def test_latlon_box():
# Create example dataset
ds = DataSubset(
None,
"aqum",
os.path.join(SAMPLEDIR, "model", "aqum_daily*"),
box=(-4, 50.4, -2.8, 51.2),
crs=ccrs.Geodetic(),
)
cube = ds.as_cube()
xcoord, ycoord = util.cubes.get_xy_coords(cube)
# Check we have the points we asked for (multiples of 2000m within
# each range)
# Strictly speaking, the transformed box would have slightly curved edges,
# and the "corner-most" gridpoints would be:
# tl: (262000, 146000)
# tr: (344000, 144000)
# br: (342000, 56000)
# bl: (258000, 58000)
# We therefore expect a slightly larger box, covering the minimum and
# maximum in both directions
assert iris.util.array_equal(xcoord.points[[0, -1]], [258000, 344000])
assert iris.util.array_equal(ycoord.points[[0, -1]], [56000, 146000])
def test_point_subset():
# Create example dataset
ds = DataSubset(
None,
"aqum",
os.path.join(SAMPLEDIR, "model", "aqum_daily*"),
point=(100, 200),
)
cube = ds.as_cube()
xcoord, ycoord = util.cubes.get_xy_coords(cube)
# Check we have the point we asked for
assert iris.util.array_equal(xcoord.points, [100])
assert iris.util.array_equal(ycoord.points, [200])
def test_box_subset():
# Create example dataset
ds = DataSubset(
None,
"aqum",
os.path.join(SAMPLEDIR, "model", "aqum_daily*"),
box=(-1000, -2000, 3000, 4000),
)
cube = ds.as_cube()
xcoord, ycoord = util.cubes.get_xy_coords(cube)
# Check we have the points we asked for (multiples of 2000m within
# each range)
assert iris.util.array_equal(xcoord.points, [0, 2000])
assert iris.util.array_equal(ycoord.points, [-2000, 0, 2000, 4000])
def test_latlon_point():
# Create example dataset
ds = DataSubset(
None,
"aqum",
os.path.join(SAMPLEDIR, "model", "aqum_daily*"),
point=(-0.1, 51.5),
crs=ccrs.Geodetic(),
)
cube = ds.as_cube()
xcoord, ycoord = util.cubes.get_xy_coords(cube)
# Check we have the point we asked for
assert iris.util.array_equal(xcoord.points.round(4), [531866.1304])
assert iris.util.array_equal(ycoord.points.round(4), [179660.9048])
def setup_class(self):
# Define a test polygon (an extremely simple representation of Exeter)
shape = shapely.geometry.Polygon([
(289271.9, 93197.0),
(289351.3, 95110.1),
(293405.1, 96855.0),
(296721.1, 94960.3),
(297165.1, 86966.9),
(294181.6, 89357.2),
(291388.0, 89272.6),
])
# Create example dataset
self.ds = DataSubset(
None,
"aqum",
os.path.join(SAMPLEDIR, "model", "aqum_hourly_o3_20200520.nc"),
shape=shape,
)
self.cube = self.ds.as_cube()
def setup_class(self):
# Define a test polygon (an extremely simple representation of Exeter)
shape = shapely.geometry.Polygon([
(-3.5690, 50.7273),
(-3.5685, 50.7445),
(-3.5115, 50.7609),
(-3.4640, 50.7445),
(-3.4555, 50.6727),
(-3.4984, 50.6936),
(-3.5379, 50.6924),
])
# Create example dataset
self.ds = DataSubset(
None,
"aqum",
os.path.join(SAMPLEDIR, "model", "aqum_hourly_o3_20200520.nc"),
shape=shape,
crs=ccrs.Geodetic(),
)
self.cube = self.ds.as_cube()
class TestPolygonSubset:
def setup_class(self):
# Define a test polygon (an extremely simple representation of Exeter)
shape = shapely.geometry.Polygon([
(289271.9, 93197.0),
(289351.3, 95110.1),
(293405.1, 96855.0),
(296721.1, 94960.3),
(297165.1, 86966.9),
(294181.6, 89357.2),
(291388.0, 89272.6),
])
# Create example dataset
self.ds = DataSubset(
None,
"aqum",
os.path.join(SAMPLEDIR, "model", "aqum_hourly_o3_20200520.nc"),
shape=shape,
)
self.cube = self.ds.as_cube()
def test_subset_mask(self):
# Define corresponding mask (note: this "looks" upside down compared
# to how it would be plotted)
expected_mask = np.array(
[[1, 1, 1, 1, 0],
[1, 1, 0, 0, 0],
[0, 0, 0, 0, 0],
[0, 0, 0, 0, 1],
[0, 0, 0, 0, 1],
[0, 0, 0, 0, 1]]
)
# Check we have the right mask (on a 2d slice of this 3d cube)
subcube = next(self.cube.slices_over("time"))
assert iris.util.array_equal(subcube.data.mask, expected_mask)
def test_subset_data(self):
# Simple data check, which, as the mask is taken into account, should
# be a pretty reliable test
assert round(self.cube.data.mean(), 8) == 57.66388811
class TestLatlonPolygonSubset:
# This is literally a copy of the unit test, with converted coordinates
def setup_class(self):
# Define a test polygon (an extremely simple representation of Exeter)
shape = shapely.geometry.Polygon([
(-3.5690, 50.7273),
(-3.5685, 50.7445),
(-3.5115, 50.7609),
(-3.4640, 50.7445),
(-3.4555, 50.6727),
(-3.4984, 50.6936),
(-3.5379, 50.6924),
])
# Create example dataset
self.ds = DataSubset(
None,
"aqum",
os.path.join(SAMPLEDIR, "model", "aqum_hourly_o3_20200520.nc"),
shape=shape,
crs=ccrs.Geodetic(),
)
self.cube = self.ds.as_cube()
def test_subset_mask(self):
# Define corresponding mask (note: this "looks" upside down compared
# to how it would be plotted)
expected_mask = np.array([[1, 1, 1, 1, 0], [1, 1, 0, 0, 0],
[0, 0, 0, 0, 0], [0, 0, 0, 0, 1],
[0, 0, 0, 0, 1], [0, 0, 0, 0, 1]])
# Check we have the right mask (on a 2d slice of this 3d cube)
subcube = next(self.cube.slices_over("time"))
assert iris.util.array_equal(subcube.data.mask, expected_mask)
def test_subset_data(self):
# Simple data check, which, as the mask is taken into account, should
# be a pretty reliable test
assert round(self.cube.data.mean(), 8) == 57.66388811