|
Server : Apache System : Linux iad1-shared-b8-43 6.6.49-grsec-jammy+ #10 SMP Thu Sep 12 23:23:08 UTC 2024 x86_64 User : dh_edsupp ( 6597262) PHP Version : 8.2.26 Disable Function : NONE Directory : /lib/python3/dist-packages/genshi/__pycache__/ |
Upload File : |
o
!�b�@ � @ s� d Z ddlmZ ddlZddlmZ ddlZddlmZ m
Z ddl m
Z
mZmZmZ ddl mZmZmZmZmZmZmZmZmZmZmZ ddl m Z m Z g d �Z d
Z d d � Z!G d
d� de"�Z#G dd� de$�Z%dd� Z&G dd� de j'e$�Z'ddd�Z(dd� Z)dS )zOSupport for constructing markup streams from files, strings, or other
sources.
� )�chainN)�expat)�
html_entities� html_parser)�Attrs�QName�Stream�
stripentities) �START�END�XML_DECL�DOCTYPE�TEXT�START_NS�END_NS� START_CDATA� END_CDATA�PI�COMMENT)�StringIO�BytesIO)�ET�
ParseError� XMLParser�XML�
HTMLParser�HTMLzrestructuredtext enc c s� � t | j�d��}tdd� | �� D ��}t||fdfV | jr&t| jdfV | D ] }t|�D ]}|V q.q(t |dfV | j
rGt| j
dfV dS dS )z�Convert a given ElementTree element to a markup stream.
:param element: an ElementTree element
:return: a markup stream
�{c S s" g | ]
\}}t |�d ��|f�qS )r )r �lstrip)�.0�attr�value� r" �./usr/lib/python3/dist-packages/genshi/input.py�
<listcomp>* s �zET.<locals>.<listcomp>�N���r& N) r �tagr r �itemsr
�textr r r �tail)�element�tag_name�attrs�child�itemr" r" r# r # s �� � �r c @ s e Zd ZdZddd�ZdS )r zWException raised when fatal syntax errors are found in the input being
parsed.
Nr& c C s< || _ |r |d| 7 }t�| |� |pd| _|| _|| _dS )a: Exception initializer.
:param message: the error message from the parser
:param filename: the path to the file that was parsed
:param lineno: the number of the line on which the error was encountered
:param offset: the column number where the error was encountered
z, in z<string>N)�msg� Exception�__init__�filename�lineno�offset)�self�messager3 r4 r5 r" r" r# r2 = s
zParseError.__init__r% )�__name__�
__module__� __qualname__�__doc__r2 r" r" r" r# r 8 s r c @ s� e Zd ZdZdd� ej�� D �Zd�e�� d�Z
d-dd�Z d d
� Z d d � Z
d
d� Zd-dd�Zdd� Zdd� Zdd� Zdd� Zdd� Zdd � Zd d � Zd d � Zd!d"� Zd#d$� Zd%d&� Zd'd(� Zd)d*� Z d+d,� Z dS ).r a� Generator-based XML parser based on roughly equivalent code in
Kid/ElementTree.
The parsing is initiated by iterating over the parser object:
>>> parser = XMLParser(StringIO('<root id="2"><child>Foo</child></root>'))
>>> for kind, data, pos in parser:
... print('%s %s' % (kind, data))
START (QName('root'), Attrs([(QName('id'), '2')]))
START (QName('child'), Attrs())
TEXT Foo
END child
END root
c C s g | ]
\}}d ||f �qS )z<!ENTITY %s "&#%d;">r" �r �namer! r" r" r# r$ ^ � zXMLParser.<listcomp>�
�utf-8Nc C s� || _ || _t�|d�}d|_t|d�rd|_d|_| j|_ | j
|_ | j |_
| j|_| j|_| j|_| j|_| j|_| j|_| j|_| j |_ | j |_ |� tj!� |�"� | j#|_$|| _g | _%dS )a� Initialize the parser for the given XML input.
