def test_complete_collapse_one_dim_using_moments_kernel(self):
self.kernel = aggregation_kernels['moments']
data1 = make_from_cube(make_5x3_lon_lat_2d_cube_with_missing_data())
data1.var_name = 'var1'
data2 = make_from_cube(make_5x3_lon_lat_2d_cube_with_missing_data())
data2.var_name = 'var2'
data2.data += 10
data = GriddedDataList([data1, data2])
output = data.collapsed(['x'], how=self.kernel)
expect_mean = numpy.array([[5.5, 8.75, 9]])
expect_stddev = numpy.array([numpy.sqrt(15), numpy.sqrt(26.25), numpy.sqrt(30)])
expect_count = numpy.array([[4, 4, 4]])
assert isinstance(output, GriddedDataList)
assert len(output) == 6
mean_1, stddev_1, count_1, mean_2, stddev_2, count_2 = output
assert mean_1.var_name == 'var1'
assert stddev_1.var_name == 'var1_std_dev'
assert count_1.var_name == 'var1_num_points'
assert mean_2.var_name == 'var2'
assert stddev_2.var_name == 'var2_std_dev'
assert count_2.var_name == 'var2_num_points'
assert_arrays_almost_equal(mean_1.data, expect_mean)
assert_arrays_almost_equal(mean_2.data, expect_mean + 10)
assert_arrays_almost_equal(stddev_1.data, expect_stddev)
assert_arrays_almost_equal(stddev_2.data, expect_stddev)
assert_arrays_almost_equal(count_1.data, expect_count)
assert_arrays_almost_equal(count_2.data, expect_count)
def test_partial_aggregation_over_more_than_one_dim_on_multidimensional_coord(
self):
from cis.data_io.gridded_data import GriddedDataList, make_from_cube
data1 = make_from_cube(
make_mock_cube(time_dim_length=7, hybrid_pr_len=5))
data2 = make_from_cube(
make_mock_cube(time_dim_length=7, hybrid_pr_len=5, data_offset=1))
datalist = GriddedDataList([data1, data2])
cube_out = datalist.collapsed(['t', 'x'], how=self.kernel)
result_data = numpy.array([[51.0, 52.0, 53.0, 54.0, 55.0],
[156.0, 157.0, 158.0, 159.0, 160.0],
[261.0, 262.0, 263.0, 264.0, 265.0],
[366.0, 367.0, 368.0, 369.0, 370.0],
[471.0, 472.0, 473.0, 474.0, 475.0]],
dtype=np.float)
multidim_coord_points = numpy.array(
[1000000., 3100000., 5200000., 7300000., 9400000.], dtype=np.float)
assert_arrays_almost_equal(cube_out[0].data, result_data)
assert_arrays_almost_equal(cube_out[1].data, result_data + 1)
assert_arrays_almost_equal(
cube_out[0].coord('surface_air_pressure').points,
multidim_coord_points)
assert_arrays_almost_equal(
cube_out[1].coord('surface_air_pressure').points,
multidim_coord_points)
def test_gridded_list_write_time_as_unlimited_dimension(self):
data = GriddedDataList(
[make_from_cube(make_mock_cube(time_dim_length=7))])
data[0].var_name = 'rain'
data.save_data(tmp_file)
self.d = Dataset(tmp_file)
assert self.d.dimensions['time'].isunlimited()
def test_gridded_list_write_no_time_has_no_unlimited_dimension(self):
data = GriddedDataList([make_from_cube(make_mock_cube())])
data[0].var_name = 'rain'
data.save_data(tmp_file)
self.d = Dataset(tmp_file)
for d in self.d.dimensions.values():
assert not d.isunlimited()
def aggregate_gridded(self, kernel):
# Make sure all coordinate have bounds - important for weighting and aggregating
# Only try and guess bounds on Dim Coords
for coord in self.data.coords(dim_coords=True):
if not coord.has_bounds() and len(coord.points) > 1:
coord.guess_bounds()
logging.warning("Creating guessed bounds as none exist in file")
new_coord_number = self.data.coord_dims(coord)
self.data.remove_coord(coord.name())
self.data.add_dim_coord(coord, new_coord_number)
coords = []
for coord in self.data.coords():
grid, guessed_axis = self.get_grid(coord)
if grid is not None:
if isnan(grid.delta):
logging.info('Aggregating on ' + coord.name() + ', collapsing completely and using ' +
kernel.cell_method + ' kernel.')
