o 0¦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 escapablecCst| ¡|_||_|jdkrt d¡|_dS|r*|js*t d|jtjtjB¡|_dSt 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_mode­s zHTMLParser.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 Cs|j}d}t|ƒ}||krU|jr;|js;| d|¡}|dkr:| dt||dƒ¡}|dkr8t d¡  ||¡s8n|}n|j   ||¡}|rI|  ¡}n|jrNn|}||kro|jrf|j rf|  t|||…ƒ¡n |  |||…¡| ||¡}||kr{nÚ|j}|d|ƒrt ||¡r| |¡} n@|d|ƒr›| |¡} n5|d|ƒr¦| |¡} n*|d|ƒr±| |¡} n|d |ƒr¼| |¡} n|d |ksÄ|rÎ|  d¡|d } nn…| dkr‰|sÙn|t ||¡ràn§|d|ƒr|d |krò|  d¡n•t ||¡rùnŽ| ||d d…¡n‚|d|ƒr0|}d D]} | | |d ¡r"|t| ƒ8}nq| ||d |…¡nW|d|ƒrF|jrF| ||dd…¡nA|||d… ¡dkr_| ||d d…¡n(|d |ƒrq| ||d d…¡n|d|ƒrƒ| ||d d…¡nt dƒ‚|} | || ¡}nÁ|d|ƒràt! ||¡}|rÃ| "¡d d…} | #| ¡| $¡} |d| d ƒs¼| d } | || ¡}q d||d…vrß|  |||d …¡| ||d ¡}nu|d|ƒrQt% ||¡}|r| "d ¡} | &| ¡| $¡} |d| d ƒs | d } | || ¡}q t' ||¡}|r;|r:| "¡||d…kr:| $¡} | |kr2|} | ||d ¡}n|d |krP|  d¡| ||d ¡}nn||ks|r€||kr€|jrq|j rq|  t|||…ƒ¡n |  |||…¡| ||¡}||d…|_dS)Nrú<ú&é"z [\t\n\r\f ;]zrr8r5r7rr1rzÚcommentabruptcloserCr?rJrS)r rWÚreportrrCrYr r r rFpszHTMLParser.parse_commentrcCsD|j}| d|d¡}|dkrdS|r| ||d|…¡|dS)Nrr1r8r)rr;rJ)r rWrarÚposr r r r^szHTMLParser.parse_bogus_commentcCsH|j}t ||d¡}|sdS| ¡}| ||d|…¡| ¡}|S)Nr1r8)rÚpicloser>r?rNrS©r rWrrCrYr r r rG‹szHTMLParser.parse_picCsd|_| |¡}|dkr|S|j}|||…|_g}t ||d¡}| ¡}| d¡ ¡|_}||krt  ||¡}|s=nS| ddd¡\} } } | sLd} n-| dd…dkr^| dd…kssn| dd…dkrq| dd…krynn| dd…} | rt | ƒ} |  |  ¡| f¡| ¡}||ks4|||…  ¡} | dvrÍ|  ¡\} }d |jvr»| |j d ¡} t|jƒ|j d ¡}n|t|jƒ}| |||…¡|S|  d ¡rÚ| ||¡|S| ||¡||jvsð|jrì|d ksð|d krù|j|d d|S||jvr|j|dd|S)Nrrr1r5ú'r8ú")rú/>Ú rgZnoscriptr"Fr T)rÚcheck_for_whole_start_tagrÚtagfind_tolerantrCrSrQr#rÚattrfind_tolerantrÚappendÚstripZgetposÚcountr:r<r@rKÚhandle_startendtagÚhandle_starttagÚCDATA_CONTENT_ELEMENTSrr)ÚRCDATA_CONTENT_ELEMENTS)r rWÚendposrÚattrsrCrZÚtagÚmÚattrnameÚrestZ attrvaluerSÚlinenoÚoffsetr r r rD—sd  &( ó   ÿ   ù ÿ þzHTMLParser.parse_starttagcCs6|j}t ||d¡}| ¡}||ddkrdS|S)Nrrr8)rÚ locatetagendrCrSrdr r r riÐs z$HTMLParser.check_for_whole_start_tagcCs´|j}| d|d¡dkrdSt ||¡s*||d|d…dkr%|dS| |¡St ||d¡}| ¡}||ddkr@dSt ||d¡}| d¡  ¡}|  |¡|  ¡|S)Nrr1rr8r5r) rr;rIrCr^r{rSrjrQr#Ú handle_endtagr*)r rWrrCrYrur r r rEÚs    zHTMLParser.parse_endtagcCs| ||¡| |¡dS©N)rpr|©r rurtr r r roøs zHTMLParser.handle_startendtagcCódSr}r r~r r r rpýózHTMLParser.handle_starttagcCrr}r )r rur r r r|r€zHTMLParser.handle_endtagcCrr}r ©r r\r r r rRr€zHTMLParser.handle_charrefcCrr}r rr r r rU r€zHTMLParser.handle_entityrefcCrr}r rr r r r@ r€zHTMLParser.handle_datacCrr}r rr r r rJr€zHTMLParser.handle_commentcCrr}r )r Zdeclr r r rMr€zHTMLParser.handle_declcCrr}r rr r r rNr€zHTMLParser.handle_picCrr}r rr r r rLr€zHTMLParser.unknown_decl)T)r)!Ú__name__Ú __module__Ú __qualname__Ú__doc__rqrrr rrrrrr)r*r,rrHrFr^rGrDrirErorpr|rRrUr@rJrMrNrLr r r r rZs@    "   9  )r…r$rZhtmlrÚ__all__r%rrVrTrPrBrIrcr_r`rjÚVERBOSErkr{Zlocatestarttagend_tolerantZ endendtagZ endtagfindrrr r r r Ús6             õ  óò