B
ÍAÒ[:M ã @ s d Z ddlmZmZmZmZ ddlT ddlmZ ddl Z ddl
Z
e d¡Z e d¡Z
e d¡Ze d ¡Ze d
¡Ze d ¡Ze d ¡Ze d
¡Ze d¡Ze d¡Ze d¡Ze de j¡Ze de j¡Ze d ¡Ze d¡ZG dd„ de ƒZ G dd„ dej ƒZ dS )zLA parser for HTML and XHTML.
Backported for python-future from Python 3.3.
é )Úabsolute_importÚdivisionÚprint_functionÚunicode_literals)Ú*)Ú _markupbaseNz[&<]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][-.a-zA-Z0-9:_]*)(?:\s|/(?!>))*z[a-zA-Z][^
/> ]*zJ\s*([a-zA-Z_][-.:a-zA-Z_0-9]*)(\s*=\s*(\'[^\']*\'|"[^"]*"|[^\s"\'=<>`]*))?z]((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*aê
<[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
aF
<[a-zA-Z][-.a-zA-Z0-9:_]* # 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 @ s" e Zd ZdZddd„Zdd„ ZdS ) ÚHTMLParseErrorz&Exception raised for all parse errors.©NNc C s& |st ‚|| _|d | _|d | _d S )Nr é )ÚAssertionErrorÚmsgÚlinenoÚoffset)Úselfr
Úposition© r úM/opt/alt/python37/lib/python3.7/site-packages/future/backports/html/parser.pyÚ__init__U s
zHTMLParseError.__init__c C s> | j }| jd k r |d| j }| jd k r:|d| jd }|S )Nz , at line %dz , column %dr )r
r r )r Úresultr r r Ú__str__[ s
zHTMLParseError.__str__N)r
)Ú__name__Ú
__module__Ú __qualname__Ú__doc__r r r r r r r R s
r c @ sì e Zd ZdZdZd:dd„Zdd„ Zdd „ Zd
d „ Zd d
„ Z dZ
dd„ Z dd„ Z dd„ Z
dd„ Zdd„ Zd;dd„Zd d „ Zd d „ Zd d!„ Zd"d#„ Zd$d%„ Zd&d'„ Zd(d)„ Zd*d+„ Zd,d-„ Zd.d/„ Zd0d1„ Zd2d3„ Z d4d5„ Z d6d7„ Z d8d9„ Z dS )<Ú
HTMLParseraÇ Find tags and other markup and call handler functions.
Usage:
p = HTMLParser()
p.feed(data)
...
p.close()
Start tags are handled by calling self.handle_starttag() or
self.handle_startendtag(); end tags by self.handle_endtag(). The
data between tags is passed from the parser to the derived class
by calling self.handle_data() with the data as argument (the data
may be split up in arbitrary chunks). Entity references are
passed by calling self.handle_entityref() with the entity
reference as the argument. Numeric character references are
passed to self.handle_charref() with the string containing the
reference as the argument.
)ÚscriptÚstyleFc C s& |rt jdtdd || _| ¡ dS )zßInitialize and reset this instance.
If strict is set to False (the default) the parser will parse invalid
markup, otherwise it will raise an error. Note that the strict mode
is deprecated.
z The strict mode is deprecated.é )Ú
stacklevelN)ÚwarningsÚwarnÚDeprecationWarningÚstrictÚreset)r r# r r r r z s
zHTMLParser.__init__c C s( d| _ d| _t| _d| _tj | ¡ dS )z1Reset this instance. Loses all unprocessed data.Ú z???N)ÚrawdataÚlasttagÚinteresting_normalÚ interestingÚ
cdata_elemr Ú
ParserBaser$ )r r r r r$ ‡ s
zHTMLParser.resetc C s | j | | _ | d¡ dS )z‘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').
