ó
Q˜[c @` sz d Z d d l m Z m Z m Z m Z d d l Td d l m 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 ƒ Z e j d e j 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( uL A parser for HTML and XHTML.
Backported for python-future from Python 3.3.
i ( t absolute_importt divisiont print_functiont unicode_literals( t *( t _markupbaseNu [&<]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][^
/> ]*uJ \s*([a-zA-Z_][-.:a-zA-Z_0-9]*)(\s*=\s*(\'[^\']*\'|"[^"]*"|[^\s"\'=<>`]*))?u] ((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*uê
<[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][-.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
u# t HTMLParseErrorc B` s# e Z d Z d d „ Z d „ Z RS( u& 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; lib/python2.7/site-packages/future/backports/html/parser.pyt __init__U s
c C` sW | j } | j d k r, | d | j } n | j d k rS | d | j d } n | S( Nu , at line %du , column %di ( R R t NoneR
( R t result( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt __str__[ s N( NN( t __name__t
__module__t __doc__R R
R ( ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR R s t
HTMLParserc B` s
e Z d Z d Z e 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 RS( uÇ 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.
u scriptu stylec C` s6 | r t j d t d d ƒn | | _ | j ƒ d S( uß 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.
u The strict mode is deprecated.t
stackleveli N( t warningst warnt DeprecationWarningt strictt reset( R R ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR
z s
c C` s8 d | _ d | _ t | _ d | _ t j j | ƒ d S( u1 Reset this instance. Loses all unprocessed data.u u ???N( t rawdatat lasttagt interesting_normalt interestingR t
cdata_elemR t
ParserBaseR ( R ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR ‡ s
c C` s! | j | | _ | j d ƒ d S( 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').
i N( R t goahead( R t data( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt feed s c C` s | j d ƒ d S( u Handle any buffered data.i N( R! ( R ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt close˜ s c C` s t | | j ƒ ƒ ‚ d S( N( R t getpos( R t message( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt errorœ s c C` s | j S( u) Return full source of start tag: '<...>'.( t _HTMLParser__starttag_text( R ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt get_starttag_text¡ s c C` s2 | j ƒ | _ t j d | j t j ƒ | _ d S( Nu ( t lowerR t ret compilet IR ( R t elem( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt set_cdata_mode¥ s c C` s t | _ d | _ d S( N( R R R R ( R ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt clear_cdata_mode© s c
C` sß | j } d } t | ƒ } xk| | k rˆ| j j | | ƒ } | rT | j ƒ } n | j ra Pn | } | | k rŠ | j | | | !ƒ n | j | | ƒ } | | k r¬ Pn | j } | d | ƒ rkt j
| | ƒ rè | j | ƒ } nÊ | d | ƒ r | j | ƒ } n© | d | ƒ r*| j
| ƒ } nˆ | d | ƒ rK| j | ƒ } ng | d | ƒ r‡| j ru| j | ƒ } q²| j | ƒ } n+ | d | k r±| j d ƒ | d } n P| d k rV| sÈPn | j rá| j d ƒ n | j d | d ƒ } | d k r5| j d | d ƒ } | d k r?| d } q?n
| d 7} | j | | | !ƒ n | j | | ƒ } q | d
| ƒ r3t j
| | ƒ } | rö| j ƒ d d !} | j | ƒ | j ƒ } | d
| d ƒ sÞ| d } n | j | | ƒ } q q…d
| | k r/| j | d d !ƒ | j | d ƒ } n Pq | d | ƒ rst j
| | ƒ } | rº| j d ƒ } | j | ƒ | j ƒ } | d
| d ƒ s¢| d } n | j | | ƒ } q n t j
| | ƒ } | r9| r5| j ƒ | | k r5| j r| j d ƒ q5| | k r | } n | j | | d ƒ } n Pq…| d | k ro| j d ƒ | j | | d ƒ } q…Pq d s t d ƒ ‚ q W| rÎ| | k rÎ| j rÎ| j | | | !ƒ | j | | ƒ } n | | | _ d S( Ni u u &#i iÿÿÿÿu ;u &u# EOF in middle of entity or char refu interesting.search() lied( R t lenR t searcht startR t handle_datat updatepost
startswitht starttagopent matcht parse_starttagt parse_endtagt
parse_commentt parse_piR t parse_declarationt parse_html_declarationR' t findt charreft groupt handle_charreft endt entityreft handle_entityreft
incompleteR (
