`test_Saver`: `DeprecationWarning: elementwise comparison failed`

[trexfeathers]

📰 Custom Issue

Running iris.tests.unit.fileformats.netcdf.test_Saver.Test_write.test_zlib() raises the following DeprecationWarning:

.../iris/fileformats/netcdf.py:2762: DeprecationWarning: elementwise comparison failed; this will raise an error in the future.
  contains_value = fill_value is not None and fill_value in data

---

This appears to stem from the fact that we are patching the netCDF4 module, which means that a MagicMock is passed as the fill_value.

Patching

iris/lib/iris/tests/unit/fileformats/netcdf/test_Saver.py

Lines 188 to 190 in 7fbc200

	api = self.patch("iris.fileformats.netcdf.netCDF4")
	with Saver("/dummy/path", "NETCDF4") as saver:
	saver.write(cube, zlib=True)

Point of warning

iris/lib/iris/fileformats/netcdf.py

Lines 2759 to 2763 in 7fbc200

	def store(data, cf_var, fill_value):
	cf_var[:] = data
	is_masked = ma.is_masked(data)
	contains_value = fill_value is not None and fill_value in data
	return is_masked, contains_value

---

To make sure the test continues to work, we will need to find a way for the MagicMock to be recognised as an appropriate value for Numpy elementwise comparison. But I have no idea how to do this when the MagicMock is coming from a patch!

[trexfeathers]

It is currently uncertain when the deprecation will expire! numpy/numpy#11521 (comment)

[trexfeathers]

#4375 nullifies the DeprecationWarning, however the problem will still need fixing if we expect the deprecation to expire (which it surely will eventually).