def test_get_feature_map_conjunction(self):
doc = self.sm_nlp(
'Richard and the man said they were entering the big house')
self.sm_rules_analyzer.initialize(doc)
feature_map = self.sm_tendencies_analyzer.get_feature_map(
Mention(doc[0], False), doc)
self.assertEqual(len(self.sm_feature_table), len(feature_map))
self.assertEqual([
0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 1, 0,
0, 1, 1, 0
], feature_map)
feature_map = self.sm_tendencies_analyzer.get_feature_map(
Mention(doc[0], True), doc)
self.assertEqual(len(self.sm_feature_table), len(feature_map))
self.assertEqual([
0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0,
0, 1, 1, 0
], feature_map)
feature_map = self.sm_tendencies_analyzer.get_feature_map(
Mention(doc[5], False), doc)
self.assertEqual(len(self.sm_feature_table), len(feature_map))
self.assertEqual([
0, 0, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 1,
1, 1, 0, 1
], feature_map)
def test_get_position_map_conjunction_second_sentence_mentions_true(self):
doc = self.sm_nlp(
'A preceding sentence. Peter and Jane spoke to him and her.')
self.sm_rules_analyzer.initialize(doc)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[4], True), doc)
self.assertEqual([0, 1, 1, 0, 0, 1, 0], position_map)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[9], True), doc)
self.assertEqual([5, 2, 1, 2, 0, 1, 0], position_map)
def test_get_position_map_second_sentence_mention(self):
doc = self.sm_nlp(
'This is a preceding sentence. Richard said he was entering the big house'
)
self.sm_rules_analyzer.initialize(doc)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[6], False), doc)
self.assertEqual([0, 1, 1, 0, 0, 0, 0], position_map)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[8], False), doc)
self.assertEqual([2, 2, 2, 0, 0, 0, 0], position_map)
def test_vectors_twoway_coordination_lg(self):
doc = self.lg_nlp('Peter and Jane arrived')
self.lg_rules_analyzer.initialize(doc)
peter_vectors = self.lg_tendencies_analyzer.get_vectors(
Mention(doc[0], False), doc)
jane_vectors = self.lg_tendencies_analyzer.get_vectors(
Mention(doc[2], False), doc)
combined_vectors = self.lg_tendencies_analyzer.get_vectors(
Mention(doc[0], True), doc)
for index in range(len(peter_vectors[0])):
self.assertAlmostEqual(
(peter_vectors[0][index] + jane_vectors[0][index]) / 2,
combined_vectors[0][index])
def test_get_position_map_root_mention(self):
doc = self.sm_nlp('Richard said he was entering the big house')
self.sm_rules_analyzer.initialize(doc)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[1], False), doc)
self.assertEqual([1, 0, 0, 0, -1, 0, 0], position_map)
def func(nlp):
if nlp.meta['name'] in excluded_nlps:
return
doc = nlp(doc_text)
rules_analyzer = RulesAnalyzerFactory.get_rules_analyzer(nlp)
rules_analyzer.initialize(doc)
assert rules_analyzer.is_independent_noun(doc[referred_index]) or \
rules_analyzer.is_potential_anaphor(doc[referred_index])
assert rules_analyzer.is_potential_anaphor(doc[referring_index])
referred_mention = Mention(doc[referred_index],
include_dependent_siblings)
self.assertEqual(
expected_truth,
rules_analyzer.is_potential_anaphoric_pair(
referred_mention, doc[referring_index], True),
nlp.meta['name'])
self.assertEqual(
expected_reflexive_truth,
rules_analyzer.is_potential_reflexive_pair(
referred_mention, doc[referring_index]), nlp.meta['name'])
self.assertEqual(
is_reflexive_anaphor_truth,
rules_analyzer.is_reflexive_anaphor(doc[referring_index]),
nlp.meta['name'])
def test_get_position_map_conjunction_first_sentence_mentions_false(self):
doc = self.sm_nlp('Peter and Jane spoke to him and her.')
