3
é3íXÃ$ ã @ sØ d dl mZmZmZ d dlmZmZmZ d dlm Z m
Z
m Z d dl m
Z
mZmZ d dlmZmZmZ G dd„ deƒZG dd „ d eƒZG d
d „ d eƒZG d d
„ d
eƒZG dd„ deƒZG dd„ deƒZG dd„ deƒZdS )é )ÚEUCTW_CHAR_TO_FREQ_ORDERÚEUCTW_TABLE_SIZEÚ EUCTW_TYPICAL_DISTRIBUTION_RATIO)ÚEUCKR_CHAR_TO_FREQ_ORDERÚEUCKR_TABLE_SIZEÚ EUCKR_TYPICAL_DISTRIBUTION_RATIO)ÚGB2312_CHAR_TO_FREQ_ORDERÚGB2312_TABLE_SIZEÚ!GB2312_TYPICAL_DISTRIBUTION_RATIO)ÚBIG5_CHAR_TO_FREQ_ORDERÚBIG5_TABLE_SIZEÚ BIG5_TYPICAL_DISTRIBUTION_RATIO)ÚJIS_CHAR_TO_FREQ_ORDERÚJIS_TABLE_SIZEÚ JIS_TYPICAL_DISTRIBUTION_RATIOc @ sL e Zd ZdZdZdZdZdd„ Zdd„ Zd d
„ Z d d „ Z
d
d„ Z dd„ Z dS )ÚCharDistributionAnalysisi g®Gáz®ï?g{®Gáz„?é c C s0 d | _ d | _d | _d | _d | _d | _| jƒ d S )N)Ú_char_to_freq_orderÚ _table_sizeÚtypical_distribution_ratioÚ_doneÚ _total_charsÚ _freq_charsÚreset)Úself© r ú&/usr/lib/python3.6/chardistribution.pyÚ__init__. s z!CharDistributionAnalysis.__init__c C s d| _ d| _d| _dS )z reset analyser, clear any stateFé N)r r r )r r r r r = s z CharDistributionAnalysis.resetc C sX |dkr| j |ƒ}nd}|dkrT| jd7 _|| jk rTd| j| krT| jd7 _dS )z"feed a character with known lengthé r r i Néÿÿÿÿ)Ú get_orderr r r r )r ÚcharZchar_lenÚorderr r r ÚfeedF 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 ÚMINIMUM_DATA_THRESHOLDÚSURE_NOr ÚSURE_YES)r Úrr r r Úget_confidenceT s
z'CharDistributionAnalysis.get_confidencec C s | j | jkS )N)r ÚENOUGH_DATA_THRESHOLD)r r r r Úgot_enough_datad s z(CharDistributionAnalysis.got_enough_datac C s dS )Nr r r )r Úbyte_strr r r r! i s z"CharDistributionAnalysis.get_orderN)
Ú__name__Ú
__module__Ú __qualname__r* r' r& r% r r r$ r) r+ r! r r r r r ( s r c s$ e Zd Z‡ fdd„Zdd„ Z‡ ZS )ÚEUCTWDistributionAnalysisc s$ t t| ƒjƒ t| _t| _t| _d S )N) Úsuperr0 r r r r r r r )r )Ú __class__r r r r s z"EUCTWDistributionAnalysis.__init__c C s0 |d }|dkr(d|d |d d S dS d S )Nr éÄ é^ r é¡ r r )r r, Ú
first_charr r r r! x s z#EUCTWDistributionAnalysis.get_order)r- r. r/ r r! Ú
__classcell__r r )r2 r r0 q s r0 c s$ e Zd Z‡ fdd„Zdd„ Z‡ ZS )ÚEUCKRDistributionAnalysisc s$ t t| ƒjƒ 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 é° 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‡ fdd„Zdd„ Z‡ ZS )ÚGB2312DistributionAnalysisc s$ t t| ƒjƒ 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 Ú second_charr r r r! ž s z$GB2312DistributionAnalysis.get_order)r- r. r/ r r! r7 r r )r2 r r: — s r: c s$ e Zd Z‡ fdd„Zdd„ Z‡ ZS )ÚBig5DistributionAnalysisc s$ t t| ƒjƒ 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 é¤ r5 é é? é@ 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< ª s r< c s$ e Zd Z‡ fdd„Zdd„ Z‡ ZS )ÚSJISDistributionAnalysisc s$ t t| ƒjƒ t| _t| _t| _d S )N) r1 rA r r r r r r r )r )r2 r r r Á 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 é éŸ é¼ éà éï é r@ é r r r )r r, r6 r; r# r r r r! Ç s z"SJISDistributionAnalysis.get_order)r- r. r/ r r! r7 r r )r2 r rA À s rA c s$ e Zd Z‡ fdd„Zdd„ Z‡ ZS )ÚEUCJPDistributionAnalysisc s$ t t| ƒjƒ t| _t| _t| _d S )N) r1 rI r r r r r r r )r )r2 r r r Ú s z"EUCJPDistributionAnalysis.__init__c C s0 |d }|dkr(d|d |d d S dS d S )Nr é r4 r5 r r r )r r, r" r r r r! à s z#EUCJPDistributionAnalysis.get_order)r- r. r/ r r! r7 r r )r2 r rI Ù s rI N) Z euctwfreqr r r Z euckrfreqr r r Z
gb2312freqr r r
Zbig5freqr r r
Zjisfreqr r r Úobjectr r0 r8 r: r< rA rI r r r r Ú s I