athena merge requestshttps://gitlab.cern.ch/atlas/athena/-/merge_requests2023-10-06T16:34:13+02:00https://gitlab.cern.ch/atlas/athena/-/merge_requests/65934emergency fix for the errors introduced by DirectInputModule (ATLASG-2602)2023-10-06T16:34:13+02:00Nils Erik Krumnackemergency fix for the errors introduced by DirectInputModule (ATLASG-2602)Well, I guess that the lesson here should be always to write unit
tests when if statements are involved, not just rely on the
integration tests. On th plus side the new infrastructure made it a
lot easier to test this, and I may have fi...Well, I guess that the lesson here should be always to write unit
tests when if statements are involved, not just rely on the
integration tests. On th plus side the new infrastructure made it a
lot easier to test this, and I may have fixed a longer-standing bug
since the previous implementation had never been tested.
Marking as ~urgent since this fixes an issue in CI tests.https://gitlab.cern.ch/atlas/athena/-/merge_requests/63609Analysis-Top: Implement JetUncertainties signed gluon fraction eta option for...2023-06-13T19:20:06+02:00Chris ScheulenAnalysis-Top: Implement JetUncertainties signed gluon fraction eta option for pPb/PbpThis MR implements in AnalysisTop the new change-over from absolute to signed gluon fraction etas for the small-R JetUncertaintiesTool recently developed by JetMET in !63355 for the asymmetric pPb/Pbp HI analysis. As a change, the signed...This MR implements in AnalysisTop the new change-over from absolute to signed gluon fraction etas for the small-R JetUncertaintiesTool recently developed by JetMET in !63355 for the asymmetric pPb/Pbp HI analysis. As a change, the signed etas are now automatically used for pPb/Pbp samples, while the absolute etas are retained for other setups.
The MR addresses a follow-up request in the discussion of ANALYSISTO-1320.Chris ScheulenChris Scheulenhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/63605AnalysisTop: fix loose electron reco SFs and experimental settings2024-02-21T08:32:53+01:00Baptiste Ravinabaptiste.ravina@cern.chAnalysisTop: fix loose electron reco SFs and experimental settingsAs reported by @cnass, we attempt to retrieve a "loose electron reco SF" that doesn't exist. Furthermore, the experimental setting forcing a different electron isolation SF was looking for the wrong decoration (the equivalent code for mu...As reported by @cnass, we attempt to retrieve a "loose electron reco SF" that doesn't exist. Furthermore, the experimental setting forcing a different electron isolation SF was looking for the wrong decoration (the equivalent code for muons is correct).https://gitlab.cern.ch/atlas/athena/-/merge_requests/63355Add signed eta support for gluon fractions in p+Pb collisions2023-06-12T18:01:10+02:00Patrycja Anna PotepaAdd signed eta support for gluon fractions in p+Pb collisionsAdd a flag 'absEtaGluonFraction' that controls if gluon fractions are read for absolute or signed eta. The flag is set to true by default to keep the default behaviour. The support for signed eta is needed in asymmetric p+Pb collisions.
...Add a flag 'absEtaGluonFraction' that controls if gluon fractions are read for absolute or signed eta. The flag is set to true by default to keep the default behaviour. The support for signed eta is needed in asymmetric p+Pb collisions.
Tag @mswiatlo @lrossini @mleblanc @camachohttps://gitlab.cern.ch/atlas/athena/-/merge_requests/62428Harmonise the tZq and ttZ parton histories2023-06-12T17:12:33+02:00Baptiste Ravinabaptiste.ravina@cern.chHarmonise the tZq and ttZ parton historiesChanges are made to two TopPartonHistories within AnalysisTop:
- for the ttZ topology: the decay products of the Z boson are taken immediately after the decay of the Z boson, instead of going down the radiation chain with `findAfterFSR`....Changes are made to two TopPartonHistories within AnalysisTop:
- for the ttZ topology: the decay products of the Z boson are taken immediately after the decay of the Z boson, instead of going down the radiation chain with `findAfterFSR`. This changes the definition of these decay products from "bare" to "Born/ME", similarly to what is done in the ttbar topology.