self.sm_rules_analyzer.initialize(doc)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[0], False), doc)
self.assertEqual([0, 1, 1, 0, 0, -1, 0], position_map)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[2], False), doc)
self.assertEqual([2, 2, 1, 1, 1, -1, 1], position_map)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[5], False), doc)
self.assertEqual([5, 2, 1, 2, 0, -1, 0], position_map)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[7], False), doc)
self.assertEqual([7, 3, 1, 1, 1, -1, 1], position_map)
def test_get_position_map_first_sentence_mention(self):
doc = self.sm_nlp('Richard said he was entering the big house')
self.sm_rules_analyzer.initialize(doc)
mention = Mention(doc[0], False)
position_map = self.sm_tendencies_analyzer.get_position_map(
mention, doc)
self.assertEqual(mention.temp_position_map, position_map)
self.assertEqual([0, 1, 1, 0, 0, 0, 0], position_map)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[2], False), doc)
self.assertEqual([2, 2, 2, 0, 0, 0, 0], position_map)
position_map = self.sm_tendencies_analyzer.get_position_map(
Mention(doc[6], False), doc)
self.assertEqual([6, 3, 2, 1, 1, 0, 0], position_map)
def test_vectors_threeway_coordination_lg(self):
doc = self.lg_nlp('They spoke to Richard, Peter and Jane.')
self.lg_rules_analyzer.initialize(doc)
richard_vectors = self.lg_tendencies_analyzer.get_vectors(
Mention(doc[3], False), doc)
peter_vectors = self.lg_tendencies_analyzer.get_vectors(
Mention(doc[5], False), doc)
jane_vectors = self.lg_tendencies_analyzer.get_vectors(
Mention(doc[7], False), doc)
combined_vectors = self.lg_tendencies_analyzer.get_vectors(
Mention(doc[3], True), doc)
for index in range(len(peter_vectors[0])):
self.assertAlmostEqual(
(peter_vectors[0][index] + jane_vectors[0][index] +
richard_vectors[0][index]) / 3,
combined_vectors[0][index],
places=3)
def test_vectors_threeway_coordination_sm(self):
doc = self.sm_nlp('Richard, Peter and Jane arrived')
self.sm_rules_analyzer.initialize(doc)
richard_vectors = self.sm_tendencies_analyzer.get_vectors(
Mention(doc[0], False), doc)
peter_vectors = self.sm_tendencies_analyzer.get_vectors(
Mention(doc[2], False), doc)
jane_vectors = self.sm_tendencies_analyzer.get_vectors(
Mention(doc[4], False), doc)
combined_vectors = self.sm_tendencies_analyzer.get_vectors(
Mention(doc[0], True), doc)
for index in range(len(peter_vectors[0])):
self.assertAlmostEqual(
(peter_vectors[0][index] + jane_vectors[0][index] +
richard_vectors[0][index]) / 3,
combined_vectors[0][index],
places=3)
def test_get_vectors_mention_without_head_lg(self):
doc = self.lg_nlp('He arrived')
self.lg_rules_analyzer.initialize(doc)
vectors = self.lg_tendencies_analyzer.get_vectors(
Mention(doc[1], False), doc)
self.assertTrue(vectors[0].any())
self.assertFalse(vectors[1].any())
self.assertEqual(len(vectors[0]), len(vectors[1]))
def test_get_cosine_similarity_lg_root_2(self):
doc = self.lg_nlp('Richard arrived. He.')
self.lg_rules_analyzer.initialize(doc)
self.assertEqual(
'[3, 1, 0, -1, 1]',
str(
self.lg_tendencies_analyzer.get_compatibility_map(
Mention(doc[0], False), doc[3])))
def test_get_compatibility_map_simple(self):
doc = self.sm_nlp('Richard said he was entering the big house')
self.sm_rules_analyzer.initialize(doc)
self.assertEqual(
'[2, 0, 1, 0.30341318, 3]',
str(
self.sm_tendencies_analyzer.get_compatibility_map(
Mention(doc[0], False), doc[2])))
def test_get_compatibility_map_same_sentence_no_governance(self):
doc = self.sm_nlp(
'After Richard arrived, he said he was entering the big house')
self.sm_rules_analyzer.initialize(doc)
self.assertEqual(
'[4, 0, 0, 0.045050576, 5]',
str(
self.sm_tendencies_analyzer.get_compatibility_map(
Mention(doc[0], False), doc[4])))
def test_get_vectors_mention_with_head_sm(self):
doc = self.sm_nlp('He arrived')
self.sm_rules_analyzer.initialize(doc)
mention = Mention(doc[0], False)
vectors = self.sm_tendencies_analyzer.get_vectors(mention, doc)
self.assertTrue(vectors[0].any())
self.assertTrue(vectors[1].any())
