".
v d dl mZmZ ddlmZ ddlmZ ddlmZm Z m
Z
ddl m Z m
Z
mZmZ G d de Zd S )
)OptionalUnion )
CharSetProber)CodingStateMachine)LanguageFilter MachineState ProbingState) HZ_SM_MODELISO2022CN_SM_MODELISO2022JP_SM_MODELISO2022KR_SM_MODELc e Zd ZdZej fdeddf fd
Zd fd Zede e
fd Z ede e
fd Z de
fd Zd
eeef defd Z xZS )
EscCharSetProberz
This CharSetProber uses a "code scheme" approach for detecting encodings,
whereby easily recognizable escape or shift sequences are relied on to
identify these encodings.
lang_filterreturnNc t | g | _ | j t j z rX| j t t | j t t | j t j
z r,| j t t | j t j z r,| j t t d| _ d | _ d | _ t" j | _ | d S )N)r r )super__init__ coding_smr r CHINESE_SIMPLIFIEDappendr r r JAPANESEr
KOREANr active_sm_count_detected_charset_detected_languager
DETECTING_statereset)selfr __class__s c/builddir/build/BUILD/cloudlinux-venv-1.0.10/venv/lib/python3.11/site-packages/chardet/escprober.pyr zEscCharSetProber.__init__0 s
[111 n ? ? J N ! !"4["A"A B B B N ! !"45G"H"H I I I n 5 5 J N ! !"45G"H"H I I I n 3 3 J N ! !"45G"H"H I I I 04 15 ",
c t | j D ] }d|_ | t | j | _ d | _ d | _ d S )NT)r r r activelenr r r )r! r r" s r# r zEscCharSetProber.reset@ sh
I #I OO "4> 2 2 !% "& r$ c | j S Nr r! s r# charset_namez EscCharSetProber.charset_nameI s %%r$ c | j S r) )r r+ s r# languagezEscCharSetProber.languageM s &&r$ c | j rdndS )NgGz?g r* r+ s r# get_confidencez EscCharSetProber.get_confidenceQ s -7tt47r$ byte_strc |D ]}| j D ]}|j s
| | }|t j k r?d|_ | xj dz c_ | j dk r t j | _ | j c c S n|t j
k rAt j | _ | | _
|j | _ | j c c S | j S )NFr r )r r&
next_stater ERRORr r
NOT_MEr stateITS_MEFOUND_ITget_coding_state_machiner r. r )r! r1 cr coding_states r# feedzEscCharSetProber.feedT s & &A !^
&
& ' ( 3 3A 6 6 <#555',I$((A-((+q00&2&9 #z))))) 1 "\%888"."7DK-6-O-O-Q-QD*.7.@D+ :%%%%% 9
& zr$ )r N)__name__
__module__ __qualname____doc__r NONEr r propertyr strr, r. floatr0 r bytes bytearrayr
r<
__classcell__)r" s @r# r r ) s 6D5H N T ' ' ' ' ' ' &hsm & & & X& '(3- ' ' ' X'8 8 8 8 8U5)#3 4 r$ r N)typingr r
charsetproberr codingstatemachiner enumsr r r
escsmr r r
r r r$ r# rN s 8 # " " " " " " " ( ( ( ( ( ( 2 2 2 2 2 2 = = = = = = = = = = = = = = =} = = = = =r$