The idea is that API.get for the HAL case will produce
a data producer which will submit a list of jobs where we will be
fetching documents incrementally.