coords.append(coord)
else:
raise NotImplementedError("Aggregation using partial collapse of "
"coordinates is not supported for GriddedData")
output = GriddedDataList([])
if isinstance(kernel, MultiKernel):
for sub_kernel in kernel.sub_kernels:
sub_kernel_out = self._gridded_full_collapse(coords, sub_kernel)
output.append_or_extend(sub_kernel_out)
else:
output.append_or_extend(self._gridded_full_collapse(coords, kernel))
return output
def test_complete_collapse_two_dims_using_moments_kernel(self):
self.kernel = aggregation_kernels['moments']
data1 = make_from_cube(make_5x3_lon_lat_2d_cube_with_missing_data())
data1.var_name = 'var1'
data2 = make_from_cube(make_5x3_lon_lat_2d_cube_with_missing_data())
data2.var_name = 'var2'
data2.data += 10
data = GriddedDataList([data1, data2])
output = data.collapsed(['x', 'y'], how=self.kernel)
expect_mean = numpy.array(7.75)
expect_stddev = numpy.array(numpy.sqrt(244.25 / 11))
expect_count = numpy.array(12)
assert isinstance(output, GriddedDataList)
assert len(output) == 6
mean_1, stddev_1, count_1, mean_2, stddev_2, count_2 = output
assert mean_1.var_name == 'var1'
assert stddev_1.var_name == 'var1_std_dev'
assert count_1.var_name == 'var1_num_points'
assert mean_2.var_name == 'var2'
assert stddev_2.var_name == 'var2_std_dev'
assert count_2.var_name == 'var2_num_points'
# Latitude area weighting means these aren't quite right so increase the rtol.
assert numpy.allclose(mean_1.data, expect_mean, 1e-3)
assert numpy.allclose(mean_2.data, expect_mean + 10, 1e-3)
assert numpy.allclose(stddev_1.data, expect_stddev)
assert numpy.allclose(stddev_2.data, expect_stddev)
assert numpy.allclose(count_1.data, expect_count)
assert numpy.allclose(count_2.data, expect_count)
def test_gridded_list_write_no_time_has_no_unlimited_dimension(self):
data = GriddedDataList([make_from_cube(make_mock_cube())])
data[0].var_name = 'rain'
data.save_data(tmp_file)
self.d = Dataset(tmp_file)
for d in self.d.dimensions.values():
assert not d.isunlimited()
def test_complete_collapse_two_dims_using_moments_kernel(self):
self.kernel = aggregation_kernels['moments']
data1 = make_from_cube(make_5x3_lon_lat_2d_cube_with_missing_data())
data1.var_name = 'var1'
data2 = make_from_cube(make_5x3_lon_lat_2d_cube_with_missing_data())
data2.var_name = 'var2'
data2.data += 10
data = GriddedDataList([data1, data2])
output = data.collapsed(['x', 'y'], how=self.kernel)
expect_mean = numpy.array(7.75)
expect_stddev = numpy.array(numpy.sqrt(244.25 / 11))
expect_count = numpy.array(12)
assert isinstance(output, GriddedDataList)
assert len(output) == 6
mean_1, stddev_1, count_1, mean_2, stddev_2, count_2 = output
assert mean_1.var_name == 'var1'
assert stddev_1.var_name == 'var1_std_dev'
assert count_1.var_name == 'var1_num_points'
assert mean_2.var_name == 'var2'
assert stddev_2.var_name == 'var2_std_dev'
assert count_2.var_name == 'var2_num_points'
# Latitude area weighting means these aren't quite right so increase the rtol.
assert numpy.allclose(mean_1.data, expect_mean, 1e-3)
assert numpy.allclose(mean_2.data, expect_mean + 10, 1e-3)
assert numpy.allclose(stddev_1.data, expect_stddev)
assert numpy.allclose(stddev_2.data, expect_stddev)
assert numpy.allclose(count_1.data, expect_count)
assert numpy.allclose(count_2.data, expect_count)
def read_data_list(self, filenames, variables, product=None, aliases=None):
"""
Read multiple data objects. Files can be either gridded or ungridded but not a mix of both.
:param filenames: One or more filenames of the files to read
:type filenames: string or list
:param variables: One or more variables to read from the files
:type variables: string or list
:param str product: Name of data product to use (optional)
:param aliases: List of variable aliases to put on each variables
data object as an alternative means of identifying them. (Optional)
:return: A list of the data read out (either a GriddedDataList or UngriddedDataList depending on the
type of data contained in the files)
"""
# if filenames or variables are not lists, make them lists of 1 element
filenames = listify(filenames)
variables = listify(variables)
aliases = listify(aliases) if aliases else None
variables = self._expand_wildcards(variables, filenames, product)
data_list = None
for idx, variable in enumerate(variables):
var_data = self._get_data_func(filenames, variable, product)
var_data.filenames = filenames
if aliases:
try:
var_data.alias = aliases[idx]
except IndexError:
raise ValueError("Number of aliases does not match number of variables")
if data_list is None:
data_list = GriddedDataList() if var_data.is_gridded else UngriddedDataList()
data_list.append(var_data)
assert data_list is not None
return data_list
def stats_cmd(main_arguments):
"""
Main routine for handling calls to the statistics command.
