GitLab

It is a widespread and reasonable, but incorrect, pattern to assume that if a handle(key) is not used then it should not be initialized.
In fact, this leaves the data dependency active in the scheduler, and no error will be thrown if the handle is never accessed.
Instead, the handle should be initialized(false) - this MR makes the change

Hello,

This MR is removing a deprecated feature in the tau-track association.
In the old days (non-MT R22 trigger), the track particle container was retrieved elsewhere (TrigTauRec) and passed as an argument to the TauTrackFinder execute method to build the tau tracks.
Now, the track particle container is retrieved from within TauTrackFinder, and we no longer need to pass it as an argument.

The reconstruction output is unchanged.

Cheers,
Bertrand

Hello,

This MR is a follow-up of !44662 that introduced ghost matching in the tau-track association procedure (ATLTAU-1442).
Ghost matching is now activated by default in the offline tau reconstruction (won't be used in the trigger).

In addition, tracks beyond R=0.25 of the tau axis and which are not ghost matched are now kept if the jet closest to the track is the seed jet of the current tau.
This condition was introduced to recover tau-track association efficiency for low-pt 3p taus, where tracks from the tau decay are not always ghost-matched with the seed jet.
This scheme will be used for the R22 reprocessing after it is carefully validated.

Sorry for adding the urgent flag again, but it would be great to have this MR reviewed and merged quickly, as it will be needed for the final Round of algorithm tuning for the R22 reprocessing.

Cheers and thanks,
Bertrand

Hello,

This MR is implementing a hybrid tau-track association that accepts all tracks within a cone of radius 0.25, and beyond 0.25, tracks are required to be ghost-matched with the seed jet.
The ghost matching requirement prevents from using tracks from another tau (a situation that leads to very problematic behaviour of tau substructure reconstruction, ATLTAU-1442).
The motivations are explained in [this talk](https://indico.cern.ch/event/1051117/contributions/4420815/attachments/2269038/3853037/TauCP_Martin_22Jun21.pdf)

The hybrid cone/ghost scheme is meant to replace the R=0.4 cone association, but it's currently disabled in tauRec.

I am adding the urgent flag because we need to validate the new scheme urgently and use it for upcoming retunings of tau algorithms for the R22 reprocessing.

Cheers,
Bertrand

Antonio De Maria authored Jun 18, 2021 and Frank Winklmeier committed Jun 18, 2021

use HLT tau vertex instead of beamspot in the TauTrackFinder + adding HypoTool monitoring histograms

Hello,

This MR is just making some trivial changes. It does not change the reconstruction output.

Cheers,
Bertrand

Bertrand Martin Dit Latour authored Mar 24, 2021 and Adam Edward Barton committed Mar 24, 2021

Hello,

As discussed in ATLASRECTS-6106 and ATLASRECTS-6104, we make an unnecessary and inappropriate use of Decorator on non-const objects, and we should have used Accessor instead.
Now fixed in TauTrackFinder.
Note for later: this could be a bit more streamlined by using the TauTrack::setDetail functionality of the EDM, extending the TrackDetail enum in TauDefs.h.

This MR does not change the reconstruction output.

Cheers,
Bertrand

Christos Anastopoulos authored Mar 23, 2021 and Adam Edward Barton committed Mar 23, 2021

ATLASRECTS-6104 . Replace un-needed Decorators with Accessors. As input is const. Also qualify auto, use empty, use nullptr

Bertrand Martin Dit Latour authored Feb 01, 2021 and Walter Lampl committed Feb 01, 2021

Hello,

This MR is adding a tau track flag to identify large radius tracks more easily, when LRTs are enabled in the tau reconstruction.
We can now retrieve specifically the LRTs associated with the tau via tau->tracks(xAOD::TauJetParameters::TauTrackFlag::LargeRadiusTrack).
This flag is agnostic w.r.t. the track classification, it is only used to tag LRTs, which will be useful e.g. to count LRTs within 0.4 of the tau axis.

As LRTs are not used by default, this MR does not change the reconstruction output.

Cheers,
Bertrand

Hello,

This MR is mostly importing !26767 from 21.2 to master, with the original MR description pasted below.

A protection is added against invalid TauTrack links when retrieving the tau tracks or the number of tau tracks.
This is required to support tau track thinning, used e.g. in DAOD_PHYS.
Now, only the tau tracks that have a valid link are considered in the TauJets interface class.
In particular, the behaviour of nAllTracks() had to be changed to be consistent with allTracks().size().
If one only keeps e.g. classifiedCharged tau tracks by thinning away the other tracks, nAllTracks() will now return the number of classifiedCharged tracks.

Something new compared to the R21 MR: to keep track of the total number of tracks associated with the tau in the reconstruction, a new Aux variable is introduced (mapped to xAOD::TauJetParameters::nAllTracks). So the user is now able to tell if tracks were thinned in the AOD->DAOD step, by comparing tau->nAllTracks() to tau->detail(xAOD::TauJetParameters::nAllTracks, nTracksTot).

