Does anybody have an example of running a flow from a serialized DAG representation? I want to be able to create a flow (in json) from a predefined set of tasks. I also want to support passing data from one task to another.

hi @Evan Curtin - you may already know this, but just as a disclaimer, prefect (post prefect 1.x anyways) discovers the graph, so in most cases you shouldn't need a static dag representation, as python control flow can typically represent the same work but there's different ways to do this, here's one. but IME these rely on at least • defining some task registry (mapping of tasks you can run) • making up some DSL to define that downstream tasks should get inputs from an upstream result that you don't know yet

ty for the link I'll take alook. I don't want the end users to be writing code so I'd prefer a DSL