Create aggregator class to handle differences in exact commands for each backend

In order to fully automate the pipeline, the actual processing needs to work with different compute backends. These backends use very similar function names but have slight differences either in implementation or exactly which packages are used.

Generic classes can be used to contain the different implementations of the processing steps for each backend. The user will pass the backend type determined by the compute module, and the correct aggregator class will be chosen and returned.