快速 API 拨打 python

Question

i am building a package in python. Package will be probably used in Jupyter notebooks. I used synchonized code for some functionality, but it is slow so i am planning to use asyncio

The question is how to implement it with async functionality without changing the code in notebook. So user will be able to eg module.orders("Example Inc") Now user has to make initial setup eg loop = asyncio.get_event_loop() loop.run_until_complete(module.orders("Example Inc"))

I want to avoid this, so it will be easier to use Thanks Or is there some better way fo make fast GET api calls

want eg module.orders("Example Inc") avoid this loop = asyncio.get_event_loop() loop.run_until_complete(module.orders("Example Inc"))

Answer 1

The solusion was relatively easy. To obtain paralelized get requests with use of asyncio and aiohttp python packages and still have option to use it as 'synchronized code' and avoid async functionality to go up, to other functions.

async def getter(params):
    # get data from server 
    c = aiohttp.TCPConnector(limit=100)
    async with aiohttp.ClientSession(connector=c) as session:
        # send requests

def orders(name):
    # no need of async
    return asyncio.run(getter(params))


dataset = orders("company")

Code above is preudocode.

for usage with jupyter notebbok it is needed to have

import nest_asyncio
nest_asyncio.apply()

GET requests are faster, but probably there is another bottleneck on the server side