def test_error_if_capping_method_quantiles_and_fold_value_not_permitted():
with pytest.raises(ValueError):
Winsorizer(capping_method="quantiles", fold=0.3)
def test_error_if_missing_values_not_permited():
with pytest.raises(ValueError):
Winsorizer(missing_values="other")
def test_error_if_fold_value_not_permitted():
with pytest.raises(ValueError):
Winsorizer(fold=-1)
def test_error_if_capping_method_not_permitted():
# test error raises
with pytest.raises(ValueError):
Winsorizer(capping_method="other")
def test_error_if_tail_value_not_permitted():
with pytest.raises(ValueError):
Winsorizer(tail="other")
def test_get_feature_names_out_input_features_is_list(df_na):
input_features = ["Age", "Marks"]
# when add_indicators is false, we've got the generic check from estimator_checks.
# We need to test only when true.
tr = Winsorizer(tail="left", add_indicators=True, missing_values="ignore")
tr.fit(df_na)
out = [f + "_left" for f in input_features]
assert tr.get_feature_names_out(input_features) == input_features + out
tr = Winsorizer(tail="right", add_indicators=True, missing_values="ignore")
tr.fit(df_na)
out = [f + "_right" for f in input_features]
assert tr.get_feature_names_out(input_features) == input_features + out
tr = Winsorizer(tail="both", add_indicators=True, missing_values="ignore")
tr.fit(df_na)
out = ["Age_left", "Age_right", "Marks_left", "Marks_right"]
assert tr.get_feature_names_out(input_features) == input_features + out
n_categories=100000000000,
replace_with=10,
ignore_format=True,
),
WoEEncoder(ignore_format=True),
PRatioEncoder(ignore_format=True),
])
def test_sklearn_compatible_encoder(estimator, check):
check(estimator)
# outliers
@parametrize_with_checks([
ArbitraryOutlierCapper(max_capping_dict={"0": 10}),
OutlierTrimmer(),
Winsorizer(),
])
def test_sklearn_compatible_outliers(estimator, check):
check(estimator)
# transformers
@parametrize_with_checks([
BoxCoxTransformer(),
LogTransformer(),
PowerTransformer(),
ReciprocalTransformer(),
YeoJohnsonTransformer(),
])
def test_sklearn_compatible_transformer(estimator, check):
check(estimator)
def test_get_feature_names_out_input_features_is_none(df_na):
original_features = df_na.columns.to_list()
input_features = ["Age", "Marks"]
# when indicators is false, we've got the generic check.
# We need to test only when true
tr = Winsorizer(tail="left", add_indicators=True, missing_values="ignore")
tr.fit(df_na)
out = [f + "_left" for f in input_features]
assert tr.get_feature_names_out() == original_features + out
tr = Winsorizer(tail="right", add_indicators=True, missing_values="ignore")
tr.fit(df_na)
out = [f + "_right" for f in input_features]
assert tr.get_feature_names_out() == original_features + out
tr = Winsorizer(tail="both", add_indicators=True, missing_values="ignore")
tr.fit(df_na)
out = ["Age_left", "Age_right", "Marks_left", "Marks_right"]
assert tr.get_feature_names_out() == original_features + out
def test_transform_raises_error_if_na_in_input_df(df_vartypes, df_na):
# test case 9: when dataset contains na, transform method
with pytest.raises(ValueError):
transformer = Winsorizer()
transformer.fit(df_vartypes)
transformer.transform(df_na[["Name", "City", "Age", "Marks", "dob"]])
def test_fit_raises_error_if_na_in_inut_df(df_na):
# test case 8: when dataset contains na, fit method
with pytest.raises(ValueError):
transformer = Winsorizer()
transformer.fit(df_na)
color="frequency")
fig.update_layout(
title="Segmentation",
width=700,
height=500,
)
fig.show()
df.drop("Customer", axis=1, inplace=True)
df.drop("Effective To Date", axis=1, inplace=True)
wind = Winsorizer(
capping_method='iqr',
tail='both',
fold=1.5,
variables=['Customer Lifetime Value', 'Income', 'Total Claim Amount'])
wind.fit(df)
df = wind.transform(df)
dummylist = []
dummy_variables = [
"State", "Response", "Coverage", "Education", "EmploymentStatus", "Gender",
"Location Code", "Policy Type", "Policy", "Renew Offer Type",
"Sales Channel", "Vehicle Class", "Vehicle Size", "Marital Status"
]
for var in dummy_variables:
dummylist.append(
def test_non_fitted_error(df_vartypes):
with pytest.raises(NotFittedError):
transformer = Winsorizer()
transformer.transform(df_vartypes)