def test_collocation_of_alt_points_on_hybrid_pressure_and_altitude_coordinates(
self):
"""
Kernel should use the auxilliary altitude dimension when altitude is present in the coordinates
"""
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3,
hybrid_pr_len=10,
geopotential_height=True))
sample_points = UngriddedData.from_points_array(
# This point actually lies outside the lower bounds for altitude at this point in space
[
HyperPoint(lat=1.0,
lon=1.0,
alt=10,
t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies in the middle of the altitude bounds at this point
HyperPoint(lat=4.0,
lon=4.0,
alt=354,
t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies outside the upper bounds for altitude at this point
HyperPoint(lat=-4.0,
lon=-4.0,
alt=1000,
t=dt.datetime(1984, 8, 27, 2, 18, 52))
])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], float(cube[2, 1, 1, 0].data))
eq_(new_data.data[1], float(cube[3, 2, 1, 4].data))
eq_(new_data.data[2], float(cube[1, 0, 0, 9].data))
def test_gridded_ungridded_lin(self):
data = make_from_cube(mock.make_mock_cube())
data.name = lambda: 'Name'
data.var_name = 'var_name'
data._standard_name = 'y_wind'
sample = UngriddedData.from_points_array([
HyperPoint(lat=1.0,
lon=1.0,
alt=12.0,
t=dt.datetime(1984, 8, 29, 8, 34)),
HyperPoint(lat=3.0,
lon=3.0,
alt=7.0,
t=dt.datetime(1984, 8, 29, 8, 34)),
HyperPoint(lat=-1.0,
lon=-1.0,
alt=5.0,
t=dt.datetime(1984, 8, 29, 8, 34))
])
constraint = None
col = GriddedUngriddedCollocator()
output = col.collocate(sample, data, constraint, 'lin')
expected_result = np.array([8.8, 10.4, 7.2])
assert len(output) == 1
assert isinstance(output, UngriddedDataList)
assert np.allclose(output[0].data, expected_result)
def test_collocation_of_pres_points_on_hybrid_pressure_coordinates_and_altitude_coordinates(
self):
"""
When only pressure coordinate is present this should be used for the collocation
"""
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))
sample_points = UngriddedData.from_points_array(
# This point actually lies outside the lower bounds for altitude at this point in space
[
HyperPoint(lat=1.0,
lon=1.0,
pres=1100000.0,
t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies in the middle of the altitude bounds at this point
HyperPoint(lat=4.0,
lon=4.0,
pres=184600000.0,
t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies outside the upper bounds for altitude at this point
HyperPoint(lat=-4.0,
lon=-4.0,
pres=63100049.0,
t=dt.datetime(1984, 8, 27, 2, 18, 52))
])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], float(cube[2, 1, 1, 0].data))
eq_(new_data.data[1], float(cube[3, 2, 1, 4].data))
eq_(new_data.data[2], float(cube[1, 0, 0, 9].data))
def test_missing_data_for_missing_sample(self):
data = make_from_cube(mock.make_mock_cube())
data.name = lambda: 'Name'
data.var_name = 'var_name'
data._standard_name = 'y_wind'
sample = UngriddedData.from_points_array([
HyperPoint(lat=1.0,
lon=1.0,
alt=12.0,
t=dt.datetime(1984, 8, 29, 8, 34)),
HyperPoint(lat=3.0,
lon=3.0,
alt=7.0,
t=dt.datetime(1984, 8, 29, 8, 34)),
HyperPoint(lat=-1.0,
lon=-1.0,
alt=5.0,
t=dt.datetime(1984, 8, 29, 8, 34))
])
constraint = None
sample_mask = [False, True, False]
sample.data = np.ma.array([0, 0, 0], mask=sample_mask)
col = GriddedUngriddedCollocator(missing_data_for_missing_sample=True)
output = col.collocate(sample, data, constraint, 'nn')
assert len(output) == 1
assert isinstance(output, UngriddedDataList)
assert np.array_equal(output[0].data.mask, sample_mask)
def test_negative_lon_points_on_hybrid_altitude_coordinates_with_0_360_grid(
self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3,
hybrid_ht_len=10,
lon_dim_length=36,
lon_range=(0., 350.)))
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=1.0,
lon=111.0,
alt=5000.0,
t=dt.datetime(1984, 8, 28, 8, 34)),
HyperPoint(lat=4.0,
lon=141.0,
alt=12000.0,
t=dt.datetime(1984, 8, 28, 8, 34)),
HyperPoint(lat=-4.0,
lon=-14.0,
alt=10000.0,
t=dt.datetime(1984, 8, 27, 2, 18, 52))
])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 2501.0) # float(cube[2,11,1,0].data))
eq_(new_data.data[1], 3675.0) # float(cube[3,14,1,4].data))
eq_(new_data.data[2], 2139.0) # float(cube[1,35,0,8].data))
def test_already_collocated_in_col_gridded_to_ungridded_in_2d(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube())
# This point already exists on the cube with value 5 - which shouldn't be a problem
sample_points = UngriddedData.from_points_array([HyperPoint(0.0, 0.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 8.0)
def test_collocation_of_pres_points_on_hybrid_pressure_coordinates(self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=0.0,
lon=0.0,
pres=111100040.5,
t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=0.0,
lon=0.0,
pres=113625040.5,
t=dt.datetime(1984, 8, 28, 12, 0, 0)),
HyperPoint(lat=5.0,
lon=2.5,
pres=177125044.5,
t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=-4.0,
lon=-4.0,
pres=166600039.0,
t=dt.datetime(1984, 8, 27))
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
# Exactly on the lat, lon, time points, interpolated over pressure
assert_almost_equal(new_data.data[0], 221.5, decimal=5)
# Exactly on the lat, lon, points, interpolated over time and pressure
assert_almost_equal(new_data.data[1], 226.5, decimal=7)
# Exactly on the lat, time points, interpolated over longitude and pressure
assert_almost_equal(new_data.data[2], 330.5, decimal=7)
# Outside of the pressure bounds - extrapolation off
assert np.ma.is_masked(new_data.data[3])
def test_collocation_of_alt_pres_points_on_hybrid_altitude_coordinates(
self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=0.0,
lon=0.0,
alt=5550.0,
pres=10000.0,
t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=4.0,
lon=4.0,
alt=6000.0,
pres=1000.0,
t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=-4.0,
lon=-4.0,
alt=6500.0,
pres=100.0,
t=dt.datetime(1984, 8, 27))
])
col = GriddedUngriddedCollocator(fill_value=np.NAN)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 222.4814815, decimal=7)
assert_almost_equal(new_data.data[1], 321.0467626, decimal=7)
# Test that points outside the cell are returned as masked, rather than extrapolated by default
assert np.ma.is_masked(new_data.data[2])
def test_wrapping_of_pres_points_on_hybrid_pressure_coordinates_on_0_360_grid(
self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3,
hybrid_pr_len=10,
lon_dim_length=36,
lon_range=(0., 350.)))
