Add new central configs

changelog.md

@@ -2,6 +2,7 @@
 
 ### [Latest]
 
+- Add new configs for central dataset [#159](https://github.com/umami-hep/umami-preprocessing/pull/159)
 - Generalise the framework beyond jets to arbitrary global objects [#156](https://github.com/umami-hep/umami-preprocessing/pull/156)
 
 ### [v0.3.1](https://github.com/umami-hep/umami-preprocessing/releases/tag/v0.3.1) (19.06.2026)

upp/configs/CentralDatasets/GN3_260728/GN3_260728.yaml

@@ -0,0 +1,185 @@
+variables: !include /home/users/r/reisch/production/slurm-scripts/umami-preprocessing/260728/configs/variables-ghost.yaml
+global_cuts: !include /home/users/r/reisch/production/slurm-scripts/umami-preprocessing/260728/configs/simple-split.yaml
+
+ttbar: &ttbar
+  name: ttbar
+  equal_jets: False
+  pattern:
+    - "user.treisch.601589.e8547_s3797_r13144_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC20d
+    - "user.treisch.601589.e8549_s4159_r15530_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23d
+    - "user.treisch.601589.e8549_s4162_r14622_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23a
+    - "user.treisch.601589.e8549_s4369_r16083_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23e
+    - "user.treisch.601589.e8549_s4618_r17610_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23g
+
+
+zprime: &zprime
+  name: zprime
+  equal_jets: False
+  pattern:
+    - "user.treisch.802818.e8514_s4618_r17610_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC20d
+    - "user.treisch.802818.e8599_s3797_r13144_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23d
+    - "user.treisch.802818.e8599_s4159_r15530_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23a
+    - "user.treisch.802818.e8599_s4162_r15540_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23e
+    - "user.treisch.802818.e8599_s4369_r16083_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23g
+    - "user.treisch.800030.e7954_s3797_r13144_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC20d
+    - "user.treisch.800030.e8514_s4159_r15530_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23d
+    - "user.treisch.800030.e8514_s4162_r14622_p7349.tdd.GN3_dev.25_2_97.26-07-14_CentralProduction_p7349_output.h5/*.h5" # MC23a
+    - "user.treisch.800030.e8514_s4369_r16083_p7349.tdd.GN3_dev.25_2_97.26-07-16_CentralProduction_p7349_output.h5/*.h5" # MC23e
+
+lowpt: &lowpt
+  name: lowpt
+  cuts:
+    - [pt_btagJes, ">", 20_000]
+    - [pt_btagJes, "<", 250_000]
+highpt: &highpt
+  name: highpt
+  cuts:
+    - [pt_btagJes, ">", 250_000]
+    - [pt_btagJes, "<", 6_000_000]
+
+components:
+  - region:
+      <<: *lowpt
+    sample:
+      <<: *ttbar
+    flavours: [ghostbjets]
+    # avail (pre-split) 246,163,504 | train uses 100% of ceiling
+    num_jets: 210_469_795
+    num_jets_test: 11_692_766
+    num_jets_val: 11_692_766
+
+  - region:
+      <<: *lowpt
+    sample:
+      <<: *ttbar
+    flavours: [ghostcjets]
+    # avail (pre-split) 56,596,467 | train uses 100% of ceiling
+    num_jets: 48_389_979
+    num_jets_test: 2_688_332
+    num_jets_val: 2_688_332
+
+  - region:
+      <<: *lowpt
+    sample:
+      <<: *ttbar
+    flavours: [ghostsjets]
+    # avail (pre-split) 44,739,813 | train uses 100% of ceiling
