Python 脚本一次向 API 发送太多请求（python-binance）

Question

我使用的脚本向 API 发送请求的速度似乎太快了。如果请求排队超过 100 次，API 会发送此错误并停止：

{'e': 'error', 'm': 'Queue overflow. Message not filled'}  

发送和接收的脚本部分如下：

while True:
    await socket.__aenter__()
    msg = await socket.recv()
    frame = createframe(msg)
    frame.to_sql(symbol, engine, if_exists="append", index=False)
    print(frame)

从打印帧的时间中可以看出，请求发送得非常快。

    Symbol                    Time     Price
0  BTCUSDT 2021-09-23 21:30:15.009  44663.17
    Symbol                    Time     Price
0  BTCUSDT 2021-09-23 21:30:15.061  44663.17
    Symbol                    Time     Price
0  BTCUSDT 2021-09-23 21:30:15.061  44663.17
    Symbol                    Time     Price
0  BTCUSDT 2021-09-23 21:30:15.133  44663.17
    Symbol                    Time     Price
0  BTCUSDT 2021-09-23 21:30:15.190  44663.17
    Symbol                    Time     Price
0  BTCUSDT 2021-09-23 21:30:15.261  44663.17
    Symbol                    Time     Price
0  BTCUSDT 2021-09-23 21:30:15.261  44663.17

在理想情况下，每秒应该有 1 个请求。但是，该脚本似乎向 API 发出请求。尝试了time.sleep()函数，但似乎无法稳定请求速度。

Answer 1

你试过睡觉吗？

import asyncio, time
...
last_request_time = 0.0
while True:
    # Calculate delay from now to the moment of previous request + 1 sec
    # btw, it's Ok to sleep negative delays as well, so we don't need an `if`
    delay = last_request_time + 1.0 - time.monotonic()
    await asyncio.sleep(delay)  # sleep
    last_request_time = time.monotonic()  # write timestamp at the moment of the new request
    await socket.__aenter__()  # That's bad, use `async with` instead 
    msg = await socket.recv()
    
    frame = createframe(msg)
    frame.to_sql(symbol, engine, if_exists="append", index=False)
    print(frame)  # <- That's bad, `print()` is blocking