编写用于 AWS Lambda 的异步代码的正确方法是什么？

Question

I wrote the following code:

import asyncio

loop = asyncio.get_event_loop()

async def get_urls(event):

    return {'msg':'Hello World'}

def lambda_handler(event,context):

    return loop.run_until_complete(get_urls(event))

I was trying to accomplish the following but faster.

def lambda_handler(event, context):
    # TODO implement
    return {'msg':'Hello World'}

What was the correct way to write this in an AWS Lambda environment?

Answer 1

Works for me... You need to choose Runtime "Python 3.6" or "Python 3.7".

import asyncio

loop = asyncio.get_event_loop()

async def get_urls(event):
    return {'msg':'Hello World'}

def lambda_handler(event, context):
    return loop.run_until_complete(get_urls(event))

Answer 2

Asynchronous execution does many things at the same time. You're only doing one thing. You can't do one thing faster than the time it takes to do one thing. Asynchronous execution allows you to perform independent tasks that would normally execute one after another(synchronously) at the same time and then return the result of all the tasks. Inherently, you have to be performing more than one operation.

Answer 3

For Python 3.7+ you can use asyncio.run() to execute your coroutines:

import asyncio

# The AWS Lambda handler
def handler(event, context):
    asyncio.run(main())

async def main():
    # Here you can await any awaitable
    await asyncio.sleep(1)
    await asyncio.gather([coroutine_1, coroutine_2])

Here's a full example of how to develop, test and deploy an asynchronous Python function using asyncio, aiohttp and aiobotocore on AWS Lambda: https://github.com/geeogi/async-python-lambda-template

Answer 4

Need to install aioboto3 locally as a part of deployment zip package `

import asyncio
import aioboto3



async def list_s3_objs (bucket):
    async with aioboto3.client("s3") as s3_client:
        objs = await s3_client.list_objects(Bucket=bucket)
    return objs['Contents']

def lambda_handler(event, context):
    loop = asyncio.get_event_loop()
    loop.run_until_complete(list_s3_objs('bucket_name'))

`