Data.mask: migrate to Dask

cf/data/data.py

@@ -6351,9 +6351,9 @@ def datetime_array(self):
 
  @property
  def mask(self):
- """The boolean missing data mask of the data array.
+ """The Boolean missing data mask of the data array.
 
- The boolean mask has True where the data array has missing data
+ The Boolean mask has True where the data array has missing data
  and False otherwise.
 
  :Returns:
@@ -6368,40 +6368,19 @@ def mask(self):
  >>> m.dtype
  dtype('bool')
  >>> m.shape
- (12, 73, 96])
+ (12, 73, 96)
 
  """
- mask = self.copy()
-
- config = mask.partition_configuration(readonly=False)
-
- for partition in mask.partitions.matrix.flat:
- partition.open(config)
- array = partition.array
-
- if partition.masked:
- # Array is masked
- partition.subarray = array.mask.copy()
- else:
- # Array is not masked
- partition.subarray = FilledArray(
- shape=array.shape,
- size=array.size,
- ndim=array.ndim,
- dtype=_dtype_bool,
- fill_value=0,
- )
-
- partition.Units = _units_None
+ mask_data_obj = self.copy()
 
- partition.close()
-
- mask._Units = _units_None
- mask.dtype = _dtype_bool
+ dx = self._get_dask()
+ mask = da.ma.getmaskarray(dx)
 
- mask._hardmask = True
+ mask_data_obj._set_dask(mask, reset_mask_hardness=True)
+ mask_data_obj.override_units(_units_None, inplace=True)
+ mask_data_obj.hardmask = True
 
- return mask
+ return mask_data_obj
 
  @staticmethod
  def mask_fpe(*arg):

cf/test/test_Data.py

@@ -35,7 +35,6 @@
 
 ones = np.ones(a.shape, dtype=float)
 
-# TODODASK: these can be moved into the lone tests that use them now
 ma = np.ma.arange(-100, 200.0, dtype=float).reshape(3, 4, 5, 5)
 ma[:, 1, 4, 4] = np.ma.masked
 ma[0, :, 2, 3] = np.ma.masked
@@ -77,7 +76,6 @@ class DataTest(unittest.TestCase):
  os.path.dirname(os.path.abspath(__file__)), "test_file2.nc"
  )
 
- # TODODASK: these can be moved into the lone tests that use them now
  a = a
  w = w
  ma = ma
@@ -535,17 +533,53 @@ def test_Data_halo(self):
  # [ 8 8 9 10 -- --]
  # [ 8 8 9 10 -- --]]
 
- @unittest.skipIf(TEST_DASKIFIED_ONLY, "no attr. 'partition_configuration'")
- def test_Data_apply_masking(self):
+ def test_Data_mask(self):
  if self.test_only and inspect.stack()[0][3] not in self.test_only:
  return
 
- a = self.ma
+ # TODODASK: once test_Data_apply_masking is passing after daskification
+ # of apply_masking, might make sense to combine this test with that?
 
+ # Test for a masked Data object (having some masked points)
+ a = self.ma
  d = cf.Data(a, units="m")
-
  self.assertTrue((a == d.array).all())
  self.assertTrue((a.mask == d.mask.array).all())
+ self.assertEqual(d.mask.shape, d.shape)
+ self.assertEqual(d.mask.dtype, bool)
+ self.assertEqual(d.mask.Units, cf.Units(None))
+ self.assertTrue(d.mask.hardmask)
+ self.assertIn(True, d.mask.array)
+
+ # Test for a non-masked Data object
+ a2 = np.arange(-100, 200.0, dtype=float).reshape(3, 4, 5, 5)
+ d2 = cf.Data(a2, units="m")
+ d2[...] = a2
+ self.assertTrue((a2 == d2.array).all())
+ self.assertEqual(d2.shape, d2.mask.shape)
+ self.assertEqual(d2.mask.dtype, bool)
+ self.assertEqual(d2.mask.Units, cf.Units(None))
+ self.assertTrue(d2.mask.hardmask)
+ self.assertNotIn(True, d2.mask.array)
+
+ # Test for a masked Data object of string type, including chunking
+ a3 = np.ma.array(["one", "two", "four"], dtype="S4")
+ a3[1] = np.ma.masked
+ d3 = cf.Data(a3, "m", chunks=(3,))
+ self.assertTrue((a3 == d3.array).all())
+ self.assertEqual(d3.shape, d3.mask.shape)
+ self.assertEqual(d3.mask.dtype, bool)
+ self.assertEqual(d3.mask.Units, cf.Units(None))
+ self.assertTrue(d3.mask.hardmask)
+ self.assertTrue(d3.mask.array[1], True)
+
+ @unittest.skipIf(TEST_DASKIFIED_ONLY, "no attr. 'partition_configuration'")
+ def test_Data_apply_masking(self):
+ if self.test_only and inspect.stack()[0][3] not in self.test_only:
+ return
+
+ a = self.ma
+ d = cf.Data(a, units="m")
 
  b = a.copy()
  e = d.apply_masking()