# Ensure the longitude coord is circular
cube.coord(standard_name='longitude').circular = True
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=0.0,
lon=355.0,
pres=1482280045.0,
t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=5.0,
lon=2.5,
pres=1879350048.0,
t=dt.datetime(1984, 8, 28, 0, 0, 0))
])
col = GriddedUngriddedCollocator(extrapolate=False)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
eq_(new_data.data[0], 2701.0011131725005)
eq_(new_data.data[1], 3266.1930161260775)
def test_collocation_of_pres_alt_points_on_hybrid_pressure_coordinates_multi_var(self):
cube_list = [make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))]
cube_list.append(make_from_cube(mock.make_mock_cube(time_dim_length=3,
hybrid_pr_len=10,
data_offset=100)))
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=0.0, lon=0.0, pres=111100040.5, alt=5000, t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=0.0, lon=0.0, pres=113625040.5, alt=4000, t=dt.datetime(1984, 8, 28, 12, 0, 0)),
HyperPoint(lat=5.0, lon=2.5, pres=177125044.5, alt=3000, t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=-4.0, lon=-4.0, pres=166600039.0, alt=3500, t=dt.datetime(1984, 8, 27))])
col = GriddedUngriddedCollocator()
outlist = col.collocate(sample_points, cube_list, None, 'lin')
# First data set:
new_data = outlist[0]
# Exactly on the lat, lon, time points, interpolated over pressure
assert_almost_equal(new_data.data[0], 221.5, decimal=5)
# Exactly on the lat, lon, points, interpolated over time and pressure
assert_almost_equal(new_data.data[1], 226.5, decimal=7)
# Exactly on the lat, time points, interpolated over longitude and pressure
assert_almost_equal(new_data.data[2], 330.5, decimal=7)
# Outside of the pressure bounds - extrapolation off
assert np.ma.is_masked(new_data.data[3])
# Second dataset:
new_data = outlist[1]
# Exactly on the lat, lon, time points, interpolated over pressure
assert_almost_equal(new_data.data[0], 321.5, decimal=5)
# Exactly on the lat, lon, points, interpolated over time and pressure
assert_almost_equal(new_data.data[1], 326.5, decimal=7)
# Exactly on the lat, time points, interpolated over longitude and pressure
assert_almost_equal(new_data.data[2], 430.5, decimal=7)
# Outside of the pressure bounds - extrapolation off
assert np.ma.is_masked(new_data.data[3])
def test_collocation_of_alt_points_on_hybrid_altitude_and_pressure_coordinates(
self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3,
hybrid_pr_len=10,
geopotential_height=True))
sample_points = UngriddedData.from_points_array(
# Test point with both pressure and altitude should interpolate over the altitude only (since that is also
# present in the data cube)
[
HyperPoint(lat=0.0,
lon=0.0,
alt=234.5,
t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=5.0,
lon=5.0,
alt=355.5,
t=dt.datetime(1984, 8, 28, 0, 0))
])
col = GriddedUngriddedCollocator(fill_value=np.NAN)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 225.5, decimal=7)
assert_almost_equal(new_data.data[1], 346.5, decimal=7)
def test_missing_data_for_missing_sample_with_no_extrapolation(self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=0.0,
lon=0.0,
alt=5550.0,
t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=4.0,
lon=4.0,
alt=6000.0,
t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=-4.0,
lon=-4.0,
alt=6500.0,
t=dt.datetime(1984, 8, 27))
])
sample_mask = [False, True, False]
sample_points.data = np.ma.array([0, 0, 0], mask=sample_mask)
col = GriddedUngriddedCollocator(fill_value=np.NAN,
missing_data_for_missing_sample=True)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 222.4814815, decimal=7)
# This point should be masked because of the sampling
assert np.ma.is_masked(new_data.data[1])
# And this one because of the extrapolation
assert np.ma.is_masked(new_data.data[2])
def test_already_collocated_in_col_gridded_to_ungridded_in_2d(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube())
# This point already exists on the cube with value 5 - which shouldn't be a problem
sample_points = UngriddedData.from_points_array([HyperPoint(0.0, 0.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 8.0)
def test_wrapping_of_alt_points_on_hybrid_height_coordinates_on_0_360_grid(
self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3,
hybrid_ht_len=10,
lon_dim_length=36,
lon_range=(0., 350.)))
