Is it possible to create parent-child streams that share data for a given run? Suppose I wanted to query an endpoint for some data but wanted to split that data into two different outputs, in a non-trivial way that involves aggregation, is this what parent-child streams are for?