def test_values(snow_fraction, phase, expected):
"""Test specific values give expected results, including all meta-data"""
input_cube = set_up_variable_cube(
np.full((2, 2), fill_value=snow_fraction, dtype=np.float32),
name="snow_fraction",
units="1",
standard_grid_metadata="uk_ens",
)
result = SignificantPhaseMask()(input_cube, phase)
assert isinstance(result, iris.cube.Cube)
assert result.name() == f"{phase}_mask"
assert str(result.units) == "1"
assert result.dtype == np.int8
assert (result.data == expected).all()
def test_basic():
"""Test that the __init__ method configures the plugin as expected."""
plugin = SignificantPhaseMask()
assert np.isclose(plugin.lower_threshold, 0.01)
assert np.isclose(plugin.upper_threshold, 0.99)
assert isinstance(plugin.phase_operator, dict)
assert all(callable(v) for v in plugin.phase_operator.values())
def test_phase_error():
"""Tests behaviour when requested phase is invalid"""
input_cube = set_up_variable_cube(
np.ones((2, 2), dtype=np.float32),
name="snow_fraction",
units="1",
standard_grid_metadata="uk_ens",
)
msg = r"Requested phase mask 'kittens' not in \['rain', 'sleet', 'snow'\]"
with pytest.raises(KeyError, match=msg):
SignificantPhaseMask()(input_cube, "kittens")
def test_units_error():
"""Tests behaviour when input cube has wrong units"""
input_cube = set_up_variable_cube(
np.ones((2, 2), dtype=np.float32),
name="snow_fraction",
units="m",
standard_grid_metadata="uk_ens",
)
msg = "Expected cube with units '1', not m"
with pytest.raises(ValueError, match=msg):
SignificantPhaseMask()(input_cube, "snow")
def test_masked_values():
"""Test specific values give expected results"""
data = np.zeros((2, 2), dtype=np.float32)
data = np.ma.masked_array(data, [[True, False], [False, False]])
input_cube = set_up_variable_cube(
data, name="snow_fraction", units="1", standard_grid_metadata="uk_ens",
)
with pytest.raises(
NotImplementedError, match="SignificantPhaseMask cannot handle masked data"
):
SignificantPhaseMask()(input_cube, "snow")
def test_data_error(snow_fraction):
"""Tests behaviour when input cube has invalid data"""
input_cube = set_up_variable_cube(
np.full((2, 2), fill_value=snow_fraction, dtype=np.float32),
name="snow_fraction",
units="1",
standard_grid_metadata="uk_ens",
)
msg = "Expected cube data to be in range 0 <= x <= 1. Found max={0}; min={0}".format(
snow_fraction)
with pytest.raises(ValueError, match=msg):
SignificantPhaseMask()(input_cube, "snow")
def test_model_id_attr(model_id_attr):
"""Test attribute handling"""
input_cube = set_up_variable_cube(
np.full((2, 2), fill_value=0.1, dtype=np.float32),
name="snow_fraction",
units="1",
standard_grid_metadata="uk_ens",
)
expected_attributes = {
"source": "Unit test",
"institution": "Met Office",
"title": "Post-Processed IMPROVER unit test",
}
input_cube.attributes.update(expected_attributes)
if model_id_attr:
expected_attributes[model_id_attr] = input_cube.attributes[model_id_attr]
result = SignificantPhaseMask(model_id_attr=model_id_attr)(input_cube, "snow")
assert result.attributes == expected_attributes
def process(cube: cli.inputcube, phase: str, *, model_id_attr: str = None):
"""
Make phase-mask cube for the specified phase.
Args:
cube (iris.cube.Cube):
The input snow-fraction data to derive the phase mask from.
phase (str):
One of "rain", "sleet" or "snow". This is the phase mask that will be
returned.
model_id_attr (str):
Name of the attribute used to identify the source model for
blending.
"""
from improver.psychrometric_calculations.significant_phase_mask import (
SignificantPhaseMask, )
return SignificantPhaseMask(model_id_attr=model_id_attr)(cube, phase)
assert all(callable(v) for v in plugin.phase_operator.values())
@pytest.mark.parametrize(
("snow_fraction", "phase", "expected"),
(
(0, "snow", 0),
(0.5, "snow", 0),
(1, "snow", 1),
(0, "sleet", 0),
(0.5, "sleet", 1),
(1, "sleet", 0),
(0, "rain", 1),
(0.5, "rain", 0),
(1, "rain", 0),
(SignificantPhaseMask().lower_threshold, "rain", 1),
(SignificantPhaseMask().lower_threshold, "sleet", 0),
(SignificantPhaseMask().upper_threshold, "snow", 1),
(SignificantPhaseMask().upper_threshold, "sleet", 0),
),
)
def test_values(snow_fraction, phase, expected):
"""Test specific values give expected results, including all meta-data"""
input_cube = set_up_variable_cube(
np.full((2, 2), fill_value=snow_fraction, dtype=np.float32),
name="snow_fraction",
units="1",
standard_grid_metadata="uk_ens",
)
result = SignificantPhaseMask()(input_cube, phase)
assert isinstance(result, iris.cube.Cube)