:param source: the XML text as a file-like object
:param filename: the name of the file, if appropriate
:param encoding: the encoding of the file; if not specified, the
encoding is assumed to be ASCII, UTF-8, or UTF-16, or
whatever the encoding specified in the XML declaration
(if any)
�}T�returns_unicodeN)&�sourcer3 r � ParserCreate� buffer_text�hasattrrB �ordered_attributes�
_handle_start�StartElementHandler� _handle_end�EndElementHandler� _handle_data�CharacterDataHandler�_handle_doctype�StartDoctypeDeclHandler�_handle_start_ns�StartNamespaceDeclHandler�_handle_end_ns�EndNamespaceDeclHandler�_handle_start_cdata�StartCdataSectionHandler�_handle_end_cdata�EndCdataSectionHandler�
_handle_pi� ProcessingInstructionHandler�_handle_xml_decl�XmlDeclHandler�_handle_comment�CommentHandler�
_handle_other�DefaultHandler�SetParamEntityParsing� XML_PARAM_ENTITY_PARSING_ALWAYS�
UseForeignDTD�_build_foreign�ExternalEntityRefHandler�_queue)r6 rC r3 �encoding�parserr" r" r# r2 b s0
zXMLParser.__init__c � � fdd�}t |� ��t�S )z�Generator that parses the XML source, yielding markup events.
:return: a markup event stream
:raises ParseError: if the XML text is not well formed
c
3 s� � zTd} d} |sDt � j�dkrD� j�| �}|s)t� d�r&� j�dd� � `d}nt|tj �r4|�
d�}� j�|d� |sDt � j�dks� jD ]}|V qGg � _|rUW d S q tj yp } zt |�}t
|� j|j|j��d }~ww ) N� F� r r � Tr@ )�lenre rC �readrF r �Parse�
isinstance�six� text_type�encode�
ExpatError�strr r3 r4 r5 )�bufsize�done�data�event�er0 �r6 r" r# � _generate� s6 �
�
���z"XMLParser.parse.<locals>._generate�r �filter� _coalesce�r6 r{ r" rz r# �parse� s zXMLParser.parsec C � t | �� �S �N��iterr� rz r" r" r# �__iter__� � zXMLParser.__iter__c C s | j �|�}|�t| j�� dS )Nrj )r �ExternalEntityParserCreate� ParseFiler �
_external_dtd)r6 �context�base�sysid�pubidrg r" r" r# rc � s zXMLParser._build_foreignc C s~ |d u r| � � }|tu r4d|v r!|�� }|d t|� d }d}n |d }|d t|� }|d ||f}| j�|||f� d S )Nr? rj r& � r )�_getposr �
splitlinesrl re �append)r6 �kindrw �pos�linesr4 r5 r" r" r# �_enqueue� s zXMLParser._enqueuec C s | j ddfS )Nr& )r3 rz r" r" r# �_getpos_unknown� r� zXMLParser._getpos_unknownc C s | j | jj| jjfS r� )r3 r �CurrentLineNumber�CurrentColumnNumberrz r" r" r# r� � s
�zXMLParser._getposc C s8 t dd� tt|�gd � D ��}| �tt|�|f� d S )Nc S s g | ]
\}}t |�|f�qS r" )r r< r" r" r# r$ � r> z+XMLParser._handle_start.<locals>.<listcomp>r� )r �zipr� r� r
r )r6 r' �attribr- r" r" r# rH � s �zXMLParser._handle_startc C s | � tt|�� d S r� )r� r r )r6 r' r" r" r# rJ � � zXMLParser._handle_endc C � | � t|� d S r� �r� r �r6 r) r" r" r# rL � � zXMLParser._handle_datac C s | � t|||f� d S r� )r� r )r6 �versionrf �
standaloner" r" r# rZ � � zXMLParser._handle_xml_declc C s | � t|||f� d S r� )r� r
