add a baseline obs_concat function that just wraps the xarray version

Author: shunt16

obsarray/concat.py

@@ -1,6 +1,5 @@
 """obsarray.concat - module with extension to xarray.concat for obs_vars and unc_vars"""
 
-import numpy as np
 import xarray as xr
 from typing import Union, Any
 from xarray.core.types import T_Dataset, T_DataArray, T_Variable
@@ -21,6 +20,8 @@ def obs_concat(
     **kwargs
 ):
     """
+    **NB: THIS FUNCTION IS NOT CURRENTLY IMPLEMENTED AND JUST ACTS AS A WRAPPER FOR XR.CONCAT**
+
     Concatenate xarray *obs_vars* along a new or existing dimension, safely handling also
     concatenating associated *unc_vars*. Extension to :py:func:`xarray.concat`.
 
@@ -47,7 +48,5 @@ def obs_concat(
     return concat_obs_vars, concat_unc_vars
 
 
-
-
 if __name__ == "__main__":
-    pass
+    pass

obsarray/test/test_concat.py

@@ -71,132 +71,45 @@ def create_test_ds():
 
     return ds
 
-def create_concat_ds():
-    c1a = np.ones((4, 3))
-    c2a = np.ones((4, 3))
-
-    c1b = np.ones((7, 5, 3))
-    c2b = np.ones((7, 5, 3))
-
-    da = np.ones((4, 3, 2)) * 1
-    da[:, : 1] = 2
-
-    s1a = np.ones((4, 3)) * 3
-    s2a = np.ones((4, 3)) * 4
-    s3a = np.ones((4, 3)) * 5
-
-    d1b = np.ones((7, 5, 3)) * 1
-
-    da_attrs = {"units": "test_units", "geometry": "a", "measurand": "d", 'm': 10}
-    s1a_attrs = {"units": "test_units", "geometry": "a", "measurand": "s", 'm': 12}
-    s2a_attrs = {"units": "test_units", "geometry": "a", "measurand": "s", 'm': 4}
-    s3a_attrs = {"units": "test_units", "geometry": "a", "measurand": "s", 'm': 5}
-    d1b_attrs = {"units": "test_units", "geometry": "b", "measurand": "d"}
-
-    ds = xr.Dataset(
-        {
-            "da": (["xa", "ya"], da, da_attrs),
-            "s1a": (["xa", "ya"], s1a, s1a_attrs),
-            "s2a": (["xa", "ya"], s2a, s2a_attrs),
-            "s3a": (["xa", "ya"], s3a, s3a_attrs),
-            "d1b": (["xb", "yb", "zb"], d1b, d1b_attrs),
-        },
-        coords={
-            "c1a": (["xa", "ya"], c1a),
-            "c2a": (["xa", "ya"], c2a),
-            "c1b": (["xb", "yb", "zb"], c1b),
-            "c2b": (["xb", "yb", "zb"], c2b),
-        },
-        attrs={
-            "history": "test_history",
-            "meas_vars": ["d1a", "d2a", "s1a", "s2a", "s3a", "d1b"],
-        },
-    )
-
-    for var in ["d1a", "d2a"]:
-        ds.unc[var]["u_r_" + var] = (["xa", "ya"], ds[var].values, {})
-
-        err_corr_def = [
-            {
-                "dim": ["xa", "ya"],
-                "form": "systematic",
-                "params": [],
-                "units": []
-            }
-        ]
-
-        ds.unc[var]["u_s_" + var] = (["xa", "ya"], ds[var].values, {"err_corr": err_corr_def})
-
-    return ds
-
-
-
-
 class TestConcat(unittest.TestCase):
 
     def test_concat_combine_unc_concat(self):
         ds = create_test_ds()
 
-        obs_vars, unc_vars = obs_concat([ds["d1a"], ds["d2a"]], "new_dim", ds, "concat")
-
-        c1a = np.ones((4, 3))
-        c2a = np.ones((4, 3))
+        concat_vars, concat_unc_vars = obs_concat([ds["d1a"], ds["d2a"]], "new_dim", ds, "concat")
 
-        c1b = np.ones((7, 5, 3))
-        c2b = np.ones((7, 5, 3))
+        # Create expected concatenated data array
+        data = np.array([[[1., 1., 1.],
+                          [1., 1., 1.],
+                          [1., 1., 1.],
+                          [1., 1., 1.]],
+                         [[2., 2., 2.],
+                          [2., 2., 2.],
+                          [2., 2., 2.],
+                          [2., 2., 2.]]])
 
-        da = np.ones((4, 3, 2)) * 1
-        da[:, : 1] = 2
+        coord_data = np.ones((4, 3))  # coords
 
-        s1a = np.ones((4, 3)) * 3
-        s2a = np.ones((4, 3)) * 4
-        s3a = np.ones((4, 3)) * 5
-
-        d1b = np.ones((7, 5, 3)) * 1
-
-        da_attrs = {"units": "test_units", "geometry": "a", "measurand": "d", 'm': 10}
-        s1a_attrs = {"units": "test_units", "geometry": "a", "measurand": "s", 'm': 12}
-        s2a_attrs = {"units": "test_units", "geometry": "a", "measurand": "s", 'm': 4}
-        s3a_attrs = {"units": "test_units", "geometry": "a", "measurand": "s", 'm': 5}
-        d1b_attrs = {"units": "test_units", "geometry": "b", "measurand": "d"}
-
-        ds = xr.Dataset(
-            {
-                "da": (["xa", "ya"], da, da_attrs),
-                "s1a": (["xa", "ya"], s1a, s1a_attrs),
-                "s2a": (["xa", "ya"], s2a, s2a_attrs),
-                "s3a": (["xa", "ya"], s3a, s3a_attrs),
-                "d1b": (["xb", "yb", "zb"], d1b, d1b_attrs),
-            },
+        exp_concat_vars = xr.DataArray(
+            data,
+            dims=("new_dim", "xa", "ya"),
             coords={
-                "c1a": (["xa", "ya"], c1a),
-                "c2a": (["xa", "ya"], c2a),
-                "c1b": (["xb", "yb", "zb"], c1b),
-                "c2b": (["xb", "yb", "zb"], c2b),
+                "c1a": (("xa", "ya"), coord_data),
+                "c2a": (("xa", "ya"), coord_data),
             },
+            name="d1a",
             attrs={
-                "history": "test_history",
-                "meas_vars": ["d1a", "d2a", "s1a", "s2a", "s3a", "d1b"],
-            },
+                "units": "test_units",
+                "geometry": "a",
+                "measurand": "d",
+                "m": 10,
+                "unc_comps": ["u_r_d1a", "u_s_d1a"]
+            }
         )
 
-        for var in ["d1a", "d2a"]:
-            ds.unc[var]["u_r_" + var] = (["xa", "ya"], ds[var].values, {})
-
-            err_corr_def = [
-                {
-                    "dim": ["xa", "ya"],
-                    "form": "systematic",
-                    "params": [],
-                    "units": []
-                }
-            ]
-
-            ds.unc[var]["u_s_" + var] = (["xa", "ya"], ds[var].values, {"err_corr": err_corr_def})
-
-        return ds
-
-
+        # Tests
+        xr.testing.assert_equal(concat_vars, exp_concat_vars)
+        self.assertIsNone(concat_unc_vars)
 
 if __name__ == "__main__":
     pass