key4hep / key4hep/k4SimDelphes
Proper setting of mother and daughter relations for MCParticles
- Dominant language
- C++
- Stars
- 6
- Forks
- 21
- PR merge metrics
- No merged PRs in 30d
Description
The current handling of setting the mothers and daughters is most probably not correct. E.g. if we look at how at [how Pythia defines these relationships](http://home.thep.lu.se/~torbjorn/pythia81html/ParticleProperties.html) then we have to handle several different cases depending on the values of `M1` and `M2`, resp. `D1` and `D2` that we obtain from a Delphes `GenCandidate`. I have opened [a ticket at the Delphes issue tracker](https://cp3.irmp.ucl.ac.be/projects/delphes/ticket/1469) describing this specific issue.
Additionally, it seems that in any case `M1` and `M2`, resp. `D1` and `D2` should in general be interpreted as ranges of indices to the mothers / daughters and not as the indices of the two mothers and daughters. This is something that is in principle easily fixed, but then clashes with the interpretation from Pythia, where there is a special meaning to cases where `M1 >= 0 && M2 >= 0 && M2 < M1` (and equivalently for the daughters). Also this would in principle be not too hard to handle in the converter, but ideally the converter does not know which reader was used to fill the initial generated particles collection. So the main question is, is the interpretation of the `M1`/`M2` and `D1`/`D2` indices something that follows a "HEP convention" and we can always interpret in the same way regardless of whether we got our MC Particles from Pythia or read them via HepMC or from a STDHEP file?
If that is not the case, then we either have to make the converter aware of how to interpret these indices to set the correct mother and daughter relations, or we maybe have to see how this could be fixed in upstream Delphes.
Contributor guide
Research direction
Start with the converter's handling of Delphes GenCandidate M1/M2 and D1/D2 values, then compare the Pythia relationship rules with the linked Delphes ticket and the HepMC and STDHEP cases mentioned in the issue. Done means determining whether one interpretation is valid across readers, or identifying the required reader-aware or upstream Delphes fix.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- data
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100