athena merge requestshttps://gitlab.cern.ch/atlas/athena/-/merge_requests2021-08-04T17:26:34+02:00https://gitlab.cern.ch/atlas/athena/-/merge_requests/45456Adding useQuantile support for truth tagging with GNN2021-08-04T17:26:34+02:00Nilotpal KakatiAdding useQuantile support for truth tagging with GNNThis MR addresses the following things -
- useQuantile implementation with GNN
- Bugfix from the previous [MR](https://gitlab.cern.ch/atlas/athena/-/merge_requests/44677). A conditional block based on strategy has been added to the setJetsThis MR addresses the following things -
- useQuantile implementation with GNN
- Bugfix from the previous [MR](https://gitlab.cern.ch/atlas/athena/-/merge_requests/44677). A conditional block based on strategy has been added to the setJetshttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45450Following up MR45421. Also add convenience LikelihoodEnums and test2021-07-28T03:03:01+02:00Christos Anastopouloschristos.anastopoulos@cern.chFollowing up MR45421. Also add convenience LikelihoodEnums and testFollowing up https://gitlab.cern.ch/atlas/athena/-/merge_requests/45421.
Also add convenience LikelihoodEnums and testFollowing up https://gitlab.cern.ch/atlas/athena/-/merge_requests/45421.
Also add convenience LikelihoodEnums and testhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45423Remove unused duplicated configuration2021-07-27T03:04:43+02:00Dan GuestRemove unused duplicated configurationRemove some confusing configuration files. These aren't used anywhere, or if they are they should not be.Remove some confusing configuration files. These aren't used anywhere, or if they are they should not be.https://gitlab.cern.ch/atlas/athena/-/merge_requests/45422DerivationFrameworkEGamma PEP8 part II2021-07-26T13:43:23+02:00Christos Anastopouloschristos.anastopoulos@cern.chDerivationFrameworkEGamma PEP8 part IIDerivationFrameworkEGamma PEP8 part IIDerivationFrameworkEGamma PEP8 part IIhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45421Add convenience EgammaPIDdefs.py and std python test for the enums2021-07-27T03:05:08+02:00Christos Anastopouloschristos.anastopoulos@cern.chAdd convenience EgammaPIDdefs.py and std python test for the enumsAdd convenience EgammaPIDdefs.py and std python test for the enums , should avoid issues due to going via cppyy etc
Mentioning @tbold as this should help solve one of the issue he has reportedAdd convenience EgammaPIDdefs.py and std python test for the enums , should avoid issues due to going via cppyy etc
Mentioning @tbold as this should help solve one of the issue he has reportedhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45418Add algorithm to b-tag with respect to the beamspot2021-07-29T17:49:19+02:00Dan GuestAdd algorithm to b-tag with respect to the beamspotOur current b-tagging impact parameter (IP) calculation relies on extrapolating the track to the perigee point with respect to the primary vertex. We might be able to do just about as well with the beamspot perigee point. There are a few...Our current b-tagging impact parameter (IP) calculation relies on extrapolating the track to the perigee point with respect to the primary vertex. We might be able to do just about as well with the beamspot perigee point. There are a few advantages to the later:
- It doesn't depend on atlas extrapolator code, which is a heavy dependency
- Many of the parameters are still defined without a primary vertex, meaning we could still b-tag without a primary vertex
The second case might be especially useful in the trigger where we might want to b-tag as a preselection step before running full-event tracking and building a primary vertex. We could also use b-tagging algorithms to estimate the primary vertex z-position.
More concretely, this merge request touches code in a few places:
- Adds another algorithm to calculate the IP with respect to the beamspot,
- Adds more flexibility in DL2 so that we can run DIPS using this IP definition
- Removes the default argument for `BTagTrackIpAccessor` since there might be a few now (and updates some code that uses this class)
- Moves `JetParticleAssociationAlgConfig.py` into `ParticleJetTools`, which is where `JetParticleAssociationAlg` is defined. This makes it accessible in standalone builds.
- Removes the configuration to write DL2 outputs to doubles. We migrated to floats a while ago now.
It should not change any variables in any ATLAS data formats.https://gitlab.cern.ch/atlas/athena/-/merge_requests/45409HIGG1D1 update jet reconstruction2021-07-30T03:04:42+02:00Anthony MorleyHIGG1D1 update jet reconstructionUpdate Jet reconstruction to use the diphoton selected vertex and some other bug fixesUpdate Jet reconstruction to use the diphoton selected vertex and some other bug fixeshttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45398FTAG release 22 b-taggers/training in reco for monitoring purpose2021-07-24T03:02:59+02:00Arnaud DuperrinFTAG release 22 b-taggers/training in reco for monitoring purposeThis MR adds b-tagging new release 22 b-taggers/training in reco for monitoring purpose.
It includes:
- Recom. DL1r on R21: recommended DL1r trained on r21 from r21 samples.
