# (C) Crown Copyright, Met Office. All rights reserved.
#
# This file is part of 'IMPROVER' and is released under the BSD 3-Clause license.
# See LICENSE in the root of the repository for full licensing details.
"""Functions to create spotdata cubes."""
from typing import List, Optional, Union
import iris
import numpy as np
from iris.coords import AuxCoord, Coord, DimCoord
from iris.cube import Cube
from numpy import ndarray
from . import UNIQUE_ID_ATTRIBUTE
[docs]
def build_spotdata_cube(
data: ndarray,
name: str,
units: str,
altitude: ndarray,
latitude: ndarray,
longitude: ndarray,
wmo_id: Union[str, List[str]],
unique_site_id: Optional[Union[List[str], ndarray]] = None,
unique_site_id_key: Optional[str] = None,
scalar_coords: Optional[List[Coord]] = None,
auxiliary_coords: Optional[List[Coord]] = None,
neighbour_methods: Optional[List[str]] = None,
grid_attributes: Optional[List[str]] = None,
additional_dims: Optional[List[Coord]] = None,
additional_dims_aux: Optional[List[List[AuxCoord]]] = None,
) -> Cube:
"""
Function to build a spotdata cube with expected dimension and auxiliary
coordinate structure.
It can be used to create spot data cubes. In this case the data is the
spot data values at each site, and the coordinates that describe each site.
It can also be used to create cubes which describe the grid points that are
used to extract each site from a gridded field, for different selection
method. The selection methods are specified by the neighbour_methods
coordinate. The grid_attribute coordinate encapsulates information required
to extract data, for example the x/y indices that identify the grid point
neighbour.
.. See the documentation for examples of these cubes.
.. include:: extended_documentation/spotdata/build_spotdata_cube/
build_spotdata_cube_examples.rst
Args:
data:
Float spot data or array of data points from several sites.
The spot index should be the last dimension if the array is
multi-dimensional (see optional additional dimensions below).
name:
Cube name (eg 'air_temperature')
units:
Cube units (eg 'K')
altitude:
Float or 1d array of site altitudes in metres
latitude:
Float or 1d array of site latitudes in degrees
longitude:
Float or 1d array of site longitudes in degrees
wmo_id:
String or list of 5-digit WMO site identifiers.
unique_site_id:
Optional list of 8-digit unique site identifiers. If provided, this
is expected to be a complete list with a unique identifier for every site.
unique_site_id_key:
String to name the unique_site_id coordinate. Required if
unique_site_id is in use.
scalar_coords:
Optional list of iris.coords.Coord instances
auxiliary_coords:
Optional list of iris.coords.Coord instances which are non-scalar.
neighbour_methods:
Optional list of neighbour method names, e.g. 'nearest'
grid_attributes:
Optional list of grid attribute names, e.g. x-index, y-index
additional_dims:
Optional list of additional dimensions to preceed the spot data dimension.
additional_dims_aux:
Optional list of auxiliary coordinates associated with each dimension in
additional_dims
Returns:
A cube containing the extracted spot data with spot data being the final dimension.
"""
# construct auxiliary coordinates
alt_coord = AuxCoord(altitude, "altitude", units="m")
lat_coord = AuxCoord(latitude, "latitude", units="degrees")
lon_coord = AuxCoord(longitude, "longitude", units="degrees")
wmo_id_coord = AuxCoord(wmo_id, long_name="wmo_id", units="no_unit")
if unique_site_id is not None:
if not unique_site_id_key:
raise ValueError(
"A unique_site_id_key must be provided if a unique_site_id is"
" provided."
)
unique_id_coord = AuxCoord(
unique_site_id,
long_name=unique_site_id_key,
units="no_unit",
attributes={UNIQUE_ID_ATTRIBUTE: "true"},
)
aux_coords_and_dims = []
# append scalar coordinates
if scalar_coords is not None:
for coord in scalar_coords:
aux_coords_and_dims.append((coord, None))
# construct dimension coordinates
if np.isscalar(data):
data = np.array([data])
spot_index = DimCoord(
np.arange(data.shape[-1], dtype=np.int32), long_name="spot_index", units="1"
)
dim_coords_and_dims = []
current_dim = 0
if neighbour_methods is not None:
neighbour_methods_coord = DimCoord(
np.arange(len(neighbour_methods), dtype=np.int32),
long_name="neighbour_selection_method",
units="1",
)
neighbour_methods_key = AuxCoord(
neighbour_methods,
long_name="neighbour_selection_method_name",
units="no_unit",
)
dim_coords_and_dims.append((neighbour_methods_coord, current_dim))
aux_coords_and_dims.append((neighbour_methods_key, current_dim))
current_dim += 1
if grid_attributes is not None:
grid_attributes_coord = DimCoord(
np.arange(len(grid_attributes), dtype=np.int32),
long_name="grid_attributes",
units="1",
)
grid_attributes_key = AuxCoord(
grid_attributes, long_name="grid_attributes_key", units="no_unit"
)
dim_coords_and_dims.append((grid_attributes_coord, current_dim))
aux_coords_and_dims.append((grid_attributes_key, current_dim))
current_dim += 1
if additional_dims is not None:
for coord, aux_coords in zip(
additional_dims, additional_dims_aux or [[] for _ in additional_dims]
):
dim_coords_and_dims.append((coord, current_dim))
for aux_coord in aux_coords:
aux_coords_and_dims.append((aux_coord, current_dim))
current_dim += 1
dim_coords_and_dims.append((spot_index, current_dim))
coords = [alt_coord, lat_coord, lon_coord, wmo_id_coord]
if unique_site_id is not None:
coords.append(unique_id_coord)
for coord in coords:
aux_coords_and_dims.append((coord, current_dim))
# append non-scalar auxiliary coordinates
if auxiliary_coords:
for coord in auxiliary_coords:
aux_coords_and_dims.append((coord, current_dim))
# create output cube
spot_cube = iris.cube.Cube(
data,
long_name=name,
units=units,
dim_coords_and_dims=dim_coords_and_dims,
aux_coords_and_dims=aux_coords_and_dims,
)
# rename to force a standard name to be set if name is valid
spot_cube.rename(name)
return spot_cube
