New lineshapes and spin factors

[phdargen]

Implemented a few more lineshapes and spin factors.

Some of the more complicated spin factors do not seem to work yet.
The expressions compile but somehow AmpGen freezes when computing them.
Are expressions like this:
Tensor term3 = term2(-alpha,-beta) * L(alpha,mu);
supposed to work with the current framework?
Those spin factors are commented for now and might be fixed in a future commit.

[tevans1260]

So the problem is that four-index tensors become very slow as currently implemented (in principal, it does work, its just impractical)
The only solution so far is to do a bit of manual algebra to simplify the vertex factors a bit such that you never actually use the spin-2 (or higher) projection operator; but this is not exactly an ideal solution .. Better solutions require lots of work though .. an appealing solution is to to try to change the tensor engine to actually do tensor algebra rather than brute forcing everything; its then fairly easy to impose simplifications like S^2 =S for the projection operators (imposing symmetries reduces the number of equations, which was the point of implementing these things, but it wasn't enough to fight the number of terms efficiently once you start having a lot of spins to track)