Merge pull request #156 from andersy005/master

Avoid setting the `.data` attribute

esmlab/core.py

@@ -84,11 +84,10 @@ def compute_time(self):
         else:
             groupby_coord = self.get_time_decoded(midpoint=False)
 
-        ds[self.time_coord_name].data = groupby_coord.data
+        ds[self.time_coord_name] = groupby_coord.data
 
         if self.time_bound is not None:
-            ds[self.tb_name].data = self.time_bound.data
-            self.time_bound[self.time_coord_name].data = groupby_coord.data
+            self.time_bound[self.time_coord_name] = groupby_coord.data
         self.time_bound_diff = self.compute_time_bound_diff(ds)
 
         self._ds_time_computed = ds
@@ -101,6 +100,7 @@ def compute_time_bound_diff(self, ds):
         if self.time_bound is not None:
             time_bound_diff.name = self.tb_name + '_diff'
             time_bound_diff.attrs = {}
+            # Compute
             time_bound_diff.data = self.time_bound.diff(dim=self.tb_dim)[:, 0]
             if self.tb_dim in time_bound_diff.coords:
                 time_bound_diff = time_bound_diff.drop(self.tb_dim)
@@ -142,7 +142,7 @@ def get_original_metadata(self):
             v: {
                 key: val
                 for key, val in self._ds[v].encoding.items()
-                if key in ['dtype', '_FillValue', 'missing_value']
+                if key in ['dtype', '_FillValue', 'missing_value', 'units', 'calendar']
             }
             for v in self._ds.variables
         }
@@ -440,7 +440,7 @@ def compute_resample_times(self, ds, temporary_time_coord_name, time_dot, method
         else:
             time_values = self._ds_time_computed[self.time_coord_name].groupby(time_dot).mean()
 
-        ds[self.time_coord_name].data = time_values.data
+        ds[self.time_coord_name] = time_values.data
 
         return ds
 
@@ -453,24 +453,13 @@ def compute_mon_climatology(self):
             self._ds_time_computed.drop(self.static_variables)
             .groupby(time_dot_month)
             .mean(self.time_coord_name)
-            .rename({'month': self.time_coord_name})
         )
-        computed_dset['month'] = computed_dset[self.time_coord_name].copy()
         attrs = {'month': {'long_name': 'Month', 'units': 'month'}}
-        encoding = {
-            'month': {'dtype': 'int32', '_FillValue': None},
-            self.time_coord_name: {'dtype': 'float', '_FillValue': None},
-        }
-
-        if self.time_bound is not None:
-            time_data = computed_dset[self.tb_name] - computed_dset[self.tb_name][0, 0]
-            computed_dset[self.tb_name] = time_data
-            computed_dset[self.time_coord_name].data = (
-                computed_dset[self.tb_name].mean(self.tb_dim).data
-            )
-            encoding[self.tb_name] = {'dtype': 'float', '_FillValue': None}
+        encoding = {'month': {'dtype': 'int32', '_FillValue': None}}
 
-        return self.restore_dataset(computed_dset, attrs=attrs, encoding=encoding)
+        if self.tb_name in computed_dset.data_vars:
+            computed_dset = computed_dset.drop(self.tb_name)
+        return self.update_metadata(computed_dset, attrs, encoding)
 
     @esmlab_xr_set_options(arithmetic_join='exact')
     def compute_mon_anomaly(self, slice_mon_clim_time=None):
@@ -504,9 +493,9 @@ def compute_mon_anomaly(self, slice_mon_clim_time=None):
         # reset month to become a variable
         computed_dset = computed_dset.reset_coords('month')
 
-        computed_dset[self.time_coord_name].data = self.time.data
+        computed_dset[self.time_coord_name] = self.time
         if self.time_bound is not None:
-            computed_dset[self.tb_name].data = self.time_bound.data
+            computed_dset[self.tb_name] = self.time_bound
 
         attrs = {'month': {'long_name': 'Month'}}
         return self.restore_dataset(computed_dset, attrs=attrs)

tests/test_core.py

@@ -33,6 +33,7 @@ def test_esmlab_accessor():
     attrs = {'calendar': 'noleap', 'units': 'days since 2000-01-01 00:00:00'}
     ds.time.attrs = attrs
     esm = ds.esmlab.set_time(time_coord_name='time')
+    xr.testing._assert_internal_invariants(esm._ds_time_computed)
     # Time and Time bound Attributes
     expected = dict(esm.time_attrs)
     attrs['bounds'] = None
@@ -96,6 +97,7 @@ def test_datetime_cftime_exception():
 
 def test_time_bound_var(dset, time_coord_name='time'):
     esm = dset.esmlab.set_time(time_coord_name=time_coord_name)
+    xr.testing._assert_internal_invariants(esm._ds_time_computed)
     results = esm.tb_name, esm.tb_dim
     expected = ('time_bound', 'd2')
     assert results == expected
@@ -114,6 +116,7 @@ def test_get_time_attrs(dset, time_coord_name='time'):
         'bounds': 'time_bound',
     }
     esm = dset.esmlab.set_time(time_coord_name=time_coord_name)
+    xr.testing._assert_internal_invariants(esm._ds_time_computed)
     results = esm.time_attrs
     assert results == expected
 
