Merge pull request #1657 from jhamman/fix/1652

fixes for warnings related to unit tests and nan comparisons

Author: shoyer

doc/whats-new.rst

@@ -85,6 +85,11 @@ Breaking changes
   disk when calling ``repr`` (:issue:`1522`).
   By `Guido Imperiale <https://github.com/crusaderky>`_.
 
+- Suppress ``RuntimeWarning`` issued by ``numpy`` for "invalid value comparisons"
+  (e.g. NaNs). Xarray now behaves similarly to Pandas in its treatment of
+  binary and unary operations on objects with ``NaN``s (:issue:`1657`).
+  By `Joe Hamman <https://github.com/jhamman>`_.
+
 - Several existing features have been deprecated and will change to new
   behavior in xarray v0.11. If you use any of them with xarray v0.10, you
   should see a ``FutureWarning`` that describes how to update your code:

xarray/core/dataarray.py

@@ -1575,7 +1575,9 @@ def __array_wrap__(self, obj, context=None):
     def _unary_op(f):
         @functools.wraps(f)
         def func(self, *args, **kwargs):
-            return self.__array_wrap__(f(self.variable.data, *args, **kwargs))
+            with np.errstate(all='ignore'):
+                return self.__array_wrap__(f(self.variable.data, *args,
+                                             **kwargs))
 
         return func
 

xarray/core/pycompat.py

@@ -25,6 +25,7 @@ def itervalues(d):
     from functools import reduce
     import builtins
     from urllib.request import urlretrieve
+    from inspect import getfullargspec as getargspec
 else:  # pragma: no cover
     # Python 2
     basestring = basestring  # noqa
@@ -43,7 +44,7 @@ def itervalues(d):
     reduce = reduce
     import __builtin__ as builtins
     from urllib import urlretrieve
-
+    from inspect import getargspec
 try:
     from cyordereddict import OrderedDict
 except ImportError:  # pragma: no cover

xarray/core/variable.py

@@ -1354,7 +1354,8 @@ def __array_wrap__(self, obj, context=None):
     def _unary_op(f):
         @functools.wraps(f)
         def func(self, *args, **kwargs):
-            return self.__array_wrap__(f(self.data, *args, **kwargs))
+            with np.errstate(all='ignore'):
+                return self.__array_wrap__(f(self.data, *args, **kwargs))
         return func
 
     @staticmethod
@@ -1364,9 +1365,10 @@ def func(self, other):
             if isinstance(other, (xr.DataArray, xr.Dataset)):
                 return NotImplemented
             self_data, other_data, dims = _broadcast_compat_data(self, other)
-            new_data = (f(self_data, other_data)
-                        if not reflexive
-                        else f(other_data, self_data))
+            with np.errstate(all='ignore'):
+                new_data = (f(self_data, other_data)
+                            if not reflexive
+                            else f(other_data, self_data))
             result = Variable(dims, new_data)
             return result
         return func
@@ -1381,7 +1383,8 @@ def func(self, other):
             if dims != self.dims:
                 raise ValueError('dimensions cannot change for in-place '
                                  'operations')
-            self.values = f(self_data, other_data)
+            with np.errstate(all='ignore'):
+                self.values = f(self_data, other_data)
             return self
         return func
 

xarray/plot/facetgrid.py

@@ -2,13 +2,13 @@
 from __future__ import division
 from __future__ import print_function
 
-import inspect
 import warnings
 import itertools
 import functools
 
 import numpy as np
 
+from ..core.pycompat import getargspec
 from ..core.formatting import format_item
 from .utils import (_determine_cmap_params, _infer_xy_labels,
                     import_matplotlib_pyplot)
@@ -228,7 +228,7 @@ def map_dataarray(self, func, x, y, **kwargs):
                        'filled': func.__name__ != 'contour',
                        }
 
-        cmap_args = inspect.getargspec(_determine_cmap_params).args
+        cmap_args = getargspec(_determine_cmap_params).args
         cmap_kwargs.update((a, kwargs[a]) for a in cmap_args if a in kwargs)
 
         cmap_params = _determine_cmap_params(**cmap_kwargs)

xarray/tests/test_accessors.py

@@ -6,7 +6,7 @@
 import numpy as np
 import pandas as pd
 
-from . import TestCase, requires_dask
+from . import TestCase, requires_dask, raises_regex
 
 
 class TestDatetimeAccessor(TestCase):
@@ -45,7 +45,7 @@ def test_not_datetime_type(self):
         nontime_data = self.data.copy()
         int_data = np.arange(len(self.data.time)).astype('int8')
         nontime_data['time'].values = int_data
-        with self.assertRaisesRegexp(TypeError, 'dt'):
+        with raises_regex(TypeError, 'dt'):
             nontime_data.time.dt
 
