Skip to main content

First-Win Parallel Steps

Use the first completed concurrent Step and cancel the remaining siblings.

Definition graph

FirstWinParallelStepsFlow

Valid

python · examples/python/dex_examples/patterns/parallel/first_win_parallel_steps_flow.py

InitStep starts multiple DoWorkStep executions. Each worker sleeps for a random short duration, creating a real race. The first execution to return a decision selects its sibling DoWorkStep executions for cancellation. The Flow keeps the first result and stops work that can no longer affect the outcome.

Core implementation​

The cancellation selector targets sibling executions of the same Step type, not unrelated branches.

class DoWorkStep(Step[int]):
    async def execute(  # type: ignore[override]
        self, context: AsyncContext, input: int
    ) -> StepDecision:
        await asyncio.sleep(random.uniform(0.05, 0.5))
        return graceful_complete(input).with_canceling_sibling_steps(DoWorkStep)


class InitStep(Step[int]):
    def execute(self, context: Context, input: int) -> StepDecision:
        return go_to_many(
            *(StepMovement.of(DoWorkStep, index) for index in range(input))
        )

Example: examples/python/dex_examples/patterns/parallel/first_win_parallel_steps_flow.py