Openjij

* improve performance of BQM and BPM
* (for BQM) initialization via numpy.ndarray

* Add source code to pypi

* modify BinaryPolynomialModel

A polynomial model is an extension of the traditional quadratic model.

How to use
Python example

import openjij as oj
sampler = oj.SASampler()

Spin variables
response = sampler.sample_hubo(J={(0,1): -1, (0,1,2): -1})
response.states
[[1, 1, 1]]

Binary variables
response = sampler.sample_hubo(J={(0,1): -1, (0,1,2): 1}, var_type = oj.BINARY)
response.states
[[1, 1, 0]]

* Speed up (x10 ~ x20) of SA.
* OpenMP parallelization of SQA (currently available on Windows and Linux)
* Some minor changes

* Use [cimod](https://github.com/OpenJij/cimod) as a backend of `BinaryQuadraticModel`.
* Modify Graph interface
* Fix bugs