:param main_arguments: The command line arguments (minus the stats command)
"""
from cis.stats import StatsAnalyzer
from cis.data_io.gridded_data import GriddedDataList
data_reader = DataReader()
data_list = data_reader.read_datagroups(main_arguments.datagroups)
analyzer = StatsAnalyzer(*data_list)
results = analyzer.analyze()
header = "RESULTS OF STATISTICAL COMPARISON:"
note = "Compared all points which have non-missing values in both variables"
header_length = max(len(header), len(note))
print(header_length * '=')
print(header)
print(header_length * '-')
print(note)
print(header_length * '=')
for result in results:
print(result.pprint())
if main_arguments.output:
cubes = GriddedDataList([result.as_cube() for result in results])
variables = []
filenames = []
for datagroup in main_arguments.datagroups:
variables.extend(datagroup['variables'])
filenames.extend(datagroup['filenames'])
history = "Statistical comparison performed using CIS version " + __version__ + \
"\n variables: " + str(variables) + \
"\n from files: " + str(set(filenames))
cubes.add_history(history)
cubes.save_data(main_arguments.output)
def test_complete_collapse_one_dim_using_moments_kernel(self):
self.kernel = aggregation_kernels['moments']
data1 = make_from_cube(make_5x3_lon_lat_2d_cube_with_missing_data())
data1.var_name = 'var1'
data2 = make_from_cube(make_5x3_lon_lat_2d_cube_with_missing_data())
data2.var_name = 'var2'
data2.data += 10
data = GriddedDataList([data1, data2])
output = data.collapsed(['x'], how=self.kernel)
expect_mean = numpy.array([[5.5, 8.75, 9]])
expect_stddev = numpy.array(
[numpy.sqrt(15), numpy.sqrt(26.25),
numpy.sqrt(30)])
expect_count = numpy.array([[4, 4, 4]])
assert isinstance(output, GriddedDataList)
assert len(output) == 6
mean_1, stddev_1, count_1, mean_2, stddev_2, count_2 = output
assert mean_1.var_name == 'var1'
assert stddev_1.var_name == 'var1_std_dev'
assert count_1.var_name == 'var1_num_points'
assert mean_2.var_name == 'var2'
assert stddev_2.var_name == 'var2_std_dev'
assert count_2.var_name == 'var2_num_points'
assert_arrays_almost_equal(mean_1.data, expect_mean)
assert_arrays_almost_equal(mean_2.data, expect_mean + 10)
assert_arrays_almost_equal(stddev_1.data, expect_stddev)
assert_arrays_almost_equal(stddev_2.data, expect_stddev)
assert_arrays_almost_equal(count_1.data, expect_count)
assert_arrays_almost_equal(count_2.data, expect_count)
def test_empty_longitude_subset_of_gridded_data_list_returns_no_data(self):
"""
Checks that the convention of returning None if subsetting results in an empty subset.
Longitude has a modulus and so uses the IRIS intersection method
"""
data = GriddedDataList([cis.test.util.mock.make_square_5x3_2d_cube()])
subset = data.subset(longitude=[1.0, 3.0])
assert (subset is None)
def test_empty_time_subset_of_gridded_data_list_returns_no_data(self):
"""
Checks that the convention of returning None if subsetting results in an empty subset.
Longitude has no modulus and so uses the IRIS extract method
"""
data = GriddedDataList([cis.test.util.mock.make_square_5x3_2d_cube_with_time()])
subset = data.subset(time=[140500, 140550])
assert (subset is None)
def test_GIVEN_GriddedDataList_WHEN_constrain_THEN_correctly_subsetted_GriddedDataList_returned(self):
gridded1 = cis.test.util.mock.make_square_5x3_2d_cube()
gridded2 = cis.test.util.mock.make_square_5x3_2d_cube()
datalist = GriddedDataList([gridded1, gridded2])
subset = datalist.subset(longitude=[0.0, 5.0], latitude=[-5.0, 5.0])
assert isinstance(subset, GriddedDataList)
assert (subset[0].data.tolist() == [[5, 6], [8, 9], [11, 12]])
assert (subset[1].data.tolist() == [[5, 6], [8, 9], [11, 12]])
def test_empty_longitude_subset_of_gridded_data_list_returns_no_data(self):
"""
Checks that the convention of returning None if subsetting results in an empty subset.
Longitude has a modulus and so uses the IRIS intersection method
"""
data = GriddedDataList([cis.test.util.mock.make_square_5x3_2d_cube()])
subset = data.subset(longitude=[1.0, 3.0])
assert (subset is None)
def test_GIVEN_GriddedDataList_WHEN_constrain_THEN_correctly_subsetted_GriddedDataList_returned(
self):
gridded1 = cis.test.util.mock.make_square_5x3_2d_cube()
gridded2 = cis.test.util.mock.make_square_5x3_2d_cube()
datalist = GriddedDataList([gridded1, gridded2])
subset = datalist.subset(longitude=[0.0, 5.0], latitude=[-5.0, 5.0])
assert isinstance(subset, GriddedDataList)
assert (subset[0].data.tolist() == [[5, 6], [8, 9], [11, 12]])
assert (subset[1].data.tolist() == [[5, 6], [8, 9], [11, 12]])
def test_empty_time_subset_of_gridded_data_list_returns_no_data(self):
"""
Checks that the convention of returning None if subsetting results in an empty subset.
