Simpy同步通信信道

Question

我是个刚开始使用Python和Simpy的新手。我希望在两个进程之间有一个同步通信通道。例如，我希望：

channel = ...
def writer(env):
    for i in range(2):
        yield env.timeout(0.75)
        yield channel.put(i)
        print("produced {} at time {}".format(i, env.now)) 

def reader(env):
    while (True):
        yield env.timeout(1.2)
        i = yield channel.get()
        print("consumed {} at time {}".format(i, env.now))

env = simpy.Environment()
env.process(writer(env))
env.process(reader(env))
env.run()

其结果应当是：

produced 0 at time 1.2
consumed 0 at time 1.2
produced 1 at time 2.4
consumed 1 at time 2.4

我应该做什么/使用什么来定义频道？

如果我使用的是Store，那么我就会得到(与上面略有不同)：

import simpy
env = simpy.Environment()
channel = simpy.Store(env)

def writer():
    for i in range(2):
        yield env.timeout(0.75)
        yield channel.put(i)
        print("produced {} at time {}".format(i, env.now))   

def reader():
    while (True):
        yield env.timeout(1.2)
        i = yield channel.get()
        print("consumed {} at time {}".format(i, env.now))

env.process(writer())
env.process(reader())
env.run()

产出如下：

produced 0 at time 0.75
consumed 0 at time 1.2
produced 1 at time 1.5
consumed 1 at time 2.4

但我应该像上面提到的那样。作者应该等到读者准备好阅读为止。

Answer 1

您想要的东西不可能直接使用内置资源。解决办法可能有以下几点：

import collections

import simpy


Message = collections.namedtuple('Message', 'received, value')


def writer(env, channel):
    for i in range(2):
        yield env.timeout(0.75)
        msg = Message(env.event(), i)
        yield channel.put(msg)
        yield msg.received
        print("produced {} at time {}".format(i, env.now))


def reader(env, channel):
    while (True):
        yield env.timeout(1.2)
        msg = yield channel.get()
        msg.received.succeed()
        print("consumed {} at time {}".format(msg.value, env.now))


env = simpy.Environment()
channel = simpy.Store(env, capacity=1)
env.process(writer(env, channel))
env.process(reader(env, channel))
env.run()

输出：

consumed 0 at time 1.2
produced 0 at time 1.2
consumed 1 at time 2.4
produced 1 at time 2.4

如果在print()之前执行yield msg.received，您将得到：

produced 0 at time 0.75
consumed 0 at time 1.2
produced 1 at time 1.95
consumed 1 at time 2.4

另一种选择是编写自己的资源类型。