def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.sentiment_comparison)
# Make sure the required NLTK data files are downloaded
for function in utils.get_initialization_functions(
self.search_algorithm, 'compare_statements').values():
function()
class SearchComparisonFunctionSpacySimilarityTests(ChatBotTestCase):
"""
Test that the search algorithm works correctly with the
spacy similarity comparison function.
"""
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.spacy_similarity)
def test_get_closest_statement(self):
"""
Note, the content of the in_response_to field for each of the
test statements is only required because the logic adapter will
filter out any statements that are not in response to a known statement.
"""
self.chatbot.storage.create_many([
Statement(text='This is a lovely bog.',
in_response_to='This is a lovely bog.'),
Statement(text='This is a beautiful swamp.',
in_response_to='This is a beautiful swamp.'),
Statement(text='It smells like a swamp.',
in_response_to='It smells like a swamp.')
])
statement = Statement(text='This is a lovely swamp.')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 2)
results_text = [result.text for result in results]
self.assertIn('This is a lovely bog.', results_text)
self.assertIn('This is a beautiful swamp.', results_text)
self.assertGreater(results[0].confidence, 0)
self.assertGreater(results[1].confidence, 0)
def test_different_punctuation(self):
self.chatbot.storage.create_many([
Statement(text='Who are you?'),
Statement(text='Are you good?'),
Statement(text='You are good')
])
statement = Statement(text='Are you good')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].text, 'Are you good?')
class SearchComparisonFunctionSentimentComparisonTests(ChatBotTestCase):
"""
Test that the search algorithm works correctly with the
sentiment comparison function by using the similarity
of sentiment polarity.
"""
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.sentiment_comparison
)
# Make sure the required NLTK data files are downloaded
for function in utils.get_initialization_functions(
self.search_algorithm,
'compare_statements'
):
function()
def test_exact_input(self):
self.chatbot.storage.create(text='What is your favorite flavor of ice cream?')
self.chatbot.storage.create(text='I enjoy raspberry ice cream.')
self.chatbot.storage.create(text='I am glad to hear that.')
self.chatbot.storage.create(text='Thank you.')
happy_statement = Statement(text='I enjoy raspberry ice cream.')
results = list(self.search_algorithm.search(happy_statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].text, 'I enjoy raspberry ice cream.')
self.assertEqual(results[0].confidence, 1)
def test_close_input(self):
self.chatbot.storage.create(text='What is your favorite flavor of ice cream?')
self.chatbot.storage.create(text='I enjoy raspberry ice cream.')
self.chatbot.storage.create(text='I am glad to hear that.')
self.chatbot.storage.create(text='Thank you, what is yours?')
self.chatbot.storage.create(text='Mine is chocolate.')
happy_statement = Statement(text='I enjoy raspberry.')
results = list(self.search_algorithm.search(happy_statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].text, 'I enjoy raspberry ice cream.')
self.assertAlmostEqual(results[0].confidence, 0.75, places=1)
class SearchComparisonFunctionSpacySimilarityTests(ChatBotTestCase):
"""
Test that the search algorithm works correctly with the
spacy similarity comparison function.
"""
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.SpacySimilarity
)
def test_get_closest_statement(self):
"""
Note, the content of the in_response_to field for each of the
test statements is only required because the logic adapter will
filter out any statements that are not in response to a known statement.
"""
self.chatbot.storage.create_many([
Statement(text='This is a lovely bog.', in_response_to='This is a lovely bog.'),
Statement(text='This is a beautiful swamp.', in_response_to='This is a beautiful swamp.'),
Statement(text='It smells like a swamp.', in_response_to='It smells like a swamp.')
])
statement = Statement(text='This is a lovely swamp.')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].text, 'This is a beautiful swamp.')
self.assertGreater(results[0].confidence, 0)
def test_different_punctuation(self):
self.chatbot.storage.create_many([
Statement(text='Who are you?'),
Statement(text='Are you good?'),
Statement(text='You are good')
])
statement = Statement(text='Are you good')
results = list(self.search_algorithm.search(statement))
self.assertEqual(len(results), 2, msg=[r.search_text for r in results])
# Note: the last statement in the list always has the highest confidence
self.assertEqual(results[-1].text, 'Are you good?')
class SearchTestCase(ChatBotTestCase):
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(self.chatbot)
def test_search_no_results(self):
"""
An exception should be raised if there is no data to return.
"""
statement = Statement(text='What is your quest?')
with self.assertRaises(StopIteration):
next(self.search_algorithm.search(statement))
def test_search_cast_to_list_no_results(self):
"""
An empty list should be returned when the generator is
cast to a list and there are no results to return.
"""
statement = Statement(text='What is your quest?')
results = list(self.search_algorithm.search(statement))
self.assertEqual(results, [])
def test_search_additional_parameters(self):
"""
It should be possible to pass additional parameters in to use for searching.
