diff options
author | Leonard Richardson <leonard.richardson@canonical.com> | 2011-02-10 16:41:10 -0500 |
---|---|---|
committer | Leonard Richardson <leonard.richardson@canonical.com> | 2011-02-10 16:41:10 -0500 |
commit | d89c8878ea86a2575c87e9fad8081cfcd81e0bcd (patch) | |
tree | cdad3f97812e658d84a611b6017b7198fd97d818 /beautifulsoup/builder/lxml_builder.py | |
parent | e1ad4220e5ca00ec0e7f77ce5087845fcb356a0e (diff) |
Added some elementary doctype handling.
Diffstat (limited to 'beautifulsoup/builder/lxml_builder.py')
-rw-r--r-- | beautifulsoup/builder/lxml_builder.py | 41 |
1 files changed, 41 insertions, 0 deletions
diff --git a/beautifulsoup/builder/lxml_builder.py b/beautifulsoup/builder/lxml_builder.py new file mode 100644 index 0000000..8336ab4 --- /dev/null +++ b/beautifulsoup/builder/lxml_builder.py @@ -0,0 +1,41 @@ +from lxml import etree +from beautifulsoup.element import Comment, Doctype +from beautifulsoup.builder import HTMLTreeBuilder + +class LXMLTreeBuilder(HTMLTreeBuilder): + + def __init__(self, parser_class=etree.HTMLParser): + self.parser = parser_class(target=self) + self.soup = None + + def feed(self, markup): + self.parser.feed(markup) + self.parser.close() + + def close(self): + pass + + def start(self, name, attrs): + self.soup.handle_starttag(name, attrs) + + def end(self, name): + self.soup.handle_endtag(name) + + def data(self, content): + self.soup.handle_data(content) + + def doctype(self, name, pubid, system): + self.soup.endData() + self.soup.handle_data(name) + self.soup.endData(Doctype) + + def comment(self, content): + "Handle comments as Comment objects." + self.soup.endData() + self.soup.handle_data(content) + self.soup.endData(Comment) + + def test_fragment_to_document(self, fragment): + """See `TreeBuilder`.""" + return u'<html><body>%s</body></html>' % fragment + |