def get_noun_chains(doc):
try:
rels = doc.relations
except ValueError:
return None
chains = collect_chains(rels)
chains = [get_noun_chain(chain) for chain in chains]
return [chain for chain in chains if chain.entities]
def test_2_chains(self):
rels = {
_create_rel(0, 1),
_create_rel(1, 2),
_create_rel(3, 4),
}
got_chains = collect_chains(rels)
expected_chains = [
_create_chain([0, 1, 2]),
_create_chain([3, 4]),
]
self.assertEqual(got_chains, expected_chains)
def apply(self, pairs, rels):
if rels is None:
return DefaultCandidateMaker().apply(pairs, rels)
chains = collect_chains(rels)
ret = []
for e1, e2 in pairs:
if any(map(lambda x: e1 in x.entities and e2 in x.entities,
chains)):
ret.append((e1, e2, self.label))
else:
ret.append((e1, e2, None))
return ret
def chain_similar_entities(
doc: Document, entities: List[Entity],
entity_comparator: Callable[[Document, Entity, Entity], bool] = compare_entities_by_tokens) \
-> List[CoreferenceChain]:
relations = set()
for i, e1 in enumerate(entities):
for e2 in entities[:i]:
if entity_comparator(doc, e1, e2):
relations.add(Relation(e1, e2, "match"))
return collect_chains(relations, entities)
def test_many_entities(self):
rels = {
_create_rel(0, 1),
_create_rel(1, 2),
_create_rel(2, 1),
_create_rel(2, 0),
_create_rel(6, 2),
}
got_chains = collect_chains(rels)
expected_chains = [
_create_chain([0, 1, 2, 6]),
]
self.assertEqual(got_chains, expected_chains)
def test_chain_common_mention(self):
rels = {
_create_rel(0, 1),
_create_rel(1, 2),
_create_rel(3, 4),
_create_rel(4, 5),
_create_rel(2, 5),
}
got_chains = collect_chains(rels)
expected_chains = [
_create_chain([0, 1, 2, 3, 4, 5]),
]
self.assertEqual(got_chains, expected_chains)
def test_1_chain(self):
entites = [
_create_entity(0),
_create_entity(1),
_create_entity(2),
_create_entity(3)
]
rels = {
_create_rel(0, 1),
_create_rel(1, 2),
}
got_chains = collect_chains(rels, entites)
expected_chains = [_create_chain([0, 1, 2]), _create_chain([3])]
self.assertEqual(got_chains, expected_chains)
def serialize_doc(self, doc: Document, fp: IO):
fp.write('#begin document (' + doc.name + '); \n')
chains = collect_chains(doc.relations, doc.entities)
group_positions = self._get_group_positions(chains)
for sentence in doc.sentences:
for i, token in enumerate(
doc.tokens[sentence.start_token:sentence.end_token]):
idx = i + sentence.start_token
group_info = self._get_group_info(idx, group_positions)
fp.write('\t'.join(
[doc.name, str(idx),
str(i), token, group_info]) + '\n')
fp.write('\n')
fp.write('#end document')
def test_3_chains(self):
rels = {
_create_rel(0, 1),
_create_rel(1, 2),
_create_rel(3, 4),
_create_rel(2, 5),
_create_rel(6, 3),
_create_rel(10, 11),
_create_rel(20, 10),
}
got_chains = collect_chains(rels)
expected_chains = [
_create_chain([0, 1, 2, 5]),
_create_chain([3, 4, 6]),
_create_chain([10, 11, 20]),
]
self.assertEqual(got_chains, expected_chains)
def test_chain_multiple_common_mention(self):
rels = {
_create_rel(0, 1),
_create_rel(1, 2),
_create_rel(3, 4),
_create_rel(4, 5),
_create_rel(2, 5),
_create_rel(10, 20),
_create_rel(20, 30),
_create_rel(100, 200),
_create_rel(40, 50),
_create_rel(50, 60),
_create_rel(60, 70),
_create_rel(70, 30),
}
got_chains = collect_chains(rels)
expected_chains = [
_create_chain([0, 1, 2, 3, 4, 5]),
_create_chain([10, 20, 30, 40, 50, 60, 70]),
_create_chain([100, 200]),
]
self.assertEqual(got_chains, expected_chains)
def _convert_to_rel_chains(rels, entities):
chains = collect_chains(rels, entities)
return chains2rels(chains)
def test_no_chains(self):
rels = set()
got_chains = collect_chains(rels)
expected_chains = []
self.assertEqual(got_chains, expected_chains)