)r6 r= r� r� �has_internal_subsetr" r" r# rN � r� zXMLParser._handle_doctypec C s | � t|pd|f� d S �Nrk )r� r )r6 �prefix�urir" r" r# rP � s zXMLParser._handle_start_nsc C s | � t|pd� d S r� )r� r )r6 r� r" r" r# rR � r� zXMLParser._handle_end_nsc C � | � t� d S r� )r� r rz r" r" r# rT � � z XMLParser._handle_start_cdatac C r� r� )r� r rz r" r" r# rV � r� zXMLParser._handle_end_cdatac C s | � t||f� d S r� )r� r )r6 �targetrw r" r" r# rX � r� zXMLParser._handle_pic C r� r� �r� r r� r" r" r# r\ � r� zXMLParser._handle_commentc C s� |� d�r@zt�tj|dd� �}| �t|� W d S ty? | �� \}}}t �
d|||f �}t j j |_
||_||_|�w d S )N�&rj r& z)undefined entity "%s": line %d, column %d)�
startswithrp �unichr�entities�name2codepointr� r �KeyErrorr� r �error�errors�XML_ERROR_UNDEFINED_ENTITY�coder4 r5 )r6 r) r3 r4 r5 r� r" r" r# r^ � s
�
��zXMLParser._handle_other�NN) r8 r9 r: r; r� r� r( � _entitydefs�joinrr r� r2 r� r� rc r� r� r� rH rJ rL rZ rN rP rR rT rV rX r\ r^ r" r" r" r# r N s2 �
+
r c C s t ttt| ����S )a? Parse the given XML source and return a markup stream.
Unlike with `XMLParser`, the returned stream is reusable, meaning it can be
iterated over multiple times:
>>> xml = XML('<doc><elem>Foo</elem><elem>Bar</elem></doc>')
>>> print(xml)
<doc><elem>Foo</elem><elem>Bar</elem></doc>
>>> print(xml.select('elem'))
<elem>Foo</elem><elem>Bar</elem>
>>> print(xml.select('elem/text()'))
FooBar
:param text: the XML source
:return: the parsed XML event stream
:raises ParseError: if the XML text is not well-formed
)r �listr r )r) r" r" r# r s r c @ s� e Zd ZdZeg d��Zd dd�Zdd� Zdd � Zd d
d �Z d d
� Z
dd� Z dd� Z dd� Z
dd� Zdd� Zdd� Zdd� ZdS ) r aU Parser for HTML input based on the Python `HTMLParser` module.
This class provides the same interface for generating stream events as
`XMLParser`, and attempts to automatically balance tags.
The parsing is initiated by iterating over the parser object:
>>> parser = HTMLParser(BytesIO(u'<UL compact><LI>Foo</UL>'.encode('utf-8')), encoding='utf-8')
>>> for kind, data, pos in parser:
... print('%s %s' % (kind, data))
START (QName('ul'), Attrs([(QName('compact'), 'compact')]))
START (QName('li'), Attrs())
TEXT Foo
END li
END ul
)
�arear� �basefont�br�col�frame�hr�img�input�isindex�link�meta�paramNc C s. t j�| � || _|| _|| _g | _g | _dS )a Initialize the parser for the given HTML input.
:param source: the HTML text as a file-like object
:param filename: the name of the file, if known
:param filename: encoding of the file; ignored if the input is unicode
N)�htmlr r2 rC r3 rf re �
_open_tags)r6 rC r3 rf r" r" r# r2 - s
zHTMLParser.__init__c rh )z�Generator that parses the HTML source, yielding markup events.
