ruạṛ
3 AL�i:M � @ s d Z ddlmZmZmZmZ ddlT ddlmZ ddl Z ddl Z e j d�Z e j d�Z e j d�Ze j d �Ze j d �Ze j d �Ze j d �Ze j d �Ze j d�Ze j d�Ze j d�Ze j de j�Ze j de j�Ze j d �Ze j d�ZG dd� de �Z G dd� dej �Z dS )zLA parser for HTML and XHTML. Backported for python-future from Python 3.3. � )�absolute_import�division�print_function�unicode_literals)�*)� _markupbaseNz[&<]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][-.a-zA-Z0-9:_]*)(?:\s|/(?!>))*z[a-zA-Z][^ /> ]*zJ\s*([a-zA-Z_][-.:a-zA-Z_0-9]*)(\s*=\s*(\'[^\']*\'|"[^"]*"|[^\s"\'=<>`]*))?z]((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*a� <[a-zA-Z][-.a-zA-Z0-9:_]* # tag name (?:\s+ # whitespace before attribute name (?:[a-zA-Z_][-.:a-zA-Z0-9_]* # attribute name (?:\s*=\s* # value indicator (?:'[^']*' # LITA-enclosed value |\"[^\"]*\" # LIT-enclosed value |[^'\">\s]+ # bare value ) )? ) )* \s* # trailing whitespace aF <[a-zA-Z][-.a-zA-Z0-9:_]* # 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 ZdZddd�Zdd� ZdS ) �HTMLParseErrorz&Exception raised for all parse errors.Nc C s& |st �|| _|d | _|d | _d S )Nr � )�AssertionError�msg�lineno�offset)�selfr �position� r �C/root/tmp/pip-build-3x0y2zol/future/future/backports/html/parser.py�__init__U s zHTMLParseError.__init__c C s> | j }| jd k r |d| j }| jd k r:|d| jd }|S )Nz , at line %dz , column %dr )r r r )r �resultr r r �__str__[ s zHTMLParseError.__str__�NN)r )�__name__� __module__� __qualname__�__doc__r r r r r r r R s r c @ s� e Zd ZdZd;Zd<dd�Zdd� Zd d � Zd d � Zd d� Z dZ dd� Z dd� Z dd� Z dd� Zdd� Zd=dd �Zd d � Zd d � Zd!d"� Zd#d$� Zd%d&� Zd'd(� Zd)d*� Zd+d,� Zd-d.� Zd/d0� Zd1d2� Zd3d4� Z d5d6� Z d7d8� Z d9d:� Z dS )>� HTMLParsera� Find tags and other markup and call handler functions. Usage: p = HTMLParser() p.feed(data) ... p.close() Start tags are handled by calling self.handle_starttag() or self.handle_startendtag(); end tags by self.handle_endtag(). The data between tags is passed from the parser to the derived class by calling self.handle_data() with the data as argument (the data may be split up in arbitrary chunks). Entity references are passed by calling self.handle_entityref() with the entity reference as the argument. Numeric character references are passed to self.handle_charref() with the string containing the reference as the argument. �script�styleFc C s& |rt jdtdd� || _| j� dS )z�Initialize and reset this instance. If strict is set to False (the default) the parser will parse invalid markup, otherwise it will raise an error. Note that the strict mode is deprecated. z The strict mode is deprecated.� )� stacklevelN)�warnings�warn�DeprecationWarning�strict�reset)r r# r r r r z s zHTMLParser.__init__c C s( d| _ d| _t| _d| _tjj| � dS )z1Reset this instance. Loses all unprocessed data.