def test_label_softclass():
# Given
problem_type = SOFTCLASS
input_labels = pd.Series([2, 4, 2, 2, 4, 1])
# Raise exception
with pytest.raises(NotImplementedError):
LabelCleaner.construct(problem_type=problem_type,
y=input_labels,
y_uncleaned=None)
def test_label_cleaner_binary():
# Given
problem_type = BINARY
input_labels_numpy = np.array(['l1', 'l2', 'l2', 'l1', 'l1', 'l2'])
input_labels = pd.Series(input_labels_numpy)
input_labels_category = input_labels.astype('category')
input_labels_with_shifted_index = input_labels.copy()
input_labels_with_shifted_index.index += 5
input_labels_new = np.array(['new', 'l1', 'l2'])
expected_output_labels = pd.Series([0, 1, 1, 0, 0, 1])
expected_output_labels_new = pd.Series([np.nan, 0, 1])
expected_output_labels_new_inverse = pd.Series([np.nan, 'l1', 'l2'])
# When
label_cleaner = LabelCleaner.construct(problem_type=problem_type,
y=input_labels)
# Raise exception
with pytest.raises(AssertionError):
LabelCleaner.construct(problem_type=problem_type, y=input_labels_new)
# Then
assert isinstance(label_cleaner, LabelCleanerBinary)
assert label_cleaner.problem_type_transform == BINARY
assert label_cleaner.cat_mappings_dependent_var == {0: 'l1', 1: 'l2'}
output_labels = label_cleaner.transform(input_labels)
output_labels_with_numpy = label_cleaner.transform(input_labels_numpy)
output_labels_category = label_cleaner.transform(input_labels_category)
output_labels_with_shifted_index = label_cleaner.transform(
input_labels_with_shifted_index)
output_labels_new = label_cleaner.transform(input_labels_new)
output_labels_inverse = label_cleaner.inverse_transform(output_labels)
output_labels_with_shifted_index_inverse = label_cleaner.inverse_transform(
output_labels_with_shifted_index)
output_labels_new_inverse = label_cleaner.inverse_transform(
output_labels_new)
assert expected_output_labels.equals(output_labels)
assert expected_output_labels.equals(output_labels_with_numpy)
assert expected_output_labels.equals(output_labels_category)
assert not expected_output_labels.equals(output_labels_with_shifted_index)
output_labels_with_shifted_index.index -= 5
assert expected_output_labels.equals(output_labels_with_shifted_index)
assert expected_output_labels_new.equals(output_labels_new)
assert input_labels.equals(output_labels_inverse)
assert input_labels_with_shifted_index.equals(
output_labels_with_shifted_index_inverse)
assert expected_output_labels_new_inverse.equals(output_labels_new_inverse)
def test_label_cleaner_regression():
# Given
problem_type = REGRESSION
input_labels_numpy = np.array([2, 4, 2, 2, 4, 1])
input_labels = pd.Series(input_labels_numpy)
input_labels_new = pd.Series([3, 5, 2])
expected_output_labels = input_labels.copy()
expected_output_labels_new = input_labels_new.copy()
expected_output_labels_new_inverse = input_labels_new.copy()
# When
label_cleaner = LabelCleaner.construct(problem_type=problem_type,
y=input_labels,
y_uncleaned=None)
# Then
assert isinstance(label_cleaner, LabelCleanerDummy)
assert label_cleaner.problem_type_transform == REGRESSION
output_labels = label_cleaner.transform(input_labels)
output_labels_with_numpy = label_cleaner.transform(input_labels_numpy)
output_labels_new = label_cleaner.transform(input_labels_new)
output_labels_inverse = label_cleaner.inverse_transform(output_labels)
output_labels_new_inverse = label_cleaner.inverse_transform(
output_labels_new)
assert expected_output_labels.equals(output_labels)
assert expected_output_labels.equals(output_labels_with_numpy)
assert expected_output_labels_new.equals(output_labels_new)
assert input_labels.equals(output_labels_inverse)
assert expected_output_labels_new_inverse.equals(output_labels_new_inverse)
def test_label_cleaner_multiclass():
# Given
problem_type = MULTICLASS
input_labels_numpy = np.array([2, 4, 2, 2, 4, 1])
input_labels = pd.Series(input_labels_numpy)
input_labels_category = input_labels.astype('category')
input_labels_with_shifted_index = input_labels.copy()
input_labels_with_shifted_index.index += 5
input_labels_new = np.array([3, 5, 2])
expected_output_labels = pd.Series([1, 2, 1, 1, 2, 0])
expected_output_labels_new = pd.Series([np.nan, np.nan, 1])
expected_output_labels_new_inverse = pd.Series([np.nan, np.nan, 2])
# When
label_cleaner = LabelCleaner.construct(problem_type=problem_type,
y=input_labels,
y_uncleaned=input_labels)
# Then
assert isinstance(label_cleaner, LabelCleanerMulticlass)
assert label_cleaner.problem_type_transform == MULTICLASS
assert label_cleaner.cat_mappings_dependent_var == {0: 1, 1: 2, 2: 4}
output_labels = label_cleaner.transform(input_labels)
output_labels_with_numpy = label_cleaner.transform(input_labels_numpy)
output_labels_category = label_cleaner.transform(input_labels_category)
output_labels_with_shifted_index = label_cleaner.transform(
input_labels_with_shifted_index)
output_labels_new = label_cleaner.transform(input_labels_new)
output_labels_inverse = label_cleaner.inverse_transform(output_labels)
