Š¹¼]?E ã @ s d d l Z d d l Z d d l Z d d l m Z d g Z e j d ƒ Z e j d ƒ Z e j d ƒ Z e j d ƒ Z
e j d ƒ Z e j d ƒ Z e j d
ƒ Z
e j d ƒ Z e j d ƒ Z e j d
e j ƒ Z e j d ƒ Z e j d ƒ Z Gd d „ d e j ƒ Z d S)é N)ÚunescapeÚ
HTMLParserz[&<]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]ú>z--\s*>z$([a-zA-Z][^
/> ]*)(?:\s|/(?!>))*z]((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*aF
<[a-zA-Z][^\t\n\r\f />\x00]* # tag name
(?:[\s/]* # optional whitespace before attribute name
(?:(?<=['"\s/])[^\s/>][^\s/=>]* # attribute name
(?:\s*=+\s* # value indicator
(?:'[^']*' # LITA-enclosed value
|"[^"]*" # LIT-enclosed value
|(?!['"])[^>\s]* # bare value
)
(?:\s*,)* # possibly followed by a comma
)?(?:\s|/(?!>))*
)*
)?
\s* # trailing whitespace
z#c @ sQ e Z d Z d9 Z d d d d „ Z d d „ Z d d
„ Z d d „ Z d
Z d d „ Z d d „ Z
d d „ Z d d „ Z d d „ Z
d d d „ Z d d „ Z d d „ Z d d „ Z d! d" „ Z d# d$ „ Z d% d& „ Z d' d( „ Z d) d* „ Z d+ d, „ Z d- d. „ Z d/ d0 „ Z d1 d2 „ Z d3 d4 „ Z d5 d6 „ Z d7 d8 „ Z d
S):r ÚscriptÚstyleÚconvert_charrefsTc C s | | _ | j ƒ d S)N)r Úreset)Úselfr © r
ú+/opt/alt/python35/lib64/python3.5/parser.pyÚ__init__W s zHTMLParser.__init__c C s8 d | _ d | _ t | _ d | _ t j j | ƒ d S)NÚ z???)ÚrawdataÚlasttagÚinteresting_normalÚ interestingÚ
cdata_elemÚ _markupbaseÚ
ParserBaser )r r
r
r r ` s
zHTMLParser.resetc C s! | j | | _ | j d ƒ d S)Nr )r Úgoahead)r Údatar
r
r Úfeedh s zHTMLParser.feedc C s | j d ƒ d S)Né )r )r r
r
r Úcloseq s zHTMLParser.closeNc C s | j S)N)Ú_HTMLParser__starttag_text)r r
r
r Úget_starttag_textw s z HTMLParser.get_starttag_textc C s2 | j ƒ | _ t j d | j t j ƒ | _ d S)Nz )Úlowerr ÚreÚcompileÚIr )r Úelemr
r
r Úset_cdata_mode{ s zHTMLParser.set_cdata_modec C s t | _ d | _ d S)N)r r r )r r
r
r Úclear_cdata_mode s zHTMLParser.clear_cdata_modec C s© | j } d } t | ƒ } xü| | k r| j r¬ | j r¬ | j d | ƒ } | d k ræ | j d t | | d ƒ ƒ } | d k r£ t j d ƒ j | | ƒ r£ P| } n: | j
j | | ƒ } | rÖ | j ƒ } n | j rà P| } | | k r<| j r%| j r%| j t
| | | … ƒ ƒ n | j | | | … ƒ | j | | ƒ } | | k r[P| j } | d | ƒ rt j | | ƒ r—| j | ƒ } n¯ | d | ƒ r¸| j | ƒ } nŽ | d | ƒ rÙ| j | ƒ } nm | d | ƒ rú| j | ƒ } nL | d | ƒ r| j | ƒ } n+ | d
| k rE| j d ƒ | d
} n P| d k rþ| sYP| j d | d
ƒ } | d k rª| j d | d
ƒ } | d k r´| d
} n
| d
7} | j rç| j rç| j t
| | | … ƒ ƒ n | j | | | … ƒ | j | | ƒ } q | d | ƒ rét j | | ƒ } | rž| j ƒ d
d … }
| j |
ƒ | j ƒ } | d | d
ƒ s†| d
} | j | | ƒ } q qd | | d … k rå| j | | | d
… ƒ | j | | d
ƒ } Pq | d | ƒ r t j | | ƒ } | rj| j d
