Need functionality to split two data packages based on an input mask

[cmutel]

In cases where we have some parameters screened out by GSA or local sensitivity, we want to use the static/"best guess" values for those parameters. Currently datapackages can only either have static (vector) or stochastic (array) values. The requested functionality here would be a new function or method that would take two input datapackages and a Numpy boolean mask, and generate a datapackage with two resource groups, one for the static values and one for the stochastic values.

For the time being, we assume that these are datapackages stored on disk, and not interfaces.