Longitude has no modulus and so uses the IRIS extract method
"""
data = GriddedDataList(
[cis.test.util.mock.make_square_5x3_2d_cube_with_time()])
subset = data.subset(time=[140500, 140550])
assert (subset is None)
def test_taylor_diagram_gridded(self):
from cis.test.util.mock import make_mock_cube
from cis.data_io.gridded_data import GriddedDataList
d = GriddedDataList([make_mock_cube(), make_mock_cube(data_offset=2)])
d[0].var_name = 'snow'
d[1].var_name = 'rain'
d.plot(how='taylor')
self.check_graphic()
def test_iris_comparative_scatter(self):
from cis.test.util.mock import make_mock_cube
from cis.data_io.gridded_data import GriddedDataList
d = GriddedDataList([make_mock_cube(), make_mock_cube(data_offset=2)])
d[0].var_name = 'snow'
d[1].var_name = 'rain'
d.plot(how='comparativescatter')
self.check_graphic()
def test_empty_time_subset_of_gridded_data_list_returns_no_data(self):
"""
Checks that the convention of returning None if subsetting results in an empty subset.
Longitude has no modulus and so uses the IRIS extract method
"""
data = GriddedDataList([cis.test.util.mock.make_square_5x3_2d_cube_with_time()])
long_coord = data.coord('time')
constraint = subset_constraint.GriddedSubsetConstraint()
constraint.set_limit(long_coord, 140500, 140550)
subset = constraint.constrain(data)
assert (subset is None)
def test_iris_multiple_scatter(self):
from cis.test.util.mock import make_mock_cube
from cis.data_io.gridded_data import GriddedDataList
# This only works with one dimensional gridded data
d = GriddedDataList([make_mock_cube(lat_dim_length=0), make_mock_cube(lat_dim_length=0, data_offset=2)])
d[0].var_name = 'snow'
d[1].var_name = 'rain'
# Will default to line plots
d.plot()
self.check_graphic()
def test_aggregate_mean(self):
from cis.data_io.gridded_data import GriddedDataList, make_from_cube
data1 = make_from_cube(make_mock_cube())
data2 = make_from_cube(make_mock_cube(data_offset=1))
datalist = GriddedDataList([data1, data2])
cube_out = datalist.collapsed(['y'], how=self.kernel)
result1 = numpy.array([7, 8, 9])
result2 = result1 + 1
assert isinstance(cube_out, GriddedDataList)
# There is a small deviation to the weighting correction applied by Iris when completely collapsing
assert_arrays_almost_equal(result1, cube_out[0].data)
assert_arrays_almost_equal(result2, cube_out[1].data)
def test_aggregate_mean(self):
from cis.data_io.gridded_data import GriddedDataList, make_from_cube
data1 = make_from_cube(make_mock_cube())
data2 = make_from_cube(make_mock_cube(data_offset=1))
datalist = GriddedDataList([data1, data2])
cube_out = datalist.collapsed(['y'], how=self.kernel)
result1 = numpy.array([7, 8, 9])
result2 = result1 + 1
assert isinstance(cube_out, GriddedDataList)
# There is a small deviation to the weighting correction applied by Iris when completely collapsing
assert_arrays_almost_equal(result1, cube_out[0].data)
assert_arrays_almost_equal(result2, cube_out[1].data)
def test_collapse_vertical_coordinate(self):
from cis.data_io.gridded_data import GriddedDataList, make_from_cube
data1 = make_from_cube(make_mock_cube(alt_dim_length=6))
data2 = make_from_cube(make_mock_cube(alt_dim_length=6, data_offset=1))
datalist = GriddedDataList([data1, data2])
cube_out = datalist.collapsed(['z'], how=self.kernel)
result1 = data1.data.mean(axis=2)
result2 = result1 + 1
assert isinstance(cube_out, GriddedDataList)
# There is a small deviation to the weighting correction applied by Iris when completely collapsing
assert_arrays_almost_equal(result1, cube_out[0].data)
assert_arrays_almost_equal(result2, cube_out[1].data)
assert numpy.array_equal(data1.coords('latitude')[0].points, cube_out.coords('latitude')[0].points)
def _make_two_gridded(self):
data1 = make_from_cube(mock.make_mock_cube())
data2 = make_from_cube(mock.make_mock_cube(data_offset=10))
data1.var_name = 'var1'
data2._var_name = 'var2'
data1.filenames = ['filename1']
data2.filenames = ['filename2']
self.data = [data1, data2]
self.data = GriddedDataList([data1, data2])
def test_collapse_vertical_coordinate(self):
from cis.data_io.gridded_data import GriddedDataList, make_from_cube
data1 = make_from_cube(make_mock_cube(alt_dim_length=6))
data2 = make_from_cube(make_mock_cube(alt_dim_length=6, data_offset=1))
datalist = GriddedDataList([data1, data2])
cube_out = datalist.collapsed(['z'], how=self.kernel)
result1 = data1.data.mean(axis=2)
result2 = result1 + 1
assert isinstance(cube_out, GriddedDataList)
# There is a small deviation to the weighting correction applied by Iris when completely collapsing
assert_arrays_almost_equal(result1, cube_out[0].data)
assert_arrays_almost_equal(result2, cube_out[1].data)
