基于python生成器封装的协程类

时间:2022-11-17 08:31:34

自从python2.2提供了yield关键字之后,python的生成器的很大一部分用途就是可以用来构建协同程序,能够将函数挂起返回中间值并能从上次离开的地方继续执行。python2.5的时候,这种生成器更加接近完全的协程,因为提供了将值和异常传递回到一个继续执行的函数中,当等待生成器的时候,生成器能返回控制。

python提供的生成器设施:

  • yield:能够将自己挂起,并提供一个返回值给等待方
  • send:唤起一个被挂起的生成器,并能够传递一个参数,可以在生成器中抛出异常
  • next:本质上相当于send(None),对每个生成器的第一次调用必须不能传递参数
  • close:主动退出一个生成器

python封装

虽然python3提供了asyncio这样的异步IO库,而且也有greenlet等其他协程库,但目前的需求并不是实际的网络IO并发操作,而是需要模拟状态机的运行,因此使用协程可以很方便的模拟,并加入认为的控制,下面是封装的一个python类。

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
class Coroutine(object):
 
  """ Base class of the general coroutine object """
 
  STATE_RUNNING = 0
  STATE_WAITING = 1
  STATE_CLOSING = 2
 
  def __init__(self):
    self.state = Coroutine.STATE_WAITING
    self.started = False
    self.args = None
    self.routine = self._co()
 
  def _co(self):
    self.ret = None
    while True:
      self.args = yield self.ret
      if not self.started:
        self.started = True
        continue
      else:
        self.state = Coroutine.STATE_RUNNING
        self.ret = self.run(self.args)
      if self.state == Coroutine.STATE_CLOSING:
        break
      self.state = Coroutine.STATE_WAITING
 
  def start(self):
    """ Start the generator """
    if self.routine is None:
      raise RuntimeError('NO task to start running!')
    self.started = True
    self.routine.next()
 
  def finish(self):
    """ Finish the execution of this routine """
    self.state = Coroutine.STATE_CLOSING
    self.routine.close()
 
  def run(self, args):
    """ The runing method to be executed every once time"""
    raise NotImplementedError
 
  def execute(self, arg_obj):
    """ Awake this routine to execute once time """
    return self.routine.send(arg_obj)

基于上述封装,下面实现了一个协同的生产者消费者示例:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
class ProducerCoroutine(Coroutine):
 
  """ The Producer concrete coroutine """
 
  def __init__(self, cnsmr):
    if not isinstance(cnsmr, Coroutine):
      raise RuntimeError('Consumer is not a Coroutine object')
    self.consumer = cnsmr
    self.consumer.start()
    super(ProducerCoroutine, self).__init__()
 
  def run(self, args):
    print 'produce ', args
    ret = self.consumer.execute(args)
    print 'consumer return:', ret
 
  def __call__(self, args):
    """ Custom method for the specific logic """
    self.start()
    while len(args) > 0:
      p = args.pop()
      self.execute(p)
    self.finish()
 
 
class ConsumerCoroutine(Coroutine):
 
  """ The Consumer concrete coroutine """
 
  def __init__(self):
    super(ConsumerCoroutine, self).__init__()
 
  def run(self, args):
    print 'consumer get args: ', args
    return 'hahaha' + repr(args)

运行结果如下:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
produce 4
consumer get args: 4
consumer return: hahaha4
produce 3
consumer get args: 3
consumer return: hahaha3
produce 2
consumer get args: 2
consumer return: hahaha2
produce 1
consumer get args: 1
consumer return: hahaha1
produce 0
consumer get args: 0
consumer return: hahaha0

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持服务器之家。

原文链接:https://blog.csdn.net/u010487568/article/details/62042709