def __init__(self, features, tolerant=False, sparse=True):
# Upgrade `features` to `Feature` instances.
features = list(map(make_feature, features))
if tolerant:
self.evaluator = TolerantFeatureEvaluator(features)
else:
self.evaluator = FeatureEvaluator(features)
self.flattener = FeatureMappingFlattener(sparse=sparse)
def test_returns_tuples_of_features_length(self):
features = [DumbFeatureA, DumbFeatureA]
ev = FeatureEvaluator(features)
ev.fit(SAMPLES)
Xt = ev.transform(SAMPLES)
x = next(Xt)
self.assertIsInstance(x, tuple)
self.assertEqual(len(x), len(features))
def test_fit_transform_does_both_things(self):
ev = FeatureEvaluator([DumbFeatureA])
Xt_1 = ev.fit_transform(SAMPLES)
Xt_2 = ev.transform(SAMPLES)
self.assertListEqual(list(Xt_1), list(Xt_2))
def test_returns_as_many_tuples_as_samples(self):
ev = FeatureEvaluator([DumbFeatureA])
ev.fit(SAMPLES)
Xt = ev.transform(SAMPLES)
self.assertEqual(len(list(Xt)), len(SAMPLES))
def test_returns_generator(self):
ev = FeatureEvaluator([DumbFeatureA])
ev.fit([])
Xt = ev.transform([])
self.assertIsInstance(Xt, types.GeneratorType)
def test_fit_creates_alive_features_tuple(self):
ev = FeatureEvaluator([DumbFeatureA])
self.assertFalse(hasattr(ev, 'alive_features'))
ev.fit([])
self.assertTrue(hasattr(ev, 'alive_features'))