ó
›çXjc @ s d Z d d l Z d d l 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 ƒ Z
e j d
e j ƒ Z e j d ƒ Z e j d ƒ Z d e f d „ ƒ YZ d e j f d „ ƒ YZ d S( s A parser for HTML and XHTML.iÿÿÿÿNs [&<]s
&[a-zA-Z#]s% &([a-zA-Z][-.a-zA-Z0-9]*)[^a-zA-Z0-9]s) &#(?:[0-9]+|[xX][0-9a-fA-F]+)[^0-9a-fA-F]s <[a-zA-Z]t >s --\s*>s( ([a-zA-Z][-.a-zA-Z0-9:_]*)(?:\s|/(?!>))*s [a-zA-Z][^
/> ]*s] ((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*s
<[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|/(?!>))*
)*
)?
\s* # trailing whitespace
s# t HTMLParseErrorc B s# e Z d Z d d „ Z d „ Z RS( s& Exception raised for all parse errors.c C s3 | s t ‚ | | _ | d | _ | d | _ d S( Ni i ( t AssertionErrort msgt linenot offset( t selfR t position( ( s" /usr/lib64/python2.7/HTMLParser.pyt __init__9 s
c C sW | j } | j d k r, | d | j } n | j d k rS | d | j d } n | S( Ns , at line %ds , column %di ( R R t NoneR ( R t result( ( s" /usr/lib64/python2.7/HTMLParser.pyt __str__? s N( NN( t __name__t
__module__t __doc__R R R ( ( ( s" /usr/lib64/python2.7/HTMLParser.pyR 6 s t
HTMLParserc B s
e Z d Z d Z d „ Z d „ Z d „ Z d „ Z d „ Z d Z
d „ Z d „ Z d
„ Z
d „ Z d „ Z d
d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d „ Z d Z d „ Z RS( sÇ 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.
t scriptt stylec C s | j ƒ d S( s# Initialize and reset this instance.N( t reset( R ( ( s" /usr/lib64/python2.7/HTMLParser.pyR _ s c C sS d | _ d | _ t | _ d | _ g | _ d | _ d | _ t j
j | ƒ d S( s1 Reset this instance. Loses all unprocessed data.t s ???i i N( t rawdatat lasttagt interesting_normalt interestingR t
cdata_elemt _pendingt _pending_lent _parse_thresholdt
markupbaset
ParserBaseR ( R ( ( s" /usr/lib64/python2.7/HTMLParser.pyR c s c C sã | j t | ƒ 7_ | j | j k r: | j j | ƒ n¥ | j sU | j | 7_ n2 | j j | ƒ | j d j | j ƒ 7_ | j 2d | _ t | j ƒ } | j d ƒ t | j ƒ | k rÍ d | _ n t | j ƒ | _ d S( s‘ 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 i i N( R t lenR R t appendR t joint goahead( R t datat n( ( s" /usr/lib64/python2.7/HTMLParser.pyt feedn s
c C sH | j r7 | j d j | j ƒ 7_ | j 2d | _ n | j d ƒ d S( s Handle any buffered data.R i i N( R R R R R! ( R ( ( s" /usr/lib64/python2.7/HTMLParser.pyt closeŒ s
c C s t | | j ƒ ƒ ‚ d S( N( R t getpos( R t message( ( s" /usr/lib64/python2.7/HTMLParser.pyt error” s c C s | j S( s) Return full source of start tag: '<...>'.( t _HTMLParser__starttag_text( R ( ( s" /usr/lib64/python2.7/HTMLParser.pyt get_starttag_text™ s c C s2 | j ƒ | _ t j d | j t j ƒ | _ d S( Ns ( t lowerR t ret compilet IR ( R t elem( ( s" /usr/lib64/python2.7/HTMLParser.pyt set_cdata_mode s c C s t | _ d | _ d S( N( R R R R ( R ( ( s" /usr/lib64/python2.7/HTMLParser.pyt clear_cdata_mode¡ s c
C st | j } d } t | ƒ } x | | k r | j j | | ƒ } | rT | j ƒ } n | j ra Pn | } | | k rŠ | j | | | !ƒ n | j | | ƒ } | | k r¬ Pn | j } | d | ƒ r7t j
| | ƒ rè | j | ƒ } n¯ | d | ƒ r | j | ƒ } nŽ | d | ƒ r*| j