assert numpy.array_equal(
data1.coords('latitude')[0].points,
cube_out.coords('latitude')[0].points)
def test_collapse_vertical_coordinate_weighted_aggregator(self):
"""
We use a weighted aggregator, though no weights should be applied since we're only summing over the vertical
"""
from cis.data_io.gridded_data import GriddedDataList, make_from_cube
data1 = make_from_cube(make_mock_cube(alt_dim_length=6))
data2 = make_from_cube(make_mock_cube(alt_dim_length=6, data_offset=1))
datalist = GriddedDataList([data1, data2])
cube_out = datalist.collapsed(['z'], how=iris.analysis.SUM)
result1 = np.sum(data1.data, axis=2)
result2 = np.sum(data2.data, axis=2)
assert isinstance(cube_out, GriddedDataList)
# There is a small deviation to the weighting correction applied by Iris when completely collapsing
assert_arrays_almost_equal(result1, cube_out[0].data)
assert_arrays_almost_equal(result2, cube_out[1].data)
assert numpy.array_equal(data1.coords('latitude')[0].points, cube_out.coords('latitude')[0].points)
def test_partial_aggregation_over_more_than_one_dim_on_multidimensional_coord(self):
from cis.data_io.gridded_data import GriddedDataList, make_from_cube
data1 = make_from_cube(make_mock_cube(time_dim_length=7, hybrid_pr_len=5))
data2 = make_from_cube(make_mock_cube(time_dim_length=7, hybrid_pr_len=5, data_offset=1))
datalist = GriddedDataList([data1, data2])
cube_out = datalist.collapsed(['t', 'x'], how=self.kernel)
result_data = numpy.array([[51.0, 52.0, 53.0, 54.0, 55.0],
[156.0, 157.0, 158.0, 159.0, 160.0],
[261.0, 262.0, 263.0, 264.0, 265.0],
[366.0, 367.0, 368.0, 369.0, 370.0],
[471.0, 472.0, 473.0, 474.0, 475.0]], dtype=np.float)
multidim_coord_points = numpy.array([1000000., 3100000., 5200000., 7300000., 9400000.], dtype=np.float)
assert_arrays_almost_equal(cube_out[0].data, result_data)
assert_arrays_almost_equal(cube_out[1].data, result_data+1)
assert_arrays_almost_equal(cube_out[0].coord('surface_air_pressure').points, multidim_coord_points)
assert_arrays_almost_equal(cube_out[1].coord('surface_air_pressure').points, multidim_coord_points)
def stats_cmd(main_arguments):
"""
Main routine for handling calls to the statistics command.
:param main_arguments: The command line arguments (minus the stats command)
"""
from cis.stats import StatsAnalyzer
from cis.data_io.gridded_data import GriddedDataList
data_reader = DataReader()
data_list = data_reader.read_datagroups(main_arguments.datagroups)
analyzer = StatsAnalyzer(*data_list)
results = analyzer.analyze()
header = "RESULTS OF STATISTICAL COMPARISON:"
note = "Compared all points which have non-missing values in both variables"
header_length = max(len(header), len(note))
print(header_length * '=')
print(header)
print(header_length * '-')
print(note)
print(header_length * '=')
for result in results:
print(result.pprint())
if main_arguments.output:
cubes = GriddedDataList([result.as_cube() for result in results])
variables = []
filenames = []
for datagroup in main_arguments.datagroups:
variables.extend(datagroup['variables'])
filenames.extend(datagroup['filenames'])
history = "Statistical comparison performed using CIS version " + __version__ + \
"\n variables: " + str(variables) + \
"\n from files: " + str(set(filenames))
cubes.add_history(history)
cubes.save_data(main_arguments.output)
def test_collapse_vertical_coordinate_weighted_aggregator(self):
"""
We use a weighted aggregator, though no weights should be applied since we're only summing over the vertical
"""
from cis.data_io.gridded_data import GriddedDataList, make_from_cube
data1 = make_from_cube(make_mock_cube(alt_dim_length=6))
data2 = make_from_cube(make_mock_cube(alt_dim_length=6, data_offset=1))
datalist = GriddedDataList([data1, data2])
cube_out = datalist.collapsed(['z'], how=iris.analysis.SUM)
result1 = np.sum(data1.data, axis=2)
result2 = np.sum(data2.data, axis=2)
assert isinstance(cube_out, GriddedDataList)
# There is a small deviation to the weighting correction applied by Iris when completely collapsing
assert_arrays_almost_equal(result1, cube_out[0].data)
assert_arrays_almost_equal(result2, cube_out[1].data)
assert numpy.array_equal(
data1.coords('latitude')[0].points,
cube_out.coords('latitude')[0].points)
def __call__(self, kernel):
from cis.data_io.gridded_data import GriddedDataList
from cis.aggregation.collapse_kernels import MultiKernel
# Make sure all coordinate have bounds - important for weighting and aggregating
# Only try and guess bounds on Dim Coords
for coord in self.data.coords(dim_coords=True):
if not coord.has_bounds() and len(coord.points) > 1:
coord.guess_bounds()