output_labels_with_shifted_index_inverse = label_cleaner.inverse_transform(
output_labels_with_shifted_index)
output_labels_new_inverse = label_cleaner.inverse_transform(
output_labels_new)
assert expected_output_labels.equals(output_labels)
assert expected_output_labels.equals(output_labels_with_numpy)
assert expected_output_labels.equals(output_labels_category)
assert not expected_output_labels.equals(output_labels_with_shifted_index)
output_labels_with_shifted_index.index -= 5
assert expected_output_labels.equals(output_labels_with_shifted_index)
assert expected_output_labels_new.equals(output_labels_new)
assert input_labels.equals(output_labels_inverse)
assert input_labels_with_shifted_index.equals(
output_labels_with_shifted_index_inverse)
assert expected_output_labels_new_inverse.equals(output_labels_new_inverse)
def test_label_cleaner_multiclass_to_binary():
# Given
problem_type = MULTICLASS
input_labels_numpy = np.array(['l1', 'l2', 'l2', 'l1', 'l1', 'l2'])
input_labels = pd.Series(input_labels_numpy)
input_labels_uncleaned = pd.Series(
['l0', 'l1', 'l2', 'l2', 'l1', 'l1', 'l2', 'l3', 'l4'])
input_labels_category = input_labels.astype('category')
input_labels_with_shifted_index = input_labels.copy()
input_labels_with_shifted_index.index += 5
input_labels_new = np.array(['l0', 'l1', 'l2'])
input_labels_proba_transformed = pd.Series([0.7, 0.2, 0.5],
index=[5, 2, 8])
expected_output_labels = pd.Series([0, 1, 1, 0, 0, 1])
expected_output_labels_new = pd.Series([np.nan, 0, 1])
expected_output_labels_new_inverse = pd.Series([np.nan, 'l1', 'l2'])
expected_output_labels_proba_transformed_inverse = pd.DataFrame(
data=[[0, 0.3, 0.7, 0, 0], [0, 0.8, 0.2, 0, 0], [0, 0.5, 0.5, 0, 0]],
index=[5, 2, 8],
columns=['l0', 'l1', 'l2', 'l3', 'l4'],
dtype=np.float64)
# When
label_cleaner = LabelCleaner.construct(problem_type=problem_type,
y=input_labels,
y_uncleaned=input_labels_uncleaned)
# Then
assert isinstance(label_cleaner, LabelCleanerMulticlassToBinary)
assert label_cleaner.problem_type_transform == BINARY
assert label_cleaner.cat_mappings_dependent_var == {0: 'l1', 1: 'l2'}
output_labels = label_cleaner.transform(input_labels)
output_labels_with_numpy = label_cleaner.transform(input_labels_numpy)
output_labels_category = label_cleaner.transform(input_labels_category)
output_labels_with_shifted_index = label_cleaner.transform(
input_labels_with_shifted_index)
output_labels_new = label_cleaner.transform(input_labels_new)
output_labels_inverse = label_cleaner.inverse_transform(output_labels)
output_labels_with_shifted_index_inverse = label_cleaner.inverse_transform(
output_labels_with_shifted_index)
output_labels_new_inverse = label_cleaner.inverse_transform(
output_labels_new)
assert expected_output_labels.equals(output_labels)
assert expected_output_labels.equals(output_labels_with_numpy)
assert expected_output_labels.equals(output_labels_category)
assert not expected_output_labels.equals(output_labels_with_shifted_index)
output_labels_with_shifted_index.index -= 5
assert expected_output_labels.equals(output_labels_with_shifted_index)
assert expected_output_labels_new.equals(output_labels_new)
assert input_labels.equals(output_labels_inverse)
assert input_labels_with_shifted_index.equals(
output_labels_with_shifted_index_inverse)
assert expected_output_labels_new_inverse.equals(output_labels_new_inverse)
output_labels_proba_transformed_inverse = label_cleaner.inverse_transform_proba(
input_labels_proba_transformed, as_pandas=True)
pd.testing.assert_frame_equal(
expected_output_labels_proba_transformed_inverse,
output_labels_proba_transformed_inverse)
label_column = 'class' # specifies which column do we want to predict train_data = train_data.head(1000) # subsample for faster demo ############################################# # Training custom model outside of task.fit # ############################################# # Separate features and labels X_train = train_data.drop(columns=[label_column]) y_train = train_data[label_column] problem_type = infer_problem_type(y=y_train) # Infer problem type (or else specify directly) naive_bayes_model = NaiveBayesModel(path='AutogluonModels/', name='CustomNaiveBayes', problem_type=problem_type) # Construct a LabelCleaner to neatly convert labels to float/integers during model training/inference, can also use to inverse_transform back to original. label_cleaner = LabelCleaner.construct(problem_type=problem_type, y=y_train) y_train_clean = label_cleaner.transform(y_train) naive_bayes_model.fit(X_train=X_train, y_train=y_train_clean) # Fit custom model # To save to disk and load the model, do the following: # load_path = naive_bayes_model.path # naive_bayes_model.save() # del naive_bayes_model # naive_bayes_model = NaiveBayesModel.load(path=load_path) # Prepare test data X_test = test_data.drop(columns=[label_column]) y_test = test_data[label_column] y_test_clean = label_cleaner.transform(y_test)
