Building output datasets

[ScottWales]

I'm taking a look at adding the CDO operations using Vistrail's automatic module generation at https://github.com/ScottWales/cwsl-mas/blob/cdo/cwsl/vt_modules/vt_cdo.py.

I've gotten up to implementing the compute() function and found myself a bit stuck - I don't understand how to create a dataset for the output files. The ProcessUnit class looks related, but it seems to support only a single input dataset for a command.

Do you have any suggestions on how I could implement compute()?