def test_raises_error_when_transform_df_with_different_n_variables(
df_datetime):
transformer = DatetimeFeatures()
transformer.fit(df_datetime)
# different number of columns than the df used to fit
with pytest.raises(ValueError):
transformer.transform(df_datetime[vars_dt])
def test_raises_error_when_nan_in_transform_df(df_datetime):
transformer = DatetimeFeatures()
transformer.fit(df_datetime)
# dataset containing nans
with pytest.raises(ValueError):
transformer.transform(dates_nan)
transformer = DatetimeFeatures(variables="index")
transformer.fit(dates_idx_dt)
with pytest.raises(ValueError):
transformer.transform(dates_idx_nan)
def test_attributes_upon_fitting(df_datetime):
transformer = DatetimeFeatures()
transformer.fit(df_datetime)
assert transformer.variables_ == vars_dt
assert transformer.features_to_extract_ == FEATURES_DEFAULT
assert transformer.n_features_in_ == df_datetime.shape[1]
transformer = DatetimeFeatures(variables="date_obj1",
features_to_extract="all")
transformer.fit(df_datetime)
assert transformer.variables_ == ["date_obj1"]
assert transformer.features_to_extract_ == FEATURES_SUPPORTED
transformer = DatetimeFeatures(
variables=["date_obj1", "time_obj"],
features_to_extract=["year", "quarter_end", "second"],
)
transformer.fit(df_datetime)
assert transformer.variables_ == ["date_obj1", "time_obj"]
assert transformer.features_to_extract_ == [
"year", "quarter_end", "second"
]
def test_raises_error_when_fitting_not_a_df(_not_a_df):
transformer = DatetimeFeatures()
# trying to fit not a df
with pytest.raises(TypeError):
transformer.fit(_not_a_df)
def test_raises_error_when_transforming_not_a_df(_not_a_df, df_datetime):
transformer = DatetimeFeatures()
transformer.fit(df_datetime)
# trying to transform not a df
with pytest.raises(TypeError):
transformer.transform(_not_a_df)