def __init__(self,
dictionary_loader: DictionaryLoader,
language="en",
filters: List[AnnotationFilter] = None):
"""
Parameters
----------
dictionary_loader: DictionaryLoader
The dictionary loader that will provide the dictionary contents
language: str
The language of the text that will processed (affects the choice of tokenner and stemmer).
filters: List[AnnotationFilter]
A list of filters to apply post recognition
"""
super().__init__(dictionary_loader, language=language, filters=filters)
self.punctuation_remove = regex.compile(
r'[\p{C}|\p{M}|\p{P}|\p{S}|\p{Z}]+', regex.UNICODE)
self.label_concept_index = {}
self.label_token_counts = {}
self.label_lengths = {}
self.trie = Trie()
if language == 'en':
import en_core_web_md
self.spacy = en_core_web_md.load()
elif language == 'fr':
import fr_core_web_md
self.spacy = fr_core_web_md.load()
else:
raise ValueError(f"Unsupported language: {language}")
def test_word_add_all_set(self):
self.trie = Trie()
self.trie.add_all({'axe', 'kick'}) #set
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('axe' in self.trie, "Word should be in trie")
self.assertTrue('kick' in self.trie, "Word should be in trie")
self.assertEqual(2, self.trie.get_word_count(), "Word count not equal")
def test_word_add_all_with_number(self):
self.trie = Trie()
self.trie.add_all(('axe', 'kick', 3)) #tuple with one integer.
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('axe' in self.trie, "Word should be in trie")
self.assertTrue('kick' in self.trie, "Word should be in trie")
self.assertEqual(2, self.trie.get_word_count(), "Word count not equal")
def test_trie_node_count(self):
self.trie = Trie()
self.trie.add_all(['ash', 'ashley'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertEqual(2, self.trie.get_word_count(), "Word count not equal")
self.assertEqual(7, len(self.trie), "Number of nodes")
def test_trie_wildcard_exception(self):
self.trie = Trie()
self.trie.add_all(['ab', 'as', 'ash', 'ashley'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertRaises(InvalidWildCardExpressionError, self.trie.search,
'#$%^a')
def test_trie_question_search(self):
self.trie = Trie()
self.trie.add_all(['ab', 'as', 'ash', 'ashley'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertEqual(sorted(self.trie.search('a?')), sorted(['ab', 'as']),
'The lists should be equal')
class TestTrieExactWordSearch(unittest.TestCase):
def test_word_in_trie(self):
self.trie = Trie()
self.trie.add_all(['ash', 'ashley'])
self.assertTrue('ash' in self.trie, "Word should be in trie")
def test_word_not_int_trie(self):
self.trie = Trie()
self.trie.add_all(['ash', 'ashley'])
self.assertFalse('salary' in self.trie, "Word should not be in trie")
def test_trie_prefix_search(self):
self.trie = Trie()
self.trie.add_all(['ashlame', 'ashley', 'askoiu', 'ashlo'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertFalse('ash' in self.trie, "Word should not be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertEqual(4, self.trie.get_word_count(), "Word count not equal")
self.assertTrue(self.trie.contains_prefix('ash'),
"Prefix should be present in Trie")
self.assertEqual(sorted(self.trie.search_with_prefix('ash')),
sorted(['ashlame', 'ashley', 'ashlo']),
'The lists should be equal')
def test_trie_node_prefix_not_exists(self):
self.trie = Trie()
self.trie.add_all(['ash', 'ashley'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertEqual(2, self.trie.get_word_count(), "Word count not equal")
self.assertFalse(self.trie.contains_prefix('xmas'),
"Prefix should be present in Trie")
self.assertFalse(self.trie.contains_prefix('xor'),
"Prefix should be present in Trie")
self.assertFalse(self.trie.contains_prefix('sh'),
"Prefix should be present in Trie")
def test_word_add_all_gen(self):
def gen_words():
a = ['ash', 'ashley', 'simpson']
for word in a:
yield word
self.trie = Trie()
self.trie.add_all(gen_words()) # generator
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertTrue('simpson' in self.trie, "Word should be in trie")
self.assertEqual(3, self.trie.get_word_count(), "Word count not equal")
def __build_tries(dataset_id: int, user_id: int) -> DatasetDictionary:
"""Creates lookup tries from levels and segments"""
col_unique_values = get_column_unique_values(dataset_id, user_id)
all_values = __condense_and_lemmatize(col_unique_values)
level_trie = Trie()
level_trie.add_all(all_values)
colname_trie = Trie()
colname_trie.add_all(list(col_unique_values.keys()))
