3
VW; @ s d Z dZdZddlZddlZejd dk rXejed ZeefZ ej
dd Z dd Z n eZ d
d Z dd l
T dd lmZmZ dd
lmZmZmZmZmZmZmZmZmZmZ y ddl Z W n( e k
r e dej d Y nX G dd dej Z G dd dej!Z"dd Z#dS )a A SAX2 driver for libxml2, on top of it's XmlReader API
USAGE
# put this file (drv_libxml2.py) in PYTHONPATH
import xml.sax
reader = xml.sax.make_parser(["drv_libxml2"])
# ...and the rest is standard python sax.
CAVEATS
- Lexical handlers are supported, except for start/endEntity
(waiting for XmlReader.ResolveEntity) and start/endDTD
- Error callbacks are not exactly synchronous, they tend
to be invoked before the corresponding content callback,
because the underlying reader interface parses
data by chunks of 512 bytes
TODO
- search for TODO
- some ErrorHandler events (warning)
- some ContentHandler events (setDocumentLocator, skippedEntity)
- EntityResolver (using libxml2.?)
- DTDHandler (if/when libxml2 exposes such node types)
- DeclHandler (if/when libxml2 exposes such node types)
- property_xml_string?
- feature_string_interning?
- Incremental parser
- additional performance tuning:
- one might cache callbacks to avoid some name lookups
- one might implement a smarter way to pass attributes to startElement
(some kind of lazy evaluation?)
- there might be room for improvement in start/endPrefixMapping
- other?
u Stéphane Bidoul <sbi@skynet.be>z0.3 N utf8 c C s | d kr | S t | d S d S )Nr )_decoder)s r #/usr/lib64/python3.6/drv_libxml2.py_d1 s r c C s | S )Nr )r r r r r 9 s )*) xmlreadersaxutils)
feature_namespacesfeature_namespace_prefixesfeature_string_interningfeature_validationfeature_external_gesfeature_external_pesproperty_lexical_handler property_declaration_handlerproperty_dom_nodeproperty_xml_stringz+libxml2 not available: import error was: %sc @ s8 e Zd ZdZdd Zdd Zdd Zdd Zd
d Zd S )
Locatorz4SAX Locator adapter for libxml2.xmlTextReaderLocatorc C s
|| _ d S )N)_Locator__locator)selflocatorr r r __init__S s zLocator.__init__c C s dS )z6Return the column number where the current event ends.r r )r r r r getColumnNumberV s zLocator.getColumnNumberc C s
| j j S )z4Return the line number where the current event ends.)r Z
LineNumber)r r r r
getLineNumberZ s zLocator.getLineNumberc C s dS )z3Return the public identifier for the current event.Nr )r r r r getPublicId^ s zLocator.getPublicIdc C s
| j j S )z3Return the system identifier for the current event.)r ZBaseURI)r r r r getSystemIdb s zLocator.getSystemIdN) __name__
__module__ __qualname____doc__r r r r r r r r r r P s r c @ s\ e Zd Zdd Zdd Zdd Zdd Zd d
Zd d Zd
d Z dd Z
dd Z dd Z dS )
LibXml2Readerc C s@ t jj| d| _d| _d| _d| _d| _d | _d | _ d | _
d S )Nr r ) r XMLReaderr _LibXml2Reader__ns_LibXml2Reader__nspfx_LibXml2Reader__validate_LibXml2Reader__extparams_LibXml2Reader__parsing_LibXml2Reader__lex_handler _LibXml2Reader__decl_handler_LibXml2Reader__errors)r r r r r h s zLibXml2Reader.__init__c C s0 | j d krg | _ | j j|t|d t|f d S )N)r. appendZSAXParseExceptionr )r argmsgseverityr r r r
_errorHandlerw s
zLibXml2Reader._errorHandlerc C sl x`| j D ]V\}}|tjtjfkr.| jj| q|rR|| j d d krR| jj| q| jj| qW d | _ d S )Nr r )r. libxml2Z PARSER_SEVERITY_VALIDITY_WARNINGZPARSER_SEVERITY_WARNING _err_handlerZwarning
fatalErrorerror)r Zfatalr2 Z exceptionr r r
