Critically examine existing preprocess/fit/project/postprocess stage breakdown

[bobkopp]

Existing modules generally break down into preprocess/fit/project/postprocess stages, though this is not religiously enforced (e.g., see the facts/* modules). In the FACTS manager, this can be flexibly specified in the pipeline file, and it is not clear that this distinction makes sense to maintain in general.

Most modules can be broken down into:

• Preparing input data (preprocess and fit stage)
• Computationally intensive execution (project stage)
• Localization (postprocess stage)

Though with the exception of the sterodynamics module, localization is done through the application of GRD fingerprints, which in the long run perhaps should be done by a separate module in a separate experiment step.

But, for example, in the fair/temperature module, there is no clear reason why emissions should be imported in the preprocess step and model calibration parameters in the fit step, nor is there a reason to have an empty postprocessing step. In this particular module, the breakdown can be critically examined as preparation for the development of the fair2 module.

[bobkopp]

The labels are hard coded in FACTS.py. It is unclear that they need to be.