U
|
¡ý°d8: ã
|
@ sò 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
|
lmZmZmZmZmZmZmZmZ dd lm Z d Z d d Z G dd deZ G dd dee Z dS )ÚMITÚLXMLTreeBuilderForXMLÚLXMLTreeBuilderé )ÚCallableN)ÚBytesIO)ÚStringIO)Úetree)ÚCommentÚDoctypeÚNamespacedAttributeÚProcessingInstructionÚXMLProcessingInstruction)ÚDetectsXMLParsedAsHTMLÚFASTÚHTMLÚHTMLTreeBuilderÚ
|
PERMISSIVEÚParserRejectedMarkupÚ TreeBuilderÚXML)ÚEncodingDetectorÚlxmlc C s t dd t| ¡ D S )zInvert a dictionary.c s s | ]\}}||fV qd S ©N© )Ú.0ÚkÚvr r úHd:\z\workplace\vscode\pyvenv\venv\Lib\site-packages\bs4/builder/_lxml.pyÚ <genexpr>' s z_invert.<locals>.<genexpr>)ÚdictÚlistÚitems)Údr r r Ú_invert% s r# c sä e Zd ZejZdZeZdZ dgZ
|
e e e e egZdZeddZeeZ fddZd d
|
Zd d Zd d Zd* fdd Zdd Zd+ddZdd Z dd Z i fddZ 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-xmlÚxmli z$http://www.w3.org/XML/1998/namespace)r$ c s t t| |¡ | | j¡ dS )zLet the BeautifulSoup object know about the standard namespace
|
mapping.
|
|
:param soup: A `BeautifulSoup`.
|
N)Úsuperr Úinitialize_soupÚ_register_namespacesÚDEFAULT_NSMAPS)ÚselfÚsoup©Ú __class__r r r& D s z%LXMLTreeBuilderForXML.initialize_soupc C s6 t | ¡ D ]$\}}|r || jjkr || jj|< q dS )aÿ Let the BeautifulSoup object know about namespaces encountered
|
while parsing the document.
|
|
This might be useful later on when creating CSS selectors.
|
|
This will track (almost) all namespaces, even ones that were
|
only in scope for part of the document. If two namespaces have
|
the same prefix, only the first one encountered will be
|
tracked. Un-prefixed namespaces are not tracked.
|
|
:param mapping: A dictionary mapping namespace prefixes to URIs.
|
N)r r! r* Ú _namespaces)r) ÚmappingÚkeyÚvaluer r r r' M s z*LXMLTreeBuilderForXML._register_namespacesc C s" | j dk r| j S tj| dd|dS )zÍFind the default parser for the given encoding.
|
|
:param encoding: A string.
|
:return: Either a parser object or a class, which
|
will be instantiated with default arguments.
|
NFT©ÚtargetZ strip_cdataZrecoverÚencoding)Ú_default_parserr Ú XMLParser©r) r3 r r r Údefault_parsere s
|
ÿz$LXMLTreeBuilderForXML.default_parserc C s( | |¡}t|tr$|| dd|d}|S )z¥Instantiate an appropriate parser for the given encoding.
|
|
:param encoding: A string.
|
:return: A parser object such as an `etree.XMLParser`.
|
FTr1 )r7 Ú
|
isinstancer )r) r3 Úparserr r r Ú
|
parser_forq s
|
|
ÿz LXMLTreeBuilderForXML.parser_forNc sL || _ |d k rt|| _d | _| jg| _t| jg| _t t
|
| j f | d S r ) r4 ÚsetÚempty_element_tagsr* ÚDEFAULT_NSMAPS_INVERTEDÚnsmapsr r( Úactive_namespace_prefixesr% r Ú__init__)r) r9 r< Úkwargsr+ r r r@ s
|
|
z LXMLTreeBuilderForXML.__init__c C s0 |d dkr$t |dd
dd¡S d |fS d S )Nr Ú{é Ú})ÚtupleÚsplit)r) Útagr r r Ú _getNsTag s z LXMLTreeBuilderForXML._getNsTagc
|
c s¼ | j }|r t| _t |¡ nt| _t|tr`t|dkrR|d dkrR|dd
}|d|dfV t|tr~| d¡d|dfV |g}|g}t
|
|||||d}|j D ]} |j | |dfV q¢dS )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.
