File: //lib/python3/dist-packages/bs4/builder/__pycache__/_lxml.cpython-38.pyc
U
t�^�/ �
@ s� d Z ddgZzddlmZ W n. ek
rJ Z zddlmZ W 5 dZ[X Y nX ddlmZ ddlm Z ddl
m Z dd l m
Z
mZmZmZmZ dd
lmZmZmZmZmZmZmZ dd lmZ d Z d
d� Z G dd� de�Z G dd� dee �Z dS )ZMIT�LXMLTreeBuilderForXML�LXMLTreeBuilder� )�CallableN)�BytesIO)�StringIO)�etree)�Comment�Doctype�NamespacedAttribute�ProcessingInstruction�XMLProcessingInstruction)�FAST�HTML�HTMLTreeBuilder�
PERMISSIVE�ParserRejectedMarkup� TreeBuilder�XML)�EncodingDetector�lxmlc C s t dd� t| �� �D ��S )zInvert a dictionary.c s s | ]\}}||fV qd S �N� )�.0�k�vr r �3/usr/lib/python3/dist-packages/bs4/builder/_lxml.py� <genexpr>& s z_invert.<locals>.<genexpr>)�dict�list�items)�dr r r �_invert$ s r! c s� e Zd ZejZdZeZdZ dgZ
e e e e
egZdZedd�Zee�Z� fdd�Zd d
� Zd d � Zd
d� Zd*� fdd� Zdd� Zd+dd�Zdd� Z dd� Z i fdd�Z d d � Z d d � Z d d!� Z!d"d#� Z"d$d%� Z#d&d'� Z$d(d)� Z%� Z&S ),r Tzlxml-xml�xmli z$http://www.w3.org/XML/1998/namespace)r" c s t t| ��|� | �| j� dS )z�Let the BeautifulSoup object know about the standard namespace
mapping.
:param soup: A `BeautifulSoup`.
N)�superr �initialize_soup�_register_namespaces�DEFAULT_NSMAPS)�self�soup�� __class__r r r$ C s z%LXMLTreeBuilderForXML.initialize_soupc C s6 t |�� �D ]$\}}|r || jjkr || jj|< q dS )z�Let the BeautifulSoup object know about namespaces encountered
while parsing the document.
This might be useful later on when creating CSS selectors.
:param mapping: A dictionary mapping namespace prefixes to URIs.
N)r r r( Z _namespaces)r' �mapping�key�valuer r r r% L s z*LXMLTreeBuilderForXML._register_namespacesc C s" | j dk r| j S tj| dd|d�S )z�Find the default parser for the given encoding.
:param encoding: A string.
:return: Either a parser object or a class, which
will be instantiated with default arguments.
NFT��targetZ strip_cdataZrecover�encoding)�_default_parserr � XMLParser�r' r0 r r r �default_parser[ s
�z$LXMLTreeBuilderForXML.default_parserc C s( | � |�}t|t�r$|| dd|d�}|S )z�Instantiate an appropriate parser for the given encoding.
:param encoding: A string.
:return: A parser object such as an `etree.XMLParser`.
FTr. )r4 �
isinstancer )r' r0 �parserr r r �
parser_forg s
�z LXMLTreeBuilderForXML.parser_forNc s> || _ |d k rt|�| _d | _| jg| _tt| �jf |� d S r ) r1 �set�empty_element_tagsr( �DEFAULT_NSMAPS_INVERTED�nsmapsr# r �__init__)r' r6 r9 �kwargsr) r r r< w s
z LXMLTreeBuilderForXML.__init__c C s0 |d dkr$t |dd � �dd��S d |fS d S )Nr �{� �})�tuple�split)r' �tagr r r � _getNsTag� s z LXMLTreeBuilderForXML._getNsTagc c s� | j }|rt| _nt| _t|t�r2|d|dfV t|t�rP|�d�d|dfV ||g}t||||�}|jD ]}|j ||dfV qldS )aA Run any preliminary steps necessary to make incoming markup
acceptable to the parser.
lxml really wants to get a bytestring and convert it to
Unicode itself. So instead of using UnicodeDammit to convert
the bytestring to Unicode using different encodings, this
implementation uses EncodingDetector to iterate over the
encodings, and tell lxml to try to parse the document as each
one in turn.
:param markup: Some markup -- hopefully a bytestring.
:param user_specified_encoding: The user asked to try this encoding.
:param document_declared_encoding: The markup itself claims to be
in this encoding.
:param exclude_encodings: The user asked _not_ to try any of
these encodings.
:yield: A series of 4-tuples:
(markup, encoding, declared encoding,
has undergone character replacement)
Each 4-tuple represents a strategy for converting the
document to Unicode and parsing it. Each strategy will be tried
in turn.
