Dispatch items from a generator to two consumer, using greenlets to parallelize operations.

green_stream_dispatching.py

import time

from greenlet import greenlet


##------------------------------------------------------------
##
## We have a producer (a simple generator) and two
## consumers (functions accepting a generator).
##
## We want to feed items to the two consumers as soon
## as items are generated by the producer.
##
## We need a workflow like this:
##
##   - get an item from producer
##   - pass item to consumer1
##     - execution is returned to main
##   - pass item to consumer2
##     - execution is returned to main
##   - keep looping / consuming items
##
##------------------------------------------------------------


def producer():
    for i in xrange(5):
        yield "Item {0}".format(i)
        time.sleep(.5)  # We're doing things, here..


def consumer1(stream):
    for item in stream:
        print("CONSUMER-1: {0}".format(item))


def consumer2(stream):
    for item in stream:
        print("CONSUMER-2: {0}".format(item))


##------------------------------------------------------------
## Greenlets machinery
##------------------------------------------------------------

class GreenGenerator(object):
    """
    Iterable object, that keep asking parent greenlet
    for the next item.
    """

    def __iter__(self):
        return self

    def __next__(self):
        g_self = greenlet.getcurrent()
        next_item = g_self.parent.switch()
        return next_item

    next = __next__  # For python 2


def dispatch_stream(stream, *consumers):
    greenlets = [greenlet(c) for c in consumers]
    for g in greenlets:
        g.switch(GreenGenerator())
    for item in stream:
        for g in greenlets:
            g.switch(item)


dispatch_stream(producer(), consumer1, consumer2)