a
XC?h�C � @ s� d dl Z d dlZd dlmZ dgZe �d�Ze �d�Ze �d�Ze �d�Z e �d�Z
e �d �Z e �d
�Z e �d �Z
e �d �Ze �d
e j�Ze �d �Ze �d�ZG dd� dej�ZdS )� N)�unescape�
HTMLParserz[&<]z
&[a-zA-Z#]z%&([a-zA-Z][-.a-zA-Z0-9]*)[^a-zA-Z0-9]z)&#(?:[0-9]+|[xX][0-9a-fA-F]+)[^0-9a-fA-F]z <[a-zA-Z]�>z--\s*>z+([a-zA-Z][^\t\n\r\f />\x00]*)(?:\s|/(?!>))*z]((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*aF
<[a-zA-Z][^\t\n\r\f />\x00]* # tag name
(?:[\s/]* # optional whitespace before attribute name
(?:(?<=['"\s/])[^\s/>][^\s/=>]* # attribute name
(?:\s*=+\s* # value indicator
(?:'[^']*' # LITA-enclosed value
|"[^"]*" # LIT-enclosed value
|(?!['"])[^>\s]* # bare value
)
\s* # possibly followed by a space
)?(?:\s|/(?!>))*
)*
)?
\s* # trailing whitespace
z#c @ s� e Zd ZdZdd�dd�Zdd� Zdd � Zd
d � Zd Zd
d� Z dd� Z
dd� Z dd� Z dd� Z
d6dd�Zdd� Zd d � Zd d � Zd d!� Zd"d#� Zd$d%� Zd&d'� Zd(d)� Zd*d+� Zd,d-� Zd.d/� Zd0d1� Zd2d3� Zd4d5� Z d S )7r )ZscriptZstyleT)�convert_charrefsc C s || _ | �� d S �N)r �reset)�selfr � r �0/opt/alt/python39/lib64/python3.9/html/parser.py�__init__V s zHTMLParser.__init__c C s( d| _ d| _t| _d | _tj�| � d S )N� z???)�rawdata�lasttag�interesting_normal� interesting�
cdata_elem� _markupbase�
ParserBaser �r r r r
r _ s
zHTMLParser.resetc C s | j | | _ | �d� d S )Nr )r
�goahead�r �datar r r
�feedg s zHTMLParser.feedc C s | � d� d S )N� )r r r r r
�closep s zHTMLParser.closeNc C s | j S r )�_HTMLParser__starttag_textr r r r
�get_starttag_textv s z HTMLParser.get_starttag_textc C s$ |� � | _t�d| j tj�| _d S )Nz )�lowerr �re�compile�Ir )r �elemr r r
�set_cdata_modez s
zHTMLParser.set_cdata_modec C s t | _d | _d S r )r r r r r r r
�clear_cdata_mode~ s zHTMLParser.clear_cdata_modec C sJ | j }d}t|�}||k �r�| jrv| jsv|�d|�}|dk r�|�dt||d ��}|dkrpt�d�� ||�sp�q�|}n*| j
� ||�}|r�|� � }n| jr��q�|}||k r�| jr�| js�| � t
|||� �� n| � |||� � | �||�}||kr��q�|j}|d|��rJt�||��r"| �|�} n�|d|��r:| �|�} nn|d|��rR| �|�} nV|d|��rj| �|�} n>|d |��r�| �|�} n&|d
|k �r�| � d� |d
} n�q�| dk �r<|�s��q�|�d |d
�} | dk �r�|�d|d
�} | dk �r|d
} n| d
7 } | j�r*| j�s*| � t
||| � �� n| � ||| � � | �|| �}q|d |��r�t�||�}|�r�|�� d
d� }
| �|
� |�� } |d| d
��s�| d
} | �|| �}qn�
rR )r �check_for_whole_start_tagr
�tagfind_tolerantr4 r= r; r r �attrfind_tolerantr �append�stripZgetpos�countr+ r- r1 �endswith�handle_startendtag�handle_starttag�CDATA_CONTENT_ELEMENTSr" )r rA �endposr
�attrsr4 rD �tag�m�attrname�restZ attrvaluer= �lineno�offsetr r r
r5 , sX
& �
�
�
zHTMLParser.parse_starttagc C s� | j }t�||�}|r�|�� }|||d � }|dkr>|d S |dkr~|�d|�rZ|d S |�d|�rjdS ||krv|S |d S |dkr�dS |dv r�dS ||kr�|S |d S td ��d S )
Nr r �/rR r( r) r z6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZzwe should not get here!)r
�locatestarttagend_tolerantr4 r= r2 �AssertionError)r rA r
ra rC �nextr r r
rT _ s. z$HTMLParser.check_for_whole_start_tagc C s | j }t�||d �}|s dS |�� }t�||�}|s�| jd urV| �|||� � |S t�||d �}|s�|||d � dkr�|d S | � |�S |�
d�� � }|� d|�� �}| �
|� |d S |�
d�� � }| jd ur�|| jkr�| �|||� � |S | �
|� | �� |S )Nr r) r( rG zr )r
� endendtagr/ r= �
endtagfindr4 r r1 rU rJ r; r r, �
handle_endtagr# )r rA r
r4 rK Z namematchZtagnamer! r r r
r6 � s6
zHTMLParser.parse_endtagc C s | � ||� | �|� d S r )r\ rl �r r` r_ r r r
r[ � s z HTMLParser.handle_startendtagc C s d S r r rm r r r
r\ � s zHTMLParser.handle_starttagc C s d S r r )r r` r r r
rl � s zHTMLParser.handle_endtagc C s d S r r �r rE r r r
r< � s zHTMLParser.handle_charrefc C s d S r r rn r r r
r? � s zHTMLParser.handle_entityrefc C s d S r r r r r r
r1 � s zHTMLParser.handle_datac C s d S r r r r r r
rL � s zHTMLParser.handle_commentc C s d S r r )r Zdeclr r r
rI � s zHTMLParser.handle_declc C s d S r r r r r r
rO � s zHTMLParser.handle_pic C s d S r r r r r r
� unknown_decl� s zHTMLParser.unknown_decl)r ) �__name__�
__module__� __qualname__r] r r r r r r r" r# r r9 rJ r8 r5 rT r6 r[ r\ rl r< r? r1 rL rI rO ro r r r r
r > s4 z
3"()r r Zhtmlr �__all__r r r@ r> r: r3 rN Z commentcloserU rV �VERBOSErg rj rk r r r r r r
� s(
��