在python里使用await关键字来等另外一个协程的实例

一个协程里可以启动另外一个协程，并等待它完成返回结果，采用await关键字，

例子如下：

import asyncio

async def outer():
 print('in outer')
 print('waiting for result1')
 result1 = await phase1()
 print('waiting for result2')
 result2 = await phase2(result1)
 return (result1, result2)

async def phase1():
 print('in phase1')
 return 'result1'

async def phase2(arg):
 print('in phase2')
 return 'result2 derived from {}'.format(arg)

event_loop = asyncio.get_event_loop()
try:
 return_value = event_loop.run_until_complete(outer())
 print('return value: {!r}'.format(return_value))
finally:
 event_loop.close()

输出结果如下：
in outer
waiting for result1
in phase1
waiting for result2
in phase2
return value: ('result1', 'result2 derived from result1')

await关键字添加了一个新的协程到循环里，而不需要明确地添加协程到这个事件循环里。

补充知识：python里使用Condition对象来唤醒指定数量的协程

在asyncio库里，定义Condition对象，它的行为与事件Event有点像，区别是事件是通知所有对象，Condition对象可以指定一定数量的协程被通知，它是通过函数notify()来实现的，如果参数里放2，就是通知两个协程，例子如下：

import asyncio

async def consumer(condition, n):
 with await condition:
   print('consumer {} is waiting'.format(n))
   await condition.wait()
   print('consumer {} triggered'.format(n))
 print('ending consumer {}'.format(n))

async def manipulate_condition(condition):
 print('starting manipulate_condition')

# pause to let consumers start
 await asyncio.sleep(0.1)

for i in range(1, 3):
   with await condition:
     print('notifying {} consumers'.format(i))
     condition.notify(n=i)
   await asyncio.sleep(0.1)

with await condition:
   print('notifying remaining consumers')
   condition.notify_all()

print('ending manipulate_condition')

async def main(loop):
 # Create a condition
 condition = asyncio.Condition()

# Set up tasks watching the condition
 consumers = [
   consumer(condition, i)
   for i in range(5)
 ]

# Schedule a task to manipulate the condition variable
 loop.create_task(manipulate_condition(condition))

# Wait for the consumers to be done
 await asyncio.wait(consumers)

event_loop = asyncio.get_event_loop()
try:
 result = event_loop.run_until_complete(main(event_loop))
finally:
 event_loop.close()

结果输出如下：

starting manipulate_condition
consumer 2 is waiting
consumer 3 is waiting
consumer 4 is waiting
consumer 1 is waiting
consumer 0 is waiting
notifying 1 consumers
consumer 2 triggered
ending consumer 2
notifying 2 consumers
consumer 3 triggered
ending consumer 3
consumer 4 triggered
ending consumer 4
notifying remaining consumers
ending manipulate_condition
consumer 1 triggered
ending consumer 1
consumer 0 triggered
ending consumer 0

来源：https://blog.csdn.net/caimouse/article/details/77867169