B
YI~d9E � @ s� d dl Z d dlZd dlZd dlmZ dgZe �d�Ze �d�Ze �d�Z e �d�Z
e �d�Z e �d �Z e �d
�Z
e �d �Ze �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 comma
)?(?:\s|/(?!>))*
)*
)?
\s* # trailing whitespace
z#</\s*([a-zA-Z][-.a-zA-Z0-9:_]*)\s*>c @ s� e Zd ZdZdd�dd�Zdd� Zdd � Zd
d � Zd Zd
d� Z dd� Z
dd� Z dd� Z dd� Z
d8dd�Zdd� Zd d � Zd d � Zd d!� Zd"d#� Zd$d%� Zd&d'� Zd(d)� Zd*d+� Zd,d-� Zd.d/� Zd0d1� Zd2d3� Zd4d5� Z d6d7� Z d S )9r )ZscriptZstyleT)�convert_charrefsc C s || _ | �� d S )N)r �reset)�selfr � r �0/opt/alt/python37/lib64/python3.7/html/parser.py�__init__W s zHTMLParser.__init__c C s( d| _ d| _t| _d | _tj�| � d S )N� z???)�rawdata�lasttag�interesting_normal� interesting�
cdata_elem� _markupbase�
ParserBaser )r r r r r ` s
zHTMLParser.resetc C s | j | | _ | �d� d S )Nr )r �goahead)r �datar r r �feedh s zHTMLParser.feedc C s | � d� d S )N� )r )r r r r �closeq s zHTMLParser.closeNc C s | j S )N)�_HTMLParser__starttag_text)r r r r �get_starttag_textw s z HTMLParser.get_starttag_textc C s$ |� � | _t�d| j tj�| _d S )Nz </\s*%s\s*>)�lowerr �re�compile�Ir )r �elemr r r �set_cdata_mode{ s
zHTMLParser.set_cdata_modec C s t | _d | _d S )N)r r r )r r r r �clear_cdata_mode s zHTMLParser.clear_cdata_modec C s@ | j }d}t|�}�x�||k �r�| jrx| jsx|�d|�}|dk r�|�dt||d ��}|dkrrt�d�� ||�srP |}n(| j
� ||�}|r�|� � }n | jr�P |}||k r�| jr�| js�| � t
|||� �� n| � |||� � | �||�}||kr�P |j}|d|��rDt�||��r | �|�} n�|d|��r8| �|�} nl|d|��rP| �|�} nT|d|��rh| �|�} n<|d |��r�| �|�} n$|d
|k �r�| � d� |d
} nP | dk �r6|�s�P |�d |d
�} | dk �r�|�d|d
�} | dk �r�|d
} n| d
7 } | j�r$| j�s$| � t
||| � �� n| � ||| � � | �|| �}q|d |��r�t�||�}|�r�|�� d
d� }
| �|
� |�� } |d| d
��s�| d
} | �|| �}qn:d||d � k�r�| � |||d
� � | �||d
�}P q|d|�rt�||�}|�rF|�d
�}
| � |
� |�� } |d| d
��s8| d
} | �|| �}qt �||�}|�r�|�r�|�� ||d � k�r�|�� } | |k�r�|} | �||d
�}P n,|d
|k �r�| � d� | �||d
�}nP qqW |�r.||k �r.| j�s.| j�r| j�s| � t
|||� �� n| � |||� � | �||�}||d � | _ d S )Nr �<�&�" z[\s;]z</z<!--z<?z<!r r z&#� ����;) r �lenr r �find�rfind�maxr r �searchr �start� handle_datar Z updatepos�
startswith� starttagopen�match�parse_starttag� parse_endtag�
parse_comment�parse_pi�parse_html_declaration�charref�group�handle_charref�end� entityref�handle_entityref�
incomplete) r r9 r �i�n�jZampposr0 r. �k�namer r r r � s�
zHTMLParser.goaheadc C s� | j }|||d � dkr$| �|�S |||d � dkrB| �|�S |||d � �� dkr�|�d|d �}|dkrvdS | �||d |� � |d
S | �|�S d S ) N� z<!--� z<![� z <!doctyper r% r$ r )r r3 Zparse_marked_sectionr r( � handle_decl�parse_bogus_comment)r r= r �gtposr r r r5 s
