More fixes to the LDO mask handling, etc.

[keflavich]

This revisits issue #392, which wasn't actually solved by that PR. However, it
turns out we need to make deeper changes.

[coveralls]

Coverage increased (+0.4%) to 75.353% when pulling dd3a86e on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[coveralls]

Coverage increased (+0.4%) to 75.353% when pulling af0f972 on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[coveralls]

Coverage increased (+0.3%) to 75.264% when pulling c616e22 on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[coveralls]

Coverage increased (+0.3%) to 75.264% when pulling 893d591 on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[coveralls]

Coverage increased (+0.3%) to 75.264% when pulling d16fccb on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[coveralls]

Coverage increased (+0.3%) to 75.264% when pulling c7d60ec on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[keflavich]

I have a case where I'm getting an error that I can't come up with a test for:

In [3]: cube[:10,20,20].max()
Out[3]: Traceback (most recent call last):
  File "/Users/adam/anaconda/envs/astropy35/lib/python3.5/site-packages/IPython/core/formatters.py", line 693, in __call__
    printer.pretty(obj)
  File "/Users/adam/anaconda/envs/astropy35/lib/python3.5/site-packages/IPython/lib/pretty.py", line 380, in pretty
    return _default_pprint(obj, self, cycle)
  File "/Users/adam/anaconda/envs/astropy35/lib/python3.5/site-packages/IPython/lib/pretty.py", line 495, in _default_pprint
    _repr_pprint(obj, p, cycle)
  File "/Users/adam/anaconda/envs/astropy35/lib/python3.5/site-packages/IPython/lib/pretty.py", line 693, in _repr_pprint
    output = repr(obj)
  File "/Users/adam/repos/spectral-cube/spectral_cube/lower_dimensional_structures.py", line 481, in __repr__
    arrstr = np.array2string(self.filled_data[:].value, separator=',',
  File "/Users/adam/repos/spectral-cube/spectral_cube/cube_utils.py", line 203, in __getitem__
    return self._func(self._other, view)
  File "/Users/adam/repos/spectral-cube/spectral_cube/base_class.py", line 273, in filled_data
    return u.Quantity(self._get_filled_data(view, fill=self._fill_value),
  File "/Users/adam/repos/astropy/astropy/units/quantity.py", line 918, in __getattr__
    attr))
AttributeError: 'OneDSpectrum' object has no '_fill_value' member

[keflavich]

How do we return a Quantity if a ufunc reducing to zero dimensions is run?

[coveralls]

Coverage increased (+0.3%) to 75.295% when pulling 6ebe24f on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[coveralls]

Coverage increased (+0.3%) to 75.333% when pulling 359d565 on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[coveralls]

Coverage increased (+0.3%) to 75.333% when pulling 359d565 on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.

[e-koch]

@keflavich -- I reverted all of the changes for 2D masking.

[e-koch]

👍

The other issues discussed here are with regards to 2D masking and WCS objects without both celestial axes defined.

[coveralls]

Coverage increased (+0.3%) to 75.333% when pulling e6230e1 on keflavich:issue392_revisited into e2b9019 on radio-astro-tools:master.