AllCombMap and AllCombMapReduce

Specializers designed to facilitated map, map-like, and map-reduce-like computations.
The computation is either a traditional map, or one which is intended to replace code that looks like

for x in ():
  for y in ():
    for z in ():
      do_something(x,y,z,...)

by replacing the loop with a version that returns a list with an item for each x,y,z.

Current Usage/API

AllCombMap(lists, func, *args).execute(nproc=n) where

• lists is a list of lists
• func is a side-effect free function that returns something
• args is the rest of arguments to pass to func
• nproc is optional argument that tells the specializer how many processes to use. If not specified, the specializer will try to detect the number of available cores.

The function signature for func must be such that it takes in a+b arguments, where a is the number of lists, and b is the number of other arguments passed to the specializer. The result will be a list that contains one item for each member in the cartesian product of lists, and that item is the result of applying func.

Example:

from allcombmap import *

def add(x,y):
  return x+y

result = AllCombMap([[1,2,3],[3,4,5]], add).execute()

result will be the list [4,5,6,5,6,7,6,7,8] in some (possibly sorted, possibly not) order.

AllCompMapReduce is not yet implemented.