File manager - Edit - /opt/saltstack/salt/lib/python3.10/site-packages/charset_normalizer/__pycache__/utils.cpython-310.pyc
Back
o ;j:- � @ s� d dl Z d dlZd dlZd dlmZ d dlmZ d dlmZ d dl m Z d dl m Z m Z mZmZmZmZ d dlmZ dd lmZmZmZmZmZmZ eed �d ed e fd d��Z eed �d ed efdd��Z eed �d ed ee fdd��Z eed �d ed e fdd��Z eed �d ed e fdd��Z!eed �d ed e fdd��Z"eed �d ed e fdd��Z#eed �d ed e fdd ��Z$eed �d ed e fd d ��Z%eed �d ed e fd d ��Z&d ed e fd!d"�Z'eed �d ed e fd#d$��Z(eed �d ed e fd%d&��Z)eed �d ed e fd'd(��Z*eed �d ed e fd)d*��Z+eed �d ed e fd+d,��Z,ee-e�d �d-ed e fd.d/��Z.eed �d ed e fd0d1��Z/d_d3e0d4e1d ee fd5d6�Z2ed7d �d8ed e fd9d:��Z3d3e0d eee e0f fd;d<�Z4d=ed e fd>d?�Z5d`dAedBe d efdCdD�Z6dEed e e fdFdG�Z7dHedIed e8fdJdK�Z9dHedIed e fdLdM�Z:dNej;dOfd8edPe1dQed dfdRdS�Z< dadTe0dUedVe=dWe1dXe dYe dZe0d[e d\ee d e eddf fd]d^�Z>dS )b� N)�IncrementalDecoder)�aliases)� lru_cache)�findall)� Generator�List�Optional�Set�Tuple�Union)�MultibyteIncrementalDecoder� )�ENCODING_MARKS�IANA_SUPPORTED_SIMILAR� RE_POSSIBLE_ENCODING_INDICATION�UNICODE_RANGES_COMBINED� UNICODE_SECONDARY_RANGE_KEYWORD�UTF8_MAXIMAL_ALLOCATION)�maxsize� character�returnc C sT zt �| �}W n ty Y dS w d|v p)d|v p)d|v p)d|v p)d|v p)d|v S )NFz WITH GRAVEz WITH ACUTEz WITH CEDILLAzWITH DIAERESISzWITH CIRCUMFLEXz WITH TILDE�� unicodedata�name� ValueError�r � description� r �L/opt/saltstack/salt/lib/python3.10/site-packages/charset_normalizer/utils.py�is_accentuated s ������r c C s. t �| �}|s | S |�d�}tt|d d��S )N� r � )r Z decomposition�split�chr�int)r Z decomposedZcodesr r r � remove_accent& s r% c C s. t | �}t�� D ] \}}||v r| S qdS )zK Retrieve the Unicode range official name from a single character. N)�ordr �items)r Z character_ord� range_nameZ ord_ranger r r � unicode_range1 s �r) c C �* z t �| �}W d|v S ty Y dS w )NFZLATINr r r r r �is_latin? s ��r+ c C s& z| � d� W dS ty Y dS w )N�asciiFT)�encode�UnicodeEncodeError�r r r r �is_asciiH s ��r0 c C s2 t �| �}d|v r dS t| �}|d u rdS d|v S )N�PTFZ Punctuation�r �categoryr) �r �character_category�character_ranger r r �is_punctuationQ s r7 c C s: t �| �}d|v s d|v rdS t| �}|d u rdS d|v S )N�S�NTFZFormsr2 r4 r r r � is_symbol` s r: c C s t | �}|d u r dS d|v S )NFZ Emoticons)r) )r r6 r r r � is_emoticono s r; c C s. | � � s| dv r dS t�| �}d|v p|dv S )N> �+u ｜�<�>T�Z> ZPdZPoZPc)�isspacer r3 �r r5 r r r � is_separatory s rB c C s | � � | �� kS �N)�islower�isupperr/ r r r �is_case_variable� s rF c C s t �| �}|dkS )NZCo)r r3 rA r r r �is_private_use_only� s rG c C r* )NFZCJKr �r Zcharacter_namer r r �is_cjk� � ��rI c C r* )NFZHIRAGANAr rH r r r � is_hiragana� rJ rK c C r* )NFZKATAKANAr rH r r r � is_katakana� rJ rL c C r* )NFZHANGULr rH r r r � is_hangul� rJ rM c C r* )NFZTHAIr rH r r r �is_thai� rJ rN r( c s t � fdd�tD ��S )Nc 3 s � | ]}|� v V qd S rC r )�.0�keyword�r( r r � <genexpr>� s � z-is_unicode_range_secondary.<locals>.