z!HTMLParser.parse_html_declarationr c C sD | j }|�d|d �}|dkr"dS |r<| �||d |� � |d S )Nr r$ r% r )r r( �handle_comment)r r= Zreportr �posr r r rF s z HTMLParser.parse_bogus_commentc C sH | j }t�||d �}|s dS |�� }| �||d |� � |�� }|S )Nr$ r% )r �picloser+ r, � handle_pir9 )r r= r r0 r? r r r r4 ! s zHTMLParser.parse_pic C s� d | _ | �|�}|dk r |S | j}|||� | _ g }t�||d �}|�� }|�d��� | _}x�||k �r t �||�}|s~P |�ddd�\} }
} |
s�d } nZ| d d� d kr�| dd � ks�n | d d� d kr�| dd � kr�n n | dd� } | �rt
| �} |� | �� | f� |�� }qbW |||� � � } | dk�r�| �
� \}
}d | j k�r||
| j �d � }
t| j �| j �d � }n|t| j � }| �|||� � |S | �d
��r�| �||� n"| �||� || jk�r�| �|� |S ) Nr r r$ rC �'r% �")r z/>�
z/>)r �check_for_whole_start_tagr �tagfind_tolerantr0 r9 r7 r r
�attrfind_tolerantr �append�stripZgetpos�countr' r) r- �endswith�handle_startendtag�handle_starttag�CDATA_CONTENT_ELEMENTSr )r r= �endposr �attrsr0 r@ �tag�mZattrname�restZ attrvaluer9 �lineno�offsetr r r r1 - sP
&(
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 |dkr�dS ||kr�|S |d S td ��d S )
Nr r �/z/>r$ r% r z6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZzwe should not get here!)r �locatestarttagend_tolerantr0 r9 r. �AssertionError)r r= r r\ r? �nextr r r rO ` s. z$HTMLParser.check_for_whole_start_tagc C s | j }t�||d �}|s dS |�� }t�||�}|s�| jd k rV| �|||� � |S t�||d �}|s�|||d � dkr�|d S | � |�S |�
d�� � }|� d|�� �}| �
|� |d S |�
d�� � }| jd k r�|| jkr�| �|||� � |S | �
|� | �� |S )Nr r% r$ rC z</>r )r � endendtagr+ r9 �
endtagfindr0 r r- rP rF r7 r r( �
handle_endtagr )r r= r r0 rG Z namematchZtagnamer r r r r2 � s6
zHTMLParser.parse_endtagc C s | � ||� | �|� d S )N)rW rf )r r[ rZ r r r rV � s z HTMLParser.handle_startendtagc C s d S )Nr )r r[ rZ r r r rW � s zHTMLParser.handle_starttagc C s d S )Nr )r r[ r r r rf � s zHTMLParser.handle_endtagc C s d S )Nr )r rA r r r r8 � s zHTMLParser.handle_charrefc C s d S )Nr )r rA r r r r; � s zHTMLParser.handle_entityrefc C s d S )Nr )r r r r r r- � s zHTMLParser.handle_datac C s d S )Nr )r r r r r rH � s zHTMLParser.handle_commentc C s d S )Nr )r Zdeclr r r rE � s zHTMLParser.handle_declc C s d S )Nr )r r r r r rK � s zHTMLParser.handle_pic C s d S )Nr )r r r r r � unknown_decl� s zHTMLParser.unknown_declc C s t jdtdd� t|�S )NzZThe unescape method is deprecated and will be removed in 3.5, use html.unescape() instead.r$ )�
stacklevel)�warnings�warn�DeprecationWarningr )r �sr r r r � s
zHTMLParser.unescape)r ) �__name__�
__module__� __qualname__rX r
r r r r r r r r r5 rF r4 r1 rO r2 rV rW rf r8 r; r- rH rE rK rg r r r r r r ? s6 z
3"()r ri r Zhtmlr �__all__r r r< r: r6 r/ rJ Z commentcloserP rQ �VERBOSEra rd re r r r r r r �<module> s&