# Shift the cube around so that the dim which isn't hybrid (time) is at the front. This breaks the fix we used
# for air pressure...
cube.transpose([2, 0, 1, 3])
# Ensure the longitude coord is circular
cube.coord(standard_name='longitude').circular = True
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=4.0,
lon=355.0,
alt=11438.0,
t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=0.0,
lon=2.0,
alt=10082.0,
t=dt.datetime(1984, 8, 28))
])
col = GriddedUngriddedCollocator(extrapolate=False)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
eq_(new_data.data[0], 3563.0)
eq_(new_data.data[1], 2185.0)
def test_extrapolation_of_pres_points_on_hybrid_pressure_coordinates_multi_var(
self):
cube_list = [
make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))
]
cube_list.append(
make_from_cube(
mock.make_mock_cube(time_dim_length=3,
hybrid_pr_len=10,
data_offset=100)))
sample_points = UngriddedData.from_points_array(
# Point interpolated in the horizontal and then extrapolated past the top vertical layer (by one layer)
[
HyperPoint(lat=-4.0,
lon=-4.0,
pres=68400050.0,
t=dt.datetime(1984, 8, 27))
])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube_list, None, 'lin')
assert_almost_equal(new_data[0].data[0], 125.0, decimal=7)
assert_almost_equal(new_data[1].data[0], 225.0, decimal=7)
def test_extrapolation_of_pres_points_on_hybrid_pressure_coordinates(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))
sample_points = UngriddedData.from_points_array(
# Point interpolated in the horizontal and then extrapolated past the top vertical layer (by one layer)
[HyperPoint(lat=-4.0, lon=-4.0, pres=68400050.0, t=dt.datetime(1984, 8, 27))])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 125.0, decimal=7)
def test_collocation_of_pres_alt_points_on_hybrid_pressure_coordinates_multi_var(
self):
cube_list = [
make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))
]
cube_list.append(
make_from_cube(
mock.make_mock_cube(time_dim_length=3,
hybrid_pr_len=10,
data_offset=100)))
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=0.0,
lon=0.0,
pres=111100040.5,
alt=5000,
t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=0.0,
lon=0.0,
pres=113625040.5,
alt=4000,
t=dt.datetime(1984, 8, 28, 12, 0, 0)),
HyperPoint(lat=5.0,
lon=2.5,
pres=177125044.5,
alt=3000,
t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=-4.0,
lon=-4.0,
pres=166600039.0,
alt=3500,
t=dt.datetime(1984, 8, 27))
])
col = GriddedUngriddedCollocator()
outlist = col.collocate(sample_points, cube_list, None, 'lin')
# First data set:
new_data = outlist[0]
# Exactly on the lat, lon, time points, interpolated over pressure
assert_almost_equal(new_data.data[0], 221.5, decimal=5)
# Exactly on the lat, lon, points, interpolated over time and pressure
assert_almost_equal(new_data.data[1], 226.5, decimal=7)
# Exactly on the lat, time points, interpolated over longitude and pressure
assert_almost_equal(new_data.data[2], 330.5, decimal=7)
# Outside of the pressure bounds - extrapolation off
assert np.ma.is_masked(new_data.data[3])
# Second dataset:
new_data = outlist[1]
# Exactly on the lat, lon, time points, interpolated over pressure
assert_almost_equal(new_data.data[0], 321.5, decimal=5)
# Exactly on the lat, lon, points, interpolated over time and pressure
assert_almost_equal(new_data.data[1], 326.5, decimal=7)
# Exactly on the lat, time points, interpolated over longitude and pressure
assert_almost_equal(new_data.data[2], 430.5, decimal=7)
# Outside of the pressure bounds - extrapolation off
assert np.ma.is_masked(new_data.data[3])
def test_basic_col_gridded_to_ungridded_using_li_in_2d(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array(
[HyperPoint(1.0, 1.0), HyperPoint(4.0, 4.0), HyperPoint(-4.0, -4.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 8.8)
assert_almost_equal(new_data.data[1], 11.2)
assert_almost_equal(new_data.data[2], 4.8)
def test_alt_extrapolation(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=-4.0, lon=-4.0, alt=6382.8, t=dt.datetime(1984, 8, 27))])
col = GriddedUngriddedCollocator(fill_value=np.NAN, extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 126.0, decimal=7)
def test_coordinates_outside_grid_in_col_gridded_to_ungridded_in_2d(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array(
[HyperPoint(5.5, 5.5), HyperPoint(-5.5, 5.5), HyperPoint(5.5, -5.5), HyperPoint(-5.5, -5.5)])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 12.0)
eq_(new_data.data[1], 6.0)
eq_(new_data.data[2], 10.0)
eq_(new_data.data[3], 4.0)
def test_collocation_of_alt_points_on_hybrid_altitude_coordinates_on_0_360_grid(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10, lon_dim_length=36,
lon_range=(0., 350.)))
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=1.0, lon=111.0, alt=5000.0, t=dt.datetime(1984, 8, 28, 8, 34)),
HyperPoint(lat=4.0, lon=141.0, alt=12000.0, t=dt.datetime(1984, 8, 28, 8, 34))])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 2501.0) # float(cube[2,11,1,0].data))
eq_(new_data.data[1], 3675.0) # float(cube[3,14,1,4].data))
def test_negative_lon_points_on_hybrid_pressure_coordinates_dont_matter(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=0.0, lon=0.0, pres=111100040.5, t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=5.0, lon=2.5, pres=177125044.5, t=dt.datetime(1984, 8, 28, 0, 0, 0))])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
# Exactly on the lat, lon, time points, interpolated over pressure
assert_almost_equal(new_data.data[0], 221.5, decimal=5)
# Exactly on the lat, time points, interpolated over latitude and pressure
assert_almost_equal(new_data.data[1], 330.5, decimal=7)
def test_negative_lon_points_in_2d_dont_matter(self):
"""
This is exactly the same test as above, except we ommit the point with negative longitude, this makes the
collocator wrap the longitude coordinate and gives a slightly different interpolation result...