+    num_jets: 38_252_540
+    num_jets_test: 2_125_141
+    num_jets_val: 2_125_141
+
+  - region:
+      <<: *lowpt
+    sample:
+      <<: *ttbar
+    flavours: [ghostudjets]
+    # avail (pre-split) 114,413,319 | train uses 100% of ceiling
+    num_jets: 97_823_387
+    num_jets_test: 5_434_632
+    num_jets_val: 5_434_632
+
+  - region:
+      <<: *lowpt
+    sample:
+      <<: *ttbar
+    flavours: [ghostgjets]
+    # avail (pre-split) 134,653,370 | train uses 100% of ceiling
+    num_jets: 115_128_631
+    num_jets_test: 6_396_035
+    num_jets_val: 6_396_035
+
+  - region:
+      <<: *lowpt
+    sample:
+      <<: *ttbar
+    flavours: [ghosttaujets]
+    # avail (pre-split) 23,657,240 | train uses 100% of ceiling
+    num_jets: 20_226_940
+    num_jets_test: 1_123_718
+    num_jets_val: 1_123_718
+
+  - region:
+      <<: *highpt
+    sample:
+      <<: *zprime
+    flavours: [ghostbjets]
+    # highpt b:  50% of lowpt (2:1 target); pre-split avail re-measured by prep (was 56.6M)
+    num_jets: 105_234_898
+    num_jets_test: 5_846_383
+    num_jets_val: 5_846_383
+
+  - region:
+      <<: *highpt
+    sample:
+      <<: *zprime
+    flavours: [ghostcjets]
+    # highpt c:  50% of lowpt (2:1 target); pre-split avail re-measured by prep (was 55.7M)
+    num_jets: 24_194_990
+    num_jets_test: 1_344_166
+    num_jets_val: 1_344_166
+
+  - region:
+      <<: *highpt
+    sample:
+      <<: *zprime
+    flavours: [ghostsjets]
+    # highpt s:  50% of lowpt (2:1 target); pre-split avail re-measured by prep (was 35.9M)
+    num_jets: 19_126_270
+    num_jets_test: 1_062_571
+    num_jets_val: 1_062_571
+
+  - region:
+      <<: *highpt
+    sample:
+      <<: *zprime
+    flavours: [ghostudjets]
+    # highpt ud: 50% of lowpt (2:1 target); pre-split avail re-measured by prep (was 39.2M)
+    num_jets: 48_911_694
+    num_jets_test: 2_717_316
+    num_jets_val: 2_717_316
+
+  - region:
+      <<: *highpt
+    sample:
+      <<: *zprime
+    flavours: [ghostgjets]
+    # highpt g:  50% of lowpt (2:1 target); pre-split avail re-measured by prep (was 76.7M)
+    num_jets: 57_564_316
+    num_jets_test: 3_198_018
+    num_jets_val: 3_198_018
+
+  - region:
+      <<: *highpt
+    sample:
+      <<: *zprime
+    flavours: [ghosttaujets]
+    # highpt tau:50% of lowpt (2:1 target); pre-split avail re-measured by prep (was 35.7M)
+    num_jets: 10_113_470
+    num_jets_test: 561_859
+    num_jets_val: 561_859
+
+resampling:
+  target: ghostcjets
+  method: countup
+  sampling_fraction: 1
+  variables:
+    pt_btagJes:
+      bins: [[20_000, 250_000, 46], [250_000, 1_000_000, 30], [1_000_000, 6_000_000, 20]]
+    absEta_btagJes:
+      bins: [[0, 2.5, 10]]
+
+# note: sensible defaults are defined in the PreprocessingConfig constructor
+global:
+  jets_name: jets
+  batch_size: 1_000_000
+  num_jets_estimate: 50_000_000
+  num_jets_estimate_plotting: 10_000_000
+  num_jets_per_output_file: 5_000_000
+  base_dir: /srv/beegfs/scratch/groups/rodem/DAOD_FTAG/p7349/upp_260728/
+  out_dir: /srv/beegfs/scratch/groups/rodem/DAOD_FTAG/p7349/upp_260728/out/
+  ntuple_dir: /srv/beegfs/scratch/groups/rodem/DAOD_FTAG/p7349/
+
+# Setting for plotting
+plotting:
+  num_jets_plotting: 10_000_000
+  show_num_jets: false

