More labels

This adds more truth labels that might be useful for boosted studies. Specifically:

• a configurable delta R for the matching. This wasn't really needed when we just decorated the nearest jet (we could add the delta R to the truth particle and cut on it later), but it's useful if we have quantities that aggregate multiple truth particles into one jet. By default it's infinity, we'll need something like 1.0 or maybe 0.8 for the boosted jet tagging.
• parentXNMatchedChildren: the number of children from a parent particle of type X matched to the jet.
• parentXIndex: rather than saving the parent barcode, I just give them an index starting with zero. So in HH events there should be 0 and 1. If there are multiple parents matched we take the closer one (as with other properties like PDGID and the truth particle link). For SH events it should always be 0, since we won't have multiple S or H in the event.
• parentXParentsMask: This is a bitmask saying which parents are matched to this particle. For SH it should always be 1 if there is a matched parent, i.e. 0b01 or 0b10, but for HH we could have 3, i.e. 0b11 if there are decay products from both higgs in one jet.

There was also some reorganization in the code, but I did all that before making the changes above and did a diff of the output to ensure that it changed nothing.

Some extra things that I might or might not do before marking this as ready for merging (depending on how keen other are to use the existing code and how long it takes me to do them):

• Add sensible deltaR for all instances. Right now there is no limit: all the truth particles go to the nearest jet.
• Add a boosted jet labeling instance. Probably only for S and H, unless people think the top labeling is useful there too.
• Add a version that works with PHYSLITE. I suspect I might not be able to follow the truth chain as far there, or maybe there's just a bug in PHYSLITE.