Repository navigation
[PWGCF,PWGHF] Add SigmaC(2455)-track femtoscopy to Charm+Track femtoscopy framework - #18692
zhangbiao-phy wants to merge 2 commits into
Conversation
Extend the existing HFC producer and pairing task with SigmaC(2455) -> LambdaC pi candidates, both LambdaC mass hypotheses, and data/MC modes. Process neutral and doubly charged candidates together, preserving charge in the output tables and using the corresponding mass for each pair. Reject tracks shared with any of the four decay daughters. Validate both tasks with syntax checks and synthetic regression tests for legacy channels, output schemas, MC filters, charge-dependent kinematics, and active workflow inputs. Real-AO2D and ONNX/CCDB inference were not run.
|
O2 linter results: ❌ 0 errors, |
mfaggin
left a comment
There was a problem hiding this comment.
Thanks @zhangbiao-phy , please find my comments, they are mostly to better understand and eventually avoid some computations where they are not needed
| } | ||
| for (int h = 0; h < NumHypotheses; ++h) { | ||
| double const deltaMass = scMass[h] - lcMass[h]; | ||
| if (!accepted[h] || !std::isfinite(deltaMass) || deltaMass < sigmacCfg.deltaMassMin.value || deltaMass > sigmacCfg.deltaMassMax.value) { |
There was a problem hiding this comment.
| if (!accepted[h] || !std::isfinite(deltaMass) || deltaMass < sigmacCfg.deltaMassMin.value || deltaMass > sigmacCfg.deltaMassMax.value) { | |
| if (!accepted[h] || !std::isfinite(deltaMass) || deltaMass < 0.135 || deltaMass > sigmacCfg.deltaMassMax.value) { |
I suggest you to hard-code this number, avoiding using a Configurable. The kinematic lower limit for the deltaMass is around 140 MeV/c2 (see deltaMass plots from the Sc(2455) analysis), therefore using e.g. 0.130 GeV/c2 or 0.135 GeV/c2 should be ok.
EDIT: if you might need it to extract the sidebands, then I understand you need a dedicated Configurable. In this case, ignore my comment.
| auto lc = sc.template prongLc_as<LcType>(); | ||
| if (lc.collisionId() != sc.collisionId()) { | ||
| continue; | ||
| } | ||
| auto p0 = lc.template prong0_as<DaughterTracks>(); |
There was a problem hiding this comment.
This in principle is not necessary, since by definition the collision index assigned to a Sigmac candidate is already the one of the daughter Lc candidate, see here https://github.com/AliceO2Group/O2Physics/blob/master/PWGHF/TableProducer/candidateCreatorSigmac0plusplus.cxx#L275
| if (softPi.globalIndex() == p0.globalIndex() || softPi.globalIndex() == p1.globalIndex() || softPi.globalIndex() == p2.globalIndex()) { | ||
| continue; | ||
| } | ||
| std::array<double, NumHypotheses> const lcMass{HfHelper::invMassLcToPKPi(lc), HfHelper::invMassLcToPiKP(lc)}; |
There was a problem hiding this comment.
Also this check here is not necessary, since it's already checked in the Sc candidate creator https://github.com/AliceO2Group/O2Physics/blob/master/PWGHF/TableProducer/candidateCreatorSigmac0plusplus.cxx#L247-L261
| if (sc.charge() != 0 && std::abs(sc.charge()) != ChargeDoublyCharged) { | ||
| continue; | ||
| } | ||
| auto const rapidity = sc.charge() == 0 ? HfHelper::ySc0(sc) : HfHelper::yScPlusPlus(sc); |
There was a problem hiding this comment.
