xrspatial.classify.reclassify#

xrspatial.classify.reclassify(agg: DataArray, bins: List[int], new_values: List[int], name: str | None = 'reclassify') → DataArray[source]#

Reclassifies data for array agg into new values based on user defined bins.

Parameters:

agg (xr.DataArray or xr.Dataset) – 2D NumPy, CuPy, NumPy-backed Dask, or Cupy-backed Dask array of values to be reclassified.
bins (array-like object) – Values or ranges of values to be changed.
new_values (array-like object) – New values for each bin.
name (str, default='reclassify') – Name of output aggregate array.

Returns:

reclass_agg – 2D aggregate array of reclassified allocations. All other input attributes are preserved. If agg is a Dataset, returns a Dataset with each variable classified independently.

Return type:

xr.DataArray or xr.Dataset

References

PySAL: https://pysal.org/mapclassify/_modules/mapclassify/classifiers.html # noqa

Examples

Reclassify works with NumPy backed xarray DataArray .. sourcecode:: python

>>> import numpy as np
>>> import xarray as xr
>>> from xrspatial.classify import reclassify

>>> data = np.array([
    [np.nan,  1.,  2.,  3.,  4.],
    [ 5.,  6.,  7.,  8.,  9.],
    [10., 11., 12., 13., 14.],
    [15., 16., 17., 18., np.inf],
])
>>> agg = xr.DataArray(data)
>>> print(agg)
<xarray.DataArray (dim_0: 4, dim_1: 5)>
array([[nan,  1.,  2.,  3.,  4.],
       [ 5.,  6.,  7.,  8.,  9.],
       [10., 11., 12., 13., 14.],
       [15., 16., 17., 18., inf]])
Dimensions without coordinates: dim_0, dim_1
>>> bins=[10, 15, np.inf]
>>> new_values=[1, 2, 3]
>>> agg_reclassify = reclassify(agg, bins=bins, new_values=new_values)
>>> print(agg_reclassify)
<xarray.DataArray 'reclassify' (dim_0: 4, dim_1: 5)>
array([[nan,  1.,  1.,  1.,  1.],
       [ 1.,  2.,  2.,  2.,  2.],
       [ 2.,  2.,  2.,  2.,  2.],
       [ 2.,  3.,  3.,  3.,  3.]], dtype=float32)
Dimensions without coordinates: dim_0, dim_1

Reclassify works with Dask with NumPy backed xarray DataArray .. sourcecode:: python

>>> import dask.array as da
>>> data_da = da.from_array(data, chunks=(3, 3))
>>> agg_da = xr.DataArray(data_da, name='agg_da')
>>> print(agg_da)
<xarray.DataArray 'agg_da' (dim_0: 4, dim_1: 5)>
dask.array<array, shape=(4, 5), dtype=float32, chunksize=(3, 3), chunktype=numpy.ndarray>
Dimensions without coordinates: dim_0, dim_1
>>> agg_reclassify_da = reclassify(agg_da, bins=bins, new_values=new_values)  # noqa
>>> print(agg_reclassify_da)
<xarray.DataArray 'reclassify' (dim_0: 4, dim_1: 5)>
dask.array<_run_numpy_bin, shape=(4, 5), dtype=float32, chunksize=(3, 3), chunktype=numpy.ndarray>
Dimensions without coordinates: dim_0, dim_1
>>> print(agg_reclassify_da.compute())  # print the computed the results
<xarray.DataArray 'reclassify' (dim_0: 4, dim_1: 5)>
array([[nan,  1.,  1.,  1.,  1.],
       [ 1.,  2.,  2.,  2.,  2.],
       [ 2.,  2.,  2.,  2.,  2.],
       [ 2.,  3.,  3.,  3.,  3.]], dtype=float32)
Dimensions without coordinates: dim_0, dim_1

Reclassify works with CuPy backed xarray DataArray. Make sure you have a GPU and CuPy installed to run this example. .. sourcecode:: python

>>> import cupy
>>> data_cupy = cupy.asarray(data)
>>> agg_cupy = xr.DataArray(data_cupy)
>>> agg_reclassify_cupy = reclassify(agg_cupy, bins, new_values)
>>> print(type(agg_reclassify_cupy.data))
<class 'cupy.core.core.ndarray'>
>>> print(agg_reclassify_cupy)
<xarray.DataArray 'reclassify' (dim_0: 4, dim_1: 5)>
array([[nan,  1.,  1.,  1.,  1.],
       [ 1.,  2.,  2.,  2.,  2.],
       [ 2.,  2.,  2.,  2.,  2.],
       [ 2.,  3.,  3.,  3.,  3.]], dtype=float32)
Dimensions without coordinates: dim_0, dim_1

Reclassify works with Dask with CuPy backed xarray DataArray.