     @requires_dask
@@ -93,4 +93,4 @@ def test_seasons(self):
                    "SON", "SON", "SON", "DJF"]
         seasons = xr.DataArray(seasons)
 
-        self.assertArrayEqual(seasons.values, dates.dt.season.values)
+        self.assertArrayEqual(seasons.values, dates.dt.season.values)

xarray/tests/test_backends.py

@@ -28,7 +28,7 @@
                requires_scipy_or_netCDF4, requires_dask, requires_h5netcdf,
                requires_pynio, requires_pathlib, has_netCDF4, has_scipy,
                assert_allclose, flaky, network, requires_rasterio,
-               assert_identical)
+               assert_identical, raises_regex)
 from .test_dataset import create_test_data
 
 from xarray.tests import mock
@@ -113,7 +113,7 @@ def __getitem__(self, key):
                 return self.array[key]
 
         array = UnreliableArray([0])
-        with self.assertRaises(UnreliableArrayFailure):
+        with pytest.raises(UnreliableArrayFailure):
             array[0]
         self.assertEqual(array[0], 0)
 
@@ -218,7 +218,7 @@ def assert_loads(vars=None):
                         self.assertTrue(v._in_memory)
                 self.assertDatasetIdentical(expected, actual)
 
-        with self.assertRaises(AssertionError):
+        with pytest.raises(AssertionError):
             # make sure the contextmanager works!
             with assert_loads() as ds:
                 pass
@@ -345,8 +345,7 @@ def test_roundtrip_datetime_data(self):
         kwds = {'encoding': {'t0': {'units': 'days since 1950-01-01'}}}
         with self.roundtrip(expected, save_kwargs=kwds) as actual:
             self.assertDatasetIdentical(expected, actual)
-            self.assertEquals(actual.t0.encoding['units'],
-                              'days since 1950-01-01')
+            assert actual.t0.encoding['units'] == 'days since 1950-01-01'
 
     def test_roundtrip_timedelta_data(self):
         time_deltas = pd.to_timedelta(['1h', '2h', 'NaT'])
@@ -528,7 +527,7 @@ def test_roundtrip_endian(self):
 
         if type(self) is NetCDF4DataTest:
             ds['z'].encoding['endian'] = 'big'
-            with self.assertRaises(NotImplementedError):
+            with pytest.raises(NotImplementedError):
                 with self.roundtrip(ds) as actual:
                     pass
 
@@ -539,7 +538,7 @@ def test_invalid_dataarray_names_raise(self):
         da = xr.DataArray(data)
         for name, e in zip([0, (4, 5), True, ''], [te, te, te, ve]):
             ds = Dataset({name: da})
-            with self.assertRaisesRegexp(*e):
+            with raises_regex(*e):
                 with self.roundtrip(ds) as actual:
                     pass
 
@@ -551,17 +550,17 @@ def test_encoding_kwarg(self):
         self.assertEqual(ds.x.encoding, {})
 
         kwargs = dict(encoding={'x': {'foo': 'bar'}})
-        with self.assertRaisesRegexp(ValueError, 'unexpected encoding'):
+        with raises_regex(ValueError, 'unexpected encoding'):
             with self.roundtrip(ds, save_kwargs=kwargs) as actual:
                 pass
 
         kwargs = dict(encoding={'x': 'foo'})
-        with self.assertRaisesRegexp(ValueError, 'must be castable'):
+        with raises_regex(ValueError, 'must be castable'):
             with self.roundtrip(ds, save_kwargs=kwargs) as actual:
                 pass
 
         kwargs = dict(encoding={'invalid': {}})
-        with self.assertRaises(KeyError):
+        with pytest.raises(KeyError):
             with self.roundtrip(ds, save_kwargs=kwargs) as actual:
                 pass
 
@@ -676,9 +675,9 @@ def test_open_group(self):
                     self.assertVariableEqual(actual['x'], expected['x'])
 
