nltk_tree(sentence) throws NameError

A pattern.text.Sentence can't be converted into an nltk.tree.Tree, at least not
with nltk version 3.0.4.

>>> import pattern
>>> from pattern.de import parsetree
>>> sentence = parsetree('Hans liebt Maria.')[0]
>>> pattern.text.tree.nltk_tree(sentence)
---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-13-7bbd6533ae36> in <module>()
----> 1 pattern.text.tree.nltk_tree(sentence)

/usr/local/lib/python2.7/dist-packages/Pattern-2.6-py2.7.egg/pattern/text/tree.pyc in nltk_tree(sentence)
   1592             v.append(ch.pnp)
   1593     T.append(')')
-> 1594     return tree.bracket_parse(' '.join(T))
   1595 
   1596 ### GRAPHVIZ DOT ###################################################################################

/usr/local/lib/python2.7/dist-packages/nltk-3.0.4-py2.7.egg/nltk/tree.pyc in bracket_parse(s)
   1488     Use Tree.read(s, remove_empty_top_bracketing=True) instead.
   1489     """
-> 1490     raise NameError("Use Tree.read(s, remove_empty_top_bracketing=True) instead.")
   1491 
   1492 def sinica_parse(s):

NameError: Use Tree.read(s, remove_empty_top_bracketing=True) instead.

>>> import nltk
>>> nltk.__version__
'3.0.4'

Start in pattern/text/tree.py at nltk_tree, using the traceback and the NLTK 3.0.4 version as context. Check the current NLTK tree parsing API and verify that converting the example pattern.text.Sentence completes without the reported NameError.