"""
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array(
[HyperPoint(1.0, 1.0), HyperPoint(4.0, 4.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 8.8)
assert_almost_equal(new_data.data[1], 11.2)
def test_basic_col_gridded_to_ungridded_in_2d(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=1.0, lon=1.0),
HyperPoint(lat=4.0, lon=4.0),
HyperPoint(lat=-4.0, lon=-4.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 8.0) # float(cube[2,1].data))
eq_(new_data.data[1], 12.0) # float(cube[3,2].data))
eq_(new_data.data[2], 4.0) # float(cube[1,0].data))
def test_collocation_of_alt_pres_points_on_hybrid_altitude_and_pressure_coordinates(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10,
geopotential_height=True))
sample_points = UngriddedData.from_points_array(
# Test point with both pressure and altitude should interpolate over the altitude only (since that is also
# present in the data cube)
[HyperPoint(lat=0.0, lon=0.0, alt=234.5, pres=1000, t=dt.datetime(1984, 8, 28, 0, 0, 0))])
col = GriddedUngriddedCollocator(fill_value=np.NAN)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 225.5, decimal=7)
def test_basic_col_gridded_to_ungridded_in_2d_with_time(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube_with_time())
sample_points = [HyperPoint(lat=1.0, lon=1.0, t=dt.datetime(1984, 8, 28, 8, 34)),
HyperPoint(lat=4.0, lon=4.0, t=dt.datetime(1984, 8, 31, 1, 23)),
HyperPoint(lat=-4.0, lon=-4.0, t=dt.datetime(1984, 9, 2, 15, 54))]
sample_points = UngriddedData.from_points_array(sample_points)
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 51.0)
eq_(new_data.data[1], 82.0)
eq_(new_data.data[2], 28.0)
def test_negative_lon_points_in_2d_dont_matter(self):
"""
This is exactly the same test as above, except we ommit the point with negative longitude, this makes the
collocator wrap the longitude coordinate and gives a slightly different interpolation result...
"""
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array(
[HyperPoint(1.0, 1.0), HyperPoint(4.0, 4.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 8.8)
assert_almost_equal(new_data.data[1], 11.2)
def test_basic_col_gridded_to_ungridded_using_li_in_2d(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array([
HyperPoint(1.0, 1.0),
HyperPoint(4.0, 4.0),
HyperPoint(-4.0, -4.0)
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 8.8)
assert_almost_equal(new_data.data[1], 11.2)
assert_almost_equal(new_data.data[2], 4.8)
def ensemble_collocate(ensemble, observations, member_dimension='job'):
"""
Efficiently collocate (interpolate) many ensemble members on to a set of (un-gridded) observations
Note
----
This function requires both Iris and CIS to be installed
Parameters
----------
ensemble: ~cis.data_io.gridded_data.GriddedData
The ensemble of (model) samples to interpolate on to the observations
observations: ~cis.data_io.ungridded_data.UngriddedData
The observations on to which the observations will be sampled
member_dimension: str
The name of the dimension which represents the ensemble members in `ensemble`
Returns
-------
col_ensemble: iris.cube.Cube
The ensemble values interpolated on to the observation locations, with the ensemble members
along the leading dimension.
"""
from iris.cube import Cube, CubeList
from iris.coords import DimCoord, AuxCoord
from cis.collocation.col_implementations import GriddedUngriddedCollocator, DummyConstraint
from cis.data_io.gridded_data import make_from_cube
col = GriddedUngriddedCollocator(missing_data_for_missing_sample=False)
col_members = CubeList()
for member in ensemble.slices_over(member_dimension):
# Use CIS to collocate each ensemble member on to the observations
# The interpolation weights are cached within col automatically
collocated_job, = col.collocate(observations, make_from_cube(member),
DummyConstraint(), 'lin')
# Turn the interpolated data in to a flat cube for easy stacking
new_c = Cube(collocated_job.data.reshape(1, -1),
long_name=collocated_job.name(),
units='1',
dim_coords_and_dims=[
(DimCoord(np.arange(collocated_job.data.shape[0]),
long_name="obs"), 1),
(DimCoord(member.coord(member_dimension).points,
long_name=member_dimension), 0)
],
aux_coords_and_dims=[
(AuxCoord(c.points, standard_name=c.standard_name), 1)
for c in collocated_job.coords()
])
col_members.append(new_c)
col_ensemble = col_members.concatenate_cube()
return col_ensemble
def test_basic_col_gridded_to_ungridded_in_2d(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=1.0, lon=1.0),
HyperPoint(lat=4.0, lon=4.0),
HyperPoint(lat=-4.0, lon=-4.0)
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 8.0) # float(cube[2,1].data))
eq_(new_data.data[1], 12.0) # float(cube[3,2].data))
eq_(new_data.data[2], 4.0) # float(cube[1,0].data))
def test_negative_lon_points_on_hybrid_altitude_coordinates_dont_matter(self):
"""This should give the same results as above"""
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array(
# This point actually lies outside the lower bounds for altitude at this point in space
[HyperPoint(lat=1.0, lon=1.0, alt=5000.0, t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies in the middle of the altitude bounds at this point
HyperPoint(lat=4.0, lon=4.0, alt=6000.0, t=dt.datetime(1984, 8, 28, 8, 34))])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 221.0) # float(cube[2,1,1,0].data))
eq_(new_data.data[1], 345.0) # float(cube[3,2,1,4].data))
def test_basic_col_with_circular_lon(self):
cube = make_from_cube(mock.make_dummy_2d_cube_with_circular_lon())
sample_points = UngriddedData.from_points_array(
[HyperPoint(0.0, 0.0), HyperPoint(0.0, 355.0), HyperPoint(0.0, 360.0),
HyperPoint(80.0, 0.0), HyperPoint(85.0, 355.0), HyperPoint(90.0, 360.0),
HyperPoint(-80.0, 0.0), HyperPoint(-85.0, 355.0), HyperPoint(-90.0, 360.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