NF�utf8)
�is_xmlr � processing_instruction_classr r5 �str�encoder Z encodings�markup) r' rJ Zuser_specified_encodingZexclude_encodingsZdocument_declared_encodingZis_htmlZ
try_encodingsZdetectorr0 r r r �prepare_markup� s(
� �
z$LXMLTreeBuilderForXML.prepare_markupc
C s� t |t�rt|�}nt |t�r&t|�}|�| j�}z\| �| jj �| _
| j
� |� t |�dkr�|�| j�}t |�dkrP| j
� |� qP| j
�
� W n2 tttjfk
r� } z t|��W 5 d }~X Y nX d S )Nr )r5 �bytesr rH r �read�
CHUNK_SIZEr7 r( �original_encodingr6 �feed�len�close�UnicodeDecodeError� LookupErrorr � ParserErrorr )r' rJ �data�er r r rP � s
zLXMLTreeBuilderForXML.feedc C s | j g| _d S r )r: r; )r' r r r rR � s zLXMLTreeBuilderForXML.closec C s t |�}d }t|�dkr4t| j�dkr4| j�d � nXt|�dkr�| �|� | j�t|�� |�� }t|�� �D ] \}}t d|d�}|||< qni }t|�� �D ]F\} }
| �
| �\}} |d kr�|
|| < q�| � |�}t || |�} |
|| < q�|}| �
|�\}}| � |�}| j �
||||� d S )Nr r? Zxmlnsz http://www.w3.org/2000/xmlns/)r rQ r; �appendr% r! �copyr r r
rD �_prefix_for_namespacer( Zhandle_starttag) r' �nameZattrsZnsmap�nsprefix�prefix� namespaceZ attributeZ new_attrs�attrr- r r r �start� s6
�
zLXMLTreeBuilderForXML.startc C s< |dkr dS t | j�D ] }|dk r||kr|| S qdS )z9Find the currently active prefix for the given namespace.N)�reversedr; )r' r^ �inverted_nsmapr r r rZ s z+LXMLTreeBuilderForXML._prefix_for_namespacec C s� | j �� | j jd }| �|�\}}d }|d k r\t| j�D ] }|d k r:||kr:|| } q\q:| j �||� t| j�dkr�| j�� d S )N���r? ) r( �endDataZtagStackrD ra r; Z
handle_endtagrQ �pop)r' r[ Z
completed_tagr^ r\ rb r r r �end s
zLXMLTreeBuilderForXML.endc C s0 | j �� | j �|d | � | j �| j� d S )N� )r( rd � handle_datarG )r' r/ rV r r r �pi s
zLXMLTreeBuilderForXML.pic C s | j �|� d S r )r( rh �r' Zcontentr r r rV ! s zLXMLTreeBuilderForXML.datac C s( | j �� t�|||�}| j �|� d S r )r( rd r Zfor_name_and_idsZobject_was_parsed)r' r[ Zpubid�system�doctyper r r rl $ s
z LXMLTreeBuilderForXML.doctypec C s& | j �� | j �|� | j �t� dS )z#Handle comments as Comment objects.N)r( rd rh r rj r r r �comment) s
z LXMLTreeBuilderForXML.commentc C s d| S )�See `TreeBuilder`.z)<?xml version="1.0" encoding="utf-8"?>
%sr �r' Zfragmentr r r �test_fragment_to_document/ s z/LXMLTreeBuilderForXML.test_fragment_to_document)NN)NNN)'�__name__�
__module__� __qualname__r r2 ZDEFAULT_PARSER_CLASSrF r rG �NAME�ALTERNATE_NAMES�LXMLr r
r �featuresrN r r& r! r: r$ r% r4 r7 r< rD rK rP rR r` rZ rf ri rV rl rm rp �
__classcell__r r r) r r ( s:
�
3 . c @ sF e Zd ZeZdgZeeeeeg Z dZ
e Z dd� Z
dd� Zdd� Zd S )
r z lxml-htmlFc C s t jS r )r Z
HTMLParserr3 r r r r4 = s z LXMLTreeBuilder.default_parserc
C sf | j j}z&| �|�| _| j�|� | j�� W n2 tttj fk
r` } z t
|��W 5 d }~X Y nX d S r ) r( rO r7 r6 rP rR rS rT r rU r )r' rJ r0 rW r r r rP @ s zLXMLTreeBuilder.feedc C s d| S )rn z <html><body>%s</body></html>r ro r r r rp J s z)LXMLTreeBuilder.test_fragment_to_documentN)rq rr rs rv rt ru r r
r rw rF r rG r4 rP rp r r r r r 4 s
) Z __license__�__all__Zcollections.abcr � ImportErrorrW � collections�ior r r r Z bs4.elementr r r
r r Z bs4.builderr
r r r r r r Z
bs4.dammitr rv r! r r r r r r �<module> s& � $