|
r u rC NFÚutf8)Úknown_definite_encodingsÚuser_encodingsÚis_htmlÚexclude_encodings) Úis_xmlr Ú processing_instruction_classr Z warn_if_markup_looks_like_xmlr r8 ÚstrÚlenÚencoder Ú encodingsÚmarkup)
|
r) rT Zuser_specified_encodingrM Zdocument_declared_encodingrL rJ rK Údetectorr3 r r r Úprepare_markup s2
|
|
|
ÿ ý
|
z$LXMLTreeBuilderForXML.prepare_markupc
|
C sÆ t |trt|}nt |tr&t|}| | j¡}z\| | jj ¡| _
|
| j
|
|¡ t |dkr| | j¡}t |dkrP| j
|
|¡ qP| j
|
¡ W n2 tttjfk
|
rÀ } z t|W 5 d }~X Y nX d S )Nr )r8 Úbytesr rP r ÚreadÚ
|
CHUNK_SIZEr: r* Úoriginal_encodingr9 ÚfeedrQ ÚcloseÚUnicodeDecodeErrorÚ LookupErrorr Ú ParserErrorr )r) rT ÚdataÚer r r r[ Û s
|
|
|
zLXMLTreeBuilderForXML.feedc C s | j g| _d S r )r= r> )r) r r r r\ ð s zLXMLTreeBuilderForXML.closec C sR t |}d }t|dkr4t| jdkr4| j d ¡ nt|dkr¾| |¡ | j t|¡ t | jd }| |¡ d|kr|d= | j |¡ | ¡ }t |
|
¡ D ] \}}t d|d}|||< q i } t |
|
¡ D ]F\}
|
} | |
|
¡\}}
|
|d krö| | |
|
< qÎ| |¡}t ||
|
|}
|
| | |
|
< qÎ| }| |¡\}}| |¡}| jj||||| jd d d S )Nr rC éÿÿÿÿÚ Úxmlnsz http://www.w3.org/2000/xmlns/)Ú
|
namespaces)r rQ r> Úappendr' r# r? ÚupdateÚcopyr r! r rH Ú_prefix_for_namespacer* Úhandle_starttag) r) ÚnameÚattrsZnsmapÚnsprefixZcurrent_mappingÚprefixÚ namespaceÚ attributeZ new_attrsÚattrr0 r r r Ústartó sL
|
|
ÿ
|
|
|
|
|
þzLXMLTreeBuilderForXML.startc C s< |dkr dS t | jD ] }|dk r||kr|| S qdS )z9Find the currently active prefix for the given namespace.N)Úreversedr> )r) ro Úinverted_nsmapr r r ri 2 s z+LXMLTreeBuilderForXML._prefix_for_namespacec C s | j ¡ | j jd }| |¡\}}d }|d k r\t| jD ] }|d k r:||kr:|| } q\q:| j ||¡ t| jdkr| j ¡ }|d k r| j ¡ d S )Nrb rC )
|
r* ÚendDataZtagStackrH rs r> Ú handle_endtagrQ Úpopr? )r) rk Z completed_tagro rm rt Zout_of_scope_nsmapr r r Úend; s
|
|
zLXMLTreeBuilderForXML.endc C s4 | j ¡ |d | }| j |¡ | j | j¡ d S )Nú )r* ru Ú handle_datarO )r) r2 r` r r r ÚpiQ s
|
zLXMLTreeBuilderForXML.pic C s | j |¡ d S r )r* rz ©r) Úcontentr r r r` W s zLXMLTreeBuilderForXML.datac C s( | j ¡ t |||¡}| j |¡ d S r )r* ru r
|
Zfor_name_and_idsZobject_was_parsed)r) rk ZpubidÚsystemÚdoctyper r r r Z s
|
z LXMLTreeBuilderForXML.doctypec C s& | j ¡ | j |¡ | j t¡ dS )z#Handle comments as Comment objects.N)r* ru rz r r| r r r Úcomment_ s
|
z LXMLTreeBuilderForXML.commentc C s d| S )úSee `TreeBuilder`.z)<?xml version="1.0" encoding="utf-8"?>
|
%sr ©r) Úfragmentr r r Útest_fragment_to_documente s z/LXMLTreeBuilderForXML.test_fragment_to_document)NN)NNN)'Ú__name__Ú
|
__module__Ú __qualname__r r5 ZDEFAULT_PARSER_CLASSrN r rO ÚNAMEÚALTERNATE_NAMESÚLXMLr r r ÚfeaturesrY r r( r# r= r& r' r7 r: r@ rH rV r[ r\ rr ri rx r{ r` r r r Ú __classcell__r r r+ r r ) s:
|
þ
|
F ? 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 Ú
|
HTMLParserr6 r r r r7 s 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* rZ r: r9 r[ r\ r] r^ r r_ r )r) rT r3 ra r r r r[ v s zLXMLTreeBuilder.feedc C s d| S )r z <html><body>%s</body></html>r r r r r r s z)LXMLTreeBuilder.test_fragment_to_documentN)r
r r r r r r r r r rN r rO r7 r[ r r r r r r j s
|
)!Ú __license__Ú__all__Úcollections.abcr Ú ImportErrorra Ú collectionsÚior r r r Z bs4.elementr r
|
r r r Z bs4.builderr r r r r r r r Z
|
bs4.dammitr r r# r r r r r r Ú<module> s& þ ( C
|
