Fix to_xarray kwargs for netcdf data object

src/earthkit/data/data/netcdf.py

@@ -59,18 +59,32 @@ def to_xarray(self, *args, xarray_open_mfdataset_kwargs=None, **kwargs):
 
         Parameters
         ----------
+        *args
+            Positional arguments to pass to the reader's to_xarray method.
+            Not used currently. It is there to allow for future extensions or
+            additional parameters that may be needed for specific use cases.
         xarray_open_mfdataset_kwargs: dict, None, optional
             Keyword arguments passed to :py:func:`xarray.open_mfdataset`.
+            Can only be used when converting multiple NetCDF files into a single
+            Xarray dataset. When specified, this argument takes precedence over
+            any other keyword arguments passed to the method.
         **kwargs
-            Keyword arguments passed to  :py:func:`xarray.open_mfdataset`
-            ``xarray_open_mfdataset_kwargs`` is not set.
+            Keyword arguments passed to :py:func:`xarray.open_dataset` for
+            single file conversion or to :py:func:`xarray.open_mfdataset` for
+            multiple file conversion. Ignored if `xarray_open_mfdataset_kwargs`
+            is specified.
 
         Returns
         -------
         :py:class:`xarray.Dataset`
             An Xarray dataset containing the NetCDF data.
         """
-        return self._reader.to_xarray(*args, **kwargs)
+        if xarray_open_mfdataset_kwargs:
+            options = dict(xarray_open_mfdataset_kwargs=xarray_open_mfdataset_kwargs)
+        else:
+            options = dict()
+
+        return self._reader.to_xarray(*args, **options, **kwargs)
 
     def to_pandas(self, *args, **kwargs):
         """Convert into a Pandas DataFrame.

src/earthkit/data/readers/netcdf/reader.py

@@ -40,12 +40,6 @@ def to_pandas(self):
 
     def to_xarray(self, **kwargs):
         return type(self).to_xarray_single_from_path(self.path, **kwargs)
-        # if isinstance(self.path, str):
-        #     return type(self).to_xarray_single_from_path(self.path, **kwargs)
-
-        #     return NetCDFReader.to_xarray_multi_from_paths([self.path], **kwargs)
-
-        # return type(self).to_xarray_multi_from_paths([self.path], **kwargs)
 
     @classmethod
     def to_xarray_multi_from_paths(cls, paths, **kwargs):
@@ -54,8 +48,13 @@ def to_xarray_multi_from_paths(cls, paths, **kwargs):
         if not isinstance(paths, list):
             paths = [paths]
 
+        kwargs = kwargs.copy()  # Make a copy to avoid modifying the original dictionary
+        xarray_open_mfdataset_kwargs = kwargs.pop("xarray_open_mfdataset_kwargs", None)
+
         options = dict()
-        options.update(kwargs.get("xarray_open_mfdataset_kwargs", {}))
+        if xarray_open_mfdataset_kwargs is not None:
+            options = dict(xarray_open_mfdataset_kwargs)
+
         if not options:
             options = dict(**kwargs)
 
@@ -71,11 +70,16 @@ def to_xarray_single_from_path(cls, path, **kwargs):
         if not isinstance(path, str):
             raise ValueError("path must be a string")
 
-        if "xarray_open_mfdataset_kwargs" in kwargs:
+        kwargs = kwargs.copy()  # Make a copy to avoid modifying the original dictionary
+        xarray_open_dataset_kwargs = kwargs.pop("xarray_open_dataset_kwargs", None)
+        xarray_open_mfdataset_kwargs = kwargs.pop("xarray_open_mfdataset_kwargs", None)
+        if xarray_open_mfdataset_kwargs:
             raise ValueError("xarray_open_mfdataset_kwargs is not supported for single file")
 
         options = dict()
-        options.update(kwargs.get("xarray_open_dataset_kwargs", {}))
+        if xarray_open_dataset_kwargs is not None:
+            options = dict(xarray_open_dataset_kwargs)
+
         if not options:
             options = dict(**kwargs)
 

tests/netcdf/test_netcdf_convert.py

@@ -9,16 +9,17 @@
 # nor does it submit to any jurisdiction.
 #
 
+
 import numpy as np
 import pytest
 
 from earthkit.data import from_source
-from earthkit.data.utils.testing import earthkit_remote_test_data_file
+from earthkit.data.utils.testing import earthkit_remote_test_data_file, earthkit_test_data_file
 
