def test_make_recipe_defence():
X, y = cars_data()
with pytest.raises(ValueError):
(X_train, X_valid, X_test, y_train, y_valid,
y_test) = mealprep.make_recipe(
X=X,
y=y,
recipe="ohe_and_standard_scaler",
splits_to_return="INVALID SPLIT COMMAND",
train_valid_prop=0.5,
)
return None
def test_make_recipe_split_prop():
X, y = cars_data()
# check OHE behavior when one categorical column
(X_train, X_valid, X_test, y_train, y_valid,
y_test) = mealprep.make_recipe(
X=X,
y=y,
recipe="ohe_and_standard_scaler",
splits_to_return="train_test",
train_valid_prop=0.5,
)
assert X_train.shape[0] == X_test.shape[0]
return None
def test_make_recipe_splits_train():
X, y = cars_data()
# Test on splits_to_return="train_test"
X_train, X_valid, X_test, y_train, y_valid, y_test = mealprep.make_recipe(
X=X, y=y, recipe="ohe_and_standard_scaler", splits_to_return="train")
assert type(X_train) == pd.DataFrame
assert X_test is None
assert type(y_train) == np.ndarray
assert y_test is None
assert X_valid is None
assert y_valid is None
assert X.shape[0] == X_train.shape[0]
return None
def test_make_recipe_splits_traintestvalid():
X, y = cars_data()
# Test on splits_to_return="train_test_valid"
X_train, X_valid, X_test, y_train, y_valid, y_test = mealprep.make_recipe(
X=X,
y=y,
recipe="ohe_and_standard_scaler",
splits_to_return="train_test_valid")
assert type(X_train) == pd.DataFrame
assert type(X_test) == pd.DataFrame
assert type(X_valid) == pd.DataFrame
assert type(y_train) == np.ndarray
assert type(y_test) == np.ndarray
assert type(y_valid) == np.ndarray
return None
def test_make_recipe_ohe():
X, y = cars_data()
# check OHE behavior when one categorical column
X_train, X_valid, X_test, y_train, y_valid, y_test = mealprep.make_recipe(
X=X,
y=y,
recipe="ohe_and_standard_scaler",
splits_to_return="train_test",
random_seed=1993,
)
assert list(X_train.columns) == [
"Miles_per_Gallon",
"Cylinders",
"Displacement",
"Horsepower",
"Weight_in_lbs",
"Acceleration",
"x0_Europe",
"x0_Japan",
"x0_USA",
]
# check OHE behavior when two categorical column
X["Cylinders"] = X["Cylinders"].apply(str)
X_train, X_valid, X_test, y_train, y_valid, y_test = mealprep.make_recipe(
X=X,
y=y,
recipe="ohe_and_standard_scaler",
splits_to_return="train_test",
random_seed=1993,
)
assert list(X_train.columns) == [
"Miles_per_Gallon",
"Displacement",
"Horsepower",
"Weight_in_lbs",
"Acceleration",
"x0_3",
"x0_4",
"x0_5",
"x0_6",
"x0_8",
"x1_Europe",
"x1_Japan",
"x1_USA",
]
# check that error is thrown if bad recipe select
with pytest.raises(ValueError):
(X_train, X_valid, X_test, y_train, y_valid,
y_test) = mealprep.make_recipe(
X=X,
y=y,
recipe="not_a_recipe",
splits_to_return="train_test",
random_seed=1993,
)
return None