logging.warning("Creating guessed bounds as none exist in file")
new_coord_number = self.data.coord_dims(coord)
self.data.remove_coord(coord.name())
self.data.add_dim_coord(coord, new_coord_number)
output = GriddedDataList([])
if isinstance(kernel, MultiKernel):
for sub_kernel in kernel.sub_kernels:
sub_kernel_out = self._gridded_full_collapse(sub_kernel)
output.append_or_extend(sub_kernel_out)
else:
output.append_or_extend(self._gridded_full_collapse(kernel))
return output
def __call__(self, kernel):
from cis.data_io.gridded_data import GriddedDataList
from cis.aggregation.collapse_kernels import MultiKernel
# Make sure all coordinate have bounds - important for weighting and aggregating
# Only try and guess bounds on Dim Coords
for coord in self.data.coords(dim_coords=True):
if not coord.has_bounds() and len(coord.points) > 1:
coord.guess_bounds()
logging.warning(
"Creating guessed bounds as none exist in file")
new_coord_number = self.data.coord_dims(coord)
self.data.remove_coord(coord.name())
self.data.add_dim_coord(coord, new_coord_number)
output = GriddedDataList([])
if isinstance(kernel, MultiKernel):
for sub_kernel in kernel.sub_kernels:
sub_kernel_out = self._gridded_full_collapse(sub_kernel)
output.append_or_extend(sub_kernel_out)
else:
output.append_or_extend(self._gridded_full_collapse(kernel))
return output
def collocate(self, points, data, constraint, kernel):
"""
:param points: cube defining the sample points
:param data: CommonData object providing data to be collocated (or list of Data)
:param constraint: instance of a Constraint subclass, which takes a data object and returns a subset of that
data based on it's internal parameters
:param kernel: instance of a Kernel subclass which takes a number of points and returns a single value
:return: GriddedDataList of collocated data
"""
if isinstance(data, list):
# If data is a list then call this method recursively over each element
output_list = []
for variable in data:
collocated = self.collocate(points, variable, constraint, kernel)
output_list.extend(collocated)
return GriddedDataList(output_list)
data_points = data.get_non_masked_points()
# Work out how to iterate over the cube and map HyperPoint coordinates to cube coordinates.
coord_map = make_coord_map(points, data)
if self.missing_data_for_missing_sample and len(coord_map) is not len(points.coords()):
raise cis.exceptions.UserPrintableException(
"A sample variable has been specified but not all coordinates in the data appear in the sample so "
"there are multiple points in the sample data so whether the data is missing or not can not be "
"determined")
coords = points.coords()
shape = []
output_coords = []
# Find shape of coordinates to be iterated over.
for (hpi, ci, shi) in coord_map:
coord = coords[ci]
if coord.ndim > 1:
raise NotImplementedError("Co-location of data onto a cube with a coordinate of dimension greater"
" than one is not supported (coordinate %s)", coord.name())
# Ensure that bounds exist.
if not coord.has_bounds():
logging.warning("Creating guessed bounds as none exist in file")
coord.guess_bounds()
shape.append(coord.shape[0])
output_coords.append(coord)
_fix_longitude_range(coords, data_points)
# Create index if constraint supports it.
data_index.create_indexes(constraint, coords, data_points, coord_map)
data_index.create_indexes(kernel, points, data_points, coord_map)
# Initialise output array as initially all masked, and set the appropriate fill value.
values = []
for i in range(kernel.return_size):
val = np.ma.zeros(shape)
val.mask = True
val.fill_value = self.fill_value
values.append(val)
if kernel.return_size == 1:
set_value_kernel = self._set_single_value_kernel
else:
set_value_kernel = self._set_multi_value_kernel
logging.info("--> Co-locating...")
if hasattr(kernel, "get_value_for_data_only") and hasattr(constraint, "get_iterator_for_data_only"):
# Iterate over constrained cells
iterator = constraint.get_iterator_for_data_only(
self.missing_data_for_missing_sample, coord_map, coords, data_points, shape, points, values)
for out_indices, data_values in iterator:
try:
kernel_val = kernel.get_value_for_data_only(data_values)
set_value_kernel(kernel_val, values, out_indices)
except ValueError:
# ValueErrors are raised by Kernel when there are no points to operate on.
# We don't need to do anything.
pass
else:
# Iterate over constrained cells
iterator = constraint.get_iterator(
self.missing_data_for_missing_sample, coord_map, coords, data_points, shape, points, values)
for out_indices, hp, con_points in iterator:
try:
kernel_val = kernel.get_value(hp, con_points)
set_value_kernel(kernel_val, values, out_indices)
except ValueError:
# ValueErrors are raised by Kernel when there are no points to operate on.
# We don't need to do anything.
pass
# Construct an output cube containing the collocated data.
kernel_var_details = kernel.get_variable_details(data.var_name, data.long_name, data.standard_name, data.units)
output = GriddedDataList([])
for idx, val in enumerate(values):
cube = self._create_collocated_cube(data, val, output_coords)
data_with_nan_and_inf_removed = np.ma.masked_invalid(cube.data)
data_with_nan_and_inf_removed.set_fill_value(self.fill_value)
cube.data = data_with_nan_and_inf_removed
cube.var_name = kernel_var_details[idx][0]
cube.long_name = kernel_var_details[idx][1]
cis.utils.set_cube_standard_name_if_valid(cube, kernel_var_details[idx][2])
try:
cube.units = kernel_var_details[idx][3]
except ValueError:
logging.warn(
"Units are not cf compliant, not setting them. Units {}".format(kernel_var_details[idx][3]))
# Sort the cube into the correct shape, so that the order of coordinates
# is the same as in the source data
coord_map = sorted(coord_map, key=lambda x: x[1])
transpose_order = [coord[2] for coord in coord_map]
cube.transpose(transpose_order)
output.append(cube)
return output
def collocate(self, points, data, constraint, kernel):
"""
:param points: cube defining the sample points
:param data: CommonData object providing data to be collocated (or list of Data)
:param constraint: instance of a Constraint subclass, which takes a data object and returns a subset of that
data based on it's internal parameters
:param kernel: instance of a Kernel subclass which takes a number of points and returns a single value
:return: GriddedDataList of collocated data
"""
log_memory_profile("GeneralGriddedCollocator Initial")
if isinstance(data, list):
# If data is a list then call this method recursively over each element
output_list = []
for variable in data:
collocated = self.collocate(points, variable, constraint, kernel)
output_list.extend(collocated)
return GriddedDataList(output_list)
data_points = data.get_non_masked_points()
log_memory_profile("GeneralGriddedCollocator Created data hyperpoint list view")
# Work out how to iterate over the cube and map HyperPoint coordinates to cube coordinates.
coord_map = make_coord_map(points, data)
if self.missing_data_for_missing_sample and len(coord_map) is not len(points.coords()):
raise cis.exceptions.UserPrintableException(
"A sample variable has been specified but not all coordinates in the data appear in the sample so "
"there are multiple points in the sample data so whether the data is missing or not can not be "
"determined")
coords = points.coords()
shape = []
output_coords = []
# Find shape of coordinates to be iterated over.
for (hpi, ci, shi) in coord_map:
coord = coords[ci]
if coord.ndim > 1:
raise NotImplementedError("Co-location of data onto a cube with a coordinate of dimension greater"
" than one is not supported (coordinate %s)", coord.name())
# Ensure that bounds exist.
if not coord.has_bounds():
logging.warning("Creating guessed bounds as none exist in file")
coord.guess_bounds()
shape.append(coord.shape[0])
output_coords.append(coord)
_fix_longitude_range(coords, data_points)
log_memory_profile("GeneralGriddedCollocator Created output coord map")
# Create index if constraint supports it.
data_index.create_indexes(constraint, coords, data_points, coord_map)
data_index.create_indexes(kernel, points, data_points, coord_map)
log_memory_profile("GeneralGriddedCollocator Created indexes")
# Initialise output array as initially all masked, and set the appropriate fill value.
values = []
for i in range(kernel.return_size):
val = np.ma.zeros(shape)
val.mask = True
val.fill_value = self.fill_value
values.append(val)
if kernel.return_size == 1:
set_value_kernel = self._set_single_value_kernel
else:
set_value_kernel = self._set_multi_value_kernel
logging.info("--> Co-locating...")
if hasattr(kernel, "get_value_for_data_only") and hasattr(constraint, "get_iterator_for_data_only"):
# Iterate over constrained cells
iterator = constraint.get_iterator_for_data_only(
self.missing_data_for_missing_sample, coord_map, coords, data_points, shape, points, values)
for out_indices, data_values in iterator:
try:
kernel_val = kernel.get_value_for_data_only(data_values)
set_value_kernel(kernel_val, values, out_indices)
except ValueError:
# ValueErrors are raised by Kernel when there are no points to operate on.
# We don't need to do anything.
pass
else:
# Iterate over constrained cells
iterator = constraint.get_iterator(
self.missing_data_for_missing_sample, coord_map, coords, data_points, shape, points, values)
for out_indices, hp, con_points in iterator:
try:
kernel_val = kernel.get_value(hp, con_points)
set_value_kernel(kernel_val, values, out_indices)
except ValueError:
# ValueErrors are raised by Kernel when there are no points to operate on.
# We don't need to do anything.
pass
log_memory_profile("GeneralGriddedCollocator Completed collocation")
# Construct an output cube containing the collocated data.
kernel_var_details = kernel.get_variable_details(self.var_name or data.var_name,
self.var_long_name or data.long_name,
data.standard_name,
self.var_units or data.units)
output = GriddedDataList([])
for idx, val in enumerate(values):
cube = self._create_collocated_cube(data, val, output_coords)
data_with_nan_and_inf_removed = np.ma.masked_invalid(cube.data)
data_with_nan_and_inf_removed.set_fill_value(self.fill_value)
cube.data = data_with_nan_and_inf_removed
cube.var_name = kernel_var_details[idx][0]
cube.long_name = kernel_var_details[idx][1]
set_standard_name_if_valid(cube, kernel_var_details[idx][2])
try:
cube.units = kernel_var_details[idx][3]
except ValueError:
logging.warn(
"Units are not cf compliant, not setting them. Units {}".format(kernel_var_details[idx][3]))
# Sort the cube into the correct shape, so that the order of coordinates
# is the same as in the source data
coord_map = sorted(coord_map, key=lambda x: x[1])
transpose_order = [coord[2] for coord in coord_map]
cube.transpose(transpose_order)
output.append(cube)
log_memory_profile("GeneralGriddedCollocator Finished")
return output
def setUp(self):
self.calc = Calculator()
self.data = GriddedDataList([make_from_cube(mock.make_mock_cube())])
self.data[0].var_name = 'var_name'
def test_gridded_list_write_time_as_unlimited_dimension(self):
data = GriddedDataList([make_from_cube(make_mock_cube(time_dim_length=7))])
data[0].var_name = 'rain'
data.save_data(tmp_file)
self.d = Dataset(tmp_file)
assert self.d.dimensions['time'].isunlimited()
def collocate(self, points, data, constraint, kernel):
"""
This collocator takes two Iris cubes, and collocates from the data cube onto the grid of the 'points' cube. The
collocator then returns another Iris cube.
