Add `postprocessing/compare_inputs.py` to compare run inputs

[clairehalloran]

Summary

I plan to add a postprocessing/compare_inputs.py script to compare files in the inputs_case folder across runs, similar to postprocessing/compare_cases.py. In contrast to the existing postprocessing/input_diff_plot.py, I envision this feature being used to explain the differences between ReEDS scenarios run on the same branch.

I've developed an initial version of this script to plot different load profiles across cases.

Proposed Changes

The primary change will be adding the postprocessing/compare_inputs.py script, which will not be part of the main ReEDS workflow.

Possible additional changes include refactoring single-case plotting functions in postprocessing/input_plots.py to be extendable to multi-case plotting, and then moving these functions to reeds/reedsplots.py or a separate, similar script.

Additional context

I would like feedback on:

• What inputs would be most helpful to include in the initial script?
• In what format should the comparisons be saved? (PowerPoint, figures saved in outputs/comparison folder)

[patrickbrown4]

Sounds great, thanks! A few quick thoughts:

• For the file format, you could copy this approach from input_plots.py to let the user specify upfront while keeping the downstream use of saveit(savename) unchanged
  • No strong preference from me on what should be the default, though if they're saved as individual files, you'd probably want to put them in a subfolder of outputs/comparison labeled similarly to the compare_cases.py output file, with the case names in the folder name
  • There might be some functions in input_plots.py that could be reused
• compare_cases.py is kind of a mess organizationally (my fault). For new plotting scripts, I'd suggest using an organization more like input_plots.py or diagnostic_plots.py (or parts of single_case_plots.py, though it also needs to be cleaned up), where self-contained plotting functions are defined above (or in an imported module) and then called inside an if __name__ == '__main__' block, rather than building the plots as one long procedure