athena merge requestshttps://gitlab.cern.ch/atlas/athena/-/merge_requests2023-09-11T10:27:23+02:00https://gitlab.cern.ch/atlas/athena/-/merge_requests/47403TriggerJobOpts: lock ConfigFlags in runHLT_standalone2023-09-11T10:27:23+02:00Frank WinklmeierTriggerJobOpts: lock ConfigFlags in runHLT_standaloneLock the `ConfigFlags` in `runHLT_standalone.py`. Had to re-arrange some parts of the job options to get this to work. Would ideally like to lock them even earlier, but that will need some changes to the ID trigger configuration.
Also h...Lock the `ConfigFlags` in `runHLT_standalone.py`. Had to re-arrange some parts of the job options to get this to work. Would ideally like to lock them even earlier, but that will need some changes to the ID trigger configuration.
Also had to make the following change to `MuonCablingConfig.py`:
- Do not modify `ConfigFlags` during configuration. They need to be set
correctly already earlier. Also re-arrange the trigger-related logic
such that the `doLVL1` flag is accessed last to avoid lookups of the
input file in the new-style auto-configuration for jobs where this may
not be set.
This is in preparation for atlas/athena!47186 that will essentially forbid the use of unlocked flags.
cc @rbielskihttps://gitlab.cern.ch/atlas/athena/-/merge_requests/61393Remove the ActsATLASConverterTool and update clients to use ActsToTrkConverte...2023-09-11T10:07:14+02:00Edward MoyseRemove the ActsATLASConverterTool and update clients to use ActsToTrkConverterToolRemove the ActsATLASConverterTool and update clients to use ActsToTrkConverterTool (and the changed API).
Tested locally with `make test` (if ACTS experts in cc can suggest other tests, I'm happy to run them - for instance, I don't thin...Remove the ActsATLASConverterTool and update clients to use ActsToTrkConverterTool (and the changed API).
Tested locally with `make test` (if ACTS experts in cc can suggest other tests, I'm happy to run them - for instance, I don't think any tests I ran actually execute the clients).
None of this code is run in production, so it will not change any output
This needs !61181 to be merged first.
cc @adye @ncalace @toyamaza @cvarni @pagessinhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/57320allow disabling systematics for EgammaCalibrationAndSmearingAlg (for ATLASG-2...2023-09-05T21:56:44+02:00Nils Erik Krumnackallow disabling systematics for EgammaCalibrationAndSmearingAlg (for ATLASG-2358)This allows to schedule it twice, once without systematics and then
again just for systematics.This allows to schedule it twice, once without systematics and then
again just for systematics.https://gitlab.cern.ch/atlas/athena/-/merge_requests/59629Update of CA config for TRT reco2023-08-23T09:44:26+02:00Thomas StreblerUpdate of CA config for TRT recoEpisode 2 Season 2 of updates of tracking CA configs, moving together the configs of the same C++ tools.
This MR:
- introduces dedicated config files for the `TRT_SegmentsToTrackTool` and `TRT_StandaloneTrackFinder` packages
- updates t...Episode 2 Season 2 of updates of tracking CA configs, moving together the configs of the same C++ tools.
This MR:
- introduces dedicated config files for the `TRT_SegmentsToTrackTool` and `TRT_StandaloneTrackFinder` packages
- updates the TRT configurations to avoid the use of the `extension` argument, in favour of well-identified config methods for the different use cases
- introduces minor updates to the main `TrackRecoConfig` relevant for the cosmics reconstruction
Checked to leave unchanged the outputs of the Run-3 q445 reco tests.https://gitlab.cern.ch/atlas/athena/-/merge_requests/63839Adding 422 clusters for GEP emulation2023-08-22T18:34:35+02:00Dylan Sheldon RankinAdding 422 clusters for GEP emulationThis MR creates a config in the `TrigGepPerf` package to add 422 clusters. It can be used as a `postInclude` with `Reco_tf.py`.
Tested with an RDO from mc21: `Reco_tf.py --CA 'all:True' --autoConfiguration 'everything' --conditionsTag ...This MR creates a config in the `TrigGepPerf` package to add 422 clusters. It can be used as a `postInclude` with `Reco_tf.py`.
