comp.lang.python | Google Groups

Re: How do you htmlentities in Python

John J. Lee <j...@pobox.com>

"Thomas Jollans" <tho...@jollans.NOSPAM.com> writes:
> "Adam Atlas" <a...@atlas.st> wrote in message
> news:1180965792.757685.132580@q75g2000hsh.googlegroups.com...
> > As far as I know, there isn't a standard idiom to do this, but it's
> > still a one-liner. Untested, but I think this should work:

> > import re
> > from htmlentitydefs import name2codepoint
> > def htmlentitydecode(s):
> > return re.sub('&(%s);' % '|'.join(name2codepoint), lambda m:
> > name2codepoint[m.group(1)], s)

> '&(%s);' won't quite work: HTML (and, I assume, SGML, but not XHTML being
> XML) allows you to skip the semicolon after the entity if it's followed by a
> white space (IIRC). Should this be respected, it looks more like this:
> r'&(%s)([;\s]|$)'

> Also, this completely ignores non-name entities as also found in XML. (eg
> %x20; for ' ' or so) Maybe some part of the HTMLParser module is useful, I
> wouldn't know. IMHO, these particular batteries aren't too commonly needed.

Here's one that handles numeric character references, and chooses to
leave entity references that are not defined in standard library
module htmlentitydefs intact, rather than throwing an exception.

It ignores the missing semicolon issue (and note also that IE can cope
with even a missing space, like "tr&eacutes mal", so you'll see that
in the wild). Probably it could be adapted to handle that (possibly
the presumably-slower htmllib-based recipe on the python.org wiki
already does handle that, not sure).

import htmlentitydefs
import re
import unittest

def unescape_charref(ref):
name = ref[2:-1]
base = 10
if name.startswith("x"):
name = name[1:]
base = 16
return unichr(int(name, base))

def replace_entities(match):
ent = match.group()
if ent[1] == "#":
return unescape_charref(ent)

repl = htmlentitydefs.name2codepoint.get(ent[1:-1])
if repl is not None:
repl = unichr(repl)
else:
repl = ent
return repl

def unescape(data):
return re.sub(r"&#?[A-Za-z0-9]+?;", replace_entities, data)

class UnescapeTests(unittest.TestCase):

def test_unescape_charref(self):
self.assertEqual(unescape_charref(u"&"), u"&")
self.assertEqual(unescape_charref(u"—"), u"\N{EM DASH}")
self.assertEqual(unescape_charref(u"—"), u"\N{EM DASH}")

def test_unescape(self):
self.assertEqual(
unescape(u"& < — — —"),
u"& < %s %s %s" % tuple(u"\N{EM DASH}"*3)
)
self.assertEqual(unescape(u"&a&"), u"&a&")
self.assertEqual(unescape(u"a&"), u"a&")
self.assertEqual(unescape(u"&nonexistent;"), u"&nonexistent;")

unittest.main()

John