def test__get_candidates_none(self):
recommender = Recommender(self.dpp_matrix, self.n_components)
X = np.array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1])
recommender.dpp_vector = X
candidates = recommender._get_candidates()
expected = None
self.assertEqual(expected, candidates)
def test__acquire_multiple(self):
recommender = Recommender(self.dpp_matrix, self.n_components)
predictions = np.array([1, 9, 9, 7])
acquired = recommender._acquire(predictions)
expected_1 = 2
expected_2 = 1
self.assertTrue(acquired == expected_1 or acquired == expected_2)
def test_predict_multiple_rankings(self):
# Run
recommender = Recommender(self.dpp_matrix, self.n_components)
# matching row is [1, 0, 0, 0, 0, 0, 0, 3, 0, 4, 6, 2, 5, 0, 8, 0]
recommender.matching_dataset = self.dpp_matrix[0]
indicies = [0, 1, 14, 13]
predictions = recommender.predict(indicies)
expected = [2, 1, 3, 1]
np.testing.assert_array_equal(predictions, expected)
def test__get_candidates_some(self):
recommender = Recommender(self.dpp_matrix, self.n_components)
X = np.array([0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0])
recommender.dpp_vector = X
expected = np.array([0, 1, 2, 4, 6, 8, 10, 11, 12, 14, 15])
candidates = recommender._get_candidates()
np.testing.assert_array_equal(
candidates,
expected,
)
def test_add_none(self, fit_mock):
recommender = Recommender(self.dpp_matrix, self.n_components)
recommender.add({})
expected_x = np.array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])
np.testing.assert_array_equal(recommender.dpp_vector, expected_x)
np.testing.assert_array_equal(
fit_mock.call_args[0][0],
expected_x.reshape(1, -1)
)
fit_mock.assert_called_once()
def test___init__(self, nmf_mock):
nmf_mock().fit_transform.return_value = self.dpp_ranked
# Run
recommender = Recommender(self.dpp_matrix, self.n_components)
np.testing.assert_array_equal(recommender.dpp_matrix, self.dpp_matrix)
np.testing.assert_array_equal(recommender.dpp_ranked, self.dpp_ranked)
assert recommender.n_components == self.n_components
assert recommender.matching_dataset is None
# assert dpp_vector has same number of entries as pipelines
assert recommender.dpp_vector.shape[0] == self.dpp_matrix.shape[1]
def test_propose_without_add(self, randint_mock):
index = 0
randint_mock.return_value = index
proposed = np.argmax(self.dpp_matrix[0]) # 14
recommender = Recommender(self.dpp_matrix, self.n_components)
np.testing.assert_array_equal(
recommender.matching_dataset,
self.dpp_matrix[index, :],
)
pipeline_index = recommender.propose()
randint_mock.assert_called_once_with(4) # 4 is self.dpp_matrix length
assert pipeline_index == proposed
def test_propose_done(self, predict_mock, get_candidates_mock):
# Set-up
recommender = Recommender(self.dpp_matrix, self.n_components)
get_candidates_mock.return_value = None
# Run
params = recommender.propose()
# Assert
expected_params = None
assert params == expected_params
def test__get_candidates_all(self):
recommender = Recommender(self.dpp_matrix, self.n_components)
X = np.array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])
recommender.X = X
candidates = recommender._get_candidates()
expected = np.array(
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]
)
np.testing.assert_array_equal(
candidates,
expected,
)
def test_fit(self, nmf_mock):
nmf_mock().fit_transform.return_value = self.dpp_ranked
n_components = 3
X = np.array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])
# Run
recommender = Recommender(self.dpp_matrix, n_components)
for i in range(self.dpp_ranked.shape[0]):
nmf_mock().transform.return_value = self.dpp_ranked[i, :]
recommender.fit(X)
np.testing.assert_array_equal(
recommender.matching_dataset,
self.dpp_matrix[i],
)
def test_propose_empty_add(self, randint_mock):
index = 0
randint_mock.return_value = index
proposed = np.argmax(self.dpp_matrix[0]) # 14
recommender = Recommender(self.dpp_matrix, self.n_components)
recommender.add({})
np.testing.assert_array_equal(
recommender.matching_dataset,
self.dpp_matrix[index, :],
)
pipeline_index = recommender.propose()
assert randint_mock.call_count == 2
assert pipeline_index == proposed
def test_propose_once(self, predict_mock, get_candidates_mock):
"""n == 1"""
# Set-up
recommender = Recommender(self.dpp_matrix, self.n_components)
get_candidates_mock.return_value = [1, 3, 7, 8]
predict_mock.return_value = np.array([1, 4, 2, 3])
# Run
pipeline_index = recommender.propose()
# Assert
expected_pipeline_index = 3
assert pipeline_index == expected_pipeline_index
def test___init__(self, nmf_mock, randint_mock):
# set-up
index = 1
nmf_mock().fit_transform.return_value = self.dpp_ranked
randint_mock.return_value = index
# run
recommender = Recommender(self.dpp_matrix, self.n_components)
# asserts
randint_mock.assert_called_once_with(4) # 4 is self.dpp_matrix length
np.testing.assert_array_equal(
recommender.matching_dataset,
self.dpp_matrix[index, :],
)
np.testing.assert_array_equal(recommender.dpp_matrix, self.dpp_matrix)
np.testing.assert_array_equal(recommender.dpp_ranked, self.dpp_ranked)
def test_add_twice(self, fit_mock):
recommender = Recommender(self.dpp_matrix, self.n_components)
recommender.add({1: 2, 3: 4, 5: 1})
expected_x_1 = np.array(
[0, 2, 0, 4, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
)
np.testing.assert_array_equal(recommender.dpp_vector, expected_x_1)
np.testing.assert_array_equal(
fit_mock.call_args[0][0],
expected_x_1.reshape(1, -1)
)
recommender.add({1: 1, 6: 3, 9: 4})
expected_x_2 = np.array(
[0, 1, 0, 4, 0, 1, 3, 0, 0, 4, 0, 0, 0, 0, 0, 0]
)
np.testing.assert_array_equal(recommender.dpp_vector, expected_x_2)
np.testing.assert_array_equal(
fit_mock.call_args[0][0],
expected_x_2.reshape(1, -1)
)
self.assertEqual(fit_mock.call_count, 2)
def test__acquire_last(self):
recommender = Recommender(self.dpp_matrix, self.n_components)
predictions = np.array([1, 3, 5, 7])
acquired = recommender._acquire(predictions)
expected = 3
self.assertEqual(acquired, expected)