| ƒ } nm | d | ƒ rK| j | ƒ } nL | d | ƒ rl| j | ƒ } n+ | d | k r–| j d ƒ | d } n P| d k r"| sPn | j d | d ƒ } | d k r| j d | d ƒ } | d k r | d } q n
| d 7} | j | | | !ƒ n | j | | ƒ } q | d | ƒ rÿt j
| | ƒ } | rÂ| j ƒ d
d !} | j | ƒ | j ƒ } | d | d ƒ sª| d } n | j | | ƒ } q qd | | k rû| j | d d
!ƒ | j | d
ƒ } n Pq | d
| ƒ rt j
| | ƒ } | r†| j d ƒ } | j | ƒ | j ƒ } | d | d ƒ sn| d } n | j | | ƒ } q n t j
| | ƒ } | rÎ| rÊ| j ƒ | | k rÊ| j d ƒ n Pq| d | k r| j d
ƒ | j | | d ƒ } qPq d s t d ƒ ‚ q W| rc| | k rc| j rc| j | | | !ƒ | j | | ƒ } n | | | _ d S( Ni t t parse_marked_sectionR+ RA t handle_declt parse_bogus_comment( R RI R t gtpos( ( s" /usr/lib64/python2.7/HTMLParser.pyR@ s
i c C s | j } | | | d !d k r0 | j d ƒ n | j d | d ƒ } | d k rV d S| rw | j | | d | !ƒ n | d S( Ni s s
( R s />( R R) t check_for_whole_start_tagR t tagfindR; R RE RC R+ R t attrfindt unescapeR t stripR& t countR t rfindR7 t endswitht handle_startendtagt handle_starttagt CDATA_CONTENT_ELEMENTSR0 ( R RI t endposR t attrsR; RK t tagt mt attrnamet restt attrvalueRE R R ( ( s" /usr/lib64/python2.7/HTMLParser.pyR< 3 sR $$ c C sý | j } t j | | ƒ } | rí | j ƒ } | | | d !} | d k rR | d S| d k r² | j d | ƒ rx | d S| j d | ƒ rŽ d S| j | | d ƒ | j d ƒ n | d k r d S| d k rÒ d S| | k râ | S| d Sn t d
ƒ ‚ d S( Ni R t /s />i iÿÿÿÿs malformed empty start tagR s6 abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZs we should not get here!( R t locatestarttagendR; RE R9 R8 R( R ( R RI R Re RJ t next( ( s" /usr/lib64/python2.7/HTMLParser.pyRW g s, c C sŠ | j } | | | d !d k s, t d ƒ ‚ t j | | d ƒ } | sL d S| j ƒ } t j | | ƒ } | s!| j d k r— | j | | | !ƒ | St
j | | d ƒ } | sâ | | | d !d k rÒ | d S| j | ƒ Sn | j ƒ j
ƒ } | j d | j ƒ ƒ } | j | ƒ | d S| j d ƒ j
ƒ } | j d k ro| | j k ro| j | | | !ƒ | Sn | j | ƒ | j ƒ | S( Ni s R ( R R t endendtagR5 RE t
endtagfindR; R R R7 t tagfind_tolerantRO RC R+ RA t
handle_endtagR1 ( R RI R R; RP t namematcht tagnameR/ ( ( s" /usr/lib64/python2.7/HTMLParser.pyR= ‡ s8 #
c C s! | j | | ƒ | j | ƒ d S( N( R` Ro ( R Rd Rc ( ( s" /usr/lib64/python2.7/HTMLParser.pyR_ ¯ s c C s d S( N( ( R Rd Rc ( ( s" /usr/lib64/python2.7/HTMLParser.pyR` ´ s c C s d S( N( ( R Rd ( ( s" /usr/lib64/python2.7/HTMLParser.pyRo ¸ s c C s d S( N( ( R RL ( ( s" /usr/lib64/python2.7/HTMLParser.pyRD ¼ s c C s d S( N( ( R RL ( ( s" /usr/lib64/python2.7/HTMLParser.pyRG À s c C s d S( N( ( R R" ( ( s" /usr/lib64/python2.7/HTMLParser.pyR7 Ä s c C s d S( N( ( R R" ( ( s" /usr/lib64/python2.7/HTMLParser.pyRQ È s c C s d S( N( ( R t decl( ( s" /usr/lib64/python2.7/HTMLParser.pyRN Ì s c C s d S( N( ( R R" ( ( s" /usr/lib64/python2.7/HTMLParser.pyRU Ð s c C s d S( N( ( R R" ( ( s" /usr/lib64/python2.7/HTMLParser.pyt unknown_declÓ s c s2 d | k r | S‡ f d † } t j d | | ƒ S( NR4 c s | j ƒ d } yZ | d d k ri | d } | d d
k rS t | d d ƒ } n t | ƒ } t | ƒ SWn t k
r† d | d SXd d l } t j d k ré i d
d 6} t _ x0 | j j ƒ D] \ } } t | ƒ | | s&