@@ -0,0 +1,26 @@
1+from HTMLParser import HTMLParser
2+import htmlentitydefs
3+
4+class HTMLTextExtractor(HTMLParser):
5+ def __init__(self):
6+ HTMLParser.__init__(self)
7+ self.result = [ ]
8+
9+ def handle_data(self, d):
10+ self.result.append(d)
11+
12+ def handle_charref(self, number):
13+ codepoint = int(number[1:], 16) if number[0] in (u'x', u'X') else int(number)
14+ self.result.append(unichr(codepoint))
15+
16+ def handle_entityref(self, name):
17+ codepoint = htmlentitydefs.name2codepoint[name]
18+ self.result.append(unichr(codepoint))
19+
20+ def get_text(self):
21+ return u''.join(self.result)
22+
23+def html_to_text(html):
24+ s = HTMLTextExtractor()
25+ s.feed(html)
26+ return s.get_text()