             # check that missing group raises appropriate exception
-            with self.assertRaises(IOError):
+            with pytest.raises(IOError):
                 open_dataset(tmp_file, group='bar')
-            with self.assertRaisesRegexp(ValueError, 'must be a string'):
+            with raises_regex(ValueError, 'must be a string'):
                 open_dataset(tmp_file, group=(1, 2, 3))
 
     def test_open_subgroup(self):
@@ -1019,7 +1018,7 @@ def create_store(self):
 
     def test_array_attrs(self):
         ds = Dataset(attrs={'foo': [[1, 2], [3, 4]]})
-        with self.assertRaisesRegexp(ValueError, 'must be 1-dimensional'):
+        with raises_regex(ValueError, 'must be 1-dimensional'):
             with self.roundtrip(ds) as roundtripped:
                 pass
 
@@ -1036,11 +1035,11 @@ def test_netcdf3_endianness(self):
 
     @requires_netCDF4
     def test_nc4_scipy(self):
-        with create_tmp_file() as tmp_file:
+        with create_tmp_file(allow_cleanup_failure=True) as tmp_file:
             with nc4.Dataset(tmp_file, 'w', format='NETCDF4') as rootgrp:
                 rootgrp.createGroup('foo')
 
-            with self.assertRaisesRegexp(TypeError, 'pip install netcdf4'):
+            with raises_regex(TypeError, 'pip install netcdf4'):
                 open_dataset(tmp_file, engine='scipy')
 
 
@@ -1096,18 +1095,18 @@ def test_write_store(self):
 
     def test_engine(self):
         data = create_test_data()
-        with self.assertRaisesRegexp(ValueError, 'unrecognized engine'):
+        with raises_regex(ValueError, 'unrecognized engine'):
             data.to_netcdf('foo.nc', engine='foobar')
-        with self.assertRaisesRegexp(ValueError, 'invalid engine'):
+        with raises_regex(ValueError, 'invalid engine'):
             data.to_netcdf(engine='netcdf4')
 
         with create_tmp_file() as tmp_file:
             data.to_netcdf(tmp_file)
-            with self.assertRaisesRegexp(ValueError, 'unrecognized engine'):
+            with raises_regex(ValueError, 'unrecognized engine'):
                 open_dataset(tmp_file, engine='foobar')
 
         netcdf_bytes = data.to_netcdf()
-        with self.assertRaisesRegexp(ValueError, 'can only read'):
+        with raises_regex(ValueError, 'can only read'):
             open_dataset(BytesIO(netcdf_bytes), engine='foobar')
 
     def test_cross_engine_read_write_netcdf3(self):
@@ -1389,12 +1388,12 @@ def test_common_coord_when_datavars_minimal(self):
     def test_invalid_data_vars_value_should_fail(self):
 
         with self.setup_files_and_datasets() as (files, _):
-            with self.assertRaises(ValueError):
+            with pytest.raises(ValueError):
                 with open_mfdataset(files, data_vars='minimum'):
                     pass
 
             # test invalid coord parameter
-            with self.assertRaises(ValueError):
+            with pytest.raises(ValueError):
                 with open_mfdataset(files, coords='minimum'):
                     pass
 
@@ -1452,7 +1451,7 @@ def test_open_mfdataset(self):
                     self.assertEqual(actual.foo.variable.data.chunks,
                                      ((3, 2, 3, 2),))
 
-        with self.assertRaisesRegexp(IOError, 'no files to open'):
+        with raises_regex(IOError, 'no files to open'):
             open_mfdataset('foo-bar-baz-*.nc', autoclose=self.autoclose)
 
     @requires_pathlib
@@ -1483,7 +1482,7 @@ def test_attrs_mfdataset(self):
                     # first dataset loaded
                     self.assertEqual(actual.test1, ds1.test1)
                     # attributes from ds2 are not retained, e.g.,
-                    with self.assertRaisesRegexp(AttributeError,
+                    with raises_regex(AttributeError,
                                                  'no attribute'):
                         actual.test2
 
