Stop cause and ccctf #12576
Stop cause and ccctf #12576
Conversation
Each TrajectoryFilter has been equipped in such a way to inject into the Trajectory or TempTrajectory the reason that stopped the propagation of the trajectory during the pattern recognition. This work is in preparation of future work that will propagate this information down the chain directly into a reco::TrackBase, so that more detailed studies can be made on the reasons of the stop.
In order to better investigate the effects of various filters during pattern recognition, a new data member has been added to the TrackBase class. That, as a consecuence, triggered the increase in the classVersion of its derived classes, which is also included into this commit. Currently the information is stored an uint8_t in TrackBase, but could be translated into a more proper enum class type as soon as ROOT6 will digest it.
The logic of the filter has been wrongly changed(inverted) in a previous commit. The correct behaviour is restored in this PR.
The main enum class has been moved to the DataFormats folder, so that it is handy available to other DataFormats products that need to use it. Also, every trajectory is now initialized with the reserved value UNINITIALIZED, so that we can better monitor all the cases in which the trajectory has not been stopped by any filter being in any case promoted to a track. All other stopping reasons have been accordingly shifted.
The CCC has been implemented as a TrajectoryFilter, in order to give more flexibility to the user in selecting possible bad (from the Cluster Charge point of view) hits. The counter of the number of the hits that would have been removed by the CCC if applied as an estimator is left as a paramter, whose current values is 3 (i.e. we tolerate up to 2 bad hits, after which we discard the trajectory). The final trajectory, if discarded due to eccessive number of bad CCC hits, will anyhow contain the would-have-been removed 2 hits, since they passed all the other estimators. The user has the possibility to disentangle lost hits from bad for CCC and apply different criteria with different thresholds during Patter Recognition and at filtering level. In order to assess the validity of this PR it would be better to turn completely of the CCC at estimator level.
In the previous implementation, the counting of the hits that should
be considered as bad for CCC was done inside the filter itself, every
time looping on all TrajectoryMeasurements associated to the
Trajectory. With this improved implentation the counting is done
directly while propagating the Trajectory. The number of bad hits for
CCC is locally cached in the {Temp,}Trajectory and updated only with
the measurements of the appended segments.
Conflicts: TrackingTools/TrajectoryFiltering/interface/CkfBaseTrajectoryFilter.h
|
A new Pull Request was created by @rovere (Marco Rovere) for CMSSW_8_0_X. It involves the following packages: DataFormats/GsfTrackReco @cmsbuild, @cvuosalo, @davidlange6, @slava77 can you please review it and eventually sign? Thanks. Following commands in first line of a comment are recognized
|
|
@cmsbuild, please test |
|
The tests are being triggered in jenkins. |
|
-1 runTheMatrix-results/134.911_RunSinglePh2015D+RunSinglePh2015D+HLTDR2_25ns+RECODR2_25nsreHLT+HARVESTDR2_25nsreHLT/step2_RunSinglePh2015D+RunSinglePh2015D+HLTDR2_25ns+RECODR2_25nsreHLT+HARVESTDR2_25nsreHLT.log ----- Begin Fatal Exception 26-Nov-2015 18:07:36 CET----------------------- An exception of category 'Configuration' occurred while [0] Constructing the EventProcessor [1] Constructing module: class=CkfTrajectoryMaker label='hltL3TrackCandidateFromL2OIState' Exception Message: MissingParameter: Parameter 'maxCCCLostHits' not found. ----- End Fatal Exception ------------------------------------------------- 135.4 step1 runTheMatrix-results/135.4_ZEE_13+ZEEFS_13+HARVESTUP15FS+MINIAODMCUP15FS/step1_ZEE_13+ZEEFS_13+HARVESTUP15FS+MINIAODMCUP15FS.log ----- Begin Fatal Exception 26-Nov-2015 18:13:06 CET----------------------- An exception of category 'Configuration' occurred while [0] Constructing the EventProcessor [1] Constructing module: class=CkfTrackCandidateMaker label='hltMu8Ele8CkfTrackCandidateMaker' Exception Message: MissingParameter: Parameter 'maxCCCLostHits' not found. ----- End Fatal Exception ------------------------------------------------- 1306.0 step2 runTheMatrix-results/1306.0_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15/step2_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15.log ----- Begin Fatal Exception 26-Nov-2015 18:14:55 CET----------------------- An exception of category 'Configuration' occurred while [0] Constructing the EventProcessor [1] Constructing module: class=CkfTrajectoryMaker label='hltL3TrackCandidateFromL2OIState' Exception