3
aj$ @ s d dl mZmZmZ d dlmZmZmZ d dlm Z m
Z
m Z d dl m
Z
mZmZ d dlmZmZmZ G dd deZG dd d eZG d
d d eZG d d
d
eZG dd deZG dd deZG dd deZdS ) )EUCTW_CHAR_TO_FREQ_ORDEREUCTW_TABLE_SIZE EUCTW_TYPICAL_DISTRIBUTION_RATIO)EUCKR_CHAR_TO_FREQ_ORDEREUCKR_TABLE_SIZE EUCKR_TYPICAL_DISTRIBUTION_RATIO)GB2312_CHAR_TO_FREQ_ORDERGB2312_TABLE_SIZE!GB2312_TYPICAL_DISTRIBUTION_RATIO)BIG5_CHAR_TO_FREQ_ORDERBIG5_TABLE_SIZE BIG5_TYPICAL_DISTRIBUTION_RATIO)JIS_CHAR_TO_FREQ_ORDERJIS_TABLE_SIZE JIS_TYPICAL_DISTRIBUTION_RATIOc @ sL e Zd ZdZdZdZdZdd Zdd Zd d
Z d d Z
d
d Z dd Z dS )CharDistributionAnalysisi gGz?g{Gz? c C s0 d | _ d | _d | _d | _d | _d | _| j d S )N)_char_to_freq_order _table_sizetypical_distribution_ratio_done _total_chars _freq_charsreset)self r &/usr/lib/python3.6/chardistribution.py__init__. s z!CharDistributionAnalysis.__init__c C s d| _ d| _d| _dS )z reset analyser, clear any stateF N)r r r )r r r r r = s z CharDistributionAnalysis.resetc C sX |dkr| j |}nd}|dkrT| jd7 _|| jk rTd| j| krT| jd7 _dS )z"feed a character with known length r r i N) get_orderr r r r )r charZchar_lenorderr r r feedF s
z CharDistributionAnalysis.feedc C sT | j dks| j| jkr | jS | j | jkrN| j| j | j | j }|| jk rN|S | jS )z(return confidence based on existing datar )r r MINIMUM_DATA_THRESHOLDSURE_NOr SURE_YES)r rr r r get_confidenceT s
z'CharDistributionAnalysis.get_confidencec C s | j | jkS )N)r ENOUGH_DATA_THRESHOLD)r r r r got_enough_datad s z(CharDistributionAnalysis.got_enough_datac C s dS )Nr r r )r byte_strr r r r! i s z"CharDistributionAnalysis.get_orderN)
__name__
__module__ __qualname__r* r' r& r% r r r$ r) r+ r! r r r r r ( s r c s$ e Zd Z fddZdd Z ZS )EUCTWDistributionAnalysisc s$ t t| j t| _t| _t| _d S )N) superr0 r r r r r r r )r ) __class__r r r r s z"EUCTWDistributionAnalysis.__init__c C s0 |d }|dkr(d|d |d d S dS d S )Nr ^ r r r )r r,
first_charr r r r! x s z#EUCTWDistributionAnalysis.get_order)r- r. r/ r r!
__classcell__r r )r2 r r0 q s r0 c s$ e Zd Z fddZdd Z ZS )EUCKRDistributionAnalysisc s$ t t| j t| _t| _t| _d S )N) r1 r8 r r r r r r r )r )r2 r r r s z"EUCKRDistributionAnalysis.__init__c C s0 |d }|dkr(d|d |d d S dS d S )Nr r4 r r5 r r )r r, r6 r r r r! s z#EUCKRDistributionAnalysis.get_order)r- r. r/ r r! r7 r r )r2 r r8 s r8 c s$ e Zd Z fddZdd Z ZS )GB2312DistributionAnalysisc s$ t t| j t| _t| _t| _d S )N) r1 r: r r r r r r
r )r )r2 r r r s z#GB2312DistributionAnalysis.__init__c C s> |d |d }}|dkr6|dkr6d|d | d S dS d S )Nr r r9 r5 r4 r r )r r, r6 second_charr r r r! s z$GB2312DistributionAnalysis.get_order)r- r. r/ r r! r7 r r )r2 r r: s r: c s$ e Zd Z fddZdd Z ZS )Big5DistributionAnalysisc s$ t t| j t| _t| _t| _d S )N) r1 r< r r r r r r
r )r )r2 r r r s z!Big5DistributionAnalysis.__init__c C sX |d |d }}|dkrP|dkr:d|d | d d S d|d | d S ndS d S ) Nr r r5 ? @ r r )r r, r6 r; r r r r! s z"Big5DistributionAnalysis.get_order)r- r. r/ r r! r7 r r )r2 r r< s r< c s$ e Zd Z fddZdd Z ZS )SJISDistributionAnalysisc s$ t t| j t| _t| _t| _d S )N) r1 rA r r r r r r r )r )r2 r r r s z!SJISDistributionAnalysis.__init__c C sr |d |d }}|dkr0|dkr0d|d }n&|dkrR|dkrRd|d d }nd S || d }|d
krnd }|S )
Nr r r@ r r r )r r, r6 r; r# r r r r! s z"SJISDistributionAnalysis.get_order)r- r. r/ r r! r7 r r )r2 r rA s rA c s$ e Zd Z fddZdd Z ZS )EUCJPDistributionAnalysisc s$ t t| j t| _t| _t| _d S )N) r1 rI r r r r r r r )r )r2 r r r s z"EUCJPDistributionAnalysis.__init__c C s0 |d }|dkr(d|d |d d S dS d S )Nr r4 r5 r r r )r r, r" r r r r! s z#EUCJPDistributionAnalysis.get_order)r- r. r/ r r! r7 r r )r2 r rI s rI N) Z euctwfreqr r r Z euckrfreqr r r Z
gb2312freqr r r
Zbig5freqr r r
Zjisfreqr r r objectr r0 r8 r: r< rA rI r r r r <module> s I