Unify ParticleData.xml between pythia and corsika

ParticleData.xml is identical in format in corsika and pythia. We originally took it from pythia 8235 and extended it with particles that apparently are not needed in pythia (for example nucleon resonances). Pythia has a mechanism to import the definition of particles at initialization. We should make it use the ParticleData.xml from corsika, then we could use pythia for every hadronic decay and do not have to revert to sibyll decay for weird resonances.

mentioned in issue #196 (closed)

mentioned in issue #263 (closed)

mentioned in merge request !311 (merged)

@riehn this was addressed, right? Please close this old Issue, or add new info.

changed title from Follow-up from "WIP: Resolve "remove internal decays from Sibyll interaction"" to Unify ParticleData.xml between pythia and corsika

changed the description

I think this issue still stands. Though the title is not very helpful.

It would also help with #255, that is the bookkeeping of decays, which is due to the fact that pythia decay can not be used for all decays (currently) so we need to keep track of which particles ought to decay through which process. Maybe we could start an 'overhaul decay' campaign at some point. I think @rprechelt also had some ideas, or misgivings about the current inmplementation.

Yes, I have the same feeling about this. Let's discuss this soon.

Yes, @riehn! I've been meaning to work on this.

Right now, you have to coordinate which Process handles the decay of each particle manually by enabling & disabling decays on each combination of Process instances. Not only is this annoying to do (and scales like N^3 for N different process that are capable of handling the decays), but it is super easy to introduce bugs since, if both processes try to decay the particle, weird things can happen (things might have changed since but at some point, this conflict was resolved by: either the process is chosen based upon its order in the process sequence, or based upon which decay lifetime is "smaller" which is basically down to random numerical precision).

Each process must also implement getLifetime which is basically copy-and-pasted since CORSIKA provides the particle lifetime (at least, that is how it used to be the last time I looked at this).

I think there are a couple of things that could be improved:

1. Calculate the decay process only once (as opposed to once per process) and then delegate to the chosen process for calculating the decay products.
2. A central registry for mapping particle decays (by PID) to the process that will decay them. This should have a default mapping (provided by CORSIKA) but the user should also be able to update it as needed.

I opened an issue about this about a year ago, #271.

I'd be open to ideas about how best to implement 2) above.

changed milestone to %Rework of particle decays