� z???N)�rawdata�lasttag�interesting_normal� interesting� cdata_elemr � ParserBaser$ )r r r r r$ � s zHTMLParser.resetc C s | j | | _ | jd� dS )z�Feed data to the parser. Call this as often as you want, with as little or as much text as you want (may include '\n'). r N)r& �goahead)r �datar r r �feed� s zHTMLParser.feedc C s | j d� dS )zHandle any buffered data.r N)r, )r r r r �close� s zHTMLParser.closec C s t || j� ��d S )N)r �getpos)r �messager r r �error� s zHTMLParser.errorNc C s | j S )z)Return full source of start tag: '<...>'.)�_HTMLParser__starttag_text)r r r r �get_starttag_text� s z HTMLParser.get_starttag_textc C s$ |j � | _tjd| 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|�}�xF||k �rZ| jj||�}|r<|j� }n | jrDP |}||k rb| j|||� � | j||�}||krxP |j}|d|��r�t j ||�r�| j |�}n�|d|�r�| j |�}n||d|�r�| j |�}nf|d|�r�| j|�}nP|d|��r| j�r| j|�}n | j|�}n$|d |k �r2| jd� |d }nP |dk �r�|�sFP | j�rX| jd� |jd |d �}|dk �r�|jd|d �}|dk �r�|d }n|d7 }| j|||� � | j||�}q|d |��rZtj ||�}|�r&|j� d d� } | j| � |j� }|d |d ��s|d }| j||�}qn2d ||d � k�rV| j|dd � � | j|d �}P q|d |��rLtj ||�}|�r�|jd�} | j| � |j� }|d |d ��s�|d }| j||�}qtj ||�}|�r |�r|j� ||d � k�r| j�r�| jd� n ||k�r |}| j||d �}P n,|d |k �rH| jd � | j||d �}nP qdstd��qW |�r�||k �r�| j �r�| j|||� � | j||�}||d � | _ d S )Nr �<z</z<!--z<?z<!r zEOF in middle of constructr z&#r �;�&z#EOF in middle of entity or char refzinteresting.search() lied���) r& �lenr) �search�startr* � handle_data� updatepos� startswith� starttagopen�match�parse_starttag� parse_endtag� parse_comment�parse_pir# �parse_declaration�parse_html_declarationr2 �find�charref�group�handle_charref�end� entityref�handle_entityref� incompleter ) r rR r& �i�nrG �jrE �k�namer r r r, � s� zHTMLParser.goaheadc C s� | j }|||d � dks"td��|||d � dkr@| j|�S |||d � dkr^| j|�S |||d � j� d kr�|jd |d �}|d kr�d S | j||d |� � |d S | j|�S d S )Nr z<!z+unexpected call to parse_html_declaration()� z<!--� z<![� z <!doctyper r r? r? )r& r rJ �parse_marked_sectionr5 rN � handle_decl�parse_bogus_comment)r rV r& �gtposr r r rM s z!HTMLParser.parse_html_declarationr c C s` | j }|||d � dks"td��|jd|d �}|dkr>d S |rX| j||d |� � |d S ) Nr �<!�</z"unexpected call to parse_comment()r r )rb rc r? r? )r& r rN �handle_comment)r rV �reportr& �posr r r r` - s z HTMLParser.parse_bogus_commentc C sd | j }|||d � dks"td��tj||d �}|s:dS |j� }| j||d |� � |j� }|S )Nr z<?z unexpected call to parse_pi()r r? )r& r �picloserA rB � handle_pirR )r rV r& rG rX r r r rK 9 s zHTMLParser.parse_pic C s2 d | _ | j|�}|dk r |S | j}|||� | _ g }tj||d �}|sPtd��|j� }|jd�j� | _ }x�||k �rJ| j r�t j||�}n t j||�}|s�P |jddd�\} } } | s�d } nb| d d� d ko�| d d � kn �s| d d� d k�o | dd � kn �r | dd� } | �r.| j | �} |j| j� | f� |j� }qnW |||� j� } | dk�r�| j� \} }d | j k�r�| | j jd � } t| j �| j jd � }n|t| j � }| j �r�| jd |||� d d � f � | j|||� � |S | jd ��r | j||� n"| j||� || jk�r.