Skip to content

[QEE] Electroweak Tupling job on "physics" 2023

Luke Grazette requested to merge lugrazet-ewAP-Coll23-iso into master

Tupling over the Spruced 2023 data. This corresponds to Jul7th-9th [~40 fb-1] + Coll23 [~40 fb-1] data. Both Turbo and Full QEE streams.

(Part of the Run3 ew-analyses framework development: see Wiki ).

Depending on stream, performs Tuplings that differ in selections AND the Full-stream also has Cone Isolation Variables on Ws/Zs included.

selection_to_hlt2line = {
    "Jpsi_Detached": "Hlt2QEE_JpsiToMuMu_Detached",
    "Jpsi_Prompt": "Hlt2QEE_JpsiToMuMu_Prompt",
    "U1S": "Hlt2QEE_Upsilon1SToMuMu"
}
selection_to_spruceline = {
    "Z": "SpruceQEE_ZToMuMu",
    "ZMuID": "SpruceQEE_ZToMuMu_SingleNoMuID",
    "Wp": "SpruceQEE_SingleHighPtMuon",
    "Wm": "SpruceQEE_SingleHighPtMuon",
}

Based mostly on AP!574 which was Hlt2-Comm23 data and AP!580 the TurboPass Example.

Understanding the Isolation: Spruce outputs extra_outputs for isolation.

  • For W decays, just any particles in the Cone DeltaR < 0.5**2 around the signal muon.
  • For Z decays, 1 cone per child muon, extra_outputs is the combination of these sets.

So at the DV level:

  • For W decays, muon = signal candidate. Cone = extra_outputs. No issue.
  • For Z decays, muon = IS_ID @ GetAllDescendants() on Signal candidate. Cone = extra_outputs + a new DR<0.5**2 cone around the specific muon being investigated. Once the weightedreltable is found, the ParticleIsolation functor collection is used.

This allows ISO050_CPT to be found per (signal) muon in our decays, which is the scalar sum of PT in the cone.

  • Could write a Vector CPT via summing all 4-momenta in the cone and then extracting the PT from that (either here or at analysis level). Found unexpected behaviour from (safe_sumcone(F.PX)**2 + safe_sumcone(F.PY)**2)**0.5, so instead have included ISO050_CPX, ISO050_CPY [which look correct] to perform this at the analysis level in process_tuples_upgrade instead.

ToDo:

  • Both streams completing Tupling jobs
  • Understand why CP, CPT, ..., are NaN Fixed F.DR2 < 0.5**2 -> F.DR2 < (0.5**2)
  • Use Vector CPT or Scalar CPT? <- Feedback from Ross. Kept CPX and CPY in tuple for analysis VecCPT calc.
  • Nicole announced Jul7th-9th data is now finished sprucing, include it in this AP.

Supersedes SOME OF AP!574 (Hlt2-Comm23) with updated files: (Spruce-Comm23) includes more events than Hlt2-Comm23 and Isolation variables for Full-stream. Doesn't supersede the VeloOpen files (but we since have realised we probably don't have any plans ourselves to use them, but I will leave them there in case we change our minds). Generates new files (Spruce-Coll23)

Edited by Luke Grazette

Merge request reports