_reportErrors~ s
zLibXml2Reader._reportErrorsc C sJ d| _ z4t|tr tj|}n&tj|}tj|j }|j |j
}|j | j d | j
r|jtjd |jtjd |jtjd |jtj| j n|jtjd | jrtji i }n
tji }g }| jj xJ|j }|dkr| jd k r| j d nN|dkr | jd k r| j d P n*| jd k r6| j d n| j j t d P |j }|dkr | jrt!|j" t!|j# f} t!|j$ }
i |_%} i |_&} g }
x|j' r>t!|j$ }t!|j( }|j)drt*|dkr|dd }nd }|
j+| | jj,|| | j-sqt!|j" t!|j# f}|| |< || |< qW |j. | jj/| |
| |j0 r| jj1| |
x&|
D ]}| jj2| qvW n
|j+|
npt!|j$ } i |_%} x,|j' rt!|j$ }t!|j( | |< qW |j. | jj3| | |j0 r| jj4| q|dkr| jrj| jj1t!|j" t!|j# ft!|j$ x4|j5 D ]}| jj2| qRW n| jj4t!|j$ q|dkr| jj6t!|j( q|d kr| jj7t!|j( q|d
kr| jj6t!|j( q|d kr,| j8d k r | j8j9 | jj6t!|j( | j8d k r| j8j: q|dkr^| j8d k rT| j;t!|j$ |j< q|d kr| j8d k r| j=t!|j$ q|d
kr| jj>t!|j$ t!|j( q|dkr| j8d k r| j8j?t!|j( q|dkrq|dkrq|dkrq|dkrqt d| qW |dkr2| jj@ |jA W d d| _ X d S )Nr r z"Read failed (no details available)Zxmlns r
zUnexpected node type %d)Br+
isinstance StringTypesr4 ZnewTextReaderFilenamer Zprepare_input_sourceZ inputBufferZ
getByteStreamZ
newTextReaderr ZSetErrorHandlerr3 r* Z
SetParserPropZPARSER_LOADDTDZPARSER_DEFAULTATTRSZPARSER_SUBST_ENTITIESZPARSER_VALIDATEr) r' r ZAttributesNSImplZAttributesImplZ
_cont_handlerZ
startDocumentZReadr. r8 r5 r6 Z SAXExceptionZNodeTyper Z NamespaceUriZ LocalNameNameZ_attrsZ_qnamesZMoveToNextAttributeZValue
startswithlenr/ ZstartPrefixMappingr( Z
MoveToElementZstartElementNSZIsEmptyElementZ endElementNSZendPrefixMappingZ startElementZ
endElementpopZ
charactersZignorableWhitespacer, Z
startCDATAZendCDATAZ startEntityZ
ResolveEntityZ endEntityZprocessingInstructionZcommentZ endDocumentZClose)r sourcereaderinputZattributesNSImplZattributesImplprefixesrZnodeTypeZeltNameZeltQNameZattrsZqnamesZ newPrefixesZqnamevalueZ newPrefixZattNameprefixr r r parse s
zLibXml2Reader.parsec C s t dd S )NzDTDHandler not supported)SAXNotSupportedException)r Zhandlerr r r
setDTDHandler8 s zLibXml2Reader.setDTDHandlerc C s t dd S )Nz EntityResolver not supported)rS )r Zresolverr r r setEntityResolver< s z LibXml2Reader.setEntityResolverc C sT |t kr| jS |tkr | jS |tkr*| jS |tkr6dS |tkrD| jS t d| d S )Nr zFeature '%s' not recognized)
r
r' r r( r r) r r r* SAXNotRecognizedException)r namer r r
getFeature@ s zLibXml2Reader.getFeaturec C s | j rtd| |tkr"|| _nZ|tkr2|| _nJ|tkrB|| _n:|tkr`|dkr|td| n |t krp|| _
n t d| d S )Nz#Cannot set feature %s while parsingr zFeature '%s' not supportedzFeature '%s' not recognized) r+ rS r
r' r r( r r) r r r* rV )r rW stater r r
setFeatureO s"
zLibXml2Reader.setFeaturec C s, |t kr| jS |tkr | jS td| d S )Nz Property '%s' not recognized)r r, r r- rV )r rW r r r getPropertyd s zLibXml2Reader.getPropertyc C s< |t kr|| _n(|tkr,td| || _n td| d S )NzProperty '%s' not supportedz Property '%s' not recognized)r r, r rS r- rV )r rW rP r r r setPropertym s zLibXml2Reader.setPropertyN)
r! r" r# r r3 r8 rR rT rU rX rZ r[ r\ r r r r r% f s , r% c C s t S )N)r% r r r r
create_parsery s r] )$r$
__author__ __version__syscodecs version_infounicode_escape_decodestrZunicoderF lookupr r Zxml.sax._exceptionsZxml.saxr r Zxml.sax.handlerr
r r r r r r r r r r4 ImportErrorZSAXReaderNotAvailableexc_infor r& r% r] r r r r <module># s0
0
All system for education purposes only. For more tools: Telegram @jackleet