def test_different_input_lengths():
fraud_cost = FraudCost(amount_col="value")
y_predicted = np.array([0, 0])
y_true = np.array([1])
with pytest.raises(ValueError, match="Inputs have mismatched dimensions"):
fraud_cost.score(y_true, y_predicted)
y_true = np.array([0, 0])
y_predicted = np.array([1, 2, 0])
with pytest.raises(ValueError, match="Inputs have mismatched dimensions"):
fraud_cost.score(y_true, y_predicted)
def test_input_contains_inf(capsys):
fraud_cost = FraudCost(amount_col="value")
y_predicted = np.array([np.inf, 0, 0])
y_true = np.array([1, 0, 0])
with pytest.raises(ValueError,
match="y_predicted contains NaN or infinity"):
fraud_cost.score(y_true, y_predicted)
y_true = np.array([np.inf, 0, 0])
y_predicted = np.array([1, 0, 0])
with pytest.raises(ValueError, match="y_true contains NaN or infinity"):
fraud_cost.score(y_true, y_predicted)
def test_input_contains_nan(X_y_binary):
fraud_cost = FraudCost(amount_col="value")
y_predicted = np.array([np.nan, 0, 0])
y_true = np.array([1, 2, 1])
with pytest.raises(ValueError,
match="y_predicted contains NaN or infinity"):
fraud_cost.score(y_true, y_predicted)
y_true = np.array([np.nan, 0, 0])
y_predicted = np.array([1, 2, 0])
with pytest.raises(ValueError, match="y_true contains NaN or infinity"):
fraud_cost.score(y_true, y_predicted)
def test_binary_more_than_two_unique_values():
fraud_cost = FraudCost(amount_col="value")
y_predicted = np.array([0, 1, 2])
y_true = np.array([1, 0, 1])
with pytest.raises(
ValueError,
match="y_predicted contains more than two unique values"):
fraud_cost.score(y_true, y_predicted)
y_true = np.array([0, 1, 2])
y_predicted = np.array([1, 0, 1])
with pytest.raises(ValueError,
match="y_true contains more than two unique values"):
fraud_cost.score(y_true, y_predicted)
def test_fraud_objective_score(X_y_binary):
X, y = X_y_binary
fraud_cost = FraudCost(amount_col="value")
y_predicted = pd.Series([.1, .5, .5])
y_true = pd.Series([True, False, True])
extra_columns = pd.DataFrame({"value": [100, 5, 250]})
out = fraud_cost.decision_function(y_predicted, 5, extra_columns)
assert isinstance(out, pd.Series)
pd.testing.assert_series_equal(out, y_true, check_names=False)
score = fraud_cost.score(y_true, out, extra_columns)
assert (score == 0.0)
out = fraud_cost.decision_function(y_predicted.to_numpy(), 5,
extra_columns)
assert isinstance(out, pd.Series)
pd.testing.assert_series_equal(out, y_true, check_names=False)
score = fraud_cost.score(y_true, out, extra_columns)
assert (score == 0.0)
out = ww.DataColumn(
fraud_cost.decision_function(y_predicted, 5, extra_columns))
pd.testing.assert_series_equal(out.to_series(),
y_true,
check_dtype=False,
check_names=False)
score = fraud_cost.score(y_true, out, extra_columns)
assert (score == 0.0)
# testing with other types of inputs
y_predicted = np.array([.1, .5, .5])
extra_columns = pd.DataFrame({"value": [100, 5, 250]})
out = fraud_cost.decision_function(y_predicted, 5, extra_columns)
pd.testing.assert_series_equal(out, y_true, check_names=False)
score = fraud_cost.score(y_true, out, extra_columns)
assert (score == 0.0)
y_predicted = pd.Series([.2, .01, .01])
extra_columns = pd.DataFrame({"value": [100, 50, 50]})
y_true = pd.Series([False, False, True])
expected_y_pred = pd.Series([True, False, False])
out = fraud_cost.decision_function(y_predicted, 10, extra_columns)
pd.testing.assert_series_equal(out, expected_y_pred, check_names=False)
score = fraud_cost.score(y_true, out, extra_columns)
assert (score == 0.255)
def test_fraud_objective_score_list(X_y_binary):
X, y = X_y_binary
fraud_cost = FraudCost(amount_col="value")
y_predicted = [.1, .5, .5]
y_true = [True, False, True]
extra_columns = pd.DataFrame({"value": [100, 5, 250]})
out = fraud_cost.decision_function(y_predicted, 5, extra_columns)
assert isinstance(out, pd.Series)
pd.testing.assert_series_equal(out, pd.Series(y_true), check_names=False)
score = fraud_cost.score(y_true, out, extra_columns)
assert (score == 0.0)
def test_zero_input_lengths():
fraud_cost = FraudCost(amount_col="value")
y_predicted = np.array([])
y_true = np.array([])
with pytest.raises(ValueError, match="Length of inputs is 0"):
fraud_cost.score(y_true, y_predicted)