<genexpr>)�anyr rQ r rQ r �is_unicode_range_secondary� s rT c C s( | � � du o| �� du o| dko| dkS )NF�u )r@ � isprintabler/ r r r �is_unprintable� s ���rW � �sequence� search_zonec C s� t | t�st�t| �}tt| dt||�� jddd��}t|�dkr$dS |D ]'}|�� � dd�}t � � D ]\}}||krB| S ||krL| S q4q&dS )zW Extract using ASCII-only decoder any specified encoding in the first n-bytes. Nr, �ignore��errorsr �-�_) � isinstance�bytes� TypeError�lenr r �min�decode�lower�replacer r' )rY rZ Zseq_len�resultsZspecified_encoding�encoding_alias� encoding_ianar r r �any_specified_encoding� s& � ��rk � r c C s | dv pt t�d�| ��jt�S )zQ Verify is a specific encoding is a multi byte one based on it IANA name > � utf_16_be� utf_32_be�utf_7Z utf_8_sig� utf_32_le� utf_16_le�utf_8�utf_32�utf_16� encodings.{})� issubclass� importlib� import_module�formatr r )r r r r �is_multi_byte_encoding� s ��rz c C sJ t D ] }t | }t|t�r|g}|D ]}| �|�r!||f S qqdS )z9 Identify and extract SIG/BOM in given sequence. )N� )r r` ra � startswith)rY � iana_encodingZmarksZmarkr r r �identify_sig_or_bom s ��r~ r} c C s | dvS )N> rs rt r )r} r r r �should_strip_sig_or_bom s r T�cp_name�strictc C sL | � � �dd�} t�� D ]\}}| ||fv r| S q |r$td�| ���| S )Nr^ r_ z Unable to retrieve IANA for '{}')rf rg r r' r ry )r� r� ri rj r r r � iana_name s �r� �decoded_sequencec C s4 t � }| D ]}t|�}|d u rq|�|� qt|�S rC )�setr) �add�list)r� Zrangesr r6 r r r � range_scan+ s r� � iana_name_a� iana_name_bc C s� t | �st |�r dS t�d�| ��j}t�d�|��j}|dd�}|dd�}d}td�D ]}t|g�}|�|�|�|�krA|d7 }q,|d S ) Ng ru r[ r\ r � r � )rz rw rx ry r �rangera re ) r� r� Z decoder_aZ decoder_bZid_aZid_bZcharacter_match_count�iZ to_be_decodedr r r � cp_similarity9 s* ���� �r� c C s | t v o |t | v S )z� Determine if two code page are at least 80% similar. IANA_SUPPORTED_SIMILAR dict was generated using the function cp_similarity. )r )r� r� r r r � is_cp_similarQ s �r� Zcharset_normalizerz)%(asctime)s | %(levelname)s | %(message)s�level� format_stringc C s: t �| �}|�|� t �� }|�t �|�� |�|� d S rC )�logging� getLogger�setLevel� StreamHandler� setFormatter� Formatter� addHandler)r r� r� �logger�handlerr r r �set_logging_handler\ s r� � sequencesrj �offsets� chunk_size�bom_or_sig_available�strip_sig_or_bom� sig_payload�is_multi_byte_decoder�decoded_payloadc c s& � |r |du r |D ]} || | | � } | s d S | V q d S |D ]p} | | } | t | �d kr/q | | | | � } |rA|du rA|| } | j||rHdndd�} |r�| dkr�t|d�} |r�| d | � |vr�t| | d d �D ]#}| || � } |r{|du r{|| } | j|dd�} | d | � |v r� nqi| V q d S ) NF� r[ r� r\ r r! � ���)rc re rd r� )r� rj r� r� r� r� r� r� r� r� �chunkZ chunk_endZ cut_sequenceZchunk_partial_size_chk�jr r r �cut_sequence_chunksi sD � � � � ��r� )rX )TrC )?rw r� r �codecsr Zencodings.aliasesr � functoolsr �rer �typingr r r r r r Z_multibytecodecr Zconstantr r r r r r �str�boolr r% r) r+ r0 r7 r: r; rB rF rG rI rK rL rM rN rc rT rW ra r$ rk rz r~ r r� r� �floatr� r� �INFOr� r� r� r r r r �<module> s� ���� ���������� � �
| ver. 1.4 |
Github
|
.
| PHP 8.2.30 | Generation time: 0 |
proxy
|
phpinfo
|
Settings