� ��fMc@sJ
dZd
dl
Z
d
dlZd
dlZejd�
Zejd�
Zejd�
Zejd�
Zejd�
Z ejd�
Z ejd �
Zejd �
Zejd�
Z ejd�
Zejd �
Zejdej�Zejdej�Zejd�
Zejd�
ZGdd�de�ZGdd�de
j�ZdS(uA parser for HTML and XHTML.iNu[&<]u &[a-zA-Z#]u%&([a-zA-Z][-.a-zA-Z0-9]*)[^a-zA-Z0-9]u)&#(?:[0-9]+|[xX][0-9a-fA-F]+)[^0-9a-fA-F]u <[a-zA-Z]u
>u--\s*>u(([a-zA-Z][-.a-zA-Z0-9:_]*)(?:\s|/(?!>))*u$([a-zA-Z][^  />]*)(?:\s|/(?!>))*uJ\s*([a-zA-Z_][-.:a-zA-Z_0-9]*)(\s*=\s*(\'[^\']*\'|"[^"]*"|[^\s"\'=<>`]*))?u]((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*u�
<[a-zA-Z][-.a-zA-Z0-9:_]* # tag name (?:\s+ # whitespace before attribute name (?:[a-zA-Z_][-.:a-zA-Z0-9_]* # attribute name (?:\s*=\s* # value indicator (?:'[^']*' # LITA-enclosed value |\"[^\"]*\" # LIT-enclosed value |[^'\">\s]+ # bare value ) )? ) )* \s* # trailing whitespace uF <[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 u#</\s*([a-zA-Z][-.a-zA-Z0-9:_]*)\s*>c

Bs5|EeZ
dZd
Zddd�
Zdd�ZdS(uHTMLParseErroru&Exception raised for all parse errors.cCs3|
st�
|
|_
|d
|_|d|_dS(Nii
(uAssertionErrorumsgulinenouoffset(uselfumsguposition((u0/opt/alt/python33/lib64/python3.3/html/parser.pyu__init__Ps



uHTMLParseError.__init__c
CsW|j}
|j
dk r,|
d
|j
}
n|jdk rS|
d|jd}
n|
S(Nu, at line %du, column %di
(umsgulinenouNoneuoffset(uselfuresult((u0/opt/alt/python33/lib64/python3.3/html/parser.pyu__str__Vs





uHTMLParseError.__str__N(NN(u__name__u __module__u__qualname__u__doc__uNoneu__init__u__str__(
u __locals__((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuHTMLParseErrorMs
uHTMLParseErrorc

Bsd
|EeZ
dZd
Zd:Zd;dd�
Zdd�Zdd �Zd d�Z dd �Z d9Zdd�Z dd�Zdd�Zdd�Zdd�Zddd�
Zdd�Zdd�Zdd �Zd!d"�Zd#d$�Zd%d&�Zd'd(�Zd)d*�Zd+d,�Zd-d.�Zd/d0�Zd1d2�Zd3d4�Zd5d6�Z d7d8�Z!d9S(<u HTMLParseru�Find tags and other markup and call handler functions. Usage: p = HTMLParser() p.feed(data) ... p.close() Start tags are handled by calling self.handle_starttag() or self.handle_startendtag(); end tags by self.handle_endtag(). The data between tags is passed from the parser to the derived class by calling self.handle_data() with the data as argument (the data may be split up in arbitrary chunks). Entity references are passed by calling self.handle_entityref() with the entity reference as the argument. Numeric character references are passed to self.handle_charref() with the string containing the reference as the argument. uscriptustylecCs6|
rtj
d
tdd�

n|
|_|j�
dS(u�Initialize and reset this instance. If strict is set to False (the default) the parser will parse invalid markup, otherwise it will raise an error. Note that the strict mode is deprecated. uThe strict mode is deprecated.u stackleveliN(uwarningsuwarnuDeprecationWarningustrictureset(uselfustrict((u0/opt/alt/python33/lib64/python3.3/html/parser.pyu__init__us 