Tested with an RDO from mc21: `Reco_tf.py --CA 'all:True' --autoConfiguration 'everything' --conditionsTag 'all:OFLCOND-MC15c-SDR-14-05' --geometryVersion 'all:ATLAS-P2-RUN4-03-00-00' --multithreaded 'True' --preInclude 'all:Campaigns.PhaseIINoPileUp' --steering 'doRAWtoALL' --inputRDOFile ../../../../mc21_14TeV.601229.PhPy8EG_A14_ttbar_hdamp258p75_SingleLep.recon.RDO.e8481_s4149_r14697/RDO.33629009._000002.pool.root.1 --outputAODFile aod.root --postInclude 'all:PyJobTransforms.UseFrontier' 'TrigGepPerf.Add422Config.Add422Cfg'`
@htorre @tstreble @mswiatlo
Previous discussion in ATLSWUPGR-194https://gitlab.cern.ch/atlas/athena/-/merge_requests/41402Update property declaration syntax for components in BeamEffects package2023-08-21T20:59:11+02:00John Derek ChapmanUpdate property declaration syntax for components in BeamEffects packageMoving from `declareProperty(...)` calls in component constructors to fully declaring properties in the class declarations in component headers.Moving from `declareProperty(...)` calls in component constructors to fully declaring properties in the class declarations in component headers.https://gitlab.cern.ch/atlas/athena/-/merge_requests/59440athena.py: argparse-based command line parser2023-08-21T17:49:20+02:00Frank Winklmeierathena.py: argparse-based command line parserRe-implement the command line parser with `arparse`. It is fully
backwards compatible. Most notable changes are:
- group options
- hide expert options (show them with `--help=all`)
- remove `--leak-check-execute` (use `--leak-check=execu...Re-implement the command line parser with `arparse`. It is fully
backwards compatible. Most notable changes are:
- group options
- hide expert options (show them with `--help=all`)
- remove `--leak-check-execute` (use `--leak-check=execute`)
The goal here is to make the parser re-usable and compatible with CA-based jobs.https://gitlab.cern.ch/atlas/athena/-/merge_requests/59221FTag training run 4 in derivation2023-08-21T16:13:29+02:00Thomas StreblerFTag training run 4 in derivationThis MR:
- renames `BTagRun3Config.py` into `BTagConfig.py`
- renames `FtagRun3DerivationConfig.py` into `FtagDerivationConfig.py`
- forward the CA config flags through the `SlimmingHelper` to be used to configure FTAG content + update a...This MR:
- renames `BTagRun3Config.py` into `BTagConfig.py`
- renames `FtagRun3DerivationConfig.py` into `FtagDerivationConfig.py`
- forward the CA config flags through the `SlimmingHelper` to be used to configure FTAG content + update accordingly the CA configs for all derivation formats available
- adds configuration to run Run 4 FTAG trainings + store associated variables in derivation
FYI @duperrin @bdong @ligang @sgoswami @khanovhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/46853TrkTrack+TrkEventTPCnv: Hold TSOS vector by pointer rather than by value.2023-08-19T23:10:04+02:00Scott SnyderTrkTrack+TrkEventTPCnv: Hold TSOS vector by pointer rather than by value.Allows giving track something derived from DataVector<TSOS>, rather than that class
exactly.
Temporary(?) change to try to track down ATLASRECTS-6331.Allows giving track something derived from DataVector<TSOS>, rather than that class
exactly.
Temporary(?) change to try to track down ATLASRECTS-6331.https://gitlab.cern.ch/atlas/athena/-/merge_requests/52412Adding fast FTagging on PFlow jets (for TLA)2023-08-17T20:51:42+02:00Stefano FranchellucciAdding fast FTagging on PFlow jets (for TLA)#### Main purpose
This MR is adding the steps for scheduling "fast" flavour tagging on PFlow jets as a simple modifier of the jets, without requesting any additional Hypo testing. The idea is to attach the b-tagger information on PFlow ...#### Main purpose
This MR is adding the steps for scheduling "fast" flavour tagging on PFlow jets as a simple modifier of the jets, without requesting any additional Hypo testing. The idea is to attach the b-tagger information on PFlow jets for TLA. (ATR-23560)
Adding @cantel.
Further details:
* Only fully calibrated PFlow jets are fast b-tagged.
* A flag is added `Trigger.Jet.fastbtagPFlow` (default `True`): If disabled, turns OFF fast b-tagging for PFlow jets.
