def test_multi_sentence_anaphora_feminine_names(self):
text = "Jane is happy. She is a carpenter"
discourse = Discourse(text)
extractor = Extractor(discourse, _knowledge)
triplets = extractor.extract()
expected_triplets = [('Jane_0', 'HAS_ROLE', 'carpenter')]
self.assertEqual(triplets, expected_triplets)
def test_pronoun_coreference(self):
sentence = 'John drove home where he has a cat.'
discourse = Discourse(sentence)
extractor = Extractor(discourse, _knowledge)
triplets = extractor.extract()
expected_triplets = [('John_0', 'OWN', 'cat')]
self.assertEqual(triplets, expected_triplets)
def test_multiple_people_same_rule(self):
text = "John is ginger. He is a carpenter. test. Jane is blond. She is a carpenter. "
discourse = Discourse(text)
extractor = Extractor(discourse, _knowledge)
triplets = extractor.extract()
expected_triplets = [('John_0', 'HAS_ROLE', 'carpenter'),
('Jane_1', 'HAS_ROLE', 'carpenter')]
self.assertTrue(triplets, expected_triplets)
def get_triplets():
if request.method != 'POST':
return []
data = json.loads(request.data)
text = data['text']
discourse = Discourse(text)
extractor = Extractor(discourse, knowledge)
triplets = extractor.extract()
return jsonify(transform_triplets_into_api_edges_and_nodes(triplets))
def test_multiple_matcing(self):
text = "John Smith is blond. He is a carpenter. There is no reason to panic. Sarah Doe is ginger. She is a carpenter."
discourse = Discourse(text)
extractor = Extractor(discourse, _knowledge)
triplets = extractor.extract()
expected_triplets = [('John_Smith_0', 'HAS_BLOND_ROLE', 'carpenter'),
('Sarah_Doe_1', 'HAS_GINGER_ROLE', 'carpenter'),
('John_0', 'HAS_ROLE', 'carpenter'),
('Jane_1', 'HAS_ROLE', 'carpenter')]
self.assertTrue(triplets, expected_triplets)
def test_asimov_wiki(self):
text = open(os.path.join(_path, '../data/wiki_asimov.txt')).read()
discourse = Discourse(text)
extractor = Extractor(discourse, _knowledge)
triplets = extractor.extract()
expected_triplets = [('Isaac_Asimov_0|Asimov_0', 'JOB_TITLE', 'writer'),
('Isaac_Asimov_0|Asimov_0', 'OWNS', 'works'),
('Isaac_Asimov_0|Asimov_0', 'OWNS', 'books')]
self.assertTrue(triplets, expected_triplets)
expected_drs = Drs.create_from_predicates_string(
"{}(a), {'word': 'Boston_University'}(b), {'type': 'at'}(a,b)")
lst = discourse._discourse.apply(DrsMatcher(expected_drs, metric))
is_match = len(lst) > 1
self.assertTrue(is_match)
import os
from pynsett.discourse import Discourse
from pynsett.extractor import Extractor
from pynsett.knowledge import Knowledge
_path = os.path.dirname(__file__)
text = "Jane was born on 10 August 1979."
knowledge = Knowledge()
knowledge.add_rules(open(os.path.join(_path, '../rules/test.rules')).read())
discourse = Discourse(text)
extractor = Extractor(discourse, knowledge)
triplets = extractor.extract()
for triplet in triplets:
print(triplet)
import os
from pynsett.discourse import Discourse
from pynsett.extractor import Extractor
_path = os.path.dirname(__file__)
if __name__ == "__main__":
import time
from pynsett.auxiliary.prior_knowedge import get_generic_knowledge
text = open(os.path.join(_path, '../data/profile.txt')).read()
discourse = Discourse(text)
extractor = Extractor(discourse, get_generic_knowledge())
start = time.time()
triplets = extractor.extract()
end = time.time()
print('total_time', (end - start), 'for', len(discourse), 'sentences')
for triplet in triplets:
print(triplet)