+
+
view Lib/html/parser.py @ 84314:347647a1f798
+
+
+
+
Clarify note and fix typo.
+
+
+
+ | author |
+ Richard Oudkerk <shibturn@gmail.com> |
+
+
+ | date |
+ Mon, 24 Jun 2013 18:12:57 +0100 |
+
+
+ | parents |
+ 0d53703b1a99 |
+
+
+ | children |
+ 20be90a3a714 |
+
+
+
+
+
+
line source
+
+
1 """A parser for HTML and XHTML."""
+
+
3 # This file is based on sgmllib.py, but the API is slightly different.
+
+
5 # XXX There should be a way to distinguish between PCDATA (parsed
+
6 # character data -- the normal case), RCDATA (replaceable character
+
7 # data -- only char and entity references and end tags are special)
+
8 # and CDATA (character data -- only end tags are special).
+
+
+
+
+
+
+
15 # Regular expressions used for parsing
+
+
17 interesting_normal = re.compile('[&<]')
+
18 incomplete = re.compile('&[a-zA-Z#]')
+
+
20 entityref = re.compile('&([a-zA-Z][-.a-zA-Z0-9]*)[^a-zA-Z0-9]')
+
21 charref = re.compile('&#(?:[0-9]+|[xX][0-9a-fA-F]+)[^0-9a-fA-F]')
+
+
23 starttagopen = re.compile('<[a-zA-Z]')
+
24 piclose = re.compile('>')
+
25 commentclose = re.compile(r'--\s*>')
+
26 tagfind = re.compile('([a-zA-Z][-.a-zA-Z0-9:_]*)(?:\s|/(?!>))*')
+
27 # see http://www.w3.org/TR/html5/tokenization.html#tag-open-state
+
28 # and http://www.w3.org/TR/html5/tokenization.html#tag-name-state
+
29 tagfind_tolerant = re.compile('[a-zA-Z][^\t\n\r\f />\x00]*')
+
+
31 # 1) the strict attrfind isn't really strict, but we can't make it
+
32 # correctly strict without breaking backward compatibility;
+
33 # 2) if you change attrfind remember to update locatestarttagend too;
+
34 # 3) if you change attrfind and/or locatestarttagend the parser will
+
35 # explode, so don't do it.
+
36 attrfind = re.compile(
+
37 r'\s*([a-zA-Z_][-.:a-zA-Z_0-9]*)(\s*=\s*'
+
38 r'(\'[^\']*\'|"[^"]*"|[^\s"\'=<>`]*))?')
+
39 attrfind_tolerant = re.compile(
+
40 r'((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*'
+
41 r'(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*')
+
42 locatestarttagend = re.compile(r"""
+
43 <[a-zA-Z][-.a-zA-Z0-9:_]* # tag name
+
44 (?:\s+ # whitespace before attribute name
+
45 (?:[a-zA-Z_][-.:a-zA-Z0-9_]* # attribute name
+
46 (?:\s*=\s* # value indicator
+
47 (?:'[^']*' # LITA-enclosed value
+
48 |\"[^\"]*\" # LIT-enclosed value
+
49 |[^'\">\s]+ # bare value
+
+
+
+
+
54 \s* # trailing whitespace
+
+
56 locatestarttagend_tolerant = re.compile(r"""
+
57 <[a-zA-Z][-.a-zA-Z0-9:_]* # tag name
+
58 (?:[\s/]* # optional whitespace before attribute name
+
59 (?:(?<=['"\s/])[^\s/>][^\s/=>]* # attribute name
+
60 (?:\s*=+\s* # value indicator
+
61 (?:'[^']*' # LITA-enclosed value
+
62 |"[^"]*" # LIT-enclosed value
+
63 |(?!['"])[^>\s]* # bare value
+
+
65 (?:\s*,)* # possibly followed by a comma
+
+
+
+
69 \s* # trailing whitespace
+
+
71 endendtag = re.compile('>')
+
72 # the HTML 5 spec, section 8.1.2.2, doesn't allow spaces between
+
73 # </ and the tag name, so maybe this should be fixed
+
74 endtagfind = re.compile('</\s*([a-zA-Z][-.a-zA-Z0-9:_]*)\s*>')
+
+
+
77 class HTMLParseError(Exception):
+
78 """Exception raised for all parse errors."""
+
+
80 def __init__(self, msg, position=(None, None)):
+
+
+
83 self.lineno = position[0]
+
84 self.offset = position[1]
+
+
+
+
88 if self.lineno is not None:
+
89 result = result + ", at line %d" % self.lineno
+
90 if self.offset is not None:
+
91 result = result + ", column %d" % (self.offset + 1)
+
+
+
+
95 class HTMLParser(_markupbase.ParserBase):
+
96 """Find tags and other markup and call handler functions.
