Async usage

Most Python API operations work synchronously and offer an asynchronous form through .aio, such as client.run() and client.run.aio(). Some methods are async only, such as run.calls() and call.children() which should be called with await. Methods that stream results use async for, such as volume.list() and client.map.aio(). Use the async interface if your application already uses async and await. While your code waits for OuterProduct to submit or finish a Call, Python can continue handling other work such as serving other web requests or waiting for several Calls at the same time.

import asyncio

import outerproduct as op


async def main() -> None:
    run: op.Run[int] | op.Call[int] = await client.run.aio(
        fn=sum_batch,
        args=([1, 2, 3],),
        image=image,
    )
    result: int = await run.wait.aio()
    print(result)  # 6


if __name__ == "__main__":
    asyncio.run(main())

await client.run.aio() returns a handle after submission. await run.wait.aio() waits for the result. See Runs, calls, and attempts for run.calls() and call.children(), and Volumes for volume.list().

Getting started

Execution

Governance