ƒ }
| j |
ƒ | j ƒ } | d | d
ƒ sU| d
} | j | | ƒ } q t j | | ƒ } | rÜ| rØ| j ƒ | | d … k rØ| j ƒ } | | k rÂ| } | j | | d
ƒ } Pq| d
| k r| j d ƒ | j | | d
ƒ } qPq q W| r’| | k r’| j r’| j ri| j ri| j t
| | | … ƒ ƒ n | j | | | … ƒ | j | | ƒ } | | d … | _ d S)Nr ú<ú&é" z[\s;]zÚ
r( r( r( )r rP )r Úcheck_for_whole_start_tagr Útagfind_tolerantr2 r; r9 r r Úattrfind_tolerantr ÚappendÚstripZgetposÚcountr) r+ r/ ÚendswithÚhandle_startendtagÚhandle_starttagÚCDATA_CONTENT_ELEMENTSr! )r r? Úendposr Úattrsr2 rB ÚtagÚmZattrnameÚrestZ attrvaluer; ÚlinenoÚoffsetr
r
r r3 - sP 00
zHTMLParser.parse_starttagc C sñ | j } t j | | ƒ } | rá | j ƒ } | | | d … } | d k rU | d S| d k r© | j d | ƒ r{ | d S| j d | ƒ r‘ d S| | k r¡ | S| d S| d k r¹ d
S| d k rÉ d S| | k rÙ | S| d St d ƒ ‚ d S) Nr r ú/z/>r& r
z6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZzwe should not get here!r( r( r( )r Úlocatestarttagend_tolerantr2 r; r0 ÚAssertionError)r r? r r_ rA Únextr
r
r rR ` s. z$HTMLParser.check_for_whole_start_tagc C ss | j } t j | | d ƒ } | s) d S| j ƒ } t j | | ƒ } | s| j d k rw | j | | | … ƒ | St j | | d ƒ } | s | | | d … d k rµ | d S| j | ƒ S| j
d ƒ j ƒ } | j d | j ƒ ƒ } | j
| ƒ | d S| j
d ƒ j ƒ } | j d k rR| | j k rR| j | | | … ƒ | S| j
| j ƒ ƒ | j ƒ | S)Nr r& rE zr r( )r Ú endendtagr- r; Ú
endtagfindr2 r r/ rS rH r9 r r* Ú
handle_endtagr" )r r? r r2 rI Z namematchZtagnamer r
r
r r4 ‚ s6
zHTMLParser.parse_endtagc C s! | j | | ƒ | j | ƒ d S)N)rZ ri )r r^ r] r
r
r rY ª s z HTMLParser.handle_startendtagc C s d S)Nr
)r r^ r] r
r
r rZ ¯ s zHTMLParser.handle_starttagc C s d S)Nr
)r r^ r
r
r ri ³ s zHTMLParser.handle_endtagc C s d S)Nr
)r rC r
r
r r: · s zHTMLParser.handle_charrefc C s d S)Nr
)r rC r
r
r r= » s zHTMLParser.handle_entityrefc C s d S)Nr
)r r r
r
r r/ ¿ s zHTMLParser.handle_datac C s d S)Nr
)r r r
r
r rJ Ã s zHTMLParser.handle_commentc C s d S)Nr
)r Zdeclr
r
r rG Ç s zHTMLParser.handle_declc C s d S)Nr
)r r r
r
r rM Ë s zHTMLParser.handle_pic C s d S)Nr
)r r r
r
r Ú unknown_declÎ s zHTMLParser.unknown_declc C s t j d t d d ƒt | ƒ S)NzZThe unescape method is deprecated and will be removed in 3.5, use html.unescape() instead.Ú
stacklevelr& )ÚwarningsÚwarnÚDeprecationWarningr )r Úsr
r
r r Ò s
zHTMLParser.unescape)r r ) Ú__name__Ú
__module__Ú __qualname__r[ r r r r r r r! r" r r7 rH r6 r3 rR r4 rY rZ ri r: r= r/ rJ rG rM rj r r
r
r
r r ? s6 z 3 " ( )r rl r Zhtmlr Ú__all__r r r> r< r8 r1 rL Z commentcloserS rT ÚVERBOSErd rg rh r r r
r
r
r Ú s&