- Recom. DL1r on R22: DL1r trained on r21 applied to r22 sampl...This MR adds b-tagging new release 22 b-taggers/training in reco for monitoring purpose.
It includes:
- Recom. DL1r on R21: recommended DL1r trained on r21 from r21 samples.
- Recom. DL1r on R22: DL1r trained on r21 applied to r22 samples.
- DL1r R22: DL1r re-trained on r22 samples applied to r22 samples (it corresponds to DL1r20210519r22).
- DL1d R22: new high-level tagger trained in rel22 using DIPS as input instead of RNNIP (it coressponds to DL1d20210528r22).
Tagging @cschiavi @vdao @mguth @fdibello @dguest @sanmay for FTAGhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45396fix IDTRKVALID and IDTIDE for MC2021-07-27T03:04:10+02:00Maximilian Emanuel Goblirsch-Kolbfix IDTRKVALID and IDTIDE for MCThis MR adds two decorations to the output list for the DAOD_IDTRKVALID and IDTIDE formats, required for MC analysis.
Also address a possible typo in the IDTIDE config, and add protection in IDPVM to prevent an uninformative athena cras...This MR adds two decorations to the output list for the DAOD_IDTRKVALID and IDTIDE formats, required for MC analysis.
Also address a possible typo in the IDTIDE config, and add protection in IDPVM to prevent an uninformative athena crash in case we run on a derivation with a slimmed track content.
CC @goetz @sroe @vcairo @npettershttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45388B-tag pflow in reconstruction2021-07-26T03:04:01+02:00Dan GuestB-tag pflow in reconstructionAlso tag pflow in reconstruction, this is mostly for monitoring.
Tagging @duperrin. I don't see any evidence that this isn't working so I think it's fine to merge.Also tag pflow in reconstruction, this is mostly for monitoring.
Tagging @duperrin. I don't see any evidence that this isn't working so I think it's fine to merge.https://gitlab.cern.ch/atlas/athena/-/merge_requests/45380Add extra jet radii and truth jets to HION5 output (ATLHI-379)2021-07-24T03:06:03+02:00Jakub KremerAdd extra jet radii and truth jets to HION5 output (ATLHI-379)This is a simple extension of the list of jet collections to store in HION5 output. The aim is to include R=0.6 reco-level jets and R=0.6, 0.8, 1.0 truth-level jets.This is a simple extension of the list of jet collections to store in HION5 output. The aim is to include R=0.6 reco-level jets and R=0.6, 0.8, 1.0 truth-level jets.https://gitlab.cern.ch/atlas/athena/-/merge_requests/45355Fix FTAG physics validation code to avoid the attempt to retrieve nonexistent...2021-07-23T03:05:14+02:00Arnaud DuperrinFix FTAG physics validation code to avoid the attempt to retrieve nonexistent aux data item SMT_mu_linkThis MR fixes the b-tagging physics validation code to avoid the attempt to retrieve nonexistent aux data item SMT_mu_link.
Tagging as urgent since it is needed by physics validation group.
On recent DAOD_PHYSVAL input with p4666, runni...This MR fixes the b-tagging physics validation code to avoid the attempt to retrieve nonexistent aux data item SMT_mu_link.
Tagging as urgent since it is needed by physics validation group.
On recent DAOD_PHYSVAL input with p4666, running the physics validation code (DAOD_PHYSVAL->NTUP_PHYSVAL) with doBtag leads to:
_PhysicsValidation 16:47:17 PhysValMonManager FATAL Standard std::exception is caught
PhysicsValidation 16:47:17 PhysValMonManager ERROR SG::ExcBadAuxVar: Attempt to retrieve nonexistent aux data item `::SMT_mu_link' (882).
PhysicsValidation 16:47:17 AthAlgSeq FATAL Standard std::excep_
SMT is deprecated, the "new" soft muon variables are all called "softMuon_*".
At the moment, SoftMu is turned off in FTAG code.
The fix has been tested on valid1:DAOD_PHYSVAL.26280787._000002.pool.root.1 using 2021-07-21T2101 (compiled with this [MR](https://gitlab.cern.ch/atlas/athena/-/merge_requests/45238)). The ouput on 500 events can be found [here](https://www.cppm.in2p3.fr/~duperrin/WORK/BtaggingValidationPackage/Fix_SoftMuLink/Histo_ttbar_doBTAG_fixMUlink/).