Cheers,
Bertrand

Guillermo Nicolas Hamity authored Oct 26, 2020 and Frank Winklmeier committed Oct 26, 2020

LRT track container included in TauTrackFinder
Using InDetLargeD0TrackParticles as LRT container
name will make this happen

Bertrand Martin Dit Latour authored Oct 05, 2020 and Vakhtang Tsulaia committed Oct 05, 2020

Hello,

This MR is harmonising the use of tau track IP and IP significance parameters across the tau reconstruction.
We now use everywhere the impact parameters w.r.t. the tau vertex instead of w.r.t. beam line (relevant for d0).
New methods are introduced to retrieve the track IP and IP significances directly from the tau track (before, we would use e.g. d0 from the associated TrackParticle, i.e. d0 w.r.t. the beam line).

Because we need track IPs early in the reconstruction (right after tau vertex finding and tau track association, and before tau track classification), I had to move the track IP calculation from TauVertexVariables to TauTrackFinder, as TauVertexVariables is scheduled late in the reconstruction chain.

Two variables have been discontinued: TauJetsAux.ipSigLeadTrk TauJetsAux.ipZ0SinThetaSigLeadTrk. These are track IPs defined for the leading track. Now that we compute these for all tau tracks, we can drop these "leading track" variables.

Using standalone tau reconstruction from ESD, I have checked that the tau reconstruction output is unchanged if I fall back to the "old" calculations for track IPs, i.e. if I set d0TJVA = track()->d0(), and z0sinthetaTJVA = z0sinThetaTJVA(tau) in Track_v1.
Meaning, this MR is mostly code reorganisation. The only modification that changes the reconstruction is the new definition of d0TJVA and z0sinthetaTJVA.
With these new definitions, we get slight changes on the BDT ID score and the RNN ID score, as well as the track classification scores, as expected.

Cheers,
Bertrand

Xiaozhong Huang authored Sep 22, 2020 and Edward Moyse committed Sep 22, 2020

The vertex correction is not applied consistently in the tau
reconstruction. In addition, it is applied to the four momentum already
corrected by PV0, indicating an over correction.

Ideally, we could apply the vertex correction to the four momentum
before the origin correction in jet reconstruction. But it is not
straight forward to obtain this information. Alternatively, we
apply the vertex correction relative to the jet vertex to fix
this bug.

Bertrand Martin Dit Latour authored Sep 09, 2020 and Vakhtang Tsulaia committed Sep 09, 2020

Hello,

This MR is dropping a variable that is not used anywhere (anymore) in the tau reconstruction.
We now use a Et-resolution weighted average between calo and substructure pt as input to the MVA pt, no longer an interpolation.
I have checked it does not change the reconstruction output, besides this variable that is now gone.
Minor code cleanup was done.

Cheers,
Bertrand

Pass tau track container to track tools via execute function instead of a link from the tau. New functions created in base classes. Updated if statements in tau algorithms to call correct execute functions. Updated TauTrackRNN calibration file.

Fixes ubsan warning.

Xiaozhong Huang authored Jan 30, 2020 and Adam Edward Barton committed Jan 30, 2020

For now, the configuration of each tauRecTool are set in tauRecFlags.py
and TauAlgorithmsHolder.py. The configuration files are not used for a
long time, thus removed.

Use ATH_CHECK when possible, and remove some commented codes which do not help understand the logic.

Xiaozhong Huang authored Jan 24, 2020 and Vakhtang Tsulaia committed Jan 24, 2020

The property of container name are replaced by ReadHandleKey, thus not
used any more. They are now removed.

We use three diferent property ( "CellCorrection", "AxisCorrection",
and "VertexCorrection") to decide whether to perform the same operation,
e.g. tau vertex correction of the cell. Now, only "VertexCorrection" is
used.

Edson Carquin Lopez authored Dec 18, 2019 and Edward Moyse committed Dec 18, 2019

Fix bug in RNN chain and with legacy code tests, get rid of some gaudi props in the mean time and harmonize a bit their naming, other cleanings applied

Basic fixes for tau trigger. Updated some tools in tau trig alg holder, reverted some tools to retrieve objects using old method for trigger. Also added line to trig tau alg to use new track finder tool


Former-commit-id: ed3a4678

Former-commit-id: 6ecc8afa

Former-commit-id: c3c564d6

Added data handles to TauTrackFinder. Required slight change to the processor alg to allow the track link to be retrieved in the tool


Former-commit-id: bc393612

	* Tagging 00-02-03
	* Fix FPE CombinedP4, don't set a TLV with eta=-1111

2017-02-09 Justin Griffiths <griffith@cern.ch>
	* Tagging 00-02-02
	* Fix bug in TauCommonCalcVars I recently introduced (IntermediateAxis is invalid in trigger)

2017-02-09 Justin Griffiths <griffith@cern.ch>
	* Tagging 00-02-01
	* CombinedP4: downgrade warnings

2017-02-08 Will Davey <will.davey@cern.ch>
	* TauWPDecorator.cxx: bugfix boundaries in score flattening

2017-02-08 Justin Griffiths <griffith@cern.ch>
	* Moving tools from TauDiscriminant to here: TauIDVarCalculator and TauEleOLRDecorator
	* Setting all wide tracks as isolation and modifiedIsolation in TauTrackFinder, bits
	properly set in TauTrackClassifier--done for benefit of trigger
	* update CMakeLists, requirements, Makefile.RootCore including xAODEgamma
	* update MvaEvauator from Terry (use combined pt)
...
(Long ChangeLog diff - truncated)


Former-commit-id: ce806001