def _clean(self, text):
"""
Cleans up MediaWiki text of markup.
Currently a "naive" version in that
it just strips *all* punctuation.
Will eventually want to strip out unnecessary
markup syntax as well, such as 'File:' and
'Category'.
Args:
| text (str) -- the MediaWiki text to cleanup.
Returns:
| str -- the replaced text.
"""
return brain.depunctuate(text)
def _clean(self, text):
"""
Cleans up MediaWiki text of markup.
Currently a "naive" version in that
it just strips *all* punctuation.
Will eventually want to strip out unnecessary
markup syntax as well, such as 'File:' and
'Category'.
Args:
| text (str) -- the MediaWiki text to cleanup.
Returns:
| str -- the replaced text.
"""
return brain.depunctuate(text)
def test_depunctuate(self):
data = "[h%e@l&l~o* (t/h>e,r.e:"
self.assertEqual(brain.depunctuate(data), " h e l l o t h e r e ")
def test_depunctuate(self):
data = '[h%e@l&l~o* (t/h>e,r.e:'
self.assertEqual(brain.depunctuate(data), ' h e l l o t h e r e ')
