WIP: 1d+2d coord plotting

[dcherian]

• Closes #xxxx
• Tests added / passed
• Passes git diff upstream/master **/*py | flake8 --diff
• Fully documented, including whats-new.rst for all changes and api.rst for new API

This PR teaches plot.contourf() to contour variables with both a 1D co-ordinate (e.g. time) and a 2D co-ordinate (e.g. time-varying depth) da:

<xarray.DataArray 'S' (z: 5, time: 85646)>
array([[       nan,        nan,        nan, ...,        nan,        nan,
               nan],
       [ 35.02816 ,  34.729567,  34.779223, ...,  34.57513 ,  34.671975,
         34.334675],
       [ 35.206943,  35.163239,  34.938674, ...,  34.780728,  34.836331,
         34.70386 ],
       [       nan,  35.184057,  35.10592 , ...,  34.656925,  34.776915,
         34.429442],
       [ 34.85562 ,  34.81994 ,  35.00963 , ...,  35.014522,  34.9747  ,
         35.033848]])
Coordinates:
    depth    (z, time) float16 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 53.5 ...
  * time     (time) datetime64[ns] 2013-12-19T06:00:01 2013-12-19T06:10:01 ...
Dimensions without coordinates: z

Now we can do da.plot(x='time', y='depth')

Couple of questions:

1. I've added a test, but am not sure how to work in an assert statement.
2. How do I test that I haven't messed up the syntax in whats-new.rst

[shoyer]

@dcherian this sounds great -- can you please comment with an example the plots you can make with this?

I've added a test, but am not sure how to work in an assert statement.

For plotting, we are often OK with just verifying that a plot of some kind is made and there are no errors.

How do I test that I haven't messed up the syntax in whats-new.rst

I'll look over it, but don't worry about this very much -- we check the output manually before releases.

[dcherian]

The second call would error out without this PR.

The practical application here is that I have instruments that are moving up and down in the water so the depth of measurement is a function of time.

[fmaussion]

This looks good to me. Thanks!

[shoyer]

Thanks @dcherian !