@@ -1513,15 +1512,15 @@ def test_save_mfdataset_roundtrip(self):
 
     def test_save_mfdataset_invalid(self):
         ds = Dataset()
-        with self.assertRaisesRegexp(ValueError, 'cannot use mode'):
+        with raises_regex(ValueError, 'cannot use mode'):
             save_mfdataset([ds, ds], ['same', 'same'])
-        with self.assertRaisesRegexp(ValueError, 'same length'):
+        with raises_regex(ValueError, 'same length'):
             save_mfdataset([ds, ds], ['only one path'])
 
     def test_save_mfdataset_invalid_dataarray(self):
         # regression test for GH1555
         da = DataArray([1, 2])
-        with self.assertRaisesRegexp(TypeError, 'supports writing Dataset'):
+        with raises_regex(TypeError, 'supports writing Dataset'):
             save_mfdataset([da], ['dataarray'])
 
 
@@ -1846,11 +1845,11 @@ def test_indexing(self):
                 # but on x and y only windowed operations are allowed, more
                 # exotic slicing should raise an error
                 err_msg = 'not valid on rasterio'
-                with self.assertRaisesRegexp(IndexError, err_msg):
+                with raises_regex(IndexError, err_msg):
                     actual.isel(x=[2, 4], y=[1, 3]).values
-                with self.assertRaisesRegexp(IndexError, err_msg):
+                with raises_regex(IndexError, err_msg):
                     actual.isel(x=[4, 2]).values
-                with self.assertRaisesRegexp(IndexError, err_msg):
+                with raises_regex(IndexError, err_msg):
                     actual.isel(x=slice(5, 2, -1)).values
 
                 # Integer indexing
@@ -1916,7 +1915,7 @@ def test_caching(self):
 
                 # Without cache an error is raised
                 err_msg = 'not valid on rasterio'
-                with self.assertRaisesRegexp(IndexError, err_msg):
+                with raises_regex(IndexError, err_msg):
                     actual.isel(x=[2, 4]).values
 
                 # This should cache everything
@@ -1976,7 +1975,7 @@ class TestEncodingInvalid(TestCase):
 
     def test_extract_nc4_variable_encoding(self):
         var = xr.Variable(('x',), [1, 2, 3], {}, {'foo': 'bar'})
-        with self.assertRaisesRegexp(ValueError, 'unexpected encoding'):
+        with raises_regex(ValueError, 'unexpected encoding'):
             _extract_nc4_variable_encoding(var, raise_on_invalid=True)
 
         var = xr.Variable(('x',), [1, 2, 3], {}, {'chunking': (2, 1)})
@@ -1992,7 +1991,7 @@ def test_extract_h5nc_encoding(self):
         # not supported with h5netcdf (yet)
         var = xr.Variable(('x',), [1, 2, 3], {},
                           {'least_sigificant_digit': 2})
-        with self.assertRaisesRegexp(ValueError, 'unexpected encoding'):
+        with raises_regex(ValueError, 'unexpected encoding'):
             _extract_nc4_variable_encoding(var, raise_on_invalid=True)
 
 
@@ -2025,17 +2024,17 @@ def new_dataset_and_coord_attrs():
             ds, attrs = new_dataset_and_attrs()
 
             attrs[123] = 'test'
-            with self.assertRaisesRegexp(TypeError, 'Invalid name for attr'):
+            with raises_regex(TypeError, 'Invalid name for attr'):
                 ds.to_netcdf('test.nc')
 
             ds, attrs = new_dataset_and_attrs()
             attrs[MiscObject()] = 'test'
-            with self.assertRaisesRegexp(TypeError, 'Invalid name for attr'):
+            with raises_regex(TypeError, 'Invalid name for attr'):
                 ds.to_netcdf('test.nc')
 
             ds, attrs = new_dataset_and_attrs()
             attrs[''] = 'test'
-            with self.assertRaisesRegexp(ValueError, 'Invalid name for attr'):
+            with raises_regex(ValueError, 'Invalid name for attr'):
                 ds.to_netcdf('test.nc')
 
             # This one should work
@@ -2046,12 +2045,12 @@ def new_dataset_and_coord_attrs():
 
             ds, attrs = new_dataset_and_attrs()
             attrs['test'] = {'a': 5}
-            with self.assertRaisesRegexp(TypeError, 'Invalid value for attr'):
+            with raises_regex(TypeError, 'Invalid value for attr'):
                 ds.to_netcdf('test.nc')
 
             ds, attrs = new_dataset_and_attrs()
             attrs['test'] = MiscObject()
-            with self.assertRaisesRegexp(TypeError, 'Invalid value for attr'):
+            with raises_regex(TypeError, 'Invalid value for attr'):
                 ds.to_netcdf('test.nc')
 
             ds, attrs = new_dataset_and_attrs()