Tagging @jferrand who reported the issue, and @dhayden @juhofer @fbeisieg @cschiavi @vdao @sanmay @dguest for their awareness.https://gitlab.cern.ch/atlas/athena/-/merge_requests/45350TElectronLikelihoodTool rm un-needed member , use unique_ptr, use constexpr2021-07-23T03:05:34+02:00Christos Anastopouloschristos.anastopoulos@cern.chTElectronLikelihoodTool rm un-needed member , use unique_ptr, use constexprTElectronLikelihoodTool rm un-needed member , use unique_ptr, use constexprTElectronLikelihoodTool rm un-needed member , use unique_ptr, use constexprhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45345Adding more DSIDs (New PW tt+bb dipole recoil) for ttHF classification2021-07-23T03:06:15+02:00Andrea Helen KnueAdding more DSIDs (New PW tt+bb dipole recoil) for ttHF classificationI added two DSIDs to the HF classification part of the MCTruthClassifier.I added two DSIDs to the HF classification part of the MCTruthClassifier.https://gitlab.cern.ch/atlas/athena/-/merge_requests/45343ElectronPhotonSelectorTools: close ROOT file after reading2021-07-22T03:04:50+02:00Frank WinklmeierElectronPhotonSelectorTools: close ROOT file after readingClose the DNN ROOT file after reading to release memory and file cache.
For an HLT job where we instantiate multiple `ElectronDNNCalculator`s
this reduces the virtual memory consumption by 800 MB(!).
Relates to ATR-23903.Close the DNN ROOT file after reading to release memory and file cache.
For an HLT job where we instantiate multiple `ElectronDNNCalculator`s
this reduces the virtual memory consumption by 800 MB(!).
Relates to ATR-23903.https://gitlab.cern.ch/atlas/athena/-/merge_requests/45322Add BTaggingBuilderAlg and other b-tagging components2021-07-27T03:03:01+02:00Dan GuestAdd BTaggingBuilderAlg and other b-tagging componentsThis adds a few algorithms to handle the plumbing around b-tagging:
- `BTaggingBuilderAlg`: creates a b-tagging object for each jet and links it to that jet. Useful for lightweight b-tagging.
- `JetToBTagLinkerAlg`: assuming that the l...This adds a few algorithms to handle the plumbing around b-tagging:
- `BTaggingBuilderAlg`: creates a b-tagging object for each jet and links it to that jet. Useful for lightweight b-tagging.
- `JetToBTagLinkerAlg`: assuming that the link from the b-tag object to the jet exist, this creates the link back. Most analysis code follows this jet to b-tag link.
- `BTagToJetLinkerAlg`: reversed version of the above. Since release 22 the b-tag object should always have a link to the jet, but in some older AODs the only link went from the jet to the b-tag.
None of these are used in current reconstruction code, so @duperrin need not worry that they will break everything. The main motivation to have them is to enable lighter b-tagging (i.e. using only DIPS) in the trigger. We should now have the tools to create a (minimal) b-tagging collection, associate tracks to it, and run a simple IP-based tagger. The existing C++ code was too monolithic to support this.https://gitlab.cern.ch/atlas/athena/-/merge_requests/45291Truth Jets: include direct photons like in R212021-07-21T03:02:38+02:00Christian GutschowTruth Jets: include direct photons like in R21Direct photons should be included in the default truth-jet collection (as was the case in R21).
cc @cdelitzs @mswiatlo @jkretzDirect photons should be included in the default truth-jet collection (as was the case in R21).
cc @cdelitzs @mswiatlo @jkretzhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45281After sign-off from MuonCP, enable Tracking CondAlgs by default.2021-08-03T03:27:17+02:00Edward MoyseAfter sign-off from MuonCP, enable Tracking CondAlgs by default.This enables the tracking geometry cond algs by default. I also updated the ref file.
Pinging @jojungge @goetz @mhodgkin @elmsheus
Also:
@gartoni and @mvanadiaThis enables the tracking geometry cond algs by default. I also updated the ref file.
Pinging @jojungge @goetz @mhodgkin @elmsheus
Also:
@gartoni and @mvanadiahttps://gitlab.cern.ch/atlas/athena/-/merge_requests/45250Adding standard truth jets to TRUTH12021-07-28T10:08:33+02:00Sebastian Tapia ArayaAdding standard truth jets to TRUTH1adding standard R=0.2 and R=0.4 truth jets to TRUTH1 following discussion in ATLHI-369adding standard R=0.2 and R=0.4 truth jets to TRUTH1 following discussion in ATLHI-369https://gitlab.cern.ch/atlas/athena/-/merge_requests/45228Add SecVtx and JFVtx btagging Vertex Containers in DAOD_PHYSVAL derivation2021-07-20T03:04:13+02:00Arnaud DuperrinAdd SecVtx and JFVtx btagging Vertex Containers in DAOD_PHYSVAL derivationThis MR add BTagging_AntiKt4EMPFlowJFVtx and BTagging_AntiKt4EMPFlowSecVtx for PFlow jets in DAOD_PHYSVAL derivation.
Tagging @cschiavi @vdao @mguth @fdibello @dguest @sanmay for FTAGThis MR add BTagging_AntiKt4EMPFlowJFVtx and BTagging_AntiKt4EMPFlowSecVtx for PFlow jets in DAOD_PHYSVAL derivation.
Tagging @cschiavi @vdao @mguth @fdibello @dguest @sanmay for FTAG