+
+
+
+
+
+
+
+
104 Start tags are handled by calling self.handle_starttag() or
+
105 self.handle_startendtag(); end tags by self.handle_endtag(). The
+
106 data between tags is passed from the parser to the derived class
+
107 by calling self.handle_data() with the data as argument (the data
+
108 may be split up in arbitrary chunks). Entity references are
+
109 passed by calling self.handle_entityref() with the entity
+
110 reference as the argument. Numeric character references are
+
111 passed to self.handle_charref() with the string containing the
+
112 reference as the argument.
+
+
+
115 CDATA_CONTENT_ELEMENTS = ("script", "style")
+
+
117 def __init__(self, strict=False):
+
118 """Initialize and reset this instance.
+
+
120 If strict is set to False (the default) the parser will parse invalid
+
121 markup, otherwise it will raise an error. Note that the strict mode
+
+
+
+
125 warnings.warn("The strict mode is deprecated.",
+
126 DeprecationWarning, stacklevel=2)
+
+
+
+
+
131 """Reset this instance. Loses all unprocessed data."""
+
+
+
134 self.interesting = interesting_normal
+
135 self.cdata_elem = None
+
136 _markupbase.ParserBase.reset(self)
+
+
138 def feed(self, data):
+
139 r"""Feed data to the parser.
+
+
141 Call this as often as you want, with as little or as much text
+
142 as you want (may include '\n').
+
+
144 self.rawdata = self.rawdata + data
+
+
+
+
148 """Handle any buffered data."""
+
+
+
151 def error(self, message):
+
152 raise HTMLParseError(message, self.getpos())
+
+
154 __starttag_text = None
+
+
156 def get_starttag_text(self):
+
157 """Return full source of start tag: '<...>'."""
+
158 return self.__starttag_text
+
+
160 def set_cdata_mode(self, elem):
+
161 self.cdata_elem = elem.lower()
+
162 self.interesting = re.compile(r'</\s*%s\s*>' % self.cdata_elem, re.I)
+
+
164 def clear_cdata_mode(self):
+
165 self.interesting = interesting_normal
+
166 self.cdata_elem = None
+
+
168 # Internal -- handle data as far as reasonable. May leave state
+
169 # and data to be processed by a subsequent call. If 'end' is
+
170 # true, force handling all data as if followed by EOF marker.
+
171 def goahead(self, end):
+
172 rawdata = self.rawdata
+
+
+
+
176 match = self.interesting.search(rawdata, i) # < or &
+
+
+
+
+
+
+
183 if i < j: self.handle_data(rawdata[i:j])
+
184 i = self.updatepos(i, j)
+
+
186 startswith = rawdata.startswith
+
187 if startswith('<', i):
+
188 if starttagopen.match(rawdata, i): # < + letter
+
189 k = self.parse_starttag(i)
+
190 elif startswith("</", i):
+
191 k = self.parse_endtag(i)
+
192 elif startswith("<!--", i):
+
193 k = self.parse_comment(i)
+
194 elif startswith("<?", i):
+
+
196 elif startswith("<!", i):
+
+
198 k = self.parse_declaration(i)
+
+
200 k = self.parse_html_declaration(i)
+
+
202 self.handle_data("<")
+
+
+
+
+
+
+
+
210 self.error("EOF in middle of construct")
+
211 k = rawdata.find('>', i + 1)
+
+
213 k = rawdata.find('<', i + 1)
+
+
+
+
+
218 self.handle_data(rawdata[i:k])
+
219 i = self.updatepos(i, k)
+
220 elif startswith("&#", i):
+
221 match = charref.match(rawdata, i)
+
+
223 name = match.group()[2:-1]
+
224 self.handle_charref(name)
+
+
226 if not startswith(';', k-1):
+
+
228 i = self.updatepos(i, k)
+
+
+
231 if ";" in rawdata[i:]: #bail by consuming &#
+
232 self.handle_data(rawdata[0:2])
+
233 i = self.updatepos(i, 2)
+
+
235 elif startswith('&', i):
+
236 match = entityref.match(rawdata, i)
+
+
238 name = match.group(1)
+
239 self.handle_entityref(name)
+
+
241 if not startswith(';', k-1):
+
+
243 i = self.updatepos(i, k)
+
+
245 match = incomplete.match(rawdata, i)
+
+
247 # match.group() will contain at least 2 chars
