Make an independent package for Global Time Series

[forsyth2]

Currently, Global Time Series is fully integrated into zppy.

zppy is a workflow toolchain / orchestrator of all the various post-processing packages used by E3SM. It coordinates all these external packages -- except for Global Time Series, which is built into it. This means the zppy package effectively functions as both a post-processing coordinator AND a global time series generator. These purposes should be separated for better maintainability.

Scoping of effort:

• Create a new repository global_time_series under E3SM-Project
• Port the following to the new repo:
  • The package-specific code from https://github.com/E3SM-Project/zppy/blob/main/zppy/templates/global_time_series.bash
  • https://github.com/E3SM-Project/zppy/blob/main/zppy/templates/coupled_global.py
  • https://github.com/E3SM-Project/zppy/blob/main/zppy/templates/ocean_month.py
  • https://github.com/E3SM-Project/zppy/blob/main/zppy/templates/readTS.py
• Update the following as needed:
  • https://github.com/E3SM-Project/zppy/blob/main/zppy/templates/default.ini
  • https://github.com/E3SM-Project/zppy/blob/main/zppy/global_time_series.py
  • Files used for testing (while the Global Time Series repository will have its own tests, the package will still be tested indirectly via zppy as NCO, E3SM Diags, MPAS-Analysis, ILAMB also are.)
  • Docs mentioning Global Time Series
• Get Global Time Series working as an independent package
  • Would likely still be using Jinja templates.
  • May want to add configuration files like E3SM Diags has. In particular, Refactor coupled_global.py #389 still requires users to add plots to modify the code -- can we make it so users can add new plots simply by editing a configuration file?
• Add tests to the new repo
  • Currently, Global Time Series is tested by having zppy run ts and mpas_analysis beforehand. Would we need zppy to be in the testing environment to run these tools? That seems like a circular dependency.
• Copy relevant discussion posts to the new repo (or at least make a post linking them): When should I use `global_time_series` or the E3SM Diags Area Mean Time Series? #312, E3Sm Diags global_time_series issue #391

[xylar]

Please choose a less generic name than global_time_series. We need something that can go on conda-forge and it can be clear what it computes global time series for. Maybe give it a non-descriptive but short and memorable name like zppy.

[forsyth2]

Resolved by #642 and E3SM-Project/zppy-interfaces#1