Use `pytest.mark.parametrize` to cut down on duplicate tests

[JoelLucaAdams]

There is an issue where we have two tests that check the same thing where the only difference is which library is used. E.g.

def test_drop_variables_original():
    with open_dataset(
        TEST_FILES_DIR / "0000.sdf",
        drop_variables=["Electric_Field/Ex", "Electric_Field/Ey"],
    ) as df:
        assert "Electric_Field_Ex" not in df
        assert "Electric_Field_Ey" not in df


def test_xr_drop_variables_original():
    with xr.open_dataset(
        TEST_FILES_DIR / "0000.sdf",
        drop_variables=["Electric_Field/Ex", "Electric_Field/Ey"],
    ) as df:
        assert "Electric_Field_Ex" not in df
        assert "Electric_Field_Ey" not in df

To avoid this duplication of tests that accomplish the same goal but with different signatures we should use the pytest decorator as suggested by @LiamPattinson.

@pytest.mark.parametrize("xrlib", (xr, sdfxr))
def test_drop_variables_original():
    with xrlib.open_dataset(
        TEST_FILES_DIR / "0000.sdf",
        drop_variables=["Electric_Field/Ex", "Electric_Field/Ey"],
    ) as df:
        assert "Electric_Field_Ex" not in df
        assert "Electric_Field_Ey" not in df