Provide support for parallelizing Python calls

[oubiwann]

This may only require:

• partitioning of input data sets
• gathering output data results

In contrast to the goal of issue #37, this ticket aims to support parallelization of a function across all Python servers.

To clarify:

• Default behaviour in py is to execute a Python function on a server, selected using the configured scheduler
• Issue #37 introduces the functionality of executing a single function on all Python servers, skipping the scheduler.
• This issue also wants to skip the scheduler, but where it differs from issue #37 is to partition the input data set so that each Python server can tackle a different part of the problem.

Maybe these functions could go in a py-para.lfe module ...

Sources for reference:

• http://ipython.org/ipython-doc/stable/parallel/parallel_multiengine.html#quick-and-easy-parallelism
• https://github.com/oubiwann/plists/blob/master/src/plists.erl
• https://github.com/cadar/pmap-lfe/blob/master/pmap.lfe