+
248 if end and match.group() == rawdata[i:]:
+
+
250 self.error("EOF in middle of entity or char ref")
+
+
+
+
+
255 i = self.updatepos(i, i + 1)
+
+
+
+
259 # not the end of the buffer, and can't be confused
+
260 # with some other construct
+
261 self.handle_data("&")
+
262 i = self.updatepos(i, i + 1)
+
+
+
+
266 assert 0, "interesting.search() lied"
+
+
268 if end and i < n and not self.cdata_elem:
+
269 self.handle_data(rawdata[i:n])
+
270 i = self.updatepos(i, n)
+
271 self.rawdata = rawdata[i:]
+
+
273 # Internal -- parse html declarations, return length or -1 if not terminated
+
274 # See w3.org/TR/html5/tokenization.html#markup-declaration-open-state
+
275 # See also parse_declaration in _markupbase
+
276 def parse_html_declaration(self, i):
+
277 rawdata = self.rawdata
+
278 assert rawdata[i:i+2] == '<!', ('unexpected call to '
+
279 'parse_html_declaration()')
+
280 if rawdata[i:i+4] == '<!--':
+
281 # this case is actually already handled in goahead()
+
282 return self.parse_comment(i)
+
283 elif rawdata[i:i+3] == '<![':
+
284 return self.parse_marked_section(i)
+
285 elif rawdata[i:i+9].lower() == '<!doctype':
+
+
287 gtpos = rawdata.find('>', i+9)
+
+
+
290 self.handle_decl(rawdata[i+2:gtpos])
+
+
+
293 return self.parse_bogus_comment(i)
+
+
295 # Internal -- parse bogus comment, return length or -1 if not terminated
+
296 # see http://www.w3.org/TR/html5/tokenization.html#bogus-comment-state
+
297 def parse_bogus_comment(self, i, report=1):
+
298 rawdata = self.rawdata
+
299 assert rawdata[i:i+2] in ('<!', '</'), ('unexpected call to '
+
+
301 pos = rawdata.find('>', i+2)
+
+
+
+
305 self.handle_comment(rawdata[i+2:pos])
+
+
+
308 # Internal -- parse processing instr, return end or -1 if not terminated
+
309 def parse_pi(self, i):
+
310 rawdata = self.rawdata
+
311 assert rawdata[i:i+2] == '<?', 'unexpected call to parse_pi()'
+
312 match = piclose.search(rawdata, i+2) # >
+
+
+
+
316 self.handle_pi(rawdata[i+2: j])
+
+
+
+
320 # Internal -- handle starttag, return end or -1 if not terminated
+
321 def parse_starttag(self, i):
+
322 self.__starttag_text = None
+
323 endpos = self.check_for_whole_start_tag(i)
+
+
+
326 rawdata = self.rawdata
+
327 self.__starttag_text = rawdata[i:endpos]
+
+
329 # Now parse the data between i+1 and j into a tag and attrs
+
+
331 match = tagfind.match(rawdata, i+1)
+
332 assert match, 'unexpected call to parse_starttag()'
+
+
334 self.lasttag = tag = match.group(1).lower()
+
+
+
337 m = attrfind.match(rawdata, k)
+
+
339 m = attrfind_tolerant.match(rawdata, k)
+
+
+
342 attrname, rest, attrvalue = m.group(1, 2, 3)
+
+
+
345 elif attrvalue[:1] == '\'' == attrvalue[-1:] or \
+
346 attrvalue[:1] == '"' == attrvalue[-1:]:
+
347 attrvalue = attrvalue[1:-1]
+
+
349 attrvalue = self.unescape(attrvalue)
+
350 attrs.append((attrname.lower(), attrvalue))
+
+
+
353 end = rawdata[k:endpos].strip()
+
354 if end not in (">", "/>"):
+
355 lineno, offset = self.getpos()
+
356 if "\n" in self.__starttag_text:
+
357 lineno = lineno + self.__starttag_text.count("\n")
+
358 offset = len(self.__starttag_text) \
+
359 - self.__starttag_text.rfind("\n")
+
+
361 offset = offset + len(self.__starttag_text)
+
+
363 self.error("junk characters in start tag: %r"
+
364 % (rawdata[k:endpos][:20],))
+
365 self.handle_data(rawdata[i:endpos])
+
+
367 if end.endswith('/>'):
+
368 # XHTML-style empty tag: <span attr="value" />
+
369 self.handle_startendtag(tag, attrs)
+
+
371 self.handle_starttag(tag, attrs)
+
372 if tag in self.CDATA_CONTENT_ELEMENTS:
+
373 self.set_cdata_mode(tag)
+
+
+
376 # Internal -- check to see if we have a complete starttag; return end
+
377 # or -1 if incomplete.