R RC R t it nR8 t jR6 t kt name( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR! ° s°
c C` sè | j } | | | d !d k s, t d ƒ ‚ | | | d !d k rP | j | ƒ S| | | d !d k rt | j | ƒ S| | | d !j ƒ d k r× | j d
| d ƒ } | d k r· d S| j | | d | !ƒ | d S| j | ƒ Sd S(
Ni u iÿÿÿÿi ( R R R; t parse_marked_sectionR* R? t handle_declt parse_bogus_comment( R RG R t gtpos( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR> s #
i c C` s{ | j } | | | d !d k s, t d ƒ ‚ | j d | d ƒ } | d k rR d S| rs | j | | d | !ƒ n | d S( Ni u iÿÿÿÿi ( u u />u
u junk characters in start tag: %ri ( u >u />( R R( t check_for_whole_start_tagR t tagfindR8 R RC RA R* R R t attrfindt attrfind_tolerantt unescapet appendt stripR% t countR1 t rfindR' R4 t endswitht handle_startendtagt handle_starttagt CDATA_CONTENT_ELEMENTSR/ ( R RG t endposR t attrsR8 RJ t tagt mt attrnamet restt attrvalueRC R R
( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR9 E s\ $$ c C` sh | j } | j r' t j | | ƒ } n t j | | ƒ } | rX| j ƒ } | | | d !} | d k rp | d S| d k rô | j d | ƒ r– | d S| j d | ƒ r¬ d S| j rÙ | j | | d ƒ | j d ƒ n | | k ré | S| d Sn | d k rd S| d k rd S| j r=| j | | ƒ | j d
ƒ n | | k rM| S| d Sn t d ƒ ‚ d S( Ni u >u /u />i iÿÿÿÿu malformed empty start tagu u6 abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZu malformed start tagu we should not get here!(
R R t locatestarttagendR8 t locatestarttagend_tolerantRC R6 R5 R' R ( R RG R Re RI t next( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyRU ~ s> c C` s· | j } | | | d !d k s, t d ƒ ‚ t j | | d ƒ } | sL d S| j ƒ } t j | | ƒ } | sH| j d k r— | j | | | !ƒ | S| j
r¾ | j d | | | !f ƒ n t j | | d ƒ } | s | | | d !d k rù | 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 r–| | j k r–| j | | | !ƒ | Sn | j | j ƒ ƒ | j ƒ | S(
Ni u u >( R R t endendtagR2 RC t
endtagfindR8 R R R4 R R' t tagfind_tolerantRN RA R* R? t
handle_endtagR0 ( R RG R R8 RO t namematcht tagnameR. ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR: © s< #
c C` s! | j | | ƒ | j | ƒ d S( N( R` Ro ( R Rd Rc ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR_ Ó s c C` s d S( N( ( R Rd Rc ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR` Ø s c C` s d S( N( ( R Rd ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyRo Ü s c C` s d S( N( ( R RK ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyRB à s c C` s d S( N( ( R RK ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyRE ä s c C` s d S( N( ( R R" ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR4 è s c C` s d S( N( ( R R" ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyRP ì s c C` s d S( N( ( R t decl( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyRM ð s c C` s d S( N( ( R R" ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyRT ô s c C` s$ | j r | j d | f ƒ n d S( Nu unknown declaration: %r( R R' ( R R" ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt unknown_decl÷ s c C` s, d | k r | Sd „ } t j d | | ƒ S( Nu &c S` s | j ƒ d } yl | d d k r{ | d } | d d k r\ t | d j d ƒ d ƒ } n t | j d ƒ ƒ } t | ƒ SWn t k
r” d | SXd d l m } | | k r¹ | | S| j d ƒ rÐ d
| SxI t d t | ƒ ƒ D]* } | | | k ræ | | | | | Sqæ Wd
| Sd S(
Ni u #i u xu Xu ;i u &#( t html5u &i ( u xu X(
t groupst intt rstript chrt
ValueErrort future.backports.html.entitiesRt R^ t rangeR1 ( t st cRt t x( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt replaceEntitiesÿ s&
u& &(#?[xX]?(?:[0-9a-fA-F]+;|\w{1,32};?))( R+ t sub( R R| R ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyRY ü s
( u scriptu styleN(! R R R Ra t FalseR
R R# R$ R' R R( R) R/ R0 R! R> RN R< R9 RU R: R_ R` Ro RB RE R4 RP RM RT Rs RY ( ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyR d s:
h 9 + * ( R t
__future__R R R R t future.builtinst future.backportsR R+ R R, R RF RD R@ R7 RS t commentcloseRV Rn RW RX t VERBOSERi Rj Rl Rm t ExceptionR R R ( ( ( s; lib/python2.7/site-packages/future/backports/html/parser.pyt s4 "