wordcount = colname_trie.get_word_count() + level_trie.get_word_count()
return DatasetDictionary(dataset_id, colname_trie, level_trie, wordcount)
class TestWordCount(unittest.TestCase):
def test_word_count_greater_than_zero(self):
self.trie = Trie()
self.trie.add_all(['ash', 'ashley', 'ashes'])
self.assertGreater(self.trie.get_word_count(), 0,
"The number of words should be greater than 0")
self.assertEqual(3, self.trie.get_word_count(), "Word count not equal")
def test_word_count_zero(self):
self.trie = Trie()
self.trie.add_all([])
self.assertEqual(0, self.trie.get_word_count(), "Word count not equal")
def test_without_count(self):
trie = Trie()
trie.add_all(['ash', 'ashley', 'ashes', 'ashes'])
expected = ['ash', 'ashley', 'ashes']
self.assertListEqual(expected, trie.search('a*'))
class TrieApproxRecognizer(ConceptRecognizer):
def __init__(self,
dictionary_loader: DictionaryLoader,
language="en",
filters: List[AnnotationFilter] = None):
"""
Parameters
----------
dictionary_loader: DictionaryLoader
The dictionary loader that will provide the dictionary contents
language: str
The language of the text that will processed (affects the choice of tokenner and stemmer).
filters: List[AnnotationFilter]
A list of filters to apply post recognition
"""
super().__init__(dictionary_loader, language=language, filters=filters)
self.punctuation_remove = regex.compile(
r'[\p{C}|\p{M}|\p{P}|\p{S}|\p{Z}]+', regex.UNICODE)
self.label_concept_index = {}
self.label_token_counts = {}
self.label_lengths = {}
self.trie = Trie()
if language == 'en':
import en_core_web_md
self.spacy = en_core_web_md.load()
elif language == 'fr':
import fr_core_web_md
self.spacy = fr_core_web_md.load()
else:
raise ValueError(f"Unsupported language: {language}")
def _load_concept_labels(self, concept_id, labels):
label_index = 0
for label in labels:
normalized = self.punctuation_remove.sub(" ", label).replace(
"-", " ").lower()
tokens, _ = span_tokenize(self.spacy, normalized)
# For each token
key = str(concept_id) + ":::" + str(label_index)
self.label_concept_index[normalized] = key
self.label_token_counts[normalized] = len(tokens)
self.label_lengths[normalized] = len(normalized)
self.trie.add(normalized, count=1)
label_index += 1
def match_mentions(
self, input_text
) -> Tuple[List[Tuple[int, int]], List[str], Set[Annotation]]:
normalized_text = self.punctuation_remove.sub(" ", input_text).replace(
"-", " ").lower()
matches = []
tokens, spans = span_tokenize(self.spacy, normalized_text)
current_token_index = 0
while current_token_index < len(tokens):
current_match_cursor = 0
while current_token_index + current_match_cursor < len(tokens):
sub_string = normalized_text[spans[current_token_index][0]:
spans[current_match_cursor][1]]
found = self.trie.search_within_distance(sub_string, dist=2)
if len(found) > 0:
# Register match
print(len(found))
current_match_cursor += 1
else:
break
current_token_index += 1
return [], [], set()
def test_word_add(self):
self.trie = Trie()
self.trie.add('axe')
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('axe' in self.trie, "Word should be in trie")
class TesTrieWordInsert(unittest.TestCase):
def test_word_add(self):
self.trie = Trie()
self.trie.add('axe')
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('axe' in self.trie, "Word should be in trie")
def test_word_add_all_list(self):
self.trie = Trie()
self.trie.add_all(['axe', 'kick']) #list
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('axe' in self.trie, "Word should be in trie")
self.assertTrue('kick' in self.trie, "Word should be in trie")
self.assertEqual(2, self.trie.get_word_count(), "Word count not equal")
def test_word_add_all_set(self):
self.trie = Trie()
self.trie.add_all({'axe', 'kick'}) #set
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('axe' in self.trie, "Word should be in trie")
self.assertTrue('kick' in self.trie, "Word should be in trie")
self.assertEqual(2, self.trie.get_word_count(), "Word count not equal")
def test_word_add_all_tuple(self):
self.trie = Trie()
self.trie.add_all(('axe', 'kick')) #tuple
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('axe' in self.trie, "Word should be in trie")
self.assertTrue('kick' in self.trie, "Word should be in trie")
self.assertEqual(2, self.trie.get_word_count(), "Word count not equal")
def test_word_add_all_with_number(self):
self.trie = Trie()
self.trie.add_all(('axe', 'kick', 3)) #tuple with one integer.