uHTMLParser.__init__c

Cs8d
|_d|_
t|_d|_tjj|�

dS(u1Reset this instance. Loses all unprocessed data.uu???N( urawdataulasttaguinteresting_normaluinterestinguNoneu cdata_elemu_markupbaseu ParserBaseureset(
uself((u0/opt/alt/python33/lib64/python3.3/html/parser.pyureset�s 



uHTMLParser.resetcCs!|j|
|_|j
d
�

dS(u�Feed data to the parser. Call this as often as you want, with as little or as much text as you want (may include '\n'). iN(urawdataugoahead(uselfudata((u0/opt/alt/python33/lib64/python3.3/html/parser.pyufeed�s
uHTMLParser.feedc

Cs|jd
�

dS(uHandle any buffered data.i
N(
ugoahead(
uself((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuclose�suHTMLParser.closecCst|
|j
���
dS(
N(uHTMLParseErrorugetpos(uselfumessage((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuerror�s
uHTMLParser.errorc


Cs|jS(
u)Return full source of start tag: '<...>'.(
u_HTMLParser__starttag_text(
uself((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuget_starttag_text�suHTMLParser.get_starttag_textcCs2|
j�|_
tjd
|j
tj�|_dS(Nu</\s*%s\s*>(uloweru cdata_elemureucompileu
Iuinteresting(uselfuelem((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuset_cdata_mode�s

uHTMLParser.set_cdata_modec

Cst|_
d|_dS(
N(uinteresting_normaluinterestinguNoneu cdata_elem(
uself((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuclear_cdata_mode�s

uHTMLParser.clear_cdata_modec Cs|j}d
}t
|�
}x�||kr�|jj||�}|rT|j�}n|jraPn|}||kr�|j|||��

n|j||�}||kr�Pn|j}|d|�rqt j ||�r�|j|�
}n�|d|�r
|j|�
}n�|d|�r-
|j |�
}n�|d|�rN
|j|�
}ng|d|�r�
|jrx
|j|�
}q�
|j|�
}n+|d|kr�
|jd�

|d}n
P|d
kr\|
s�
Pn|jr�
|jd�

n|jd |d�}|d
kr8|jd|d�}|d
krB|d}qBn |d7}|j|||��

n|j||�}q|d |�rMtj ||�}|r�|j�dd�} |j| �

|j�}|d|d�s�|d}n|j||�}qq�d||d�krI|j|||d��

|j||d�}nPq|d |�r�tj ||�}|r�|jd�
} |j| �

|j�}|d|d�s�|d}n|j||�}qntj ||�}|re|
ra|j�||d�kra|jr'|jd�

qa|j�}||k
rH|}n|j||d�}nPq�|d|kr�|jd �

|j||d�}q�Pqd
std�
�
qW|
r�||kr�|jr�|j|||��

|j||�}n||d�|_dS(Niu
<u</u<!--u<?u<!i
uEOF in middle of constructu
>u&#iu
;u
&u#EOF in middle of entity or char refuinteresting.search() liedi����(urawdataulenuinterestingusearchustartu cdata_elemuhandle_datau updateposu startswithustarttagopenumatchuparse_starttaguparse_endtagu parse_commentuparse_piustrictuparse_declarationuparse_html_declarationuerrorufinducharrefugroupuhandle_charrefuendu entityrefuhandle_entityrefu incompleteuAssertionError( uselfuendurawdatau
iu
numatchu
ju startswithu
kuname((u0/opt/alt/python33/lib64/python3.3/html/parser.pyugoahead�s�

































































"









uHTMLParser.goaheadcCs�|j}||
|
d
�dks/t
d�
�
||
|
d�dkrV|j|
�
S||
|
d�dkr}|j|
�
S||
|
d�j�d kr�|jd |
d�}|dkr�d S|j||
d
|��

|dS|j|
�
SdS(Niu<!u+unexpected call to parse_html_declaration()iu<!--iu<![i u <!doctypeu
>i
i����i����(urawdatauAssertionErroru parse_commentuparse_marked_sectionulowerufinduhandle_decluparse_bogus_comment(uselfu
iurawdataugtpos((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuparse_html_declaration
s

