a ´iÌRã@sÞdZddlZddlZddlmZdgZe d¡Ze d¡Ze d¡Z e d¡Z e d ¡Z e d ¡Z e d ¡Z e d ¡Ze d ¡Ze d¡Ze dej¡Ze dej¡Ze dej¡Ze d ¡Ze d¡ZGdd„dejƒZdS)zA parser for HTML and XHTML.éN)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]z z--!?>z-?>z0([a-zA-Z][^\t\n\r\f />]*)(?:[\t\n\r\f ]|/(?!>))*a{ ( (?<=['"\t\n\r\f /])[^\t\n\r\f />][^\t\n\r\f /=>]* # attribute name ) ([\t\n\r\f ]*=[\t\n\r\f ]* # value indicator ('[^']*' # LITA-enclosed value |"[^"]*" # LIT-enclosed value |(?!['"])[^>\t\n\r\f ]* # bare value ) )? (?:[\t\n\r\f ]|/(?!>))* # possibly followed by a space a [a-zA-Z][^\t\n\r\f />]* # tag name [\t\n\r\f /]* # optional whitespace before attribute name (?:(?<=['"\t\n\r\f /])[^\t\n\r\f />][^\t\n\r\f /=>]* # attribute name (?:[\t\n\r\f ]*=[\t\n\r\f ]* # value indicator (?:'[^']*' # LITA-enclosed value |"[^"]*" # LIT-enclosed value |(?!['"])[^>\t\n\r\f ]* # bare value ) )? [\t\n\r\f /]* # possibly followed by a space )* >? 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 space )?(?:\s|/(?!>))* )* )? \s* # trailing whitespace z#c@seZdZdZdZdZdddœdd„Zd d „Zd d „Zd d„Z dZ dd„Z ddœdd„Z dd„Z d>dd„Zdd„Zdd„Zd?dd„Zd@d d!„Zd"d#„Zd$d%„Zd&d'„Zd(d)„Zd*d+„Zd,d-„Zd.d/„Zd0d1„Zd2d3„Zd4d5„Zd6d7„Zd8d9„Zd:d;„Zd'.)Ú_HTMLParser__starttag_textrr r r Úget_starttag_textszHTMLParser.get_starttag_text©Ú escapablecCsp| ¡|_||_|jdkr(t d¡|_nD|rP|jsPt d|jtjtjB¡|_nt d|jtjtjB¡|_dS)NÚ plaintextz\Zz&|])z])) ÚlowerrrÚreÚcompilerrÚ IGNORECASEÚASCII)r Úelemr"r r r Úset_cdata_mode¡s     ÿ  ÿzHTMLParser.set_cdata_modecCst|_d|_d|_dS)NT)rrrrrr r r Úclear_cdata_mode­szHTMLParser.clear_cdata_modecCs ||_dS)aEnable or disable support of the CDATA sections. If enabled, "<[CDATA[" starts a CDATA section which ends with "]]>". If disabled, "<[CDATA[" starts a bogus comments which ends with ">". This method is not called by default. Its purpose is to be called in custom handle_starttag() and handle_endtag() methods, with value that depends on the adjusted current node. See https://html.spec.whatwg.org/multipage/parsing.html#markup-declaration-open-state for details. N)r)r Úflagr r r Ú_set_support_cdata²s zHTMLParser._set_support_cdatac Cs4|j}d}t|ƒ}||krÌ|jrv|jsv| d|¡}|dkr | dt||dƒ¡}|dkrpt d¡  ||¡spqÌ|}n*|j   ||¡}|r’|  ¡}n|jrœqÌ|}||krÞ|jrÌ|j rÌ|  t|||…ƒ¡n|  |||…¡| ||¡}||kröqÌ|j}|d|ƒrrr9r6r8rr2rzÚ rhZnoscriptr#Fr!T)rÚcheck_for_whole_start_tagrÚtagfind_tolerantrDrTrRr$rÚattrfind_tolerantrÚappendÚstripZgetposÚcountr;r=rArLÚhandle_startendtagÚhandle_starttagÚCDATA_CONTENT_ELEMENTSrr*ÚRCDATA_CONTENT_ELEMENTS)r rXÚendposrÚattrsrDr[ÚtagÚmÚattrnameÚrestZ attrvaluerTÚlinenoÚoffsetr r r rE—sh   & ÿ ÿ       ÿ   ÿÿþ zHTMLParser.parse_starttagcCs6|j}t ||d¡}| ¡}||ddkr2dS|S)Nrrr9)rÚ locatetagendrDrTrer r r rjÐs z$HTMLParser.check_for_whole_start_tagcCs´|j}| d|d¡dkrdSt ||¡sT||d|d…dkrJ|dS| |¡St ||d¡}| ¡}||ddkr€dSt ||d¡}| d¡  ¡}|  |¡|  ¡|S)Nrr2rr9r6r) rr<rJrDr_r|rTrkrRr$Ú handle_endtagr+)r rXrrDrZrvr r r rFÚs    zHTMLParser.parse_endtagcCs| ||¡| |¡dS©N)rqr}©r rvrur r r rpøs zHTMLParser.handle_startendtagcCsdSr~r rr r r rqýszHTMLParser.handle_starttagcCsdSr~r )r rvr r r r}szHTMLParser.handle_endtagcCsdSr~r ©r r]r r r rSszHTMLParser.handle_charrefcCsdSr~r r€r r r rV szHTMLParser.handle_entityrefcCsdSr~r rr r r rA szHTMLParser.handle_datacCsdSr~r rr r r rKszHTMLParser.handle_commentcCsdSr~r )r Zdeclr r r rNszHTMLParser.handle_declcCsdSr~r rr r r rOszHTMLParser.handle_picCsdSr~r rr r r rMszHTMLParser.unknown_decl)T)T)r)!Ú__name__Ú __module__Ú __qualname__Ú__doc__rrrsr r rrrr r*r+r-rrIrGr_rHrErjrFrprqr}rSrVrArKrNrOrMr r r r rZs>     "   9 )r„r%rZhtmlrÚ__all__r&rrWrUrQrCrJrdr`rarkÚVERBOSErlr|Zlocatestarttagend_tolerantZ endendtagZ endtagfindrrr r r r Ús4            õ  óò