wanted = np.asarray([325.0, 342.5, 325.0,
613.0, (630.5 + 666.5) / 2, 649.0,
37.0, (54.5 + 18.5) / 2, 1.0])
assert_almost_equal(new_data.data, wanted)
def test_order_of_coords_doesnt_matter(self):
from iris.cube import Cube
from iris.coords import DimCoord
import numpy as np
from cis.data_io.gridded_data import make_from_cube
from cis.data_io.ungridded_data import UngriddedCoordinates, Metadata
from cis.data_io.Coord import Coord
cube_lat = DimCoord(np.linspace(-90, 90, 18),
standard_name='latitude',
units='degrees')
cube_lon = DimCoord(np.linspace(0, 359, 36),
standard_name='longitude',
units='degrees',
circular=True)
cube_alt = DimCoord(np.linspace(0, 10000, 10),
standard_name='altitude',
units='meters')
times = np.linspace(0, 30, 12)
cube_time = DimCoord(times,
standard_name='time',
units='days since 1970-01-01 00:00:00')
data = np.arange(12 * 18 * 36 * 10).reshape(12, 18, 36, 10)
source = make_from_cube(
Cube(data,
dim_coords_and_dims=[(cube_time, 0), (cube_lat, 1),
(cube_lon, 2), (cube_alt, 3)]))
n = 10
sample_lats = np.linspace(-9.1, 9.9, n)
sample_longs = np.linspace(-9.1, 9.9, n)
sample_alts = np.linspace(99, 599, n)
sample_times = np.linspace(0, 30, n)
sample = UngriddedCoordinates([
Coord(sample_lats, Metadata('latitude')),
Coord(sample_longs, Metadata('longitude')),
Coord(sample_alts, Metadata('altitude')),
Coord(sample_times,
Metadata('time', units='days since 1970-01-01 00:00:00'))
])
col = GriddedUngriddedCollocator()
output = col.collocate(sample, source, None, 'nn')[0]
source.transpose()
col = GriddedUngriddedCollocator()
assert_equal(
col.collocate(sample, source, None, 'nn')[0].data, output.data)
source.transpose((2, 1, 0, 3))
col = GriddedUngriddedCollocator()
assert_equal(
col.collocate(sample, source, None, 'nn')[0].data, output.data)
def test_coordinates_outside_grid_in_col_gridded_to_ungridded_in_2d(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array([
HyperPoint(5.5, 5.5),
HyperPoint(-5.5, 5.5),
HyperPoint(5.5, -5.5),
HyperPoint(-5.5, -5.5)
])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 12.0)
eq_(new_data.data[1], 6.0)
eq_(new_data.data[2], 10.0)
eq_(new_data.data[3], 4.0)
def test_wrapping_of_pres_points_on_hybrid_pressure_coordinates_on_0_360_grid(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10, lon_dim_length=36,
lon_range=(0., 350.)))
# Ensure the longitude coord is circular
cube.coord(standard_name='longitude').circular = True
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=0.0, lon=355.0, pres=1482280045.0, t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=5.0, lon=2.5, pres=1879350048.0, t=dt.datetime(1984, 8, 28, 0, 0, 0))])
col = GriddedUngriddedCollocator(extrapolate=False)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
eq_(new_data.data[0], 2701.0011131725005)
eq_(new_data.data[1], 3266.1930161260775)
def test_alt_extrapolation(self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=-4.0,
lon=-4.0,
alt=6382.8,
t=dt.datetime(1984, 8, 27))
])
col = GriddedUngriddedCollocator(fill_value=np.NAN, extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 126.0, decimal=7)
def test_basic_col_gridded_to_ungridded_in_2d_with_time(self):
cube = make_from_cube(mock.make_square_5x3_2d_cube_with_time())
sample_points = [
HyperPoint(lat=1.0, lon=1.0, t=dt.datetime(1984, 8, 28, 8, 34)),
HyperPoint(lat=4.0, lon=4.0, t=dt.datetime(1984, 8, 31, 1, 23)),
HyperPoint(lat=-4.0, lon=-4.0, t=dt.datetime(1984, 9, 2, 15, 54))
]
sample_points = UngriddedData.from_points_array(sample_points)
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 51.0)
eq_(new_data.data[1], 82.0)
eq_(new_data.data[2], 28.0)
def test_collocation_of_alt_pres_points_on_hybrid_altitude_coordinates(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=0.0, lon=0.0, alt=5550.0, pres=10000.0, t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=4.0, lon=4.0, alt=6000.0, pres=1000.0, t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=-4.0, lon=-4.0, alt=6500.0, pres=100.0, t=dt.datetime(1984, 8, 27))])
col = GriddedUngriddedCollocator(fill_value=np.NAN)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 222.4814815, decimal=7)
assert_almost_equal(new_data.data[1], 321.0467626, decimal=7)
# Test that points outside the cell are returned as masked, rather than extrapolated by default
assert np.ma.is_masked(new_data.data[2])
def test_negative_lon_points_dont_matter_with_0_360_grid_in_2d(self):
# This cube is defined over a 0-360 longitude grid
cube = make_from_cube(mock.make_dummy_2d_cube())
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=1.0, lon=1.0),
HyperPoint(lat=19.0, lon=44.0),
HyperPoint(lat=-4.0, lon=-14.0),
HyperPoint(lat=-4.0, lon=-44.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 325.0) # float(cube[9,0].data)
eq_(new_data.data[1], 365.0) # float(cube[10,4].data))
eq_(new_data.data[2], 324.0) # float(cube[8,35].data))
eq_(new_data.data[3], 321.0) # float(cube[8,32].data))
def test_lon_points_over_360_dont_matter_with_0_360_grid_in_2d(self):
# This cube is defined over a 0-360 longitude grid
cube = make_from_cube(mock.make_dummy_2d_cube())
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=1.0, lon=0.0),
HyperPoint(lat=1.0, lon=20.0),
HyperPoint(lat=1.0, lon=361.0),
HyperPoint(lat=1.0, lon=381.0)])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 325.0) # float(cube[9,0].data))
eq_(new_data.data[1], 327.0) # float(cube[9,0].data))
eq_(new_data.data[2], 325.0) # float(cube[9,0].data))
eq_(new_data.data[3], 327.0) # float(cube[9,2].data))
def test_negative_lon_points_on_hybrid_altitude_coordinates_dont_matter(self):
"""
This is exactly the same test as above, except we ommit the point with negative longitude, this makes the
collocator wrap the longitude coordinate and gives a slightly different interpolation result...
