def test_scale_with_exclude_cols():
df = _some_df1()
scale_stage = Scale("StandardScaler", exclude_columns=["lbl"])
res_df = scale_stage(df)
assert list(res_df.columns) == ["ph", "gt", "lbl"]
assert "ph" in res_df.columns
assert "gt" in res_df.columns
assert res_df["ph"][1] < df["ph"][1]
# see only transform (no fit) when already fitted
df2 = _some_df1b()
res_df2 = scale_stage(df2)
assert "ph" in res_df2.columns
assert "gt" in res_df2.columns
assert df['ph'][1] < df2['ph'][1]
assert res_df2["ph"][1] < df2["ph"][1]
assert res_df["ph"][1] < res_df2["ph"][1]
# check fit_transform when already fitted
df3 = _some_df1b()
res_df3 = scale_stage.fit_transform(df2)
assert "ph" in res_df3.columns
assert "gt" in res_df3.columns
assert res_df3["ph"][1] < df3["ph"][1]
assert res_df3["ph"][1] < res_df2["ph"][1]
def test_scale_with_exclude():
"""Basic binning test."""
df = _some_df2()
scale_stage = Scale("StandardScaler", with_std=False)
res_df = scale_stage(df)
assert 'ph' in res_df.columns
assert 'gt' in res_df.columns
def test_scale_transform_exception():
df1 = _some_df1()
scale_stage = Scale("StandardScaler", exmsg="ERR")
scale_stage(df1)
# test transform exception
df2 = _bad_dtype_df1()
with pytest.raises(PipelineApplicationError):
scale_stage(df2)
def test_scale():
df = _some_df2()
scale_stage = Scale("StandardScaler")
res_df = scale_stage(df)
assert 'ph' in res_df.columns
assert 'gt' in res_df.columns
assert res_df['ph'][1] < df['ph'][1]
# see only transform (no fit) when already fitted
df2 = _some_df2b()
res_df2 = scale_stage(df2)
assert 'ph' in res_df2.columns
assert 'gt' in res_df2.columns
assert res_df2['ph'][1] < df2['ph'][1]
assert res_df['ph'][1] < res_df2['ph'][1]
# check fit_transform when already fitted
df3 = _some_df2b()
res_df3 = scale_stage.fit_transform(df2)
assert 'ph' in res_df3.columns
assert 'gt' in res_df3.columns
assert res_df3['ph'][1] < df3['ph'][1]
assert res_df3['ph'][1] < res_df2['ph'][1]
def test_scale_with_exclude_cols():
df = _some_df1()
scale_stage = Scale("StandardScaler", exclude_columns=['lbl'], exmsg='AA')
res_df = scale_stage(df)
assert list(res_df.columns) == ['ph', 'gt', 'lbl']
assert 'ph' in res_df.columns
assert 'gt' in res_df.columns
assert res_df['ph'][1] < df['ph'][1]
# see only transform (no fit) when already fitted
df2 = _some_df1b()
res_df2 = scale_stage(df2)
assert 'ph' in res_df2.columns
assert 'gt' in res_df2.columns
assert res_df2['ph'][1] < df2['ph'][1]
assert res_df['ph'][1] < res_df2['ph'][1]
# check fit_transform when already fitted
df3 = _some_df1b()
res_df3 = scale_stage.fit_transform(df2)
assert 'ph' in res_df3.columns
assert 'gt' in res_df3.columns
assert res_df3['ph'][1] < df3['ph'][1]
assert res_df3['ph'][1] < res_df2['ph'][1]
def test_scale_app_exception():
df1 = _some_df1()
scale_stage = Scale("StandardScaler",
exclude_columns=[],
exclude_object_columns=False)
with pytest.raises(PipelineApplicationError):
scale_stage(df1)
df2 = _some_df2()
res_df = scale_stage(df2)
assert "ph" in res_df.columns
assert "gt" in res_df.columns
# test transform exception
with pytest.raises(PipelineApplicationError):
scale_stage(df1)
def test_scale_app_exception():
df = _some_df1()
scale_stage = Scale("StandardScaler", exclude_columns=[])
with pytest.raises(PipelineApplicationError):
scale_stage(df)
def test_scale_fit_transform_exception():
df1 = _some_df1()
scale_stage = Scale("StandardScaler", columns=['ph', 'lbl'])
with pytest.raises(PipelineApplicationError):
scale_stage(df1)