How could roocs cope with orchestrate requests for multiple dataset IDs?

[agstephens]

Example request:

• Take a collection of CMIP6 ensemble members for a given expt+model+variable:
  • maybe defined with a wildcard, e.g. foo.baa.foo2.baa2.*.foo3.baa3
• Subset both in space/time (operation 1: subset)
• Regrid both to the same grid (operation 2: regrid)
• Calculate the ensemble average of all of them (operation 3: ensemble_average)

Implications for "roocs" stack:

• rook needs to consolidate a list of files (as output from operation 1) into:
  • a sequence of FileMapper objects
• daops can receive and consolidate multiple FileMappers as inputs
• all interfaces can handle receiving multiple dataset IDs (somehow)
• interfaces need to be able to receive wildcards in dataset IDs and resolve them via search/lookup (of catalog?)