- for the tZq topology: do the same, except the radiation method there is called `findAfterGamma`. Also remove calls to `findAfterGamma` in the arguments of `TauIsHadronic`, since it itself calls `findAfterFSR` internally.
Tagging @mdubovsk @thsteven @dellasta @alhroob for the ttZ and tZq analyses.https://gitlab.cern.ch/atlas/athena/-/merge_requests/60390Fixing Z jet truth label2023-02-06T17:42:29+01:00Jason Robert VeatchFixing Z jet truth labelAdding a split12 requirement to the R10TruthLabel_R21Precision_2022v1 Z jet truth label definition.
Tagging @jcantero and @brle for their information.Adding a split12 requirement to the R10TruthLabel_R21Precision_2022v1 Z jet truth label definition.
Tagging @jcantero and @brle for their information.https://gitlab.cern.ch/atlas/athena/-/merge_requests/60098AthenaPoolCnvSvc: Make dataHeaderID file-aware to fix DHForm in multi-stream ...2023-01-29T15:43:21+01:00Alaettin Serhan MeteAthenaPoolCnvSvc: Make dataHeaderID file-aware to fix DHForm in multi-stream legacy SharedWriter jobsThis needs to be independently checked/verified but at least it seems to work in my local tests. In a nutshell, by making the `dataHeaderID` file-aware, we avoid corrupting the DHForm when producing multiple streams with the legacy Share...This needs to be independently checked/verified but at least it seems to work in my local tests. In a nutshell, by making the `dataHeaderID` file-aware, we avoid corrupting the DHForm when producing multiple streams with the legacy SharedWriter.Marcin NowakPeter Van GemmerenMarcin Nowakhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/60069change handling of JVT in CP algorithms (ATLASG-2447)2023-02-02T23:16:34+01:00Nils Erik Krumnackchange handling of JVT in CP algorithms (ATLASG-2447)There was a difference between sequence and block configuration in the
nightlies, which caused the test failures in the nightlies. I changed a number of minor things about the jvt configuration:
* add SuppressInputDependence flag to Jet...There was a difference between sequence and block configuration in the
nightlies, which caused the test failures in the nightlies. I changed a number of minor things about the jvt configuration:
* add SuppressInputDependence flag to JetVertexTaggerTool
* disable runJvtUpdate for the test sequence
* don't include the JVT selection in the global jet selection, but introduce a selection named JVT.
The last is indeed what fixed the discrepancy, though I'm unsure if this is the right thing to do, or if instead I ought to have a baseline for all jets that includes JVT. I leave that for the jet experts.
Also reordered configuration blocks to match configuration sequences. This is mostly meant to allow easier comparisons of the generated
sequence to spot differences more easily.https://gitlab.cern.ch/atlas/athena/-/merge_requests/59863Correct space cascade structure2023-01-20T14:59:33+01:00Vadim KostyukhinCorrect space cascade structureCure cases when daughter vertex happened to be in front of mother vertexCure cases when daughter vertex happened to be in front of mother vertexhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/59760Adding support for Hybrid Truth Tagging with GNN2023-01-31T10:42:12+01:00Nilotpal KakatiAdding support for Hybrid Truth Tagging with GNN- Added the additional option of Hybrid Truth Tagging with GNN
- Hybrid Truth tagging = Direct tag the given flavor of jets and Truth tag the rest of the jets in the event.
- It first gets the efficiencies of all the jets the usual way, ...- Added the additional option of Hybrid Truth Tagging with GNN
- Hybrid Truth tagging = Direct tag the given flavor of jets and Truth tag the rest of the jets in the event.
- It first gets the efficiencies of all the jets the usual way, and if we are in the Hybrid tagging mode, it will replace the efficiencies of the jets for the given flavor with the Direct Tagging scores.
- The rest of the computation remains same.
tagging @eschopf @hanar @yama @fdibellohttps://gitlab.cern.ch/atlas/athena/-/merge_requests/59576Significant speed-up to switch systematics in BTagging tools2023-01-20T13:25:17+01:00Nicolas MorangeSignificant speed-up to switch systematics in BTagging toolsTwo changes that can produce significant speedup in xAOD processing codes:
* `applySystematicVariation` was quite slow as it always treated the complicated cases, while it can be much faster for the simple and (common) case where the use...Two changes that can produce significant speedup in xAOD processing codes:
* `applySystematicVariation` was quite slow as it always treated the complicated cases, while it can be much faster for the simple and (common) case where the user is passing the right thing.