&






u!HTMLParser.parse_html_declarationi
cCs�|j}||
|
d
�dks/t
d�
�
|jd|
d
�}|dkrUd S|ry|j||
d
|��

n|dS( Niu<!u</u"unexpected call to parse_comment()u
>i
(u<!u</i����i����(urawdatauAssertionErrorufinduhandle_comment(uselfu
iureporturawdataupos((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuparse_bogus_comment)
s

&




uHTMLParser.parse_bogus_commentcCs�|j}||
|
d
�dks/t
d�
�
tj||
d
�}|sOdS|j�}|j||
d
|��

|j�}|S(Niu<?uunexpected call to parse_pi()i
i����(urawdatauAssertionErrorupicloseusearchustartu handle_piuend(uselfu
iurawdataumatchu
j((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuparse_pi5
s

&






uHTMLParser.parse_picCsd|_
|j|
�
}|d
kr(|S|j}||
|�|_
g}|jrltj||
d�}ntj||
d�}|s�td�
�
|j �}|j d�
j�|_}x'
||kr�
|jr�t j||�}ntj||�}|s
Pn|j ddd�\} } }| s2
d}ns|dd�dko]
|d d�kn
s�
|dd�dko�
|dd�kn
r�
|dd�}n|r�
|j|�
}n|j| j�|f�

|j �}q�W|||�j�}|dkr�|j�\} }d |j
kra| |j
jd �
} t|j
�
|j
jd �
}n|t|j
�
}|jr�|jd|||�dd�f
�

n|j||
|��

|S|jd �
r�|j||�
n/|j||�
||jkr|j|�

n|S(Nii
u#unexpected call to parse_starttag()iiu
'u
"u
>u/>u
u junk characters in start tag: %rii����i����i����(u
>u/>(uNoneu_HTMLParser__starttag_textucheck_for_whole_start_tagurawdataustrictutagfindumatchutagfind_tolerantuAssertionErroruendugroupulowerulasttaguattrfinduattrfind_tolerantuunescapeuappendustripugetposucountulenurfinduerroruhandle_datauendswithuhandle_startendtaguhandle_starttaguCDATA_CONTENT_ELEMENTSuset_cdata_mode(uselfu
iuendposurawdatauattrsumatchu
kutagu
muattrnameurestu attrvalueuendulinenouoffset((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuparse_starttagA
s`



















0
0













"





uHTMLParser.parse_starttagcCsk
|j}|j
r'tj||
�}ntj||
�}|r[
|j�}|||d
�}|dkrs|d
S|dkr�|jd|�r�|dS|jd|�r�dS|j
r�|j|
|d
�
|jd�

n||
kr�|S|
d
Sn|dkr
dS|dkr
d S|j
r@
|j|
|�
|jd �

n||
krP
|S|
d
Snt d �
�
dS(Ni
u
>u
/u/>iumalformed empty start taguu6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZumalformed start taguwe should not get here!i����i����i����( urawdataustrictulocatestarttagendumatchulocatestarttagend_tolerantuendu startswithu updateposuerroruAssertionError(uselfu
iurawdatau
mu
junext((u0/opt/alt/python33/lib64/python3.3/html/parser.pyucheck_for_whole_start_tag}
s>























u$HTMLParser.check_for_whole_start_tagcCs�
|j}||
|
d
�dks/t
d�
�
tj||
d�}|sOd S|j�}tj||
�}|sW
|jdk r�|j ||
|��

|S|j r�|jd||
|�f
�

ntj||
d
�}|s
||
|
d�dkr
|
dS|j |
�
Sn|jd�
j�}|jd|j��}|j|�

|dS|jd�
j�}|jdk r�
||jkr�
|j ||
|��

|Sn|j|j��

|j�
|S( Niu</uunexpected call to parse_endtagi
ubad end tag: %riu</>u
>i����(urawdatauAssertionErroru endendtagusearchuendu endtagfindumatchu cdata_elemuNoneuhandle_dataustrictuerrorutagfind_tolerantuparse_bogus_commentugroupulowerufindu handle_endtaguclear_cdata_mode(uselfu
iurawdataumatchugtposu namematchutagnameuelem((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuparse_endtag�
s<

