CompEcon Sub-Package (API and Python Supporting Code)

Add a compecon subpackage which provides an interface to the same functionality provided by the matlab CompEcon toolkit in python. The aim of this sub-package is to provide a mixture of numba accelerated functions / possibly cython functions etc. with wrappers that provide the same interfaces as what is provided in CompEcon. This will assist with migrating some users who are familiar with matlab to python and julia.

@cc7768 has also made a good point in that this sub project may provide some good opportunities when putting together a sprint session for a conference.

QuantEcon.jl already provides a subset of CompEcon.

This is related to Issue #21 with a discussion of what is available in python from the quantecon toolkit.

There is potential that down the track, Julia code may be compiled to a library that can be called by other languages (like python). Therefore it is possible that down the track we may want to change the approach taken to supporting the CompEcon toolkit.

Start by reviewing Issue #21 and the subset of CompEcon already provided by QuantEcon.jl. Then determine the intended Python subpackage scope and how closely its interfaces should match the MATLAB CompEcon toolkit, including the role of accelerated Python code and possible future Julia integration. Done would mean an agreed implementation scope and compatible API.