* loop on systematics in truth tagging tool was super inefficient as it called `applySystematicVariation` twice per *jet*, always putting syst back to nominal. Now it's the responsibility of the systematics loop to do so.
These changes should be ported to r22, I presume.https://gitlab.cern.ch/atlas/athena/-/merge_requests/59552Fix Cascade vertex fit divergence2023-01-06T18:52:17+01:00Vadim KostyukhinFix Cascade vertex fit divergenceDon't use covariance matrix scaling for cascade vertices made of pseudo-tracks only.Don't use covariance matrix scaling for cascade vertices made of pseudo-tracks only.https://gitlab.cern.ch/atlas/athena/-/merge_requests/59296Bug Fix for external files added to tarball sent to the grid when using `EL::...2023-01-05T19:17:29+01:00Romain BouquetBug Fix for external files added to tarball sent to the grid when using `EL::PrunDriver`I found an issue when someone wants to add external files to be added to the tarball sent to the grid with the `EL::PrunDriver`
It consists in a one line change.
The
```
sample.meta().castString('nc_EventLoop_UserFiles')
```
returns a ...I found an issue when someone wants to add external files to be added to the tarball sent to the grid with the `EL::PrunDriver`
It consists in a one line change.
The
```
sample.meta().castString('nc_EventLoop_UserFiles')
```
returns a string provided by the user that contains the file names to be added to the tarball (additionally to the content of the `build` directory).
And so the call of
```
",".join (sample.meta().castString('nc_EventLoop_UserFiles'))
```
was adding a comma between each character contained in the string.
While we want to add a comma between each file names provided by the user (in case file names are space separated instead of comma separated).
The change is precisely doing that by replacing all space in the `nc_EventLoop_UserFiles` string by comma.
This was causing instability when submitting jobs with the `EL::PrunDriver` when adding external files
(the tarball building was sometimes failing I believe because of that issue)
The changes have been successfully tested with the `CxAODFramework`
Tagging @krumnack on that issue
(And just for information tagging @yama, @eschopf and @hanar)https://gitlab.cern.ch/atlas/athena/-/merge_requests/58552Switch kError to kRange in checkAbsEta2022-12-05T10:40:55+01:00Bingxuan LiuSwitch kError to kRange in checkAbsEtaThis is supposed to fix the uni test failure reported here: http://atlas-computing.web.cern.ch/atlas-computing/links/distDirectory/gitwww/MASTERWebArea/ardoc_web_areamaster64BC7G11AthenaOpt/ARDOC_TestLog_2022-11-20T2101/PhysicsAnalysis_A...This is supposed to fix the uni test failure reported here: http://atlas-computing.web.cern.ch/atlas-computing/links/distDirectory/gitwww/MASTERWebArea/ardoc_web_areamaster64BC7G11AthenaOpt/ARDOC_TestLog_2022-11-20T2101/PhysicsAnalysis_Algorithms_AnalysisAlgorithmsConfig___AnalysisAlgorithmsConfigConf__AnalysisAlgorithmsConfigTest__m.html
Fixes ATLASG-2418.https://gitlab.cern.ch/atlas/athena/-/merge_requests/58444Add ID track radiusOfFirstHit to BPHY202022-11-24T16:21:06+01:00Semen TurchikhinAdd ID track radiusOfFirstHit to BPHY20The MR adds the radiusOfFirstHit variable for `InDetTrackParticles` in BPHY20. The variable is necessary for selecting the conversion photon candidates in the Bc* -> Bc gamma analysis.
Attaching it to ATLBPHYS-115.
cc @tlyubush , @glad...The MR adds the radiusOfFirstHit variable for `InDetTrackParticles` in BPHY20. The variable is necessary for selecting the conversion photon candidates in the Bc* -> Bc gamma analysis.
Attaching it to ATLBPHYS-115.
cc @tlyubush , @gladilinhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/58416Fix support for multiple CA arguments2022-12-05T10:40:55+01:00Tadej Novaktadej.novak@cern.chFix support for multiple CA argumentsFix support for multiple CA arguments.
