� J�g��z�ddlmZmZmZmZmZddlmZddlm Z m Z m Z Gd�de��Z Gd�de��Z d S) �)�Dict�List� NamedTuple�Optional�Union�)� CharSetProber)�CharacterCategory� ProbingState�SequenceLikelihoodc��eZdZUeed<eed<eeefed<eeeeeffed<eed<eed<eed<dS) �SingleByteCharSetModel� charset_name�language�char_to_order_map�language_model�typical_positive_ratio�keep_ascii_letters�alphabetN) �__name__� __module__� __qualname__�str�__annotations__r�int�float�bool���g/home/asafur/pinokio/api/open-webui.git/app/env/lib/python3.11/site-packages/chardet/sbcharsetprober.pyrr#sx������������M�M�M��C��H�~�%�%�%���d�3��8�n�,�-�-�-�-�!�!�!�!������M�M�M�M�Mrrc ����eZdZdZdZdZdZ ddeded e e d df�fd � Z d�fd � Z e d e efd ���Ze d e efd���Zdeeefd efd�Zd efd�Z�xZS)�SingleByteCharSetProber�@igffffff�?g�������?FN�model� is_reversed� name_prober�returnc����t�����||_||_||_d|_g|_d|_d|_d|_ d|_ |� ��dS�N��r) �super�__init__�_model� _reversed� _name_prober� _last_order� _seq_counters� _total_seqs� _total_char� _control_char� _freq_char�reset)�selfr$r%r&� __class__s �r r,z SingleByteCharSetProber.__init__3sq��� ���������� �$���'������(*��������������� � � � � � � � rc����t�����d|_dgtj��z|_d|_d|_d|_d|_ dSr)) r+r6r0r �get_num_categoriesr1r2r3r4r5)r7r8s �r r6zSingleByteCharSetProber.resetGsX��� ��� � ��������S�#5�#H�#J�#J�J����������������rc�@�|jr |jjS|jjS�N)r/rr-�r7s r rz$SingleByteCharSetProber.charset_nameRs$�� � � 2��$�1� 1��{�'�'rc�@�|jr |jjS|jjSr<)r/rr-r=s r rz SingleByteCharSetProber.languageXs$�� � � .��$�-� -��{�#�#r�byte_strc���|jjs|�|��}n|�|��}|s|jS|jj}|jj}|D]�}|�|tj ��}|tj kr|xj dz c_ ||j krs|xj dz c_ |j|j krS|xjdz c_|js||j|}n|||j}|j|xxdz cc<||_��|jj}|jt&jkr�|j|jkr�|���}||jkr.|j�d||��t&j|_n>||jkr3|j�d|||j��t&j|_|jS)Nrz$%s confidence = %s, we have a winnerz8%s confidence = %s, below negative shortcut threshold %s)r-r�filter_international_words�remove_xml_tags�staterr�getr � UNDEFINED�CONTROLr3� SAMPLE_SIZEr5r0r2r.r1rr � DETECTING�SB_ENOUGH_REL_THRESHOLD�get_confidence�POSITIVE_SHORTCUT_THRESHOLD�logger�debug�FOUND_IT�_state�NEGATIVE_SHORTCUT_THRESHOLD�NOT_ME) r7r?rr�char�order�lm_catr� confidences r �feedzSingleByteCharSetProber.feed^s ���{�-� 6��6�6�x�@�@�H�H��+�+�H�5�5�H�� ��:� � �K�9����3��� %� %�D�%�)�)�$�0A�0K�L�L�E� �(�0�0�0�� � �A�%� � ��t�'�'�'����1�$����#�d�&6�6�6��$�$��)�$�$��>�I�!/��0@�!A�%�!H���!/��!6�t�7G�!H���&�v�.�.�.�!�3�.�.�.�$�D� � ��{�/� � �:��/� /� /���$�">�>�>�!�0�0�2�2� ��� @�@�@��K�%�%�>� �j����#/�"7�D�K�K��$�"B�B�B��K�%�%�R�$�"��8� ��� #/�"5�D�K��z�rc��d}|jdkr{|jtjd|jtjzz|jz |jjz }||j|jz z|jz }||j z|jz }|dkrd}|S)Ng{�G�z�?rg�?g�?g�G�z��?) r2r1r �POSITIVE�LIKELYr-rr3r4r5)r7�rs r rJz&SingleByteCharSetProber.get_confidence�s��� �� � �a� � ��&�'9�'B�C��T�/�0B�0I�J�J�K��"� #� �+�4� 5� ��T�%��(:�:�;�d�>N�N�A��D�O�#�d�&6�6�A��C�x�x����r)FN)r'N)rrrrGrIrKrPrrrr r,r6�propertyrrrr�bytes� bytearrayr rVrrJ� __classcell__)r8s@r r"r"-sH��������K�"��"&��"&�� "�/3� ��%�����m�,� � � ������( � � � � � ��(�h�s�m�(�(�(��X�(� �$�(�3�-�$�$�$��X�$� 0�U�5�)�#3�4�0��0�0�0�0�d���������rr"N)�typingrrrrr� charsetproberr �enumsr r r rr"rrr �<module>rbs���:;�:�:�:�:�:�:�:�:�:�:�:�:�:�(�(�(�(�(�(�F�F�F�F�F�F�F�F�F�F������Z����u�u�u�u�u�m�u�u�u�u�ur
Memory