&










!










uHTMLParser.parse_endtagcCs!|j|
|�
|j
|
�

dS(
N(uhandle_starttagu handle_endtag(uselfutaguattrs((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuhandle_startendtag�
s

uHTMLParser.handle_startendtagc
CsdS(
N((uselfutaguattrs((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuhandle_starttag�
s
uHTMLParser.handle_starttagc
CsdS(
N((uselfutag((u0/opt/alt/python33/lib64/python3.3/html/parser.pyu handle_endtag�
s
uHTMLParser.handle_endtagc
CsdS(
N((uselfuname((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuhandle_charref�
s
uHTMLParser.handle_charrefc
CsdS(
N((uselfuname((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuhandle_entityref�
s
uHTMLParser.handle_entityrefc
CsdS(
N((uselfudata((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuhandle_data�
s
uHTMLParser.handle_datac
CsdS(
N((uselfudata((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuhandle_comment�
s
uHTMLParser.handle_commentc
CsdS(
N((uselfudecl((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuhandle_decl�
s
uHTMLParser.handle_declc
CsdS(
N((uselfudata((u0/opt/alt/python33/lib64/python3.3/html/parser.pyu handle_pi�
s
uHTMLParser.handle_picCs$|jr |j
d
|
f
�

ndS(Nuunknown declaration: %r(ustrictuerror(uselfudata((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuunknown_decl�
s

uHTMLParser.unknown_declcCs8d
|
kr|
Sdd�}tj
d||
dtj�
S(Nu
&c
 SsB
|j�d
}yx|d
dkr�|dd�}|d
dkrht
|dd�jd�
d�}
nt
|jd�
�
}
t|
�
SWntk r�


d|SYn�Xd
d lm}
||kr�||S|jd�
r�d |Sx[tdt |�
�D]<}|d|�|kr�||d|�||d�Sq�Wd |SdS( Niu
#i
u
xu
Xu
;iu&#(
uhtml5u
&i(u
xu
X( ugroupsuinturstripuchru ValueErroru html.entitiesuhtml5uendswithurangeulen(u
su
cuhtml5u
x((u0/opt/alt/python33/lib64/python3.3/html/parser.pyureplaceEntities�
s&





%









$u,HTMLParser.unescape.<locals>.replaceEntitiesu&&(#?[xX]?(?:[0-9a-fA-F]+;|\w{1,32};?))uflags(ureusubuASCII(uselfu
sureplaceEntities((u0/opt/alt/python33/lib64/python3.3/html/parser.pyuunescape�
s



uHTMLParser.unescapeN(uscriptustyleF("u__name__u __module__u__qualname__u__doc__uCDATA_CONTENT_ELEMENTSuFalseu__init__uresetufeeducloseuerroruNoneu_HTMLParser__starttag_textuget_starttag_textuset_cdata_modeuclear_cdata_modeugoaheaduparse_html_declarationuparse_bogus_commentuparse_piuparse_starttagucheck_for_whole_start_taguparse_endtaguhandle_startendtaguhandle_starttagu handle_endtaguhandle_charrefuhandle_entityrefuhandle_datauhandle_commentuhandle_declu handle_piuunknown_decluunescape(
u __locals__((u0/opt/alt/python33/lib64/python3.3/html/parser.pyu HTMLParser_s:  i<+*u HTMLParser(u__doc__u_markupbaseureuwarningsucompileuinteresting_normalu incompleteu entityrefucharrefustarttagopenupicloseucommentcloseutagfindutagfind_tolerantuattrfinduattrfind_tolerantuVERBOSEulocatestarttagendulocatestarttagend_tolerantu endendtagu endtagfindu ExceptionuHTMLParseErroru ParserBaseu HTMLParser(((u0/opt/alt/python33/lib64/python3.3/html/parser.pyu<module>
s0