| j|� |S )Nr r z#unexpected call to parse_starttag()r r\ �'�"r �/>� z junk characters in start tag: %r� r? r? r? )r rk )r3 �check_for_whole_start_tagr& �tagfindrG r rR rP r5 r' r# �attrfind�attrfind_tolerant�unescape�append�stripr0 �countr@ �rfindr2 rC �endswith�handle_startendtag�handle_starttag�CDATA_CONTENT_ELEMENTSr: )r rV �endposr& �attrsrG rY �tag�m�attrname�rest� attrvaluerR r r r r r rH E s\ *, zHTMLParser.parse_starttagc C s | j }| jrtj||�}n tj||�}|� r�|j� }|||d � }|dkrT|d S |dkr�|jd|�rp|d S |jd|�r�d S | jr�| j||d � | jd� ||kr�|S |d S |dkr�d S |dkr�d S | jr�| j||� | jd � ||kr�|S |d S t d ��d S )Nr r �/z/>r zmalformed empty start tagr% z6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZzmalformed start tagzwe should not get here!r? r? r? ) r& r# �locatestarttagendrG �locatestarttagend_tolerantrR rE rD r2 r )r rV r& r~ rX �nextr r r rn ~ s> z$HTMLParser.check_for_whole_start_tagc C sP | j }|||d � dks"td��tj||d �}|s:d S |j� }tj||�}|� s�| jd k rt| j|||� � |S | j r�| j d|||� f � t j||d �}|s�|||d � dkr�|d S | j |�S |j � j� }|jd|j� �}| j|� |d S |j d�j� }| jd k �r6|| jk�r6| j|||� � |S | j|j� � | j� |S ) Nr z</z unexpected call to parse_endtagr zbad end tag: %rr\ z</>r r? )r& r � endendtagrA rR � endtagfindrG r* rC r# r2 �tagfind_tolerantr` rP r5 rN � handle_endtagr; )r rV r& rG ra � namematch�tagnamer9 r r r rI � s< zHTMLParser.parse_endtagc C s | j ||� | j|� d S )N)ry r� )r r} r| r r r rx � s z HTMLParser.handle_startendtagc C s d S )Nr )r r} r| r r r ry � s zHTMLParser.handle_starttagc C s d S )Nr )r r} r r r r� � s zHTMLParser.handle_endtagc C s d S )Nr )r rZ r r r rQ � s zHTMLParser.handle_charrefc C s d S )Nr )r rZ r r r rT � s zHTMLParser.handle_entityrefc C s d S )Nr )r r- r r r rC � s zHTMLParser.handle_datac C s d S )Nr )r r- r r r rd � s zHTMLParser.handle_commentc C s d S )Nr )r �declr r r r_ � s zHTMLParser.handle_declc C s d S )Nr )r r- r r r rh � s zHTMLParser.handle_pic C s | j r| jd|f � d S )Nzunknown declaration: %r)r# r2 )r r- r r r � unknown_decl� s zHTMLParser.unknown_declc C s" d|kr |S dd� }t jd||�S )Nr> c S s� | j � d } yX| d dkrb| dd � } | d d krLt| dd � jd�d�}nt| jd��}t|�S W n tk r| d| S X dd lm} | |kr�|| S | jd�r�d | S xJtd t | ��D ]0}| d |� |kr�|| d |� | |d � S q�W d | S d S ) Nr �#r �x�Xr= � z&#)�html5r> r )r� r� ) �groups�int�rstrip�chr� ValueErrorZ future.backports.html.entitiesr� rw �ranger@ )�s�cr� r� r r r �replaceEntities� s& z,HTMLParser.unescape.<locals>.replaceEntitiesz&&(#?[xX]?(?:[0-9a-fA-F]+;|\w{1,32};?)))r6 �sub)r r� r� r r r rr � s zHTMLParser.unescape)r r )F)r ) r r r r rz r r$ r. r/ r2 r3 r4 r: r; r, rM r` rK rH rn rI rx ry r� rQ rT rC rd r_ rh r� rr r r r r r d s: h 9+*r ) r � __future__r r r r Zfuture.builtinsZfuture.backportsr r6 r r7 r( rU rS rO rF rg � commentclosero r� rp rq �VERBOSEr� r� r� r� � Exceptionr r+ r r r r r �<module> s4
cải xoăn