def merge_arrays(
dataarrays: Iterable[DataArray],
bounds: Optional[Tuple] = None,
res: Optional[Tuple] = None,
nodata: Optional[float] = None,
precision: Optional[float] = None,
method: Union[str, Callable, None] = None,
crs: Optional[CRS] = None,
parse_coordinates: bool = True,
) -> DataArray:
"""
Merge data arrays geospatially.
Uses rasterio.merge.merge:
https://rasterio.readthedocs.io/en/stable/api/rasterio.merge.html#rasterio.merge.merge
.. versionadded:: 0.2 crs
Parameters
----------
dataarrays: list
List of xarray.DataArray's with all geo attributes.
The first one is assumed to have the same
CRS, dtype, and dimensions as the others in the array.
bounds: tuple, optional
Bounds of the output image (left, bottom, right, top).
If not set, bounds are determined from bounds of input DataArrays.
res: tuple, optional
Output resolution in units of coordinate reference system.
If not set, the resolution of the first DataArray is used.
If a single value is passed, output pixels will be square.
nodata: float, optional
nodata value to use in output file.
If not set, uses the nodata value in the first input DataArray.
precision: float, optional
Number of decimal points of precision when computing inverse transform.
method: str or callable, optional
See rasterio docs.
crs: rasterio.crs.CRS, optional
Output CRS. If not set, the CRS of the first DataArray is used.
parse_coordinates: bool, optional
If False, it will disable loading spatial coordinates.
Returns
-------
:obj:`xarray.DataArray`:
The geospatially merged data.
"""
input_kwargs = dict(bounds=bounds,
res=res,
nodata=nodata,
precision=precision,
method=method)
if crs is None:
crs = dataarrays[0].rio.crs
if res is None:
res = tuple(abs(res_val) for res_val in dataarrays[0].rio.resolution())
# prepare the duck arrays
rioduckarrays = []
for dataarray in dataarrays:
da_res = tuple(abs(res_val) for res_val in dataarray.rio.resolution())
if da_res != res or dataarray.rio.crs != crs:
rioduckarrays.append(
RasterioDatasetDuck(
dataarray.rio.reproject(dst_crs=crs, resolution=res)))
else:
rioduckarrays.append(RasterioDatasetDuck(dataarray))
# use rasterio to merge
merged_data, merged_transform = _rio_merge(
rioduckarrays,
**{key: val
for key, val in input_kwargs.items() if val is not None},
)
# generate merged data array
representative_array = rioduckarrays[0]._xds
if parse_coordinates:
coords = _make_coords(
representative_array,
merged_transform,
merged_data.shape[-1],
merged_data.shape[-2],
)
else:
coords = _get_nonspatial_coords(representative_array)
xda = DataArray(
name=representative_array.name,
data=merged_data,
coords=coords,
dims=tuple(representative_array.dims),
attrs=representative_array.attrs,
)
xda.rio.write_nodata(
nodata if nodata is not None else representative_array.rio.nodata,
inplace=True)
xda.rio.write_crs(representative_array.rio.crs, inplace=True)
xda.rio.write_transform(merged_transform, inplace=True)
return xda
def merge_arrays(
dataarrays: Iterable[DataArray],
bounds: Union[Tuple, None] = None,
res: Union[Tuple, None] = None,
nodata: Union[float, None] = None,
precision: Union[float, None] = None,
method: Union[str, Callable, None] = None,
parse_coordinates: bool = True,
) -> DataArray:
"""
Merge data arrays geospatially.
Uses rasterio.merge.merge:
https://rasterio.readthedocs.io/en/stable/api/rasterio.merge.html#rasterio.merge.merge
Parameters
----------
dataarrays: list
List of xarray.DataArray's with all geo attributes.
The first one is assumed to have the same
CRS, dtype, and dimensions as the others in the array.
bounds: tuple, optional
Bounds of the output image (left, bottom, right, top).
If not set, bounds are determined from bounds of input DataArrays.
res: tuple, optional
Output resolution in units of coordinate reference system.
If not set, the resolution of the first DataArray is used.
If a single value is passed, output pixels will be square.
nodata: float, optional
nodata value to use in output file.
If not set, uses the nodata value in the first input DataArray.
precision: float, optional
Number of decimal points of precision when computing inverse transform.
method: str or callable, optional
See rasterio docs.
parse_coordinates: bool, optional
If False, it will disable loading spatial coordinates.
Returns
-------
DataArray: The geospatially merged data.
"""
input_kwargs = dict(bounds=bounds,
res=res,
nodata=nodata,
precision=precision,
method=method)
merged_data, merged_transform = _rio_merge(
[RasterioDatasetDuck(dataarray) for dataarray in dataarrays],
**{key: val
for key, val in input_kwargs.items() if val is not None},
)
merged_shape = merged_data.shape
representative_array = dataarrays[0]
merged_crs = representative_array.rio.crs
if parse_coordinates:
coords = _make_coords(
representative_array,
merged_transform,
merged_shape[-1],
merged_shape[-2],
merged_crs,
)
else:
coords = _get_nonspatial_coords(representative_array)
out_attrs = representative_array.attrs
out_attrs["transform"] = tuple(merged_transform)[:6]
xda = DataArray(
name=dataarrays[0].name,
data=merged_data,
coords=coords,
dims=tuple(representative_array.dims),
attrs=out_attrs,
)
out_nodata = nodata if nodata is not None else representative_array.rio.nodata
xda.rio.write_nodata(out_nodata, inplace=True)
xda.rio.write_crs(representative_array.rio.crs, inplace=True)
return xda