,]([{numbers}])z([;:@#\$%&{}{}])z([{alphanum}])\/([{alphanum}])z$1 \@\/\@ $2([^.])([.])([\]\)}>"']*) ?$\1 \2\3([?!])([\]\[\(\){}<>])\(-LRB-\)-RRB-\[-LSB-\]-RSB-\{-LCB-\}-RCB--- -- ^$" '' ([^'])' \1 ' ([^'])''([sSmMdD]) '\1 'll 'll 're 're 've 've n't n't 'LL 'LL 'RE 'RE 'VE 'VE N'T N'T
([Cc])annot
\1an not ([Dd])'ye \1' ye ([Gg])imme \1im me ([Gg])onna \1on na ([Gg])otta \1ot ta ([Ll])emme \1em me
([Mm])ore'n
\1ore 'n '([Tt])is '\1 is '([Tt])was '\1 was ([Ww])anna \1an na *^ * *$&&\||<<>>\''\""[]]z([^{alpha}])[']([^{alpha}]))alphaz\1 ' \2z([^{alpha}{isn}])[']([{alpha}]))r isnz([{alpha}])[']([^{alpha}])z([{alpha}])[']([{alpha}])z\1 '\2z([{isn}])[']([s]))r z([^{alpha}])[']([{alpha}])z\1' \2 ' \.' ?$ . ' z<\/?\S+\/?>z#<\S+( [a-zA-Z0-9]+\="?[^"]")+ ?\/?>z#<\S+( [a-zA-Z0-9]+\='?[^']')+ ?\/?>z'[\w\-\_\.]+\@([\w\-\_]+\.)+[a-zA-Z]{2,}z/(http[s]?|ftp):\/\/[^:\/\s]+(\/\w+)*\/[\w\-\.]+z"((https?|ftp|rsync)://|www\.)[^ ]*z@[a-zA-Z0-9_]+z#[a-zA-Z0-9_]+enNc s t t j | _dd tj|D _|rg _t|dB}x:|D ]2}|j }|rF|j d rF| jkrF jj
| qFW W d Q R X fdd jD _ jdkrd
} jdkr|td
j
tjd7 } jdk r|td
j
tjd7 } jdkrF|td
j
tjd
7 }|td
j
tjd7 }|td
j
tjd7 } j|7 _ j|7 _dj jdf _dj jddf _dj jddf _d S )Nc S s g | ]}|j qS )strip).0Z_nbpr r =/usr/local/lib/python3.6/site-packages/sacremoses/tokenize.py
1 s z+MosesTokenizer.__init__..r#c s$ g | ]} j |r|jd d qS )r r )has_numeric_only
rpartition)r w)selfr r r ? s zhjakocjkr ZHangulHanHiraganaKatakanaz([^{}\s\.'\`\,\-])z \1 z ([{alphanum}])\-(?=[{alphanum}]))r z\1 @-@ z([{alphanum}])\/([{alphanum}])z$1 \@\/\@ $2)r r r r )r r )r r )r r )superr __init__langnonbreaking_prefixeswordsNONBREAKING_PREFIXESopenr
startswithappendNUMERIC_ONLY_PREFIXESr joinperlunipropscharsr r formatPAD_NOT_ISALNUMAGGRESSIVE_HYPHEN_SPLITINTRATOKEN_SLASHES)r r Z custom_nonbreaking_prefixes_fileZfinlineZ cjk_chars) __class__)r r r * s<