def test_incorrect_input_key():
with pytest.raises(KeyError):
outlier = HandleOutlier()
outlier.handle_outliers(params={
"train_df": train_df,
"cols": ["Place"]
})
def test_replace_output(test_input):
outlier = HandleOutlier()
outlier.handle_outliers(params=test_input)
assert test_input["train_df"].shape == train_df.shape
assert -999 in test_input["train_df"]["Distance"].values
assert "-999" not in test_input["train_df"]["Capitals"].values
assert "-999" not in test_input["train_df"]["Other Capitals"].values
def test_true_arguments():
with pytest.raises(ArgumentsError):
outlier = HandleOutlier()
outlier.handle_outliers(params={
"train_df": train_df,
"cols": ["Distance"],
"replace": True,
})
def test_true_arguments():
with pytest.raises(ArgumentsError):
outlier = HandleOutlier()
outlier.handle_outliers(
params={
"train_df": train_df,
"cat_cols": ["Capitals", "Other Capitals"],
"replace": True,
})
def test_false_arguments():
with pytest.warns(UserWarning):
outlier = HandleOutlier()
outlier.handle_outliers(
params={
"train_df": train_df,
"cols": ["Distance"],
"remove_outliers": False,
"replace": False,
})
def __init__(
self,
train_df_path=None,
test_df_path=None,
steps=None,
config_file=None,
params=None,
custom_reader=None,
):
steps = [
Parser().parse_dataset,
NullValuesHandler().execute,
Encoder().encode,
HandleOutlier().handle_outliers,
Scaler().execute,
SelectKBest().fit_transform,
Split().train_test_split,
]
super().__init__(
train_df_path=train_df_path,
test_df_path=test_df_path,
steps=steps,
config_file=config_file,
params=params,
custom_reader=custom_reader,
)
def test_incorrect_input_type(test_input):
with pytest.raises(TypeError):
outlier = HandleOutlier()
outlier.handle_outliers(params=test_input)
def test_all(test_input):
outlier = HandleOutlier()
outlier.handle_outliers(params=test_input)
assert -999 in test_input["train_df"]["Distance"].values
assert -999 in test_input["test_df"]["Distance"].values
assert test_input["train_df"].equals(test_input["test_df"])
def test_removeoutliers_output(test_input):
outlier = HandleOutlier()
outlier.handle_outliers(params=test_input)
assert test_input["train_df"].shape != train_df.shape
