Provide clear instructions to users how a BestPV should be determined.

Triggered by a question in the Upgrade HLT2 channel by @cprouve

We should document how the "BestPV" can/should be obtained in our software.

The original Question was how to adapt the following code

LHCb::Event::PV::bestPV( allPVs, bCand.endVertex()->position(), bCand.momentum() )

to a charged particle. The suggested solution would be to use:

LHCb::Event::PV::bestPV( allPVs, referencePoint(basic_particle), threeMomentum(basic_particle) )

Is this a valid suggestion for all kinds of basic_particle?

Aside from the above LHCb::Event::PV::bestPV, we also have Sel::getBestPV

It's not clear to me why we have both, and we currently don't provide any guidance when to use which.

The first performs a simple linear extrapolation to the common z coordinate and associates based on minimal rho distance, while the latter does a IPCHI2 based association.

cc @cprouve @gligorov