"""
self.chatbot.storage.create_many([
Statement(text='A', conversation='test_1'),
Statement(text='A', conversation='test_2')
])
statement = Statement(text='A')
results = list(self.search_algorithm.search(
statement, conversation='test_1'
))
self.assertIsLength(results, 1)
self.assertEqual(results[0].text, 'A')
self.assertEqual(results[0].conversation, 'test_1')
class SearchTestCase(ChatBotTestCase):
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(self.chatbot)
def test_search_no_results(self):
"""
An exception should be raised if there is no data to return.
"""
statement = Statement(text='What is your quest?')
with self.assertRaises(StopIteration):
next(self.search_algorithm.search(statement))
def test_search_cast_to_list_no_results(self):
"""
An empty list should be returned when the generator is
cast to a list and there are no results to return.
"""
statement = Statement(text='What is your quest?')
results = list(self.search_algorithm.search(statement))
self.assertEqual(results, [])
def test_search_additional_parameters(self):
"""
It should be possible to pass additional parameters in to use for searching.
"""
self.chatbot.storage.create_many([
Statement(text='A', conversation='test_1'),
Statement(text='A', conversation='test_2')
])
statement = Statement(text='A')
results = list(self.search_algorithm.search(
statement, conversation='test_1'
))
self.assertIsLength(results, 1)
self.assertEqual(results[0].text, 'A')
self.assertEqual(results[0].conversation, 'test_1')
def __init__(self, name, **kwargs):
self.name = name
storage_adapter = kwargs.get('storage_adapter', 'chatterbot.storage.SQLStorageAdapter')
logic_adapters = kwargs.get('logic_adapters', [
'chatterbot.logic.BestMatch'
])
# Check that each adapter is a valid subclass of it's respective parent
utils.validate_adapter_class(storage_adapter, StorageAdapter)
# Logic adapters used by the chat bot
self.logic_adapters = []
self.storage = utils.initialize_class(storage_adapter, **kwargs)
primary_search_algorithm = IndexedTextSearch(self, **kwargs)
text_search_algorithm = TextSearch(self, **kwargs)
self.search_algorithms = {
primary_search_algorithm.name: primary_search_algorithm,
text_search_algorithm.name: text_search_algorithm
}
for adapter in logic_adapters:
utils.validate_adapter_class(adapter, LogicAdapter)
logic_adapter = utils.initialize_class(adapter, self, **kwargs)
self.logic_adapters.append(logic_adapter)
preprocessors = kwargs.get(
'preprocessors', [
'chatterbot.preprocessors.clean_whitespace'
]
)
self.preprocessors = []
self.postprocessors = []
for preprocessor in preprocessors:
self.preprocessors.append(utils.import_module(preprocessor))
postprocessors = kwargs.get(
'postprocessors', [
'chatterbot.postprocessors.joint_sentence'
]
)
for postprocessor in postprocessors:
self.postprocessors.append(utils.import_module(postprocessor))
self.logger = kwargs.get('logger', logging.getLogger(__name__))
# Allow the bot to save input it receives so that it can learn
self.read_only = kwargs.get('read_only', False)
def __init__(self, name, **kwargs):
self.name = name
self.logger = kwargs.get(
'logger',
log.Logger('chatterbot',
costum_format='[%Y-%m-%d %H:%M:%S] chatterbot',
debug=kwargs.get('debug', False)))
storage_adapter = kwargs.get('storage_adapter',
'chatterbot.storage.SQLStorageAdapter')
logic_adapters = kwargs.get('logic_adapters',
['chatterbot.logic.BestMatch'])
# Check that each adapter is a valid subclass of it's respective parent
utils.validate_adapter_class(storage_adapter, StorageAdapter)
# Logic adapters used by the chat bot
self.logic_adapters = []
self.storage = utils.initialize_class(storage_adapter, **kwargs)
primary_search_algorithm = IndexedTextSearch(self, **kwargs)
text_search_algorithm = TextSearch(self, **kwargs)
self.search_algorithms = {
primary_search_algorithm.name: primary_search_algorithm,
text_search_algorithm.name: text_search_algorithm
}
for adapter in logic_adapters:
utils.validate_adapter_class(adapter, LogicAdapter)
logic_adapter = utils.initialize_class(adapter, self, **kwargs)
self.logic_adapters.append(logic_adapter)
preprocessors = kwargs.get(
'preprocessors', ['chatterbot.preprocessors.clean_whitespace'])
self.preprocessors = []
for preprocessor in preprocessors:
self.preprocessors.append(utils.import_module(preprocessor))
self.trainer = ChatterBotCorpusTrainer(self)
self.trainer.train("chatterbot.corpus.custom")
log.add_logger('chatterbot', self.logger)
self.this_file = os.path.dirname(__file__)
self.learn_YAML = os.path.join(
''.join(os.path.split(self.this_file)[0]), 'chatterbot_corpus',
'data', 'custom', 'myown.yml')
# Allow the bot to save input it receives so that it can learn
self.read_only = kwargs.get('read_only', False)
class SearchTestCase(ChatBotTestCase):
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(self.chatbot)
def test_search_no_results(self):
"""
An exception should be raised if there is no data to return.