:return: a markup event stream
:raises ParseError: if the HTML text is not well formed
c
3 s* � � j rt�� j �} | � j�}n� j}z`d}d} |sHt� j�dkrH|�|�}|s0� �� d}nt|t j
�s:t d��� � |� |sHt� j�dks"� jD ] \}}}|||fV qKg � _|rs� j
}|�� |D ]
}tt|�|fV qeW d S q tjy� } zd| j| j| jf }
t|
� j| j| j��d } ~ ww )Nri Frj r Tz0source returned bytes, but no encoding specifiedz%s: line %d, column %d)rf �codecs� getreaderrC rl re rm �closero rp rq � UnicodeError�feedr� �reverser r r� �HTMLParseErrorr0 r4 r5 r r3 ) �readerrC ru rv rw r� r� � open_tagsr' ry r0 rz r" r# r{ A sB �
� ���z#HTMLParser.parse.<locals>._generater| r r" rz r# r� ; s zHTMLParser.parsec C r� r� r� rz r" r" r# r� b r� zHTMLParser.__iter__c C s& |d u r| � � }| j�|||f� d S r� )r� re r� )r6 r� rw r� r" r" r# r� e s zHTMLParser._enqueuec C s | � � \}}| j||fS r� )�getposr3 )r6 r4 �columnr" r" r# r� j s zHTMLParser._getposc C sz g }|D ]\}}|d u r|}|� t|�t|�f� q| �tt|�t|�f� || jv r5| �tt|�� d S | j� |� d S r� ) r� r r r� r
r � _EMPTY_ELEMSr r� )r6 r' r� � fixed_attribr= r! r" r" r# �handle_starttagn s
zHTMLParser.handle_starttagc C sP || j vr$| jr&| j�� }| �tt|�� |�� |�� kr d S | jsd S d S d S r� )r� r� �popr� r r �lower)r6 r' �open_tagr" r" r# �
handle_endtag{ s
��zHTMLParser.handle_endtagc C r� r� r� r� r" r" r# � handle_data� r� zHTMLParser.handle_datac C sF |� � �d�rt�t|dd � d��}nt�t|��}| �t|� d S )N�xrj � )r� r� rp r� �intr� r �r6 r= r) r" r" r# �handle_charref� s zHTMLParser.handle_charrefc C s@ z
t �tj| �}W n
ty d| }Y nw | �t|� d S )Nz&%s;)rp r� r� r� r� r� r r� r" r" r# �handle_entityref� s �zHTMLParser.handle_entityrefc C sb |� d�r |d d� }z
|�d d�\}}W n
ty" |}d}Y nw | �t|�� |�� f� d S )N�?r& rj rk )�endswith�split�
ValueErrorr� r �strip)r6 rw r� r" r" r# � handle_pi� s
� zHTMLParser.handle_pic C r� r� r� r� r" r" r# �handle_comment� r� zHTMLParser.handle_commentr� r� )r8 r9 r: r; � frozensetr� r2 r� r� r� r� r� r� r� r� r� r� r� r" r" r" r# r s
'
r c C s< t | tj�rtttt| �dd���S tttt| �|d���S )aM Parse the given HTML source and return a markup stream.
Unlike with `HTMLParser`, the returned stream is reusable, meaning it can be
iterated over multiple times:
>>> html = HTML('<body><h1>Foo</h1></body>', encoding='utf-8')
>>> print(html)
<body><h1>Foo</h1></body>
>>> print(html.select('h1'))
<h1>Foo</h1>
>>> print(html.select('h1/text()'))
Foo
:param text: the HTML source
:return: the parsed XML event stream
:raises ParseError: if the HTML text is not well-formed, and error recovery
fails
N)rf )ro rp rq r r� r r r )r) rf r" r" r# r � s r c c sz � g }d}t | dg�D ]/\}}}|tu r |�|� |du r |}q |r2td�|�|fV |dd�= d}|r:|||fV q dS )z3Coalesces adjacent TEXT events into a single event.N)NNNrk )r r r� r� )�stream�textbuf�textposr� rw r� r" r" r# r~ � s" �
�
��r~ r� )*r; � itertoolsr r� � xml.parsersr rp � six.movesr r� r r� � genshi.corer r r r r
r r r
r r r r r r r �
genshi.compatr r �__all__�
__docformat__r r1 r �objectr r r r r~ r" r" r" r# �<module> s(
4 5