Message: MissingParameter: Parameter 'maxCCCLostHits' not found. ----- End Fatal Exception ------------------------------------------------- 1330.0 step2 runTheMatrix-results/1330.0_ZMM_13+ZMM_13+DIGIUP15+RECOUP15+HARVESTUP15/step2_ZMM_13+ZMM_13+DIGIUP15+RECOUP15+HARVESTUP15.log ----- Begin Fatal Exception 26-Nov-2015 18:19:55 CET----------------------- An exception of category 'Configuration' occurred while [0] Constructing the EventProcessor [1] Constructing module: class=CkfTrajectoryMaker label='hltL3TrackCandidateFromL2OIState' Exception Message: MissingParameter: Parameter 'maxCCCLostHits' not found. ----- End Fatal Exception ------------------------------------------------- 25202.0 step2 runTheMatrix-results/25202.0_TTbar_13+TTbar_13+DIGIUP15_PU25+RECOUP15_PU25+HARVESTUP15_PU25/step2_TTbar_13+TTbar_13+DIGIUP15_PU25+RECOUP15_PU25+HARVESTUP15_PU25.log ----- Begin Fatal Exception 26-Nov-2015 18:27:37 CET----------------------- An exception of category 'Configuration' occurred while [0] Constructing the EventProcessor [1] Constructing module: class=CkfTrajectoryMaker label='hltL3TrackCandidateFromL2OIState' Exception Message: MissingParameter: Parameter 'maxCCCLostHits' not found. ----- End Fatal Exception ------------------------------------------------- 50202.0 step2 runTheMatrix-results/50202.0_TTbar_13+TTbar_13+DIGIUP15_PU50+RECOUP15_PU50+HARVESTUP15_PU50/step2_TTbar_13+TTbar_13+DIGIUP15_PU50+RECOUP15_PU50+HARVESTUP15_PU50.log ----- Begin Fatal Exception 26-Nov-2015 18:29:39 CET----------------------- An exception of category 'Configuration' occurred while [0] Constructing the EventProcessor [1] Constructing module: class=CkfTrajectoryMaker label='hltL3TrackCandidateFromL2OIState' Exception Message: MissingParameter: Parameter 'maxCCCLostHits' not found. ----- End Fatal Exception ------------------------------------------------- you can see the results of the tests here: The following merge commits were also included on top of IB + this PR after doing git cms-merge-topic: |
|
|
||
| explicit MaxCCCLostHitsTrajectoryFilter ( | ||
| const edm::ParameterSet & pset, edm::ConsumesCollector& iC) : | ||
| theMaxCCCLostHits_(pset.getParameter<int>("maxCCCLostHits")), |
|
please test (compile and run the short matrix) before updating the PR or requesting another jenkins tests. |
| * patter recognition. Used mainly as a criteria for abandoning a | ||
| * trajectory candidate during trajectory building. | ||
| */ | ||
| int CCCBadHits() const { return theNumberOfCCCBadHits_;} |
There was a problem hiding this comment.
as in TempTrajectory:
start with a lower case
cccBadHits is just as good
... in general, I find it a bad idea to define a category based on one particular cut. If we redo this cut to do something else (not just cluster charge), the name will become inappropriate.
|
@rovere |
|
@cmsbuild please test |
|
The tests are being triggered in jenkins. |
cmsbuild
added
tests-pending
comparison-pending
and removed
comparison-available
tests-approved
labels
|
Pull request #12576 was updated. @cmsbuild, @cvuosalo, @davidlange6, @slava77 can you please check and sign again. |
|
+1 for #12576 265e83b
|
cmsbuild
added
reconstruction-approved
fully-signed
and removed
pending-signatures
reconstruction-pending
labels
|
This pull request is fully signed and it will be integrated in one of the next CMSSW_8_0_X IBs after it passes the integration tests. This pull request requires discussion in the ORP meeting before it's merged. @slava77, @davidlange6, @Degano, @smuzaffar |
|
+1 |
Stop reason for a track/trajectory
Each TrajectoryFilter has been equipped in such a way to
inject into the Trajectory or TempTrajectory the reason that
stopped the propagation of the trajectory during the pattern
recognition. This work is in preparation of future work that
will propagate this information down the chain directly into
a reco::TrackBase, so that more detailed studies can be made
on the reasons of the stop.
CCC as a Trajectory Filter
The CCC has been implemented as a TrajectoryFilter, in order to give
more flexibility to the user in selecting possible bad (from the
Cluster Charge point of view) hits. The final trajectory, if discarded due
to eccessive number of bad CCC hits, will anyhow contain the
would-have-been removed 2 hits, since they passed all the other estimators.
The user has the possibility to disentangle lost hits from bad for CCC
and apply different criteria with different thresholds during Patter
Recognition and at filtering level.
NOTE
The CCC as Trajectory filter has been turned off, so that we should expect no changes and have this PR integrated quickly. The usage and tuning of this new features will be done on top of this. Private results are anyway extremely encouraging (w.r.t to the hit collection efficiency).