[PWGHF] Correlated backgrounds in derived data creators - #18043
Conversation
|
O2 linter results: ❌ 0 errors, |
|
|
||
| Filter filterSelectCandidates = (aod::hf_sel_candidate_dplus::isSelDplusToPiKPi & static_cast<int>(BIT(aod::SelectionStep::RecoMl - 1))) != 0; // select candidates which passed all cuts at least up to RecoMl - 1 | ||
| Filter filterMcGenMatching = nabs(aod::hf_cand_mc_flag::flagMcMatchGen) == static_cast<int8_t>(hf_decay::hf_cand_3prong::DecayChannelMain::DplusToPiKPi); | ||
| Filter filterMcGenMatching = (nabs(aod::hf_cand_mc_flag::flagMcMatchGen) == static_cast<int8_t>(hf_decay::hf_cand_3prong::DecayChannelMain::DplusToPiKPi)) || (fillCorrelatedBkgs && aod::hf_cand_mc_flag::flagMcMatchGen != 0); |
There was a problem hiding this comment.
I'm not sure this is correct. It seems to me that this will select all flagged 3-prong-candidate decays, not just D+.
There was a problem hiding this comment.
Same applies to the rec. level.
There was a problem hiding this comment.
I think that this is correct, i.e. one gets all candidates passing e.g. the Dplus selection but some of them come from correlated background sources which can be flagged and selected at the analysis level. Am I missing something?
There was a problem hiding this comment.
Where do you reject all the other species (Ds, Lc, Dstar, Xic, cd)?
There was a problem hiding this comment.
If I understood correctly, one should run the candidateCreator3Prong with createDplus only switched on. This way, one accepts only candidates reconstructed as D+ in the skimming. Some of these may stem from correlated bkgs and are tagged by the MC flags, but with different flag with respect to the signal. In order to access all sources, i just put the condition != 0
There was a problem hiding this comment.
Yes, so when one wants to study the correlated backgrounds he can run the data creator with acceptCorrelatedBkgs=true and have access to both the true signal and the backgrounds using the same train. This way, one can compute ratios of number of reconstructed candidates with respect to the signals and can anchor the various sources without needing the generated information and full reconstruction efficiencies.
There was a problem hiding this comment.
OK, thanks for the clarification. I understand better now. But then, don't you need the same for the gen. level?
There was a problem hiding this comment.
As far as I know, it is currently not used in analysis. But I can add it back for consistency if you think it would be clearer
There was a problem hiding this comment.
I added the same implementation for the generated
|
Hi @vkucera, I see that all checks are successful. If the logic of the implementation is clarified, would it be possible to proceed with the merge? Thanks in advance! |
4 participants
Follows the strategy currently adopted in tree creators