Add histogram plot type and ocean histogram task

[cbegeman]

Adds the ability to create histogram plots from 2-d ocean variables.

[cbegeman]

I've tested this out on chrysalis.

Using this cfg section:

[oceanHistogram]
obsList = ['AVISO']
variableList = ['ssh', 'pressureAdjustedSSH']
weightList = ['areaCell','areaCell']
regionGroups = ['Ocean Basins']
regionNames = ['Atlantic_Basin','Pacific_Basin']
seasons =  ['ANN','JFM']
bins = 40

You can create the figures shown here:
https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.cbegeman/analysis/A_WCYCL1850.ne4_oQU480.anvil/clim_3-5_ts_1-5_histRegionWeight/ocean/index.html

[xylar]

@cbegeman, this is really great work, and it will be such a nice capability to have!

I tested this out and ran across a few things that could be done to improve the task a bit. The main one is that you need to make sure the plot subtasks only run after the weight subtasks.

But it also seems like it would be a good idea to make sure the subtask names and file names include the region group, since some regions are in multiple region groups so there could be conflicts.

I'll re-test as soon as these changes are in place.

[cbegeman]

@xylar Thanks for your review! I made the changes you suggested. After rebasing, there were still problematic typos in sea ice climatology addressed by this commit dd29f0c. Feel free to pull that out of this branch since it is not related to the histogram task.

[xylar]

@cbegeman, oh, great, thanks for fixing the sea-ice typo! I think that's fine to fix here. I can't believe we missed that!

[xylar]

I'll re-test now and then run the full test suite if everything goes well.

[xylar]

Great! Just 2 small things I noticed when I ran my test, now that the web page came up. Almost done...

[xylar]

I thought we had it but that's why we do testing...

Two more little things. Please double check my suggestions, I'm not 100% sure they're right, well formatted and typo free.

[xylar]

One more request would be to add some documentation, both in the API:
https://github.com/MPAS-Dev/MPAS-Analysis/blob/develop/docs/developers_guide/api.rst
and in the user's guide, similar to:
https://github.com/MPAS-Dev/MPAS-Analysis/blob/develop/docs/users_guide/tasks/regionalTSDiagrams.rst

[cbegeman]

@xylar I added the task to the docs and fixed up a couple config-related issues in the code. This is ready for your review.

[xylar]

Running one last set of tests, then I think this will be ready to merge!

[xylar]

I'm seeing:

/gpfs/fs1/home/ac.xylar/mpas-work/MPAS-Analysis/add_ocean_hisogram_task/docs/users_guide/tasks/oceanHistogram.rst:113: ERROR: Unexpected indentation.

But this line looks perfectly fine to me. I'm going to ignore it but figured I'd note this in case you can spot the problem for a later clean-up PR.

[xylar]

I thought we were there but I'm seeing an error when testing this without ncclimo:

Traceback (most recent call last):
  File "/gpfs/fs1/home/ac.xylar/chrysalis/mambaforge/envs/test_mpas_analysis_py3.9/lib/python3.9/site-packages/mpas_analysis/shared/analysis_task.py", line 322, in run
    self.run_task()
  File "/gpfs/fs1/home/ac.xylar/chrysalis/mambaforge/envs/test_mpas_analysis_py3.9/lib/python3.9/site-packages/mpas_analysis/ocean/histogram.py", line 539, in run_task
    xLabel = f"{ds[var_name].attrs['long_name']} " \
KeyError: 'long_name'

I don't think this error has anything directly to do with this PR. I think the long_name attribute isn't getting retained when we compute climatologies with python code. I'll try to fix this soon. But since this isn't a mode we actually run with very often, I'm going to let this go in with this note.

[xylar]

Results of the test suite on Chrysalis are here:
https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.xylar/analysis_testing/chrysalis/add_ocean_hisogram_task/
All runs except no_ncclimo ran great, and the new documentation looks good.

[cbegeman]

Thanks, @xylar!