It seems there's a python limitation that in case of arguments with arbitrary number it's impossible to set the default value when argument is provided without values (i.e. `--CA`). I had to use ac...Fix support for multiple CA arguments.
It seems there's a python limitation that in case of arguments with arbitrary number it's impossible to set the default value when argument is provided without values (i.e. `--CA`). I had to use actions which do the full parsing so I created `argAction` class and `argActionFactory` class.
/cc @nstyles @jcatmore @christos @mhodgkinhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/58130ATLASRECTS-7350,ATLASRECTS-7351 Fix mem leak and WARNING->DEBUG2022-11-08T14:30:19+01:00Christos Anastopouloschristos.anastopoulos@cern.chATLASRECTS-7350,ATLASRECTS-7351 Fix mem leak and WARNING->DEBUGATLASRECTS-7350,ATLASRECTS-7351
Fix mem leak and WARNING->DEBUG
ping @tamartin , @amete , @sroeATLASRECTS-7350,ATLASRECTS-7351
Fix mem leak and WARNING->DEBUG
ping @tamartin , @amete , @sroehttps://gitlab.cern.ch/atlas/athena/-/merge_requests/57838Fixing missing variables in CA version of PHYS2022-10-25T21:42:49+02:00James CatmoreFixing missing variables in CA version of PHYSATLASDPD-1779
This fixes an issue where some variables needed for data and MC were put in the MC-only cateogory in PHYS
In the migration of DAOD_PHYS to CA we did a clean up of variables that are only found in MC, to reduce the number ...ATLASDPD-1779
This fixes an issue where some variables needed for data and MC were put in the MC-only cateogory in PHYS
In the migration of DAOD_PHYS to CA we did a clean up of variables that are only found in MC, to reduce the number of warnings. Unfortunately some variables that are also needed in data were accidentally put behind the isMC if statement, leading to missing variables from data. The variable comparison tests were done with MC so this was missed.
This is fixed in this MR. Marked urgent because no further PHYS production can be done on data without this fix.https://gitlab.cern.ch/atlas/athena/-/merge_requests/57751eflowRec and DerivationFramework: fix useMuonTopoClusters configuration2022-10-24T16:00:02+02:00Bertrand Martin Dit LatoureflowRec and DerivationFramework: fix useMuonTopoClusters configurationHello,
This MR is fixing a bug introduced in !57581. The `UseMuonTopoClusters` property of the `PFMuonFlowElementAssoc` algorithm was steered in the python configuration by arguments either called `UseMuonTopoClusters` or `useMuonTopoCl...Hello,
This MR is fixing a bug introduced in !57581. The `UseMuonTopoClusters` property of the `PFMuonFlowElementAssoc` algorithm was steered in the python configuration by arguments either called `UseMuonTopoClusters` or `useMuonTopoClusters`, and only the latter was properly treated. In this MR, we now use consistently `useMuonTopoClusters` everywhere. I also renamed the algorithm property to avoid confusion.
Adding the urgent flag as it fixes a bug in the DAOD production.
Tagging @mhodgkin .
Cheers,
Bertrandhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/57743VrtSecInclusive: Fix bug in calculation of nPixelLayers in secondary vertexing2022-11-15T16:23:39+01:00Zachary PollockVrtSecInclusive: Fix bug in calculation of nPixelLayers in secondary vertexingFixes a bug in the VrtSecInclusive, when selecting if a secondary vertex passes fake rejection based on the number of hits in pixel layers. Previously, nPixelLayers was calculated incorrectly by adding 2^(layer index) when a hit is found...Fixes a bug in the VrtSecInclusive, when selecting if a secondary vertex passes fake rejection based on the number of hits in pixel layers. Previously, nPixelLayers was calculated incorrectly by adding 2^(layer index) when a hit is found instead of simply incrementing.
The VrtSecInclusive algorithm runs at the derivation level, so I ran a test job on lxplus redoing the DAOD_RPVLL to SUSY15 derivation. The job ran with no issues.
This bug is present in both r21 and r22. For r22, this should also be target master. I'm not sure where this should be targeting for r21. VrtSecInclusive is run at both the reconstruction and derivation step, so should it target both 21.0 and 21.2?