self.assertEqual(len(vectors[0]), len(vectors[1]))
self.assertEqual(vectors, mention.temp_vectors)
def test_get_cosine_similarity_lg_root_1(self):
doc = self.lg_nlp('Richard. He said he was entering the big house')
self.lg_rules_analyzer.initialize(doc)
self.assertEqual(
'[2, 1, 0, -1, 1]',
str(
self.lg_tendencies_analyzer.get_compatibility_map(
Mention(doc[0], False), doc[2])))
def test_get_compatibility_map_different_sentences(self):
doc = self.sm_nlp(
'Richard called. He said he was entering the big house')
self.sm_rules_analyzer.initialize(doc)
self.assertEqual(
'[3, 1, 0, 0.507851, 6]',
str(
self.sm_tendencies_analyzer.get_compatibility_map(
Mention(doc[0], False), doc[3])))
def test_get_compatibility_map_coordination(self):
doc = self.sm_nlp(
'Richard and Jane said he was entering the big house')
self.sm_rules_analyzer.initialize(doc)
self.assertEqual(
'[4, 0, 1, 0.23422348, 3]',
str(
self.sm_tendencies_analyzer.get_compatibility_map(
Mention(doc[0], True), doc[4])))
def test_get_cosine_similarity_lg(self):
doc = self.lg_nlp(
'After Richard arrived, he said he was entering the big house')
self.lg_rules_analyzer.initialize(doc)
self.assertEqual(
'[4, 0, 0, 0.3336621, 5]',
str(
self.lg_tendencies_analyzer.get_compatibility_map(
Mention(doc[0], False), doc[4])))
def test_get_feature_map_simple_mention(self):
doc = self.sm_nlp('Richard said he was entering the big house')
self.sm_rules_analyzer.initialize(doc)
mention = Mention(doc[0], False)
feature_map = self.sm_tendencies_analyzer.get_feature_map(mention, doc)
self.assertEqual(len(self.sm_feature_table), len(feature_map))
self.assertEqual(mention.temp_feature_map, feature_map)
self.assertEqual([
0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 1, 0,
0, 1, 1, 0
], feature_map)
feature_map = self.sm_tendencies_analyzer.get_feature_map(
Mention(doc[2], False), doc)
self.assertEqual(len(self.sm_feature_table), len(feature_map))
self.assertEqual([
0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 1, 0, 1,
1, 1, 0, 1
], feature_map)
def test_get_compatibility_map_same_sentence_lefthand_sibling_governance(
self):
doc = self.lg_nlp(
'Richard said Peter and he were entering the big house')
self.lg_rules_analyzer.initialize(doc)
self.assertEqual(
'[4, 0, 1, 0.15999001, 3]',
str(
self.sm_tendencies_analyzer.get_compatibility_map(
Mention(doc[0], False), doc[4])))
def test_get_compatibility_map_same_sentence_lefthand_sibling_no_governance(
self):
doc = self.sm_nlp(
'After Richard arrived, Peter and he said he was entering the big house'
)
self.sm_rules_analyzer.initialize(doc)
self.assertEqual(
'[5, 0, 0, 0.41453412, 1]',
str(
self.sm_tendencies_analyzer.get_compatibility_map(
Mention(doc[1], False), doc[6])))
def test_object_access(self):
doc = self.sm_nlp(
'I saw Peter. Richard and he came in. They had arrived')
self.assertEqual('1: [4, 6], [10]', str(doc._.coref_chains[1]))
self.assertEqual(2, len(doc._.coref_chains))
self.assertEqual(2, len(doc._.coref_chains[1]))
self.assertEqual('[4, 6]', str(doc._.coref_chains[1][0]))
self.assertEqual(2, len(doc._.coref_chains[1][0]))
self.assertEqual(4, doc._.coref_chains[1][0][0])
found = False
for chain in doc._.coref_chains:
for mention in chain.mentions:
if mention == Mention(doc[10], False):
found = True
self.assertTrue(found)
def func(nlp):
if nlp.meta['name'] in excluded_nlps:
return
doc = nlp(doc_text)
rules_analyzer = RulesAnalyzerFactory.get_rules_analyzer(nlp)
rules_analyzer.initialize(doc)
assert rules_analyzer.is_independent_noun(doc[referred_index]) or \
rules_analyzer.is_potential_anaphor(doc[referred_index])
assert rules_analyzer.is_potential_anaphor(doc[referring_index])
referred_mention = Mention(doc[referred_index], include_dependent_siblings)
if consider_syntax:
self.assertEqual(expected_truth,
rules_analyzer.language_independent_is_potential_anaphoric_pair(
referred_mention, doc[referring_index]), nlp.meta['name'])
else:
self.assertEqual(expected_truth, rules_analyzer.is_potential_anaphoric_pair(
referred_mention, doc[referring_index], False), nlp.meta['name'])