Fix future warning from xarray

[xylar]

Iterating over data sets will no longer iterate over coordinates in addition to data variables. We want both when we write out NetCDF files, so we switch to doing so explicitly.

[xylar]

A potential future problem I noticed in the process of testing #280 under python 3 (because I'm spending a lot of time in the log files).

@milenaveneziani, a quick check to make sure you don't see any future warnings in the log files when NetCDF files get written out (e.g. during remapping) would be all that this would need. And obviously it would be good to make sure it doesn't break anything, but I don't expect any trouble since it passed my usual QU240 test.

[xylar]

@vanroekel, is this a PR you would have time to test for me? @milenaveneziani has quite a bit on her plate right now on the A-Prime side. If not, no problem. Just say so.

[vanroekel]

@xylar sure I can test this. Do you have steps to reproduce the warning with develop? I tried on titan just now and don't see any warnings about x-array. There is a numpy warning about masked arrays. Do I need to be using the new e3sm conda environment? I'm guessing that the default environment (acme) doesn't have python 3.

[xylar]

@vanroekel, you don't need to check with python 3 but presumably you do need to check with the latest e3sm-unified environment, which is currently only available on Edison (module load e3sm-unified/1.1.2-py2-nox) or acme1. Or you can install the required libraries listed in the MPAS-Analysis documentation into a new anaconda environment of your own, in which case you should get new enough versions of numpy, xarray, etc. that you should see the issue.

[vanroekel]

@xylar I am unable to reproduce the future warning. It looks like this is because I can't get xarray v0.10. On all machines, including the environment you reference (e3sm-unified) xarray is version 0.9.6. If I do conda update -c conda-forge xarray I still get xarray 0.9.6. Any suggestions on how to update to v0.10?

[xylar]

I can’t explain it, then. I must be getting the future warnings in 0.9.6. Could it be that you’re simply not looking in the right log files? I’ll double check tomorrow but I’m sure I saw the warning again today on my laptop.

…

On Mon, Dec 4, 2017 at 10:44 PM Luke Van Roekel ***@***.***> wrote: @xylar <https://github.com/xylar> I am unable to reproduce the future warning. It looks like this is because I can't get xarray v0.10. On all machines, including the environment you reference (e3sm-unified) xarray is version 0.9.6. If I do conda update -c conda-forge xarray I still get xarray 0.9.6. Any suggestions on how to update to v0.10? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#281 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AD_EeGS84eoTMG5jM_WfpRsOQBX9yCeLks5s9Ge_gaJpZM4Qu6qg> .

[vanroekel]

I did grep -i future *.log and grep -i warn *.log in the logs directory, future turned up no hits and warn gave me

climatologyMapSeaIceThickNH.log:/global/project/projectdirs/acme/software/anaconda_envs/edison_e3sm_unified_1.1.2_py2_nox/lib/python2.7/site-packages/numpy/ma/core.py:2344: RuntimeWarning: invalid value encountered in less_equal
climatologyMapSeaIceThickSH.log:/global/project/projectdirs/acme/software/anaconda_envs/edison_e3sm_unified_1.1.2_py2_nox/lib/python2.7/site-packages/numpy/ma/core.py:2344: RuntimeWarning: invalid value encountered in less_equal
timeSeriesSST.log:/global/project/projectdirs/acme/software/anaconda_envs/edison_e3sm_unified_1.1.2_py2_nox/lib/python2.7/site-packages/matplotlib/axes/_axes.py:545: UserWarning: No labelled objects found. Use label='...' kwarg on individual plots.
timeSeriesSST.log:  warnings.warn("No labelled objects found. "
timeSeriesSeaIceAreaVol.log:/global/project/projectdirs/acme/software/anaconda_envs/edison_e3sm_unified_1.1.2_py2_nox/lib/python2.7/site-packages/matplotlib/axes/_axes.py:545: UserWarning: No labelled objects found. Use label='...' kwarg on individual plots.
timeSeriesSeaIceAreaVol.log:  warnings.warn("No labelled objects found. "

but these seem unrelated. When you have a chance tomorrow, could you post the text of the warning that you see?

[xylar]

Sorry, @vanroekel. It does seem like I'm seeing this with xarray 0.10.0 on my personal laptop and not on Edison as I thought. I see the same warning more often in python 3 than in python 2 but I see it in both. I will set up a python 3 environment on Edison with xarray 0.10.0.

[xylar]

@vanroekel, if you could do the following:

export PATH=/global/project/projectdirs/acme/software/anaconda_envs/edison_mpas_analysis_py3/bin:$PATH

you should be able to run with the python 3 environment I'm setting up right now. I would suggest running with this config file:
https://github.com/MPAS-Dev/MPAS-Analysis/blob/develop/configs/edison/config.20170807.beta1.G_oQU240.edison
so that everything goes pretty quickly. If you could verify that you see the future warning with develop but not with this branch, that would be great.

[xylar]

@vanroekel, I also had to rebase and make a small fix to this PR to make it work with python 3, so please update your local branch.

[xylar]

Testing

I can confirm that I see the warnings with develop when I use a python 3 environment with xarray 0.10.0 but I do not see the warnings with the same environment and this branch.

[vanroekel]

@xylar I just tried the python environment from your comment and get permission denied errors

PermissionError: [Errno 13] Permission denied: '/global/project/projectdirs/acme/software/anaconda_envs/edison_mpas_analysis_py3/lib/python3.6/site-packages/fastcache-1.0.2-py3.6.egg-info'

[xylar]

@vanroekel, sorry about that. Could you try again?

[vanroekel]

@xylar I can now confirm the FutureWarning with develop. Using your branch the Warning is gone. Thanks for this fix @xylar

[xylar]

Great, sorry this was so much more trouble to test than it should have been. But it was helpful in the end since I fixed a python 3 problem in the process and set up some environments for testing that should be helpful in the future.