class TestUtterance(unittest.TestCase):
"""Tests correct working of the Utterance class."""
def setUp(self):
self.barbara = Utterance('b a r b a r a')
self.ararat = Utterance('a r a r a t')
def test_index(self):
test_pairs = ((['b', 'a', 'r'], 0),
(['b', 'a', 'r', 'a'], 3),
(['a', 'r', 'a'], 4))
for phrase, idx in test_pairs:
self.assertEqual(self.barbara.index(phrase), idx)
self.assertRaises(ValueError, self.barbara.index, ['r', 'a', 'b'])
self.assertEqual(self.ararat.index(['a', 'r', 'a', 't']), 2)
def test_ngram_iterator(self):
# Normal use case.
trigrams = [['b', 'a', 'r'],
['a', 'r', 'b'],
['r', 'b', 'a'],
['b', 'a', 'r'],
['a', 'r', 'a'],
]
trigrams_with_boundaries = [
[SENTENCE_START, 'b', 'a'],
['b', 'a', 'r'],
['a', 'r', 'b'],
['r', 'b', 'a'],
['b', 'a', 'r'],
['a', 'r', 'a'],
['r', 'a', SENTENCE_END],
]
act_trigrams = list(self.barbara.iter_ngrams(3))
act_trigrams_with_boundaries = list(
self.barbara.iter_ngrams(3, with_boundaries=True))
self.assertItemsEqual(trigrams, act_trigrams)
self.assertItemsEqual(trigrams_with_boundaries,
act_trigrams_with_boundaries)
# Corner cases.
self.assertItemsEqual(list(self.barbara.iter_ngrams(7)),
[['b', 'a', 'r', 'b', 'a', 'r', 'a']])
self.assertItemsEqual(list(self.barbara.iter_ngrams(8)), [])
self.assertItemsEqual(
list(self.barbara.iter_ngrams(8, with_boundaries=True)),
[[SENTENCE_START, 'b', 'a', 'r', 'b', 'a', 'r', 'a'],
['b', 'a', 'r', 'b', 'a', 'r', 'a', SENTENCE_END]])
self.assertItemsEqual(
list(self.barbara.iter_ngrams(9, with_boundaries=True)),
[[SENTENCE_START, 'b', 'a', 'r', 'b', 'a', 'r', 'a',
SENTENCE_END]])
self.assertItemsEqual(
list(self.barbara.iter_ngrams(10, with_boundaries=True)),
[])