@@ -122,12 +125,14 @@ def test_compute_time_var(dset, time_coord_name='time'):
     idx = dset.indexes[time_coord_name]
     assert isinstance(idx, pd.core.indexes.numeric.Index)
     esm = dset.esmlab.set_time(time_coord_name=time_coord_name)
+    xr.testing._assert_internal_invariants(esm._ds_time_computed)
     results = esm.get_time_decoded()
     assert isinstance(results, xr.DataArray)
 
 
 def test_uncompute_time_var(dset, time_coord_name='time'):
     esm = dset.esmlab.set_time(time_coord_name=time_coord_name)
+    xr.testing._assert_internal_invariants(esm._ds_time_computed)
     ds = esm.compute_time_var()
     assert ds[time_coord_name].dtype == np.dtype('O')
     dset_with_uncomputed_time = esm.uncompute_time_var()
@@ -137,6 +142,7 @@ def test_uncompute_time_var(dset, time_coord_name='time'):
 # For some strange reason, this case fails when using pytest parametrization
 def test_sel_time_(dset):
     esm = dset.esmlab.set_time()
+    xr.testing._assert_internal_invariants(esm._ds_time_computed)
     dset = esm.sel_time(indexer_val=slice('1850-01-01', '1850-12-31'), year_offset=1850)
     assert len(dset.time) == 12
 
@@ -155,10 +161,10 @@ def test_sel_time_(dset):
 def test_mon_climatology(ds_name, decoded, variables, time_coord_name):
     ds = esmlab.datasets.open_dataset(ds_name, decode_times=decoded)
     esm = ds.esmlab.set_time(time_coord_name=time_coord_name)
+    xr.testing._assert_internal_invariants(esm._ds_time_computed)
     computed_dset = esmlab.climatology(ds, freq='mon')
-    esmlab_res = computed_dset.drop(esm.static_variables).to_dataframe()
+    esmlab_res = computed_dset.to_dataframe()
     esmlab_res = esmlab_res.groupby('month').mean()[variables]
-
     df = (
         esm._ds_time_computed.drop(esm.static_variables)
         .to_dataframe()
@@ -171,11 +177,6 @@ def test_mon_climatology(ds_name, decoded, variables, time_coord_name):
 
     assert_both_frames_equal(esmlab_res, pd_res)
 
-    assert computed_dset[esm.time_coord_name].dtype == ds[esm.time_coord_name].dtype
-    for key, value in ds[esm.time_coord_name].attrs.items():
-        assert key in computed_dset[esm.time_coord_name].attrs
-        assert value == computed_dset[esm.time_coord_name].attrs[key]
-
 
 @pytest.mark.parametrize(
     'ds_name, decoded, variables, time_coord_name, time_bound_name',
@@ -196,9 +197,10 @@ def test_mon_climatology_drop_time_bounds(
     ds = ds.drop(ds_time_bound)
     del ds[time_coord_name].attrs[time_bound_name]
     esm = ds.esmlab.set_time(time_coord_name=time_coord_name)
+    xr.testing._assert_internal_invariants(esm._ds_time_computed)
     computed_dset = esmlab.climatology(ds, freq='mon')
 
-    esmlab_res = computed_dset.drop(esm.static_variables).to_dataframe()
+    esmlab_res = computed_dset.to_dataframe()
     esmlab_res = esmlab_res.groupby('month').mean()[variables]
 
     df = (
@@ -213,11 +215,6 @@ def test_mon_climatology_drop_time_bounds(
 
     assert_both_frames_equal(esmlab_res, pd_res)
 
-    assert computed_dset[esm.time_coord_name].dtype == ds[esm.time_coord_name].dtype
-    for key, value in ds[esm.time_coord_name].attrs.items():
-        assert key in computed_dset[esm.time_coord_name].attrs
-        assert value == computed_dset[esm.time_coord_name].attrs[key]
-
 
 def test_anomaly_with_monthly_clim(dset):
     computed_dset = esmlab.anomaly(dset, clim_freq='mon')
@@ -276,7 +273,6 @@ def test_resample_mon_mean(dset):
     for method in {'left', 'right'}:
         ds = esmlab.resample(dset, freq='mon', method=method)
         assert len(ds.time) == 12
-
     computed_dset = esmlab.resample(dset, freq='mon')
     res = computed_dset.variable_1.data
     expected = np.full(shape=(12, 2, 2), fill_value=0.5, dtype=np.float32)