Add condition strain node to make workflow

[cdcapano]

There should be a condition strain node in the workflow. The condition strain will do all of the data loading, then write out a (hdf?) file of the Fourier domain data. All subsequent nodes will then simply load the FD data, instead of trying to repeat data conditioning. This will ensure that the ROQ weights, Bayeswave (when it's added) and gwin are all using exactly the same data.

[ahnitz]

@cdcapano This is a good idea. I think it almost does what is needed, but probably we need to add a flag to the fourier domain output. I guess some logic will need to be added to the workflow then to take over keeping track of segment lengths.