Skip to content

Charm developments towards July 2024 deadline

Alessia Anelli requested to merge Charm2024-July into 2024-patches

Dependencies

This merge request depends on:

Overall impact on bandwidth

imagen

Taken from the latest test after https://gitlab.cern.ch/lhcb/Moore/-/commit/b4b5722d9580b583faa2474224af9b6c0372a1d1

Goals

General Charm MR to include modifications for the period after the MD happening in August. The internal deadline is set to the 17th of July. These changes include tuning of lines (with a special focus on downstream tracks), bug fixes in the decay descriptor, implemented MVA-based lines, etc.

Contributing

If you want to contribute to this MR, please open a new MR and target the Charm2024-July branch and explain your goals.

Any contributions to this MR should be ready by the 23th of July, in such a way that the impact on the rate, bandwidth and throughput has been studied and is within the thresholds.

The tests must be run using this snippet.

All the additions need to be tested locally, including:

  • Impact on the bandwidth and rate: moderate rates and bandwidths should be below 100 Hz and 10 MB/s, respectively. This must be tested using expected_2024_min_bias_hlt1_filtered_v2.
  • (optional) Impact on the throughput: comparing the time with the execution of 2024-patches with no extra modifications.
  • (optional) Determine efficiencies: if there is simulation already available, it will be interesting to know the efficiencies so we ensure that there is no bug in the selections.

Do not rebase or force-push to this branch without the consent of the liaison @lpica @aanelli and @cacochat, since this might lead to losses on the changes

Handling Git conflicts

Since the Charm2024-July branch is based on 2024-patches, and due to the continuous development state of the latter, we need to do frequent rebase operations to keep it up-to-date. If you are collaborating in an independent branch e.g. $username/charm_etc you will need to have as a base Charm2024-July. You can do so via

git checkout Charm2024-July
git checkout -b $username/charm_etc

Note that if you have been initially basing the branch on 2024-patches at the point of changing the target branch in your merge request you will have a lot of commits from 2024-patches that will be added to Charm2024-July, which is not desirable (it will just add noise to the set of changes and be a source of potential conflicts). Something similar might happen if we rebase Charm2024-July; the base commit of $username/charm_etc will be lost because the whole history is reorganized during this operation (i.e. the base commit identifier, the hash, will no longer be in the history of Charm2024-July). In any of these two cases, we recommend to backup your branch with a new name (just in case) and then make use of the --onto argument of git rebase.

Changes

Expected changes:

https://gitlab.cern.ch/lhcb/Moore/-/merge_requests?scope=all&state=opened&target_branch=Charm2024-July

Official Changes:

Please add your changes:

cbaryon_‎to_sl @ziyiw !3650 (merged)

  • Add WS lines to inclusive Lc/Xic/Oc semi-leptonic decays
  • Tune the cut and replace CORMM with MASS according to the study based on Run2 and Run3 samples.

d0_to_kshh @mamartin !3685 (merged)

  • Added HLT1 filtering on newly introduced D2KsHH HLT1 line and TwoTrackMVA for the nobias line.
  • Added a DOCA<0.2mm cut on the two hadrons in the D0 candidate.

d_to_ksh @smaccoli !3673 (merged)

  • Reproduced what has been done in !3626 (merged) (cuts on KS IPCHI2, Dp FD RHO, Dp END VZ, KS FD Z, KS FD XY)
  • Further improvements: [looser cut] KS daughters PIDK < 5 --> 15, [looser cut] Dp DOCA < 0.1 mm --> 2 mm (only for DD candidated), [tighter cut] KS FD Z > 5 mm --> 50 mm (only for DD candidated), [tighter cut] KS FD T (transverse) > 1 mm --> 10 mm (only for DD candidates)

hyperon_TT @ziyiw !3674 (merged)

  • Enlarge the mass windows for c-baryons
  • Add GhostProb selection to Long/Downstream tracks
  • Fine-tune the remaining cuts based on the Simulation samples and the preliminary results from 2024 data.

d_to_ksh, d_to_hhh @ldufour !3626 (merged)

hyperons @mstahl !3688 (merged) !3733 (merged)

  • General tuning of lines after looking at post TS 2024 data
  • Moving to OWNPV and ProbNN functors
  • Move to selective persistence for lines that had persist reco
  • Add new lines: Hlt2Charm_SpToPpMumMup_VeloMatch_SP, Hlt2Charm_XimToL0MumNu_VeloMatch_SP, Hlt2Charm_Xic0ToSmPip_LongSm_Inclb_SP, Hlt2Charm_Xic0ToSmKsPip_{LL,DD}_LongSm_Inclb_SP
  • Merge some Xicp and Lcp, as well as Xic0 and Oc0 lines by widening mass windows. Merged lines are Hlt2Charm_LcpXicpToL0Kp_{LL,DD}, Hlt2Charm_Xic0Oc0ToL0KmPip_{LL,DD,LL_Inclb_SP,DD_Inclb_SP}, Hlt2Charm_Xic0Oc0ToL0KmKp_{LL,DD}, Hlt2Charm_Xic0Oc0ToXimKsPip_{LLLLL,LLLDD,DDLLL,DDLDD}, Hlt2Charm_LcpXicpToL0KmKpPip_{LL,DD} and Hlt2Charm_Xic0Oc0ToOmKpPimPip_{LLL,DDL}
  • Removed Hlt2Charm_Xib0ToXicpPim_XicpToSpKmPip_LongSp_SP as Sigma+ does not live long enough to have sufficient statistics with this method of reconstruction

cbaryon_‎mva @shuyu !3661 (merged)

  • Decouple the MVA-based lines for Lcp, Xicp and Xic0 from the cut-based lines.
  • Improve Lcp, Xicp and Xic0 MVA lines with new BDT trained with 2024 data.
  • The MVA-based lines are: Hlt2Charm_LcpToPpKmPip_MVA_PR, Hlt2Charm_XicpToPpKmPip_MVA_PR, Hlt2Charm_Xic0ToPpKmKmPip_MVA_PR
Edited by Shuyu Han

Merge request reports