This project lets you use Google's SyntaxNet (AKA Parsey McParseface) from python code and without hours of installing and training models.
import pyparseface
parsed_sentence_dict = pyparseface.parse_sentence(
"Maybe there was once a human who looked like you, and somewhere along the "
"line you killed him and took his place. And your superiors don't know.")
print("OrderedDict: %s\n" % parsed_sentence_dict)
pyparseface.pretty_print_dict(parsed_sentence_dict)
OrderedDict: OrderedDict([(u'was VBD ROOT', OrderedDict([(u'Maybe RB advmod', OrderedDict()), (u'there EX expl', OrderedDict()), (u'once RB advmod', OrderedDict()), (u'human NN nsubj', OrderedDict([(u'a DT det', OrderedDict()), (u'looked VBD rcmod', OrderedDict([(u'who WP nsubj', OrderedDict()), (u'like IN prep', OrderedDict([(u'you PRP pobj', OrderedDict())]))]))])), (u', , punct', OrderedDict()), (u'and CC cc', OrderedDict()), (u'killed VBD conj', OrderedDict([(u'somewhere RB advmod', OrderedDict([(u'along IN prep', OrderedDict([(u'line NN pobj', OrderedDict([(u'the DT det', OrderedDict())]))]))])), (u'you PRP nsubj', OrderedDict()), (u'him PRP dobj', OrderedDict()), (u'and CC cc', OrderedDict()), (u'took VBD conj', OrderedDict([(u'place. NN dobj', OrderedDict([(u'his PRP$ poss', OrderedDict())]))])), (u'And CC cc', OrderedDict()), (u'know VB conj', OrderedDict([(u'superiors NNS nsubj', OrderedDict([(u'your PRP$ poss', OrderedDict())])), (u'do VBP aux', OrderedDict()), (u"n't RB neg", OrderedDict())]))])), (u'. . punct', OrderedDict())]))])
was VBD ROOT
+-- Maybe RB advmod
+-- there EX expl
+-- once RB advmod
+-- human NN nsubj
| +-- a DT det
| +-- looked VBD rcmod
| +-- who WP nsubj
| +-- like IN prep
| +-- you PRP pobj
+-- , , punct
+-- and CC cc
+-- killed VBD conj
| +-- somewhere RB advmod
| | +-- along IN prep
| | +-- line NN pobj
| | +-- the DT det
| +-- you PRP nsubj
| +-- him PRP dobj
| +-- and CC cc
| +-- took VBD conj
| | +-- place. NN dobj
| | +-- his PRP$ poss
| +-- And CC cc
| +-- know VB conj
| +-- superiors NNS nsubj
| | +-- your PRP$ poss
| +-- do VBP aux
| +-- n't RB neg
+-- . . punct
Most of the hard stuff was written by Google. Ben Plowman made it slightly easier to get up and running.