"""
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=4.0, lon=4.0, alt=6000.0, t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=0.0, lon=0.0, alt=5550.0, t=dt.datetime(1984, 8, 28))])
col = GriddedUngriddedCollocator(fill_value=np.NAN)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 321.0467626, decimal=7)
assert_almost_equal(new_data.data[1], 222.4814815, decimal=7)
def test_lon_points_over_360_dont_matter_with_0_360_grid_in_2d(self):
# This cube is defined over a 0-360 longitude grid
cube = make_from_cube(mock.make_dummy_2d_cube())
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=1.0, lon=0.0),
HyperPoint(lat=1.0, lon=20.0),
HyperPoint(lat=1.0, lon=361.0),
HyperPoint(lat=1.0, lon=381.0)
])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 325.0) # float(cube[9,0].data))
eq_(new_data.data[1], 327.0) # float(cube[9,0].data))
eq_(new_data.data[2], 325.0) # float(cube[9,0].data))
eq_(new_data.data[3], 327.0) # float(cube[9,2].data))
def test_negative_lon_points_dont_matter_with_0_360_grid_in_2d(self):
# This cube is defined over a 0-360 longitude grid
cube = make_from_cube(mock.make_dummy_2d_cube())
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=1.0, lon=1.0),
HyperPoint(lat=19.0, lon=44.0),
HyperPoint(lat=-4.0, lon=-14.0),
HyperPoint(lat=-4.0, lon=-44.0)
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 325.0) # float(cube[9,0].data)
eq_(new_data.data[1], 365.0) # float(cube[10,4].data))
eq_(new_data.data[2], 324.0) # float(cube[8,35].data))
eq_(new_data.data[3], 321.0) # float(cube[8,32].data))
def test_collocation_of_alt_points_on_hybrid_altitude_coordinates(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10,
geopotential_height=True))
sample_points = UngriddedData.from_points_array(
# This point actually lies outside the lower bounds for altitude at this point in space
[HyperPoint(lat=1.0, lon=1.0, alt=5000.0, t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies in the middle of the altitude bounds at this point
HyperPoint(lat=4.0, lon=4.0, alt=6000.0, t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies outside the upper bounds for altitude at this point
HyperPoint(lat=-4.0, lon=-4.0, alt=6500.0, t=dt.datetime(1984, 8, 27, 2, 18, 52))])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 221.0) # float(cube[2,1,1,0].data))
eq_(new_data.data[1], 345.0) # float(cube[3,2,1,4].data))
eq_(new_data.data[2], 100.0) # float(cube[1,0,0,9].data))
def test_collocation_over_scalar_coord(self):
# A scalar time coordinate should make no difference when collocating points with a time value.
cube = make_from_cube(mock.make_square_5x3_2d_cube_with_scalar_time())
sample_points = UngriddedData.from_points_array(
[HyperPoint(1.0, 1.0, t=dt.datetime(1984, 8, 22, 0, 0, 0)),
HyperPoint(4.0, 4.0, t=dt.datetime(1984, 8, 28, 0, 0, 0)),
# Note that it doesn't even matter if the point is outside the bounds of the scalar time...
# TODO: Is this correct behaviour?
HyperPoint(-4.0, -4.0, t=dt.datetime(1984, 10, 1, 0, 0, 0))])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 8.8)
assert_almost_equal(new_data.data[1], 11.2)
assert_almost_equal(new_data.data[2], 4.8)
def test_coordinates_exactly_between_points_in_col_gridded_to_ungridded_in_2d(self):
"""
This works out the edge case where the points are exactly in the middle or two or more datapoints.
Iris seems to count a point as 'belonging' to a datapoint if it is greater than a datapoint cell's lower
bound and less than or equal to it's upper bound. Where a cell is an imaginary boundary around a
datapoint which divides the grid.
"""
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array(
[HyperPoint(2.5, 2.5), HyperPoint(-2.5, 2.5), HyperPoint(2.5, -2.5), HyperPoint(-2.5, -2.5)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 8.0)
eq_(new_data.data[1], 5.0)
eq_(new_data.data[2], 7.0)
eq_(new_data.data[3], 4.0)
def test_wrapping_of_alt_points_on_hybrid_height_coordinates_on_0_360_grid(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10, lon_dim_length=36,
lon_range=(0., 350.)))
# Shift the cube around so that the dim which isn't hybrid (time) is at the front. This breaks the fix we used
# for air pressure...
cube.transpose([2, 0, 1, 3])
# Ensure the longitude coord is circular
cube.coord(standard_name='longitude').circular = True
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=4.0, lon=355.0, alt=11438.0, t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=0.0, lon=2.0, alt=10082.0, t=dt.datetime(1984, 8, 28))])
col = GriddedUngriddedCollocator(extrapolate=False)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
eq_(new_data.data[0], 3563.0)
eq_(new_data.data[1], 2185.0)
def test_collocation_over_scalar_coord(self):
# A scalar time coordinate should make no difference when collocating points with a time value.
cube = make_from_cube(mock.make_square_5x3_2d_cube_with_scalar_time())
