Integrating a 2-dimensional angular distribution

[jgalan]

This issue raised at PR #22. The question is what happens if we need to describe a more arbitrary angDist angular distribution with distributions for both angles, theta or phi.

In principle to motivate the development of this upgrade one should find a case where this is needed. But I let it here anyway.

A possibility would be to directly provide a TH2D type angular distribution. Or to provide a file for a TH1D Theta distribution, and another one, TH1D, for Phi angle distribution. Assuming both are independent, they could be randomised independently.

Perhaps that would need to come out with a bit more complex way to describe the angular distribution?

<angularDist type="TH1D" direction="(0,0,-1)"/>
       <thetaDist file="CosmicAngles.root" spctName="Theta2"/>
       <phiDist file="CosmicAngles.root" spctName="Phi2"/>
</angularDist>

If we use the previous way to define angularDist, i.e. if phiDist is not defined then it will assume phiDist is uniformly distributed. In the new implementation we could provide either thetaDist or phiDist, and if not they are just uniformly distributed.

We could have also the new angular distribution type TH2D.

<angularDist type="TH1D" direction="(0,0,-1)"  file="CosmicAngles.root" spctName="SolidAngle"/>

For more complex generators I believe a file of the style used by Decay0 output could be used.

Or perhaps a new class in the way described at the following post.

https://rest-forum.unizar.es/t/doubt-can-you-use-two-generators-in-one-simulation/327/12