r N)r& Úgoahead)r Údatar r r Úfeed s zHTMLParser.feedc C s | d¡ dS )zHandle any buffered data.r N)r, )r r r r Úclose˜ s zHTMLParser.closec C s t || ¡ ƒ‚d S )N)r Úgetpos)r Úmessager r r Úerrorœ s zHTMLParser.errorNc C s | j S )z)Return full source of start tag: '<...>'.)Ú_HTMLParser__starttag_text)r r r r Úget_starttag_text¡ s z HTMLParser.get_starttag_textc C s$ | ¡ | _t d| j tj¡| _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}xF||k rZ| j ||¡}|r<| ¡ }n | jrDP |}||k rb| |||… ¡ | ||¡}||krxP |j}|d|ƒr¾t
||¡r¢| |¡}n’|d|ƒr¸| |¡}n||d|ƒrÎ|
|¡}nf|d|ƒrä| |¡}nP|d|ƒr| jr| |¡}n
| |¡}n$|d |k r2| d¡ |d }nP |dk r°|sFP | jrX| d¡ | d |d ¡}|dk r–| d|d ¡}|dk rž|d }n|d7 }| |||… ¡ | ||¡}q|d
|ƒrZt
||¡}|r&| ¡ d d … } | | ¡ | ¡ }|d
|d ƒs|d }| ||¡}qn2d
||d … krV| |dd … ¡ | |d ¡}P q|d|ƒrLt
||¡}|rº| d¡} | | ¡ | ¡ }|d
|d ƒs¬|d }| ||¡}qt
||¡}|r |r| ¡ ||d … kr| jrü| d¡ n ||kr
|}| ||d ¡}P n,|d |k rH| d¡ | ||d ¡}nP qdstdƒ‚qW |r’||k r’| js’| |||… ¡ | ||¡}||d … | _ d S )Nr údS |rX| ||d |… ¡ |d S )Nr )zÚ
z junk characters in start tag: %ré z/>)r3 Úcheck_for_whole_start_tagr& ÚtagfindrG r rQ rO r5 r' r# ÚattrfindÚattrfind_tolerantÚunescapeÚappendÚstripr0 Úcountr@ Úrfindr2 rC ÚendswithÚhandle_startendtagÚhandle_starttagÚCDATA_CONTENT_ELEMENTSr: )r rU Úendposr& ÚattrsrG rX ÚtagÚmÚattrnameÚrestZ attrvaluerQ r r r r r rH E s\
(,
zHTMLParser.parse_starttagc C s | j }| jrt ||¡}n t ||¡}|rú| ¡ }|||d … }|dkrR|d S |dkr²| d|¡rn|d S | d|¡r~dS | jrž| ||d ¡ | d¡ ||krª|S |d S |dkr¾dS |d krÊdS | jræ| ||¡ | d
¡ ||krò|S |d S t d ƒ‚d S ) Nr r ú/z/>r r= zmalformed empty start tagr% z6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZzmalformed start tagzwe should not get here!)
r& r# ÚlocatestarttagendrG Úlocatestarttagend_tolerantrQ rE rD r2 r )r rU r& ry rW Únextr r r ri ~ s>
z$HTMLParser.check_for_whole_start_tagc C sN | j }|||d … dks"tdƒ‚t ||d ¡}|s:dS | ¡ }t ||¡}|sø| jd k rr| |||… ¡ |S | j r|
d|||… f ¡ t ||d ¡}|sÊ|||d … dkrÀ|d S | |¡S |
¡ ¡ }| d | ¡ ¡}| |¡ |d S |
d¡ ¡ }| jd k r4|| jkr4| |||… ¡ |S | | ¡ ¡ | ¡ |S )
Nr zr )r& r Ú endendtagrA rQ Ú
endtagfindrG r* rC r# r2 Útagfind_tolerantr^ rO r5 rM Ú
handle_endtagr; )r rU r& rG r_ Z namematchZtagnamer9 r r r rI © s<
zHTMLParser.parse_endtagc C s | ||¡ | |¡ d S )N)rt rƒ )r rx rw r r r rs Ó s z HTMLParser.handle_startendtagc C s d S )Nr )r rx rw r r r rt Ø s zHTMLParser.handle_starttagc C s d S )Nr )r rx r r r rƒ Ü s zHTMLParser.handle_endtagc C s d S )Nr )r rY r r r rP à s zHTMLParser.handle_charrefc C s d S )Nr )r rY r r r rS ä s zHTMLParser.handle_entityrefc C s d S )Nr )r r- r r r rC è s zHTMLParser.handle_datac C s d S )Nr )r r- r r r r` ì s zHTMLParser.handle_commentc C s d S )Nr )r Zdeclr r r r] ð s zHTMLParser.handle_declc C s d S )Nr )r r- r r r rd ô s zHTMLParser.handle_pic C s | j r| d|f ¡ d S )Nzunknown declaration: %r)r# r2 )r r- r r r Ú unknown_decl÷ s zHTMLParser.unknown_declc C s" d|kr |S dd„ }t d||¡S )Nr? c S sü | ¡ d } yX| d dkrb| dd … } | d dkrLt| dd … d¡dƒ}nt| d¡ƒ}t|ƒS W n tk
r| d| S X ddlm} | |krš|| S | d¡r¬d | S xJtd
t | ƒƒD ]0}| d |… |kr¼|| d |… | |d … S q¼W d | S d S ) Nr ú#r )ÚxÚXr> é z&#)Úhtml5r? r )
ÚgroupsÚintÚrstripÚchrÚ
ValueErrorZ future.backports.html.entitiesr‰ rr Úranger@ )ÚsÚcr‰ r† r r r ÚreplaceEntitiesÿ s&
z,HTMLParser.unescape..replaceEntitiesz&&(#?[xX]?(?:[0-9a-fA-F]+;|\w{1,32};?)))r6 Úsub)r r r’ r r r rm ü s
zHTMLParser.unescape)F)r ) r r r r ru r r$ r. r/ r2 r3 r4 r: r; r, rL r^ rK rH ri rI rs rt rƒ rP rS rC r` r] rd r„ rm r r r r r d s:
h
9+*r ) r Ú
__future__r r r r Zfuture.builtinsZfuture.backportsr r6 r r7 r( rT rR rN rF rc Z commentcloserj r‚ rk rl ÚVERBOSEr} r~ r€ r Ú Exceptionr r+ r r r r r Ú s4