Workflow: Water Availability System

[rburghol]

• Basic objectives (may be workflow modules, TBD):
  • Sensitivity scenarios (varying levels of demand, meteorology, or permit operational run_mode)
  • Statistical model building for each reach / reservoir / facility
  • Generate Lookup Tables / functions / json configurations for use in R
    • Hey @COBrogan : I know that the json exports of statistical models was a little hard to manage in the met data project, however, passing statistical models via JSON could be a very powerful and secure way of implementing the framework that we have been discussing. Just wanted to put that into your brain to get your impressions of upsides/downsides/ROI

[COBrogan]

I know that the json exports of statistical models was a little hard to manage in the met data project

Didn't we end up "fixing" the issue? If I recall, the R json packages failed to handle them completely by themselves, but the only work we had to do was to make sure the output from the package was wrapped in brackets. I'm fairly confident we can rely on this working, but we need to be aware that one day this "fix" could break the export if the JSON R packages are ever updated to handle these kinds of S objects.

[rburghol]

@COBrogan I recall that "fix" now that you mention it. I think if we wrap any of these parser/unparsers in an object class we should be able to pretty seamlessly manage the case.

Also, if I recall, part of the issue can be we were trying direct instantiation of an object from json, which while being super convenient, is something that we handle automatically in the Rom* classes with the from_list() function which takes in a list or a json structure and populates its properties from there.

Requires a little more work but not crazy amounts and could be a tad more secure and gives us a place to handle weird stuff like geometries.