def test(self):
v = FeatureVector()
entity = Entity('1', EntityType.GPE, 'New York City', EntityOrigin.GEO)
entity.names = {'NYC', 'New York'}
chain = MentionChain(
[Mention('Nueva York', 'doc1', (), (), EntityType.GPE)])
SharedTokensFeature().extract(chain, entity, None, v)
self.assertAlmostEqual(0.5, v.data[0])
def test_match(self):
v = FeatureVector()
entity = Entity('1', EntityType.GPE, 'New York', EntityOrigin.GEO)
entity.names = {'New York', 'New York City', 'NYC'}
chain = MentionChain(
[Mention('Nueva York', 'doc1', (), (), EntityType.GPE)])
chain.mentions[0].translate_string = 'new york'
ExactMatchFeature().extract(chain, entity, None, v)
self.assertTrue(v.data[0])
def test_positive(self):
v = FeatureVector()
entity = Entity('1',
EntityType.GPE,
'New York',
EntityOrigin.GEO,
urls=['http://en.wikipedia.org/wiki/New_York_City'])
entity.names = {'New York', 'New York City', 'NYC'}
chain = MentionChain(
[Mention('New York City', 'doc1', (), (), EntityType.GPE)])
WikipediaFeature().extract(chain, entity, None, v)
self.assertTrue(v.data[0])
def test_one_match(self):
doc = unittest.mock.Mock()
doc.doc_id = 'doc1'
doc.mention_chains = [
MentionChain(
[Mention('John Smith', 'doc1', (4, 8), (), EntityType.PER)]),
]
entity1 = Entity('122', EntityType.PER, 'John', EntityOrigin.WLL)
entity1.names = {'John', 'J. Smith', 'john smith'}
doc.mention_chains[0].candidates = [
entity1,
Entity('123', EntityType.PER, 'Not John', EntityOrigin.WLL)
]
ExactNameResolver().resolve(doc)
self.assertEqual(1, len(doc.mention_chains))
self.assertEqual(2, len(doc.mention_chains[0].candidates))
self.assertEqual(entity1, doc.mention_chains[0].entity)