"""
statement = Statement(text='What is your quest?')
with self.assertRaises(StopIteration):
next(self.search_algorithm.search(statement))
def test_search_cast_to_list_no_results(self):
"""
An empty list should be returned when the generator is
cast to a list and there are no results to return.
"""
statement = Statement(text='What is your quest?')
results = list(self.search_algorithm.search(statement))
self.assertEqual(results, [])
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.levenshtein_distance)
class SearchComparisonFunctionLevenshteinDistanceComparisonTests(
ChatBotTestCase):
"""
Test that the search algorithm works correctly with the
Levenshtein distance comparison function.
"""
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.levenshtein_distance)
def test_get_closest_statement(self):
"""
Note, the content of the in_response_to field for each of the
test statements is only required because the logic adapter will
filter out any statements that are not in response to a known statement.
"""
self.chatbot.storage.create_many([
Statement(text='Who do you love?',
in_response_to='I hear you are going on a quest?'),
Statement(text='What is the meaning of life?',
in_response_to='Yuck, black licorice jelly beans.'),
Statement(text='I am Iron Man.',
in_response_to='What... is your quest?'),
Statement(text='What... is your quest?',
in_response_to='I am Iron Man.'),
Statement(text='Yuck, black licorice jelly beans.',
in_response_to='What is the meaning of life?'),
Statement(text='I hear you are going on a quest?',
in_response_to='Who do you love?'),
])
statement = Statement(text='What is your quest?')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].text, 'What... is your quest?')
def test_confidence_exact_match(self):
self.chatbot.storage.create(text='What is your quest?',
in_response_to='What is your quest?')
statement = Statement(text='What is your quest?')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].confidence, 1)
def test_confidence_half_match(self):
from unittest.mock import MagicMock
# Assume that the storage adapter returns a partial match
self.chatbot.storage.filter = MagicMock(
return_value=[Statement(text='xxyy')])
statement = Statement(text='wwxx')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].confidence, 0.5)
def test_confidence_no_match(self):
from unittest.mock import MagicMock
# Assume that the storage adapter returns a partial match
self.search_algorithm.chatbot.storage.filter = MagicMock(
return_value=[Statement(text='xxx', in_response_to='xxx')])
statement = Statement(text='yyy')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 0)
class SearchComparisonFunctionLevenshteinDistanceComparisonTests(ChatBotTestCase):
"""
Test that the search algorithm works correctly with the
Levenshtein distance comparison function.
"""
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.LevenshteinDistance
)
def test_get_closest_statement(self):
"""
Note, the content of the in_response_to field for each of the
test statements is only required because the search process will
filter out any statements that are not in response to something.
"""
self.chatbot.storage.create_many([
Statement(text='What is the meaning of life?', in_response_to='...'),
Statement(text='I am Iron Man.', in_response_to='...'),
Statement(text='What... is your quest?', in_response_to='...'),
Statement(text='Yuck, black licorice jelly beans.', in_response_to='...'),
Statement(text='I hear you are going on a quest?', in_response_to='...'),
])
statement = Statement(text='What is your quest?')
results = list(self.search_algorithm.search(statement))
self.assertEqual(len(results), 1, msg=[r.text for r in results])
self.assertEqual(results[0].text, 'What... is your quest?')
def test_confidence_exact_match(self):
self.chatbot.storage.create(text='What is your quest?', in_response_to='What is your quest?')
statement = Statement(text='What is your quest?')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].confidence, 1)
def test_confidence_half_match(self):
from unittest.mock import MagicMock
# Assume that the storage adapter returns a partial match
self.chatbot.storage.filter = MagicMock(return_value=[
Statement(text='xxyy')
])
statement = Statement(text='wwxx')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 1)
self.assertEqual(results[0].confidence, 0.5)
def test_confidence_no_match(self):
from unittest.mock import MagicMock
# Assume that the storage adapter returns a partial match
self.search_algorithm.chatbot.storage.filter = MagicMock(return_value=[
Statement(text='xxx', in_response_to='xxx')
])
statement = Statement(text='yyy')
results = list(self.search_algorithm.search(statement))
self.assertIsLength(results, 0)
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(self.chatbot)
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.SpacySimilarity
)
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.spacy_similarity)
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(self.chatbot)
def setUp(self):
super().setUp()
self.search_algorithm = IndexedTextSearch(
self.chatbot,
statement_comparison_function=comparisons.LevenshteinDistance
)