sample_points = UngriddedData.from_points_array([
HyperPoint(1.0, 1.0, t=dt.datetime(1984, 8, 22, 0, 0, 0)),
HyperPoint(4.0, 4.0, t=dt.datetime(1984, 8, 28, 0, 0, 0)),
# Note that it doesn't even matter if the point is outside the bounds of the scalar time...
# TODO: Is this correct behaviour?
HyperPoint(-4.0, -4.0, t=dt.datetime(1984, 10, 1, 0, 0, 0))
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 8.8)
assert_almost_equal(new_data.data[1], 11.2)
assert_almost_equal(new_data.data[2], 4.8)
def test_collocation_of_pres_alt_points_on_hybrid_pressure_coordinates_nn(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3))
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=0.0, lon=0.0, t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=0.0, lon=0.0, t=dt.datetime(1984, 8, 28, 12, 0, 0)),
HyperPoint(lat=5.0, lon=2.5, t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=-4.0, lon=-4.0, t=dt.datetime(1984, 8, 27))])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
# Exactly on the lat, lon, time points, interpolated over pressure
assert_almost_equal(new_data.data[0], 23.0, decimal=5)
# Exactly on the lat, lon, points, interpolated over time and pressure
assert_almost_equal(new_data.data[1], 23.5, decimal=7)
# Exactly on the lat, time points, interpolated over longitude and pressure
assert_almost_equal(new_data.data[2], 33.5, decimal=7)
# Outside of the pressure bounds - extrapolation off
assert_almost_equal(new_data.data[3], 12.4, decimal=7)
def test_guessing_the_bounds_on_a_cube_doesnt_matter_for_negative_lon_points_on_a_0_360_grid_in_2d(self):
"""This should be identical to above but there was an issue in iris where this caused a problem"""
# This cube is defined over a 0-360 longitude grid
cube = make_from_cube(mock.make_dummy_2d_cube())
cube.coord(standard_name='longitude').guess_bounds()
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=1.0, lon=1.0),
HyperPoint(lat=19.0, lon=44.0),
HyperPoint(lat=-4.0, lon=-14.0),
HyperPoint(lat=-4.0, lon=-44.0)])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 325.0) # float(cube[9,0].data)
eq_(new_data.data[1], 365.0) # float(cube[10,4].data))
eq_(new_data.data[2], 324.0) # float(cube[8,35].data))
eq_(new_data.data[3], 321.0) # float(cube[8,32].data))
def test_collocation_of_pres_points_on_hybrid_pressure_coordinates_and_altitude_coordinates(self):
"""
When only pressure coordinate is present this should be used for the collocation
"""
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))
sample_points = UngriddedData.from_points_array(
# This point actually lies outside the lower bounds for altitude at this point in space
[HyperPoint(lat=1.0, lon=1.0, pres=1100000.0, t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies in the middle of the altitude bounds at this point
HyperPoint(lat=4.0, lon=4.0, pres=184600000.0, t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies outside the upper bounds for altitude at this point
HyperPoint(lat=-4.0, lon=-4.0, pres=63100049.0, t=dt.datetime(1984, 8, 27, 2, 18, 52))])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], float(cube[2, 1, 1, 0].data))
eq_(new_data.data[1], float(cube[3, 2, 1, 4].data))
eq_(new_data.data[2], float(cube[1, 0, 0, 9].data))
def test_gridded_ungridded_lin(self):
data = make_from_cube(mock.make_mock_cube())
data.name = lambda: 'Name'
data.var_name = 'var_name'
data._standard_name = 'y_wind'
sample = UngriddedData.from_points_array(
[HyperPoint(lat=1.0, lon=1.0, alt=12.0, t=dt.datetime(1984, 8, 29, 8, 34)),
HyperPoint(lat=3.0, lon=3.0, alt=7.0, t=dt.datetime(1984, 8, 29, 8, 34)),
HyperPoint(lat=-1.0, lon=-1.0, alt=5.0, t=dt.datetime(1984, 8, 29, 8, 34))])
constraint = None
col = GriddedUngriddedCollocator()
output = col.collocate(sample, data, constraint, 'lin')
expected_result = np.array([8.8, 10.4, 7.2])
assert len(output) == 1
assert isinstance(output, UngriddedDataList)
assert np.allclose(output[0].data, expected_result)
def test_missing_data_for_missing_sample_with_no_extrapolation(self):
cube = make_from_cube(mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array(
[HyperPoint(lat=0.0, lon=0.0, alt=5550.0, t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=4.0, lon=4.0, alt=6000.0, t=dt.datetime(1984, 8, 28)),
HyperPoint(lat=-4.0, lon=-4.0, alt=6500.0, t=dt.datetime(1984, 8, 27))])
sample_mask = [False, True, False]
sample_points.data = np.ma.array([0, 0, 0], mask=sample_mask)
col = GriddedUngriddedCollocator(fill_value=np.NAN, missing_data_for_missing_sample=True)
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
assert_almost_equal(new_data.data[0], 222.4814815, decimal=7)
# This point should be masked because of the sampling
assert np.ma.is_masked(new_data.data[1])
# And this one because of the extrapolation
assert np.ma.is_masked(new_data.data[2])
def test_basic_col_with_circular_lon(self):
cube = make_from_cube(mock.make_dummy_2d_cube_with_circular_lon())
sample_points = UngriddedData.from_points_array([
HyperPoint(0.0, 0.0),
HyperPoint(0.0, 355.0),
HyperPoint(0.0, 360.0),
HyperPoint(80.0, 0.0),
HyperPoint(85.0, 355.0),
HyperPoint(90.0, 360.0),
HyperPoint(-80.0, 0.0),
HyperPoint(-85.0, 355.0),
HyperPoint(-90.0, 360.0)
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
wanted = np.asarray(
[325.0, 360.0, 325.0, 613.0, 648.0, 649.0, 37.0, 36.0, 1.0])
assert_almost_equal(new_data.data, wanted)
def test_no_missing_data_for_missing_sample(self):
data = make_from_cube(mock.make_mock_cube())
data.name = lambda: 'Name'
data.var_name = 'var_name'
data._standard_name = 'y_wind'
sample = UngriddedData.from_points_array(
[HyperPoint(lat=1.0, lon=1.0, alt=12.0, t=dt.datetime(1984, 8, 29, 8, 34)),
HyperPoint(lat=3.0, lon=3.0, alt=7.0, t=dt.datetime(1984, 8, 29, 8, 34)),
HyperPoint(lat=-1.0, lon=-1.0, alt=5.0, t=dt.datetime(1984, 8, 29, 8, 34))])
constraint = None
sample_mask = [False, True, False]
sample.data = np.ma.array([0, 0, 0], mask=sample_mask)
col = GriddedUngriddedCollocator(missing_data_for_missing_sample=False)
output = col.collocate(sample, data, constraint, 'nn')
assert len(output) == 1
assert isinstance(output, UngriddedDataList)
assert not any(output[0].data.mask)
def test_guessing_the_bounds_on_a_cube_doesnt_matter_for_negative_lon_points_on_a_0_360_grid_in_2d(
self):
"""This should be identical to above but there was an issue in iris where this caused a problem"""
# This cube is defined over a 0-360 longitude grid
cube = make_from_cube(mock.make_dummy_2d_cube())
cube.coord(standard_name='longitude').guess_bounds()
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=1.0, lon=1.0),
HyperPoint(lat=19.0, lon=44.0),
HyperPoint(lat=-4.0, lon=-14.0),
HyperPoint(lat=-4.0, lon=-44.0)
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 325.0) # float(cube[9,0].data)
eq_(new_data.data[1], 365.0) # float(cube[10,4].data))
eq_(new_data.data[2], 324.0) # float(cube[8,35].data))
eq_(new_data.data[3], 321.0) # float(cube[8,32].data))
def test_coordinates_exactly_between_points_in_col_gridded_to_ungridded_in_2d(
self):
"""
This works out the edge case where the points are exactly in the middle or two or more datapoints.