This check here in principle is not necessary because it is already done in the Sigmac candidate creator, throwing a fatal in case of issues, see here https://github.com/AliceO2Group/O2Physics/blob/master/PWGHF/TableProducer/candidateCreatorSigmac0plusplus.cxx#L263-L270
| Partition<FilteredCharmCandScs> partitionCharmHadronSc = ((aod::fdhf::bdtBkg < 0.f && aod::fdhf::bdtPrompt < 0.f && aod::fdhf::bdtFD < 0.f) || | ||
| (aod::fdhf::bdtBkg < charmSel.charmHadBkgBDTmax && aod::fdhf::bdtFD < charmSel.charmHadFdBDTmax && aod::fdhf::bdtFD > charmSel.charmHadFdBDTmin && |
There was a problem hiding this comment.
here I do not understand: why are you keeping in the Partition these Sigmac candidates (if any)? Wouldn't they be meaningless, given negative values for the BDT scores?
There was a problem hiding this comment.
I'd add a control check, in case, requesting that all the scores are >=0
| Partition<FilteredCharmMcCand3ProngsXic> partitionMcCharmHadron3ProngXic = aod::fdhf::originMcRec == OriginRecPrompt || aod::fdhf::originMcRec == OriginRecFD; | ||
| Partition<FilteredCharmMcCand2Prongs> partitionMcCharmHadron2Prong = aod::fdhf::originMcRec == OriginRecPrompt || aod::fdhf::originMcRec == OriginRecFD; | ||
| Partition<FilteredCharmMcCandDstars> partitionMcCharmHadronDstar = aod::fdhf::originMcRec == OriginRecPrompt || aod::fdhf::originMcRec == OriginRecFD; | ||
| Partition<FilteredCharmMcCandScs> partitionMcCharmHadronSc = (aod::fdhf::originMcRec == OriginRecPrompt || aod::fdhf::originMcRec == OriginRecFD) && ((aod::fdhf::bdtBkg < 0.f && aod::fdhf::bdtPrompt < 0.f && aod::fdhf::bdtFD < 0.f) || |
There was a problem hiding this comment.
here again you keep cases with negative BDT scores. Ah, is it a way to keep candidates in analyses w/o the usage of BDT?
| } charmSel; | ||
| struct : ConfigurableGroup { | ||
| std::string prefix = "sigmac"; | ||
| Configurable<int> particleSign{"particleSign", 0, "1: particle, -1: antiparticle, 0: both (also for neutral SigmaC)"}; |
There was a problem hiding this comment.
minor comment: for better readibility, you might consider defining theses cases -1, 1 and 0 with a dedicated enum
| selectedTrackHisto.init(®istry, binmultTempFit, binMulPercentile, binpTTrack, binEta, binPhi, binTempFitVarTrack, binNSigmaTPC, binNSigmaTOF, binNSigmaTPCTOF, binTPCClusters, dummy, dummy, isMc, trackSel.pdgCodeTrack1, true); | ||
|
|
||
| // SigmaC masses are updated from each candidate's charge when filling pairs. | ||
| int const charmPdg = processSc ? Pdg::kSigmaCPlusPlus : charmSel.charmHadPDGCode.value; |
There was a problem hiding this comment.
Here when you run the Sigmac analysis you always consider Pdg::kSigmaCPlusPlus, and never Pdg::kSigmaC0.
I see that here we are in the init function, not in a candidate loop where you can check candidate by candidate the charge. I just raise this point in case it might be relevant for these same/mixed event containers, otherwise ignore it.
| if constexpr (Channel == DecayChannel::ScToLcPi) { | ||
| if (p1.trackId() == p2.prong0Id() || p1.trackId() == p2.prong1Id() || | ||
| p1.trackId() == p2.prong2Id() || p1.trackId() == p2.softPiId()) { | ||
| continue; | ||
| } | ||
| } |
There was a problem hiding this comment.
Is this check meaningful here, given that you are studying the mixed-event case in this piece of code?
Extend the existing HFC producer and pairing task with SigmaC(2455) -> LambdaC pi candidates. Process neutral and doubly charged candidates together, preserving charge in the output tables and using the corresponding mass for each pair. Reject tracks shared with any of the four decay daughters.
Tag @mfaggin and @fgrosa