upp/configs/CentralDatasets/GN3_260728/simple-split.yaml

@@ -0,0 +1,8 @@
+# simple split on event number: 90/05/05 train/val/test
+# no k-folding
+train:
+  - [eventNumber, "%20<=", 17]
+val:
+  - [eventNumber, "%20==", 18]
+test:
+  - [eventNumber, "%20==", 19]

upp/configs/CentralDatasets/GN3_260728/variables-ghost.yaml

@@ -0,0 +1,184 @@
+jets:
+  inputs:
+    - pt_btagJes
+    - eta_btagJes
+  labels:
+    - HadronConeExclTruthLabelID
+    - HadronConeExclExtendedTruthLabelID
+    - HadronConeExclTruthLabelPt
+    - HadronConeExclTruthLabelLxy
+    - HadronConeExclTruthLabelDR
+    - HadronGhostTruthLabelID
+    - HadronGhostExtendedTruthLabelID
+    - HadronGhostTruthLabelPt
+    - HadronGhostTruthLabelLxy
+    - HadronGhostTruthLabelDR
+    - HadronGhostInitialTruthLabelID
+    - HadronGhostInitialTruthLabelPdgId
+    - HadronGhostInitialExtendedTruthLabelID
+    - PartonExtendedTruthLabelID
+    - PartonTruthLabelID
+    - pt
+    - eta
+    - mass
+    - n_tracks_ghost
+    - n_truth_promptLepton
+    - eventNumber
+    - jetFoldHash
+    - ptFromTruthJet
+    - matchedToTruthJet
+    - ptFromTruthDressedWZJet
+    - matchedToTruthDressedWZJet
+    - decayModeFromTruthTaus
+    - matchedToTruthTaus
+    - pt_visFromTruthTaus
+    - matchedToTauJet
+    - ptFinalCalibFromTauJet
+
+tracks_ghost:
+  inputs:
+    - d0
+    - z0SinTheta
+    - dphi
+    - deta
+    - qOverP
+    - lifetimeSignedD0Significance
+    - lifetimeSignedZ0SinThetaSignificance
+    - phiUncertainty
+    - thetaUncertainty
+    - qOverPUncertainty
+    - numberOfPixelHits
+    - numberOfSCTHits
+    - numberOfInnermostPixelLayerHits
+    - numberOfNextToInnermostPixelLayerHits
+    - numberOfInnermostPixelLayerSharedHits
+    - numberOfInnermostPixelLayerSplitHits
+    - numberOfPixelSharedHits
+    - numberOfPixelSplitHits
+    - numberOfSCTSharedHits
+    # Soft Muon variables
+    - leptonID
+    - muon_quality
+    - muon_qOverPratio
+    - muon_momentumBalanceSignificance
+    - muon_scatteringNeighbourSignificance
+
+  labels:
+    - ftagTruthOriginLabel
+    - ftagTruthTypeLabel
+    - ftagTruthSourceLabel
+    - ftagTruthVertexIndex
+
+flows:
+  inputs:
+    - pt
+    - energy
+    - deta
+    - dphi
+    - isCharged
+
+calo:
+  inputs:
+    - "usedInChargedFlow"
+    - "ENG_BAD_CELLS"
+    - "ISOLATION"
+    - "CENTER_MAG"
+    - "CELL_SIGNIFICANCE"
+    - "ENG_FRAC_MAX"
+    - "LATERAL"
+    - "SIGNIFICANCE"
+    - "LONGITUDINAL"
+    - "ENG_POS"
+    - "EM_PROBABILITY"
+    - "CENTER_LAMBDA"
+    - "SECOND_LAMBDA"
+    - "FIRST_ENG_DENS"
+    - "SECOND_R"
+    - "AVG_LAR_Q"
+    - "MASS"
+    - "rawPhi"
+    - "calPhi"
+    - "rawEta"
+    - "calEta"
+    - "rawE"
+    - "calE"
+    - "PreSamplerB"
+    - "EMB1"
+    - "EMB2"
+    - "EMB3"
+    - "PreSamplerE"
+    - "EME1"
+    - "EME2"
+    - "EME3"
+    - "HEC0"
+    - "HEC1"
+    - "HEC2"
+    - "HEC3"
+    - "TileBar0"
+    - "TileBar1"
+    - "TileBar2"
+    - "TileGap1"
+    - "TileGap2"
+    - "TileGap3"
+    - "TileExt0"
+    - "TileExt1"
+    - "TileExt2"
+    - "FCAL0"
+    - "FCAL1"
+    - "FCAL2"
+
+electrons:
+  inputs:
+    - pt
+    - ptfrac
+    - ptrel
+    - dr
+    - deta
+    - dphi
+    - ftag_et
+    - qOverP
+    - d0RelativeToBeamspot
+    - d0RelativeToBeamspotSignificance
+    - ftag_ptVarCone30OverPt
+    - numberOfPixelHits
+    - numberOfSCTHitsInclDead
+    - ftag_deltaPOverP
+    - eProbabilityHT
+    - deltaEta1
+    - deltaPhiRescaled2
+    - ftag_energyOverP
+    - Rhad
+    - Rhad1
+    - Eratio
+    - weta2
+    - Rphi
+    - Reta
+    - wtots1
+    - f1
+    - f3
+
+  labels:
+    - ftagTruthOriginLabel
+    - ftagTruthTypeLabel
+    - ftagTruthSourceLabel
+    - ftagTruthVertexIndex
+
+muons:
+  inputs:
+    - quality
+    - pt
+    - eta
+    - phi
+    - momentumBalanceSignificance
+    - scatteringNeighbourSignificance
+    - ptfrac
+    - ptrel
+    - dr
+    - qOverP
+    - qOverPratio
+    - numberOfPixelHits
+    - eProbabilityHT
+    - lifetimeSignedD0
+    - lifetimeSignedZ0SinTheta
+    - lifetimeSignedD0Significance
+    - lifetimeSignedZ0SinThetaSignificance