Xiaozhong Huang authored 4 years ago and Vakhtang Tsulaia committed 4 years ago

For Topo seed jets, the cluster links in the tau candidate point to the
clusters in a shallow copy container, which will not be available in AOD
of R22. To fix this, the cluster links point to the cluster in the
original container.

For PFlow seed jets, the cluster links in the tau candidate point to the
PFOs(neutral or charged). It is confusing. Now the cluster links point
to the corresponding cluter in the PFO object.

Previously, only clusters within 0.2 cone of tau candidate is stored. In
the tau reconstruction, we need all the clusters. Thus the cut of dR is now
removed.

The clusters could have energy <=0, which will be skimmed in AOD. The
link then becomes invalid. Since we do not want an invalid link, all
the clusters are requied to have positive energy.

TrigT1CaloCalibConditions: Fix cppcheck warnings.

See merge request atlas/athena!37963

Reconstruction: cmake cleanup of tracking-related packages

See merge request atlas/athena!37905

tauRecTools: drop "TauEtaCalib" 4-momentum calibration

See merge request atlas/athena!37952

Bertrand Martin Dit Latour authored 4 years ago and Edward Moyse committed 4 years ago

Hello,

This MR is dropping the "TauEtaCalib" 4-momentum calibration from the tau reconstruction.
This is a calo-only calibration which is almost identical to the "TauEnergyScale" calibration, except that a tiny correction is applied to the reconstructed calo-only eta to compensate for a tiny average bias (correction derived w.r.t. truth tau visible eta).
Given the poor calo-only angular resolution, an absolute eta correction of at most 0.005 on eta is irrelevant in terms of tau performance.
We could consider deploying an MVA regression at xAOD level to improve the angular resolution, so we don't need this intermediate momentum calibration in the reconstruction.
Benefit: there will be one less tool scheduled (we used to call this tool twice), and a bunch of floats are no longer saved in the xAOD.
In the few places where we were using the TauEtaCalib calibration, we are now using the TauEnergyScale calibration.

This MR affects a number of xAOD output variables, but the changes are very small and under control.
For instance, for our final MVA pt, the relative change in the average pt is ~2.5e-4.
The "TauEtaCalib" variables are now 0. Same for TauJetParameters::TESCalibConstant and TauJetParameters::TESOffset (not useful to keep).
The rest is noise.
```
Py:diff-root         INFO  [TauJetsAux.BDTJetScoreSigTrans]: 38 leaves differ
Py:diff-root         INFO  [TauJetsAux.RNNEleScore]: 2 leaves differ
Py:diff-root         INFO  [TauJetsAux.RNNEleScoreSigTrans]: 30 leaves differ
Py:diff-root         INFO  [TauJetsAux.RNNJetScore]: 3 leaves differ
Py:diff-root         INFO  [TauJetsAux.RNNJetScoreSigTrans]: 35 leaves differ
Py:diff-root         INFO  [TauJetsAux.TESCalibConstant]: 64 leaves differ
Py:diff-root         INFO  [TauJetsAux.eta]: 14 leaves differ
Py:diff-root         INFO  [TauJetsAux.etaFinalCalib]: 14 leaves differ
Py:diff-root         INFO  [TauJetsAux.etaPanTauCellBased]: 14 leaves differ
Py:diff-root         INFO  [TauJetsAux.etaTauEtaCalib]: 64 leaves differ
Py:diff-root         INFO  [TauJetsAux.mPanTauCellBased]: 8 leaves differ
Py:diff-root         INFO  [TauJetsAux.phiTauEtaCalib]: 64 leaves differ
Py:diff-root         INFO  [TauJetsAux.pt]: 63 leaves differ
Py:diff-root         INFO  [TauJetsAux.ptFinalCalib]: 63 leaves differ
Py:diff-root         INFO  [TauJetsAux.ptPanTauCellBased]: 14 leaves differ
Py:diff-root         INFO  [TauJetsAux.ptTauEtaCalib]: 64 leaves differ
Py:diff-root         INFO  [TauJetsAuxDyn.eta_combined]: 40 leaves differ
Py:diff-root         INFO  [TauJetsAuxDyn.m_combined]: 39 leaves differ
Py:diff-root         INFO  [TauJetsAuxDyn.pt_combined]: 63 leaves differ
```

I've minimally updated TrigTauAlgorithmsHolder.py. Further cleanup of deprecated TauCalibrateLC options could be done in TrigTauRecConfig.py and TrigTauRecConfigMT.py, but I would leave this to trigger experts.
Let me just mention the tool name should be unique to distinguish the "caloOnly=True" and "caloOnly=False" cases: currently, a first TauCalibrateLC tool will be configured, and then reused regardless of the caloOnly setting.
https://gitlab.cern.ch/atlas/athena/-/blob/master/Trigger/TrigAlgorithms/TrigTauRec/python/TrigTauAlgorithmsHolder.py#L113

Cheers,
Bertrand

Updating r22 JetTagging: DL1 and cleanup

See merge request atlas/athena!37842

GaudiSvcConf import fixes for Gaudi v35

See merge request !37947

Frank Winklmeier authored 4 years ago and Edward Moyse committed 4 years ago

Gaudi v35 drops the backwards compatiblity imports of `GaudiSvcConf`.
Import the services from their actual location.

SCT_ReadoutGeometry+CommissionEvent: cmake fix

See merge request atlas/athena!37965

AnalysisTop: migrate AodMetaDataReader.py to py3

See merge request atlas/athena!37949

AnalysisTop: Restoring code with btagging links for jets

See merge request atlas/athena!37951

Update muon trigger view data verifier contents

See merge request atlas/athena!37972

Replace direct TagInfo access in IOVDbSvc with use of TagInfoMgr

See merge request atlas/athena!37919

Marcin Nowak authored 4 years ago and Edward Moyse committed 4 years ago

Trying to group access to TagInfo object exclusively to the
TagInfoMgr so it can be better managed during input file transitions
in MT

ATLASRECTS-5743  fix the rest of the ART tests

See merge request atlas/athena!37969

Fixing mu retrieving for legacy TauMonitoring

See merge request atlas/athena!37938

Python 3 cleanup for AthenaCommon

See merge request atlas/athena!37948

Removed references to eleBDT

See merge request atlas/athena!37940

Sweep TruthClassificationTool to master

See merge request atlas/athena!37819

ATLIDTRKCP-306 Make ACTS primary vertexing default

See merge request atlas/athena!37283

HLT framework monitoring updates for Technical Run 14

See merge request atlas/athena!37962

MuonRecExample: fix MuonRecTools.AtlasTrackingGeometrySvc

Closes ATLASSIM-4944

See merge request atlas/athena!37959

LArByteStream: Make converters const.

See merge request atlas/athena!37926

Manual Sweep: Update default cross talk to 20 % and 4 %

See merge request atlas/athena!37934

tauRec: drop TauEleOLRDecorator tool

See merge request atlas/athena!37937

chainDump.py: add option to write a condensed counts yaml file

See merge request atlas/athena!37918

Add xAOD::getOriginalObjectLink(...), master branch (2020.11.05.)

See merge request atlas/athena!37950

Add selection for the muon Iso tracking to the muon chains

See merge request atlas/athena!37960

CxxUtils: Small improvements for CachedValue::isValid().

See merge request atlas/athena!37967

Fix standalone usage of RecExCommon

Closes ATLASRECTS-5747

See merge request atlas/athena!37989

Disable configuration check in CutFlowSvc for now

See merge request atlas/athena!37976

Removing some unnecessary objects from the muon view data verifiers (mostly objects that were being loaded that were not produced or used anywhere)