def test_update(self):
"""Test _FeatureExtractor update method."""
hook = Hook(key='key', func=lambda: None)
_prep = _FeatureExtractor().update(hook)
self.assertTrue('key' in _prep.keys)
def test_hook(self):
"""Test Hook initialization and key handling."""
hook = Hook(key='key', func=lambda: 'test')
self.assertEqual(hook.key, 'key')
# w/o args
self.assertEqual(hook(), 'test')
# check invalid key
with pytest.raises(ValueError):
_ = Hook(key='key', func=lambda: None)
# hook with args
hook_args = Hook(key='key_', func=lambda x: x)
self.assertEqual(Hook.get_current_keys(), {'key', 'key_'})
self.assertEqual(hook_args.key, 'key_')
def test_init(self):
hook = Hook(key='label', func=lambda x: x)
attributes = ['project']
label_prep = LabelPreprocessor(feed_attributes=attributes, hook=hook)
self.assertIsInstance(label_prep, LabelPreprocessor)
self.assertIsInstance(label_prep._hook, Hook) # pylint: disable=protected-access
def test_feed(self):
"""Test _FeatureExtractor feed method."""
hook = Hook(key='key', func=lambda x: x)
_prep = _FeatureExtractor().update(hook)
# feed the extractor with skip=True
result = _prep.feed({'x': 'test'}, skip_unfed_hooks=True)
self.assertIsInstance(result, dict)
key, value = list(*result.items())
self.assertEqual(key, 'key')
self.assertEqual(value, 'test')
# feed and disable skip
with self.assertRaises(TypeError):
result = _prep.feed({'x': 'test'}, skip_unfed_hooks=False)
key, value = list(*result.items())
self.assertEqual(key, 'key')
self.assertEqual(value, 'test')
def test_fit_transform(self):
"""Test LabelPreprocessor `fit_transform` method."""
hook = Hook(key='label', func=lambda p, d: 'label')
attributes = ['project', 'description']
test_data = [TEST_CVE]
# prepare the data for label preprocessor
feed_prep = NVDFeedPreprocessor(attributes, skip_duplicity=True)
test_data = feed_prep.transform(X=test_data)
label_prep = LabelPreprocessor(feed_attributes=attributes,
output_attributes=['description'],
hook=hook)
result = label_prep.fit_transform(test_data)
# check not empty
self.assertTrue(any(result))
# check that correct label is returned by the Hook
label = result[0].label
self.assertEqual(label, 'label')
The cmp function must be of signature (a: int, b: int) -> bool
:param limit: int, limit to compare to (the `b` argument in `cmp`)
"""
word, _ = features[pos] # type: str
if cmp is None:
cmp = operator.gt
return cmp(len(word), limit)
# Feature Hooks
# -------------
has_uppercase_hook: Hook = Hook(key="has_uppercase", func=__has_uppercase)
"""Hook: Return whether the word in `tagged` contains uppercase letter."""
is_alnum_hook: Hook = Hook(key="is_alnum", func=__is_alnum)
"""Hook: Return whether the word contains only alphanumeric characters."""
ver_follows_hook: Hook = Hook(key="ver_follows", func=__ver_follows)
"""Hook: Return whether the given word is followed by a version string."""
ver_precedes_hook: Hook = Hook(key="ver_precedes", func=__ver_precedes)
"""Hook: Return whether the given word is preceded by a version string."""
ver_pos_hook: Hook = Hook(key="ver_pos", func=__ver_pos)
"""Hook: Return version position in the `tagged` w.r.t the given word."""
vendor_product_match_hook: Hook = Hook(key="vendor_product_match",