 
 @pytest.mark.long_test
 @pytest.mark.download
-def test_netcdf_to_xarray_args():
+def test_netcdf_fieldlist_to_xarray_args():
     # The JD variable in the NetCDF is defined as follows:
     #
     # short JD(time, lat, lon) ;
@@ -55,6 +56,86 @@ def test_netcdf_to_xarray_args():
     r["JD"].shape == (1, 20880, 28440)
 
 
+@pytest.mark.long_test
+@pytest.mark.download
+def test_netcdf_single_to_xarray_args():
+    # The JD variable in the NetCDF is defined as follows:
+    #
+    # short JD(time, lat, lon) ;
+    #   string JD:long_name = "Date of the first detection" ;
+    #   string JD:units = "days since 2022-01-01" ;
+    #   string JD:comment = "Possible values: 0  when the pixel is not burned; 1 to 366 day of
+    #       the first detection when the pixel is burned; -1 when the pixel is not observed
+    #       in the month; -2 when pixel is not burnable: water bodies, bare areas, urban areas,
+    #       and permanent snow and ice.
+    #
+    # when loaded with xarray.open_dataset/xarray.open_mdataset without any kwargs the
+    # type of the JD variable is datetime64[ns], which is wrong. The correct type should
+    # be int16.
+
+    ds = from_source(
+        "url",
+        earthkit_remote_test_data_file("20220401-C3S-L3S_FIRE-BA-OLCI-AREA_3-fv1.1.nc"),
+    )
+
+    r = ds.to_xarray(decode_cf=False, decode_times=False)
+    assert r["JD"].dtype == "int16"
+    r["JD"].shape == (1, 20880, 28440)
+    assert np.isclose(r["JD"].values.min(), -2)
+    assert np.isclose(r["JD"].values.max(), 120)
+
+    r = ds.to_xarray(decode_cf=False, decode_times=False)
+    assert r["JD"].dtype == "int16"
+    r["JD"].shape == (1, 20880, 28440)
+    assert np.isclose(r["JD"].values.min(), -2)
+    assert np.isclose(r["JD"].values.max(), 120)
+
+    r = ds.to_xarray()
+    assert r["JD"].dtype == "<M8[ns]"
+    r["JD"].shape == (1, 20880, 28440)
+
+
+def test_netcdf_single_to_xarray_1():
+    d = from_source(
+        "file",
+        earthkit_test_data_file("era5_2t_1.nc"),
+    )
+
+    r = d.to_xarray(xarray_open_dataset_kwargs=dict(decode_cf=False, decode_times=False))
+    assert r["t2m"].shape == (1, 9, 18)
+
+    r = d.to_xarray(decode_cf=False, decode_times=False)
+    assert r["t2m"].shape == (1, 9, 18)
+
+    with pytest.raises(TypeError):
+        d.to_xarray(decode_cf=False, decode_times=False, wrong_kwarg=True)
+
+    with pytest.raises(ValueError):
+        d.to_xarray(xarray_open_mfdataset_kwargs=dict(decode_cf=False, decode_times=False))
+
+
+def test_netcdf_multi_to_xarray_1():
+    d = from_source(
+        "file",
+        [
+            earthkit_test_data_file("era5_2t_1.nc"),
+            earthkit_test_data_file("era5_2t_2.nc"),
+        ],
+    )
+
+    r = d.to_xarray(xarray_open_mfdataset_kwargs=dict(decode_cf=False, decode_times=False))
+    assert r["t2m"].shape == (2, 9, 18)
+
+    with pytest.raises(TypeError):
+        d.to_xarray(xarray_open_dataset_kwargs=dict(decode_cf=False, decode_times=False))
+
+    r = d.to_xarray(decode_cf=False, decode_times=False)
+    assert r["t2m"].shape == (2, 9, 18)
+
+    with pytest.raises(TypeError):
+        d.to_xarray(decode_cf=False, decode_times=False, wrong_kwarg=True)
+
+
 if __name__ == "__main__":
     from earthkit.data.utils.testing import main