def trees(self, parse, nbest=1, post_process=False, filter=None, mrepr='tokens'):
"""
Return a genartor of pretty printed Tree structures of the
classification.
The output for a word such as `aandeelhoudersvergadering` is::
>>> trees = classifier.trees(results)
>>> for tree in trees:
... print tree
(NOU
(NOU (NOU (ADP aan) (NOU deel)) (VRB houd) (SUF er))
(LE s)
(NOU (VRB vergader) (SUF ing)))
Args:
- parse: the return value of :func:`mbmp.MBMA.classify` for a given word.
- nbest (int): specify how many parses should be returned
- post_process (bool): perform some cleaning operation after parsing
- filter (function): a function used to filter certain trees
(function must return True or False)
- mrepr (str): specify how the leaves should be represented in the
Tree: options are: token forms ('tokens'), lemma forms
('lemmas') and tokens and lemmas ('tokens-and-lemmas')
Yields:
the nbest parses
"""
self.parser.set_grammar(make_grammar(parse, mrepr))
return self.parser.nbest_parse(
[m.pprint(mrepr) for m in parse], nbest, post_process, filter)
def demo():
"""
A demo showing some basic functionality.
"""
from mbmp.parse.util import make_grammar
from mbmp.datatypes import Morpheme
parse = [
Morpheme(pos='V|*V', token='ver', lemma='ver'),
Morpheme(pos='V', token='koop', lemma='koop'),
Morpheme(pos='A|V*', token='baar', lemma='baar')
]
print 'Parse produced by MBMA:'
print['(%s %s)' % (m.pos, m) for m in parse]
print
grammar = cPickle.load(open(PCFG))
print 'Parsing word "ver koop baar" with default CKY Parser'
parser = CKYParser(grammar)
for tree in parser.nbest_parse('ver koop baar'.split()):
print tree
print
print 'Parsing word "ver koop baar" with Mbma CKY Parser'
print 'Compiling grammar rules from parse...'
g = make_grammar(parse)
for prod in g:
print prod
print
parser = MbmaCKYParser(grammar)
parser.set_grammar(g)
for tree in parser.nbest_parse('ver koop baar'.split(), n=5):
print tree
def demo():
"""
A demo showing some basic functionality.
"""
from mbmp.parse.util import make_grammar
from mbmp.datatypes import Morpheme
parse = [Morpheme(pos='V|*V', token='ver', lemma='ver'),
Morpheme(pos='V', token='eis', lemma='eis'),
Morpheme(pos='V|VINFL', token='t', lemma='t')]
print 'Parse produced by MBMA:'
print parse
print
print 'Compiling grammar rules from parse...'
productions = make_grammar(parse)
for prod in productions:
print prod
print
print 'Parsing word "ver eis t"'
parser = MbmaParser(productions)
for tree in parser.nbest_parse('ver eis t'.split()):
print tree
def trees(self,
parse,
nbest=1,
post_process=False,
filter=None,
mrepr='tokens'):
"""
Return a genartor of pretty printed Tree structures of the
classification.
The output for a word such as `aandeelhoudersvergadering` is::
>>> trees = classifier.trees(results)
>>> for tree in trees:
... print tree
(NOU
(NOU (NOU (ADP aan) (NOU deel)) (VRB houd) (SUF er))
(LE s)
(NOU (VRB vergader) (SUF ing)))
Args:
- parse: the return value of :func:`mbmp.MBMA.classify` for a given word.
- nbest (int): specify how many parses should be returned
- post_process (bool): perform some cleaning operation after parsing
- filter (function): a function used to filter certain trees
(function must return True or False)
- mrepr (str): specify how the leaves should be represented in the
Tree: options are: token forms ('tokens'), lemma forms
('lemmas') and tokens and lemmas ('tokens-and-lemmas')
Yields:
the nbest parses
"""
self.parser.set_grammar(make_grammar(parse, mrepr))
return self.parser.nbest_parse([m.pprint(mrepr) for m in parse], nbest,
post_process, filter)
from mbmp.parse.util import make_grammar
from mbmp.datatypes import Morpheme
parse = [Morpheme(pos='V|*V', token='ver', lemma='ver'),
Morpheme(pos='V', token='koop', lemma='koop'),
Morpheme(pos='A|V*', token='baar', lemma='baar')]
print 'Parse produced by MBMA:'
print ['(%s %s)' % (m.pos, m) for m in parse]
print
grammar = cPickle.load(open(PCFG))
print 'Parsing word "ver koop baar" with default CKY Parser'
parser = CKYParser(grammar)
for tree in parser.nbest_parse('ver koop baar'.split()):
print tree
print
print 'Parsing word "ver koop baar" with Mbma CKY Parser'
print 'Compiling grammar rules from parse...'
g = make_grammar(parse)
for prod in g:
print prod
print
parser = MbmaCKYParser(grammar)
parser.set_grammar(g)
for tree in parser.nbest_parse('ver koop baar'.split(), n=5):
print tree
__all__ = ['CKYParser', 'MbmaCKYParser']
if __name__ == '__main__':
demo()