* Shrinking cone track association is used.
* Configs for fixed cone association are added but not used here (considered as alternative, as network training was performed with fixed cone association)
* 2 neural networks are run: network (a) trained on b-jet precision tracks, network (b) trained on FS fast tracks.
* Additional jet info is added to the EDM: associated track links (for validation), b-tag outputs.
* Added to `TriggerEDMRun3`
* Added to `TriggerEDMAuxAccessors.h`
* Required modification in `TriggerEDMCLIDs.h` to save track links
#### Additional bug fix for b-jet presel jets
* fast b-tag outputs recorded for the preselection calo jets
* associated tracks for fast b-tag recorded for the preselection calo jets, for validation/commissioning.
* As noted above: Required modification of `TriggerEDMCLIDs.h`.https://gitlab.cern.ch/atlas/athena/-/merge_requests/49919Getting dips selection into jet hypo2023-08-17T20:51:42+02:00Marco ValenteGetting dips selection into jet hypoThis includes the necessary modifications to apply a dips selection starting from the decorations applied to EMTopo trigger jets in https://gitlab.cern.ch/atlas/athena/-/merge_requests/49668. The necessary steps to add this selection int...This includes the necessary modifications to apply a dips selection starting from the decorations applied to EMTopo trigger jets in https://gitlab.cern.ch/atlas/athena/-/merge_requests/49668. The necessary steps to add this selection into a jet trigger pre-selection will come later in another MR.Marco ValenteMarco Valentehttps://gitlab.cern.ch/atlas/athena/-/merge_requests/58812add V0 container to FTAG1 from BPHY102023-08-16T19:08:42+02:00Ligang Xiaadd V0 container to FTAG1 from BPHY10
- add V0 container for training a charm tagger
- add a flag to turn on this V0 finder
Here are the testing samples (tag @wleinone ):
- /afs/cern.ch/user/l/ligang/public/FTag_R22/ftag1_V0_v2/
jira: AFT-611
- add V0 container for training a charm tagger
- add a flag to turn on this V0 finder
Here are the testing samples (tag @wleinone ):
- /afs/cern.ch/user/l/ligang/public/FTag_R22/ftag1_V0_v2/
jira: AFT-611https://gitlab.cern.ch/atlas/athena/-/merge_requests/50644Avoid division by zero if line surface and track parameters are parallel.2023-08-08T16:40:45+02:00Goetz GayckenAvoid division by zero if line surface and track parameters are parallel.Addresses ATLASRECTS-6841.Addresses ATLASRECTS-6841.https://gitlab.cern.ch/atlas/athena/-/merge_requests/57869pickle-support with athena --CA2023-08-04T18:55:34+02:00Walter Lamplpickle-support with athena --CAThis MR brings support for the `--config-only` argument and running from a pickled ComponentAccumulator to `athena --CA`
With this, one can dump the result of the configuration process to a pickle file like
`athena --CA --config-only=m...This MR brings support for the `--config-only` argument and running from a pickled ComponentAccumulator to `athena --CA`
With this, one can dump the result of the configuration process to a pickle file like
`athena --CA --config-only=myConfig.pkl favoriteConfig.py`
and the run from the pickle file:
`athena --CA myConfig.pkl`
This is achieved by intercepting the `ComponentAccumulator.run()` method via the environment variable `PICKLECAFILE`. If this variable is set, the run method actually doesn't run but stores the `ComponentAccumlator` to this file.https://gitlab.cern.ch/atlas/athena/-/merge_requests/63754First steps to use ITk geometry in GPU accelerated FTF2023-08-02T08:49:53+02:00Aleksandra PorebaFirst steps to use ITk geometry in GPU accelerated FTF- Parse volume and layer disk id to ITk geometry
- Pass ITk global layer ID to SILICON_LAYER.m_subdet
- Update the sizes of GPU accelerated track setting EDM:
```
MAX_SILICON_LAYERS 50 -> 216
MAX_NUMBER_PIX_MODULES 2100 -> 6300 (3 times ...- Parse volume and layer disk id to ITk geometry
- Pass ITk global layer ID to SILICON_LAYER.m_subdet
- Update the sizes of GPU accelerated track setting EDM:
```
MAX_SILICON_LAYERS 50 -> 216
MAX_NUMBER_PIX_MODULES 2100 -> 6300 (3 times more)
MAX_NUMBER_SCT_MODULES 8200 -> 24600 (3 times more)
MAX_NUMBER_SPACEPOINTS 100000 -> 300000 (3 times more)
```
- The new geometry can be toggled by setting the property TrigInDetAccelerationSvc.useITkGeometry
Next steps: evaluate the current performance, remove dependency in Strips readout
Tested with:
```
Reco_tf.py --maxEvents=100 --CA --inputRDOFile /afs/cern.ch/work/a/aporeba/ph2-input/mc21_14TeV/RDO.32628466._000001.pool.root.1 --outputAODFile AOD.root --steering doRAWtoALL --preInclude InDetConfig.ConfigurationHelpers.OnlyTrackingPreInclude --preExec "flags.Tracking.useITkFTF=True " 2>&1 | tee reco.log
runIDPVM.py --filesInput AOD.root --doTightPrimary --doExpertPlots --truthMinPt 1000 --outputFile IDPVM.root > output_runIDPVM.log 2>&1
```
Related to EFTRACK-76
Tagging @demelian @bwynne @markhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/63822Update cosmic hanconfig for ID trigger2023-07-26T22:50:04+02:00Julie KirkUpdate cosmic hanconfig for ID triggerupdate cosmic hanconfig for ID triggerupdate cosmic hanconfig for ID triggerhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/36911NewCovarianceMatrix, simplify and use Eigen2023-07-18T21:57:26+02:00Christos Anastopouloschristos.anastopoulos@cern.chNewCovarianceMatrix, simplify and use Eigen- Use Eigen operations
- It should be a bit faster.
- But makes also the code a bit simpler and easier to read .
- A quick test https://godbolt.org/z/G48xbP
- New math alone https://godbolt.org/z/jxofxj
- Old math alone https://godbolt...- Use Eigen operations
- It should be a bit faster.
- But makes also the code a bit simpler and easier to read .
- A quick test https://godbolt.org/z/G48xbP
- New math alone https://godbolt.org/z/jxofxj
- Old math alone https://godbolt.org/z/eo3oT6
I have seen changes at the exact binary level,
kind of expected given a slightly different order of operations
between doing things 100% explicit and calling Eigen method.
An example from my local tests :
```
019.GSFConversionVerticesAux.covariance.1.40 -1.1598064020290622e-06 -> -1.1598065157158999e-06 => diff= [-0.00000245%]
```
So I expect some minor changes
Mentiong @npetters @goblirsc , @amorley , @ssnyderhttps://gitlab.cern.ch/atlas/athena/-/merge_requests/49003[ATR-24689] Change the trigger menu metadata merging strategy2023-07-17T15:43:47+02:00Tim Martin[ATR-24689] Change the trigger menu metadata merging strategyChange the trigger menu metadata merging strategy to double-merge (both input file -> internal store, and internal store -> output file)
This is to resolve an issue in MP AOD->DAOD processing in release 22 where some JSON were missing f...Change the trigger menu metadata merging strategy to double-merge (both input file -> internal store, and internal store -> output file)
This is to resolve an issue in MP AOD->DAOD processing in release 22 where some JSON were missing from the output file due to the lack of this second merging pass.
A global mutex is put in just for safety.https://gitlab.cern.ch/atlas/athena/-/merge_requests/63124electron-to-photon fake background estimation2023-07-17T15:43:20+02:00Yong Zhouelectron-to-photon fake background estimationAdd support for electron fake photon background estimation in FakeBkgTool.
tag @ligang
New MR !62954Add support for electron fake photon background estimation in FakeBkgTool.
tag @ligang
New MR !62954https://gitlab.cern.ch/atlas/athena/-/merge_requests/64125Bump Acts Tag to v27.1.02023-07-17T15:42:44+02:00Carlo Varnicarlo.varni@cern.chBump Acts Tag to v27.1.0This bumps ACTS from v27.0.0 to v27.1.0.
Release notes of in-between versions:
* [v27.1.0](https://github.com/acts-project/acts/releases/tag/v27.1.0)
Validation ticket: ATLIDTRKCP-474This bumps ACTS from v27.0.0 to v27.1.0.
Release notes of in-between versions:
* [v27.1.0](https://github.com/acts-project/acts/releases/tag/v27.1.0)
Validation ticket: ATLIDTRKCP-474