How to pass arguments to an Op without creating a dependency

[AfonsoTellesD4D]

Im trying to call an Op from a job, but the value inside a particular variable that can be there sometimes and not in others. I've tried with Optionals, default_values and is_required. But none of them seem to work, because I always have to send the variable and that creates a dependency.

This is a representation of my code:

@job(...)
def job1():
   a, b = op1() #Variable b serves as trigger for wether or not to execute op2
   c = op2(b)
   op3(a,c)

@op( out={"a": Out(Dataframe), "b": Out(Optional[DataFrame], is_required=False)
def op1(context: OpExecutionContext):
   ...
   yield Output(someVar1, output_name="a")
   #Im not goint to return b (its optional)

@op( ins={"start": In(Nothing)}, out={"c": Out(DataFrame, is_required= False)})
def op2(context: OpExecutionContext):
   return c

@op( ins={"a": In(SparkDF), "c": In(SparkDF, is_required=False)})
def op3(context: OpExecutionContext, a: SparkDF, c: SparkDF):
   ...

In this job, Im calling op1() and if the output b is return it triggers the op2. What Im trying to do is start op3() even if op2 is skipped.

I've tried using config_schema in the decorator but SparkDF is a custom class and not serialized by default.

Any suggestions?