Mercurial > public > sg101
comparison gpp/core/html.py @ 9:b3b11edf91d8
News: removed the lxml stuff. Based on Jacob Kaplan-Moss suggestion, use html5lib to clean html. Added that functionality in a new core.html module.
| author | Brian Neal <bgneal@gmail.com> |
|---|---|
| date | Sun, 12 Apr 2009 02:03:03 +0000 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| 8:d6f3c38e8f50 | 9:b3b11edf91d8 |
|---|---|
| 1 import html5lib | |
| 2 from html5lib import sanitizer, treebuilders, treewalkers, serializer | |
| 3 | |
| 4 def sanitizer_factory(*args, **kwargs): | |
| 5 san = sanitizer.HTMLSanitizer(*args, **kwargs) | |
| 6 # This isn't available yet | |
| 7 # san.strip_tokens = True | |
| 8 return san | |
| 9 | |
| 10 def clean_html(buf): | |
| 11 """Cleans HTML of dangerous tags and content.""" | |
| 12 buf = buf.strip() | |
| 13 if not buf: | |
| 14 return buf | |
| 15 | |
| 16 p = html5lib.HTMLParser(tree=treebuilders.getTreeBuilder("dom"), | |
| 17 tokenizer=sanitizer_factory) | |
| 18 dom_tree = p.parseFragment(buf) | |
| 19 | |
| 20 walker = treewalkers.getTreeWalker("dom") | |
| 21 stream = walker(dom_tree) | |
| 22 | |
| 23 s = serializer.htmlserializer.HTMLSerializer( | |
| 24 omit_optional_tags=False, | |
| 25 quote_attr_values=True) | |
| 26 return s.render(stream) | |
| 27 | |
| 28 # vim: ts=4 sw=4 |