Iris seems to count a point as 'belonging' to a datapoint if it is greater than a datapoint cell's lower
bound and less than or equal to it's upper bound. Where a cell is an imaginary boundary around a
datapoint which divides the grid.
"""
cube = make_from_cube(mock.make_square_5x3_2d_cube())
sample_points = UngriddedData.from_points_array([
HyperPoint(2.5, 2.5),
HyperPoint(-2.5, 2.5),
HyperPoint(2.5, -2.5),
HyperPoint(-2.5, -2.5)
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 8.0)
eq_(new_data.data[1], 5.0)
eq_(new_data.data[2], 7.0)
eq_(new_data.data[3], 4.0)
def test_negative_lon_points_on_hybrid_pressure_coordinates_dont_matter(
self):
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_pr_len=10))
sample_points = UngriddedData.from_points_array([
HyperPoint(lat=0.0,
lon=0.0,
pres=111100040.5,
t=dt.datetime(1984, 8, 28, 0, 0, 0)),
HyperPoint(lat=5.0,
lon=2.5,
pres=177125044.5,
t=dt.datetime(1984, 8, 28, 0, 0, 0))
])
col = GriddedUngriddedCollocator()
new_data = col.collocate(sample_points, cube, None, 'lin')[0]
# Exactly on the lat, lon, time points, interpolated over pressure
assert_almost_equal(new_data.data[0], 221.5, decimal=5)
# Exactly on the lat, time points, interpolated over latitude and pressure
assert_almost_equal(new_data.data[1], 330.5, decimal=7)
def test_negative_lon_points_on_hybrid_altitude_coordinates_dont_matter(
self):
"""This should give the same results as above"""
cube = make_from_cube(
mock.make_mock_cube(time_dim_length=3, hybrid_ht_len=10))
sample_points = UngriddedData.from_points_array(
# This point actually lies outside the lower bounds for altitude at this point in space
[
HyperPoint(lat=1.0,
lon=1.0,
alt=5000.0,
t=dt.datetime(1984, 8, 28, 8, 34)),
# This point lies in the middle of the altitude bounds at this point
HyperPoint(lat=4.0,
lon=4.0,
alt=6000.0,
t=dt.datetime(1984, 8, 28, 8, 34))
])
col = GriddedUngriddedCollocator(extrapolate=True)
new_data = col.collocate(sample_points, cube, None, 'nn')[0]
eq_(new_data.data[0], 221.0) # float(cube[2,1,1,0].data))
eq_(new_data.data[1], 345.0) # float(cube[3,2,1,4].data))
def test_order_of_coords_doesnt_matter(self):
from iris.cube import Cube
from iris.coords import DimCoord
import numpy as np
from cis.data_io.gridded_data import make_from_cube
from cis.data_io.ungridded_data import UngriddedCoordinates, Metadata
from cis.data_io.Coord import Coord
cube_lat = DimCoord(np.linspace(-90, 90, 18), standard_name='latitude', units='degrees')
cube_lon = DimCoord(np.linspace(0, 359, 36), standard_name='longitude', units='degrees', circular=True)
cube_alt = DimCoord(np.linspace(0, 10000, 10), standard_name='altitude', units='meters')
times = np.linspace(0, 30, 12)
cube_time = DimCoord(times, standard_name='time', units='days since 1970-01-01 00:00:00')
data = np.arange(12 * 18 * 36 * 10).reshape(12, 18, 36, 10)
source = make_from_cube(Cube(data,
dim_coords_and_dims=[(cube_time, 0), (cube_lat, 1), (cube_lon, 2), (cube_alt, 3)]))
n = 10
sample_lats = np.linspace(-9.1, 9.9, n)
sample_longs = np.linspace(-9.1, 9.9, n)
sample_alts = np.linspace(99, 599, n)
sample_times = np.linspace(0, 30, n)
sample = UngriddedCoordinates(
[Coord(sample_lats, Metadata('latitude')), Coord(sample_longs, Metadata('longitude')),
Coord(sample_alts, Metadata('altitude')),
Coord(sample_times, Metadata('time', units='days since 1970-01-01 00:00:00'))])
col = GriddedUngriddedCollocator()
output = col.collocate(sample, source, None, 'nn')[0]
source.transpose()
col = GriddedUngriddedCollocator()
assert_equal(col.collocate(sample, source, None, 'nn')[0].data, output.data)
source.transpose((2, 1, 0, 3))
col = GriddedUngriddedCollocator()
assert_equal(col.collocate(sample, source, None, 'nn')[0].data, output.data)
