관리-도구
편집 파일: sbcharsetprober.cpython-310.pyc
o ƚRe� � @ sN d dl mZ ddlmZ ddlmZmZmZ edg d��ZG dd� de�Z d S ) � )� namedtuple� )� CharSetProber)�CharacterCategory� ProbingState�SequenceLikelihood�SingleByteCharSetModel)� charset_name�language�char_to_order_map�language_model�typical_positive_ratio�keep_ascii_letters�alphabetc sb e Zd ZdZdZdZdZd� fdd� Z� fd d �Ze d d � �Z e d d� �Z dd� Z dd� Z � ZS )�SingleByteCharSetProber�@ i gffffff�?g�������?FNc sJ t t| ��� || _|| _|| _d | _d | _d | _d | _ d | _ | � � d S �N) �superr �__init__�_model� _reversed� _name_prober� _last_order� _seq_counters� _total_seqs� _total_char� _freq_char�reset)�self�model�reversed� name_prober�� __class__� ��/builddir/build/BUILDROOT/alt-python310-pip-21.3.1-3.el8.x86_64/opt/alt/python310/lib/python3.10/site-packages/pip/_vendor/chardet/sbcharsetprober.pyr 3 s z SingleByteCharSetProber.__init__c s: t t| ��� d| _dgt�� | _d| _d| _d| _ d S )N� r ) r r r r r �get_num_categoriesr r r r �r r" r$ r% r A s z SingleByteCharSetProber.resetc C � | j r| j jS | jjS r )r r r r( r$ r$ r% r K � z$SingleByteCharSetProber.charset_namec C r) r )r r r r( r$ r$ r% r R r* z SingleByteCharSetProber.languagec C sF | j js | �|�}|s| jS | j j}| j j}|D ]L}|�|tj�}|tj k r-| j d7 _ || j k ra| j d7 _ | j | j k ra| jd7 _| jsQ|| j | }n|| | j }| j| d7 < || _ q| j j}| jtjkr�| j| jkr�| �� }|| jkr�| j�d||� tj| _| jS || jk r�| j�d||| j� tj | _| jS )Nr z$%s confidence = %s, we have a winnerz9%s confidence = %s, below negative shortcut threshhold %s) r r �filter_international_words�stater r �getr � UNDEFINED�CONTROLr � SAMPLE_SIZEr r r r r r r � DETECTING�SB_ENOUGH_REL_THRESHOLD�get_confidence�POSITIVE_SHORTCUT_THRESHOLD�logger�debug�FOUND_IT�_state�NEGATIVE_SHORTCUT_THRESHOLD�NOT_ME) r �byte_strr r �char�order�lm_catr � confidencer$ r$ r% �feedY sJ � ��z SingleByteCharSetProber.feedc C sL d}| j dkr$d| jtj | j | jj }|| j | j }|dkr$d}|S )Ng{�G�z�?r g �?g�G�z��?)r r r �POSITIVEr r r r )r �rr$ r$ r% r3 � s ��z&SingleByteCharSetProber.get_confidence)FN)�__name__� __module__� __qualname__r0 r2 r4 r9 r r �propertyr r r@ r3 � __classcell__r$ r$ r" r% r - s 0r N) � collectionsr � charsetproberr �enumsr r r r r r$ r$ r$ r% �<module> s �