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('axe' in self.trie, "Word should be in trie")
self.assertTrue('kick' in self.trie, "Word should be in trie")
self.assertEqual(2, self.trie.get_word_count(), "Word count not equal")
def test_word_add_all_gen(self):
def gen_words():
a = ['ash', 'ashley', 'simpson']
for word in a:
yield word
self.trie = Trie()
self.trie.add_all(gen_words()) # generator
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertTrue('simpson' in self.trie, "Word should be in trie")
self.assertEqual(3, self.trie.get_word_count(), "Word count not equal")
def test_word_add_all_file_path(self):
self.trie = Trie()
self.trie.add_all(small_dataset) # From a file
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertTrue('simpson' in self.trie, "Word should be in trie")
self.assertEqual(8, self.trie.get_word_count(), "Word count not equal")
def test_word_not_int_trie(self):
self.trie = Trie()
self.trie.add_all(['ash', 'ashley'])
self.assertFalse('salary' in self.trie, "Word should not be in trie")
def test_word_in_trie(self):
self.trie = Trie()
self.trie.add_all(['ash', 'ashley'])
self.assertTrue('ash' in self.trie, "Word should be in trie")
def test_word_count_zero(self):
self.trie = Trie()
self.trie.add_all([])
self.assertEqual(0, self.trie.get_word_count(), "Word count not equal")
class TestWildCardSearch(unittest.TestCase):
def test_trie_asterisk_search(self):
self.trie = Trie()
self.trie.add_all(['ash', 'ashley'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertEqual(sorted(self.trie.search('a*')),
sorted(['ash', 'ashley']),
'The lists should be equal')
def test_trie_question_search(self):
self.trie = Trie()
self.trie.add_all(['ab', 'as', 'ash', 'ashley'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertEqual(sorted(self.trie.search('a?')), sorted(['ab', 'as']),
'The lists should be equal')
def test_trie_wildcard_search(self):
self.trie = Trie()
self.trie.add_all(['ab', 'as', 'ash', 'ashley'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertEqual(sorted(self.trie.search('*a******?')),
sorted(['ab', 'as', 'ash', 'ashley']),
'The lists should be equal')
def test_trie_wildcard_exception(self):
self.trie = Trie()
self.trie.add_all(['ab', 'as', 'ash', 'ashley'])
self.assertIsInstance(self.trie, Trie,
"Object should be of type `lexpy.trie.Trie`")
self.assertTrue('ash' in self.trie, "Word should be in trie")
self.assertTrue('ashley' in self.trie, "Word should be in trie")
self.assertRaises(InvalidWildCardExpressionError, self.trie.search,
'#$%^a')
def test_with_count(self):
trie = Trie()
trie.add_all(['ash', 'ashley', 'ashes', 'ashes'])
expected = [('ash', 1), ('ashley', 1), ('ashes', 2)]
self.assertListEqual(expected, trie.search('a*', with_count=True))
