def test_search():
"""Test that searching for a longform in the trie works correctly"""
rec = AdeftRecognizer('ER', grounding_map)
example = [
'for', 'women', ',', 'mandatory', 'hmo', 'programs', 'reduce', 'some',
'types', 'of', 'non', 'emergency', 'room'
]
assert rec._search(example) == 'emergency room'
def test_recognizer():
"""Test the recognizer end to end"""
rec = AdeftRecognizer('ER', grounding_map)
for text, result in [example1, example2, example3, example4, example5]:
longform = rec.recognize(text)
assert longform.pop() == result
# Case where defining pattern appears at the start of the fragment
assert not rec.recognize('(ER) stress')
def test_recognizer():
"""Test the recognizer end to end"""
rec = AdeftRecognizer('ER', grounding_map)
for text, expected in [example1, example2, example3, example4, example5]:
result = rec.recognize(text)
assert result.pop()['grounding'] == expected
# Case where defining pattern appears at the start of the fragment
assert not rec.recognize('(ER) stress')
def test_strip_defining_patterns():
rec = AdeftRecognizer('ER', grounding_map)
test_cases = [
'The endoplasmic reticulum (ER) is a transmembrane',
'The endoplasmic reticulum (ER)is a transmembrane',
'The endoplasmic reticulum (ER)-is a transmembrane'
]
results = (['The ER is a transmembrane'] * 2 +
['The ER -is a transmembrane'])
for case, result in zip(test_cases, results):
assert rec.strip_defining_patterns(case) == result
null_case = 'Newly developed extended release (ER) medications'
null_result = 'Newly developed extended release ER medications'
assert rec.strip_defining_patterns(null_case) == null_result
def __init__(self, classifier, grounding_dict, names):
self.classifier = classifier
self.shortforms = classifier.shortforms
self.recognizers = [AdeftRecognizer(shortform,
grounding_map)
for shortform,
grounding_map in grounding_dict.items()]
self.grounding_dict = grounding_dict
self.names = names
self.labels = set(value for grounding_map in grounding_dict.values()
for value in grounding_map.values())
self.pos_labels = classifier.pos_labels
def test_init():
"""Test that the recognizers internal trie is initialized correctly"""
rec = AdeftRecognizer('ER', grounding_map)
trie = rec._trie
for longform, grounding in grounding_map.items():
edges = tuple(_stemmer.stem(token)
for token, _ in tokenize(longform))[::-1]
current = trie
for index, token in enumerate(edges):
assert token in current.children
if index < len(edges) - 1:
assert current.children[token].longform is None
else:
assert current.children[token].longform == longform
current = current.children[token]
def test_exclude():
"""Test that using excluded words works"""
rec = AdeftRecognizer('ER', grounding_map, exclude=['emergency'])
assert not rec.recognize(example3[0])
def __init__(self, grounding_dict):
self.grounding_dict = grounding_dict
self.recognizers = [
AdeftRecognizer(shortform, grounding_map)
for shortform, grounding_map in grounding_dict.items()
]