:param points: An Iris cube with the sampling grid to collocate onto.
:param data: The Iris cube with the data to be collocated.
:param constraint: None allowed yet, as this is unlikely to be required for gridded-gridded.
:param kernel: The kernel to use, current options are gridded_gridded_nn and gridded_gridded_li.
:return: An Iris cube with the collocated data.
"""
self._check_for_valid_kernel(kernel)
# Force the data longitude range to be the same as that of the sample grid.
_fix_longitude_range(points.coords(), data)
# Initialise variables used to create an output mask based on the sample data mask.
sample_coord_lookup = {} # Maps coordinate in sample data -> location in dimension order
for idx, coord in enumerate(points.coords()):
sample_coord_lookup[coord] = idx
sample_coord_transpose_map = [] # For coords in both sample and data, contains the position in the sample
other_coord_transpose_map = [] # For coords in data but not in sample, records that coord's position in data.
repeat_size = 1
output_mask = np.ma.nomask
# Make a list of the coordinates we have, with each entry containing a list with the name of the coordinate and
# the number of points along its axis. One is for the sample grid, which contains the points where we
# interpolate too, and one is for the output grid, which will additionally contain any dimensions missing in the
# sample grid.
coord_names_and_sizes_for_sample_grid = []
coord_names_and_sizes_for_output_grid = []
for idx, coord in enumerate(data.coords(dim_coords=True)):
# First try and find the coordinate in points, the sample grid. If an exception is thrown, it means that
# name does not appear in the sample grid, and instead take the coordinate name and length from the original
# data, as this is what we will be keeping.
try:
sample_coord = points.coords(coord.name())[0]
coord_names_and_sizes_for_sample_grid.append([coord.name(), len(sample_coord.points)])
# Find the index of the sample coordinate corresponding to the data coordinate.
sample_coord_transpose_map.append(sample_coord_lookup[sample_coord])
except IndexError:
coord_names_and_sizes_for_output_grid.append([coord.name(), len(coord.points)])
repeat_size *= len(coord.points)
other_coord_transpose_map.append(idx)
# Now we sort the sample coordinates so that they are in the same order as in the sample file,
# rather than the order of the data file (that's the order we want the output dimensions).
coord_names_and_sizes_for_sample_grid = [x[0] for x in sorted(zip(coord_names_and_sizes_for_sample_grid,
sample_coord_transpose_map),
key=lambda t: t[1])]
# Adding the lists together in this way ensures that the coordinates not in the sample grid appear in the final
# position, which is important for adding the points from the Iris interpolater to the new array. The data
# returned from the Iris interpolater method will have dimensions of these missing coordinates, which needs
# to be the final dimensions in the numpy array, as the iterator will give the position of the other dimensions.
coord_names_and_sizes_for_output_grid = coord_names_and_sizes_for_sample_grid + \
coord_names_and_sizes_for_output_grid
# An array for the collocated data, with the correct shape
output_shape = tuple(i[1] for i in coord_names_and_sizes_for_output_grid)
new_data = np.zeros(output_shape)
if self.missing_data_for_missing_sample:
output_mask = self._make_output_mask(coord_names_and_sizes_for_sample_grid, output_shape,
points, repeat_size)
# Now recreate the points cube, while ignoring any DimCoords in points that are not in the data cube
new_dim_coord_list = []
new_points_array_shape = []
for i in range(0, len(coord_names_and_sizes_for_output_grid)):
# Try and find the coordinate in the sample grid
coord_found = points.coords(coord_names_and_sizes_for_output_grid[i][0])
# If the coordinate exists in the sample grid then append the new coordinate to the list. Iris requires
# this be given as a DimCoord object, along with a axis number, in a tuple pair.
if len(coord_found) != 0:
new_dim_coord_list.append((coord_found[0], len(new_dim_coord_list)))
new_points_array_shape.append(coord_found[0].points.size)
new_points_array = np.zeros(tuple(new_points_array_shape))
# Use the new_data array to recreate points, without the DimCoords not in the data cube
points = iris.cube.Cube(new_points_array, dim_coords_and_dims=new_dim_coord_list)
output_cube = self._iris_interpolate(coord_names_and_sizes_for_output_grid,
coord_names_and_sizes_for_sample_grid, data,
kernel, output_mask, points, self.extrapolate)
if not isinstance(output_cube, list):
return GriddedDataList([output_cube])
else:
return output_cube