+
378 def check_for_whole_start_tag(self, i):
+
379 rawdata = self.rawdata
+
+
381 m = locatestarttagend.match(rawdata, i)
+
+
383 m = locatestarttagend_tolerant.match(rawdata, i)
+
+
+
386 next = rawdata[j:j+1]
+
+
+
+
390 if rawdata.startswith("/>", j):
+
+
392 if rawdata.startswith("/", j):
+
+
+
+
+
397 self.updatepos(i, j + 1)
+
398 self.error("malformed empty start tag")
+
+
+
+
+
+
+
+
406 if next in ("abcdefghijklmnopqrstuvwxyz=/"
+
407 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"):
+
408 # end of input in or before attribute value, or we have the
+
409 # '/' from a '/>' ending
+
+
+
+
413 self.error("malformed start tag")
+
+
+
+
+
418 raise AssertionError("we should not get here!")
+
+
420 # Internal -- parse endtag, return end or -1 if incomplete
+
421 def parse_endtag(self, i):
+
422 rawdata = self.rawdata
+
423 assert rawdata[i:i+2] == "</", "unexpected call to parse_endtag"
+
424 match = endendtag.search(rawdata, i+1) # >
+
+
+
+
428 match = endtagfind.match(rawdata, i) # </ + tag + >
+
+
430 if self.cdata_elem is not None:
+
431 self.handle_data(rawdata[i:gtpos])
+
+
+
434 self.error("bad end tag: %r" % (rawdata[i:gtpos],))
+
435 # find the name: w3.org/TR/html5/tokenization.html#tag-name-state
+
436 namematch = tagfind_tolerant.match(rawdata, i+2)
+
+
438 # w3.org/TR/html5/tokenization.html#end-tag-open-state
+
439 if rawdata[i:i+3] == '</>':
+
+
+
442 return self.parse_bogus_comment(i)
+
443 tagname = namematch.group().lower()
+
444 # consume and ignore other stuff between the name and the >
+
445 # Note: this is not 100% correct, since we might have things like
+
446 # </tag attr=">">, but looking for > after tha name should cover
+
447 # most of the cases and is much simpler
+
448 gtpos = rawdata.find('>', namematch.end())
+
449 self.handle_endtag(tagname)
+
+
+
452 elem = match.group(1).lower() # script or style
+
453 if self.cdata_elem is not None:
+
454 if elem != self.cdata_elem:
+
455 self.handle_data(rawdata[i:gtpos])
+
+
+
458 self.handle_endtag(elem.lower())
+
459 self.clear_cdata_mode()
+
+
+
462 # Overridable -- finish processing of start+end tag: <tag.../>
+
463 def handle_startendtag(self, tag, attrs):
+
464 self.handle_starttag(tag, attrs)
+
465 self.handle_endtag(tag)
+
+
467 # Overridable -- handle start tag
+
468 def handle_starttag(self, tag, attrs):
+
+
+
471 # Overridable -- handle end tag
+
472 def handle_endtag(self, tag):
+
+
+
475 # Overridable -- handle character reference
+
476 def handle_charref(self, name):
+
+
+
479 # Overridable -- handle entity reference
+
480 def handle_entityref(self, name):
+
+
+
483 # Overridable -- handle data
+
484 def handle_data(self, data):
+
+
+
487 # Overridable -- handle comment
+
488 def handle_comment(self, data):
+
+
+
491 # Overridable -- handle declaration
+
492 def handle_decl(self, decl):
+
+
+
495 # Overridable -- handle processing instruction
+
496 def handle_pi(self, data):
+
+
+
499 def unknown_decl(self, data):
+
+
501 self.error("unknown declaration: %r" % (data,))
+
+
503 # Internal -- helper to remove special character quoting
+
504 def unescape(self, s):
+
+
+
507 def replaceEntities(s):
+
+
+
+
+
512 if s[0] in ['x','X']:
+
513 c = int(s[1:].rstrip(';'), 16)
+
+
515 c = int(s.rstrip(';'))
+
+
+
+
+
520 from html.entities import html5
+
+
+
523 elif s.endswith(';'):
+
+
525 for x in range(2, len(s)):
+
+
527 return html5[s[:x]] + s[x:]
+
+
+
+
531 return re.sub(r"&(#?[xX]?(?:[0-9a-fA-F]+;|\w{1,32};?))",
+
532 replaceEntities, s, flags=re.ASCII)
+
+
+
