Channel Multiplier

A = Channel('A') * <multiplier>

>>> A= Channel('A') * 3
>>> A[0].name
'A'
>>> A[1].name
'A1'
>>> A[2].name
'A2'

Process Multiplier

Parallel(
  P(<parameters>) * <multiplier>
)

>>> @process
... def P():
...     pass
... 
>>> P()
<Process(Thread-1, initial)>
>>> P() * 2
[<Process(Thread-2, initial)>, <Process(Thread-3, initial)>]

>>> @process
... def P():
...   print 'Hello', current_process_id()
... 
>>> Parallel(P() * 5)
Hello 0.2122764007741289306638.27
 Hello 0.3402057569491289306638.27
Hello 0.630868661511289306638.27
Hello 0.2323905151721289306638.27
 Hello 0.02407685114261289306638.27

When multiplying processes the argumentlist is traversed for channel ends and new channel ends are instantiated for new processes, to avoid having two processes referencing the same channel end.

@process
def P(cin):
  print current_process_id() + ' ' + str(cin())

chan = Channel('X')
Spawn(
  P(chan.reader(), ) * 10
)

cout = chan.writer()
for i in range(10):
  cout(i)

The above code produces the following output:

0.1083452839941289306925.94 1
0.9701965731871289306925.94 2
0.09972212802521289306925.94 0
0.08811378840181289306925.94 3
 0.9911784182721289306925.94 6
0.02519160333881289306925.94 5
0.7082968559831289306925.94 7
0.1502696612091289306925.94 8
 0.2164145215991289306925.94 4
0.9314496638651289306925.94 9

Reader (+) and Writer (-)

>>> A = Channel()
>>> +A
<ChannelEndRead ...>
>>> -A
<ChannelEndWrite ...>

replacing

>>> A = Channel()
>>> A.reader()
<ChannelEndRead ...>
>>> A.writer()
<ChannelEndWrite ...>

In use:

chan = Channel()
Parallel(
  P1(+chan)
  P2(-chan)
)