Add Cirq MaxCut solver

[schweikart]

Replaces #19. I've decided to rebuild the cirq max-cut solver integration since we've refactored our solver setup in the last couple of weeks. With that being said, I'm happy that the integration with our new solver structure went on seamlessly :)

[schweikart]

It is worth mentioning that the quality of the results of the Cirq MaxCut QAOA implementation is far from perfect. I haven't found out why the algorithm performs so poorly and we should maybe consider investigating this before releasing this solver.

[Elscrux]

In general, this looks good. I'm pleased to see adding new solvers can be this simple now.
But for some reason, the result of our example returns everything in partition 2, which can definitely not be correct

graph [
    id 42
    node [
        id 1
        label "1"
    ]
    node [
        id 2
        label "2"
    ]
    node [
        id 3
        label "3"
    ]
    edge [
        source 1
        target 2
    ]
    edge [
        source 2
        target 3
    ]
    edge [
        source 3
        target 1
    ]
]