Async Logic Callables

Pipeline logic callables can also be coroutines. The best part about coroutines is that they can be run in parallel. You parallel execution in a pipeline easily with a tuple, i.e., a parallel block.

async def async_step1(...):
    ...

async def async_step2_1(...):
    ...

async def async_step2_2(...):
    ...

async def async_step3(...):
    ...

from pipeline_views import BasePipelineView
from .services import async_step1, async_step2_1, async_step2_2, async_step3


class SomeView(BasePipelineView):

    pipelines = {
        "GET": [
            async_step1,
            (
                async_step2_1,
                async_step2_2,
            ),
            async_step3,
        ],
    }

Now async_step2_1 and async_step2_2 will be executed in parallel, and will both receive all the arguments from async_step1. async_step2_1 and async_step2_2 should return dictionaries like standard pipeline logic methods. async_step3 will then receive a union of the results of these coroutines. If the return dictionaries of the two coroutines have any common keys, the value from the coroutine defined last in the parallel block will be used.

Return values from the callable before the parallel excecution can be passed to the next step after the parallel execution by setting ... (ellipses) in the tuple. If there are any common keys in this case, the values from the logic block will be prioritized.

from pipeline_views import BasePipelineView
from .services import async_step1, async_step2_1, async_step2_2, async_step3


class SomeView(BasePipelineView):

    pipelines = {
        "GET": [
            async_step1,
            (
                async_step2_1,
                async_step2_2,
                ...
            ),
            async_step3,
        ],
    }