Add support for large ensembles

[forsyth2]

zppy can currently be run on multiple simulations/members of an ensemble, by using different configuration files to specify different data paths.

It would be nice to add automation such that a single zppy configuration file could be used to post-process all members of an ensemble. Perhaps a summary pdf could be generated as well, displaying mean, median, etc.

[forsyth2]

One potential issue may be submitting/queueing the large number of jobs that would be generated, considering post-processing just one simulation/member launches many jobs.

[forsyth2]

From #186 (a duplicate of this issue): Add support for large ensembles. Add loop to run zppy on multiple experiments.

[wagmanbe]

I ran zppy on a 250-member ensemble using a code that loops over each case and modifies the .cfg file, then submits the runs if I have less than some threshold number of runs in the queue. It worked great!

[rljacob]

Is that the solution for all zppy users?

[wagmanbe]

Sorry, I didn't mean to "close" the issue. The method I used works perfectly well, but zppy could certainly come up with its own way.