Hi, I have the following code in a job:

...
val = foo(val)
val = bar(val)
val = baz(val)
...

(foo, bar and baz are ops) Now,

bar

is optional - there is a configuration that allows the user to skip

bar

and then the output of

foo

will go strait to

baz.

I can implement it naively with a simple condition in the beginning of

bar

(and I did), but - Can I visualize it better? represent it somehow in the DAG? maybe gray out the op during execution? Thanks!

Just move the condition to foo and output a Tuple[Optional[…],Optional[…])

optBazInput, optBarInput = foo(val)
bazInput = bar(optBarInput[required])
bazRes = baz(optBazInput[optional], bazInput[optional])

And definitely rename them. But just returning both possible outcomes, functional slitly like an Either Tuple and then consume that tuple baz. Of course requires that in baz to have the logic coalesce the 2 optional inputs to one. So its a bit more code but you get it visualized in the graph. Think of it as functional style code with hard/statically typed outputs than it usually maps to a valid DAG I think.

Thanks! I can actually make it a generic sub-graph, renaming foo to is_enabled() and baz to coalesce() or something like that