Question

我正在尝试运行一组代码，这些代码在UTC时间的5秒内完全启动，从偶数分钟开始。

例如，它将完全执行每个样本：

11:20:45
11:20:50
11:20:55
11:21:00
11:21:05
11:21:10

我希望无论代码块的执行时间如何都会发生这种情况，如果运行代码是即时的或3秒钟，我仍然希望以5秒的UTC时间间隔执行。

不完全确定怎么做，虽然我认为datetime.datetime.utcnow().timestamp() - (datetime.datetime.utcnow().timestamp() % 5.0) + 5让我下一个即将到来的开始时间？

Answer 1

您可以使用python的scheduler模块：

from datetime import datetime
import sched, time

s = sched.scheduler(time.time, time.sleep)

def execute_something(start_time):
    print("starting at: %f" % time.time())
    time.sleep(3)  # simulate a task taking 3 seconds
    print("Done at: %f" % time.time())
    # Schedule next iteration
    next_start_time = start_time + 5
    s.enterabs(next_start_time, 1, execute_something, argument=(next_start_time,))

next_start_time = round(time.time() + 5, -1)  # align to next to 10sec
s.enterabs(next_start_time, 1, execute_something, argument=(next_start_time,))
print("Starting scheduler at: %f" % time.time())
s.run()


# Starting scheduler at: 1522031714.523436

# starting at: 1522031720.005633
# Done at: 1522031723.008825

# starting at: 1522031725.002102
# Done at: 1522031728.005263

# starting at: 1522031730.002157
# Done at: 1522031733.005365

# starting at: 1522031735.002160
# Done at: 1522031738.005370

Answer 2

使用time.sleep等到所需的时间。请注意，这是近似值;特别是当系统处于高负载时，您的过程可能无法及时唤醒。你可以increase process priority增加你的机会。

为了避免阻塞等待线程，可以在单独的线程中运行任务，方法是为每个任务构建一个新线程或使用（更快）thread pool，如下所示：

import concurrent.futures
import time

def do_something():  # Replace this with your real code
    # This code outputs the time and then simulates work between 0 and 10 seconds
    import datetime
    import random
    print(datetime.datetime.utcnow())
    time.sleep(10 * random.random())

pool = concurrent.futures.ThreadPoolExecutor()
while True:
    now = time.time()
    time.sleep(5 - now % 5)
    pool.submit(do_something)

Python - 使用UTC时间每{5}块运行代码

2 个答案: