compile script reorganization

[btobers]

Closes #113.

Dictionary structures used to initialize and aggregate regional results.

[yelizy]

Neat! I haven't tried running the code but it looked okay to me. I am wondering if there is a way to reduce the use of elif statements for each variables to assign their attributes.

[btobers]

Neat! I haven't tried running the code but it looked okay to me. I am wondering if there is a way to reduce the use of elif statements for each variables to assign their attributes.

Yes, maybe a dictionary structure with the per-variable metadata that can be indexed when building the datasets. That'd certainly help clean up this script. I'm sure it can be further improved in the future.

[btobers]

As of now, there's no way to autmoatically test the postproc_compile_simulations script, since the test notebooks which are run automatically only run single glaciers and don't compile any regional results. Per #50, we should add a test for this script. In the meantime, I have run compilation locally to verify this is working as expected.

[btobers]

Decided to create a compile test script (tests/test_04_postproc.py) which is now included in the Git CLI test workflow. This relatively simple test script post-processes the monthly mass (calling postproc/postproc_monthly_mass.py and then compiles region 15 results (calling postproc/postproc_compile_simulations.py) using the results of PyGEM-notebooks/simple_test.ipynb for RGI60-15.03733. So we now hit two more postprocessing scripts in our CLI workflow testing. I think this can now be approved by @PyGEM-Community/dev and we can continue to build on it and improve in the future.