HLTriggerOffline/BJet Package description

The offline trigger efficiency code is located in HLTriggerOffline/BJet.

The core component is the HLTBtagLifetimeAnalyzer module, which will analyse a single HLT trigger path and measure the rates after each "level". Levels corresponds to any number of filters in HLT path (currently, "L1", "L2", "L2.5", "L3") and are described by:

  • name and title
  • a filter
  • a jet collection
  • a jet-tracks association collection

For each level, the rate of events passing the given filter is measured. If a jet collection is given, plots are made for the energy, ET, eta and phi distributions of the jets in the collection, passing the filter. If a track association is also given, plots are made for the energy, pT, eta and phi distributions of the tracks associated to the jets, passing the filter.

Optionally, jet Monte Carlo truth can be specified, in the form of a jet-to-parton association map. The map needs not have as keys the same jets present in the HLT collection, as an association is performed between the two, based on minimum pseudo-angular distance (within a maximum of 0.1, configurable via .cfg). Different parton flavours can be requested, and rates are computed (and plots are made) for each group (currently, "udsg" aka "light", "uds", "g", "c", "b", "other").

Optionally, matching to offline btagged jets is performed. This requires a collection of JetTag objects and a list of discriminator cuts: rates are computed (and plots are made) only for jets matching an offline one with a discriminator above the given value. Currently, the jetProbabilityBJetTags collection is used, with cuts of 0.3, 0.6, 1.1, corresponding do roughly 80%, 50% and 20% btagging efficiencies.

Finally, differential efficiencies are computed for each step (each filter w.r.t. the former one), and cumulative (each filter w.r.t. the first collection). Jets and tracks plots are made, if requested. MC truth and offline-matching efficiencies and plots are made, if requested.

I'm in the process of moving all the "analysis" code (computing and plotting cumulative and step-by-step efficiencies) in an external (set of) ROOT macro(s) that will reside in the test directory. This allows splitting the heavy part of the work in jobs (e.g. to be submitted to the grid o a batch system). The output .root files can then be merged into a single .root file (hadd final.root fragment_1.root fragment_2.root ...), and analyzed together.

Currently, only plot.cc is available, and despite its name it's main purpose is to compute the efficiencies at each level (L1, L2, L2.5, L3).

Description of the code

Getting the code

For CMSSW 2.0.0 and later:

cvs co -r V03-02-01    HLTriggerOffline/BJet

For CMSSW 1.8.4:

cvs co -r V02-02-01    HLTriggerOffline/BJet

For CMSSW 1.7.6:

cvs co -r V01-02-01    HLTriggerOffline/BJet
cvs co -r V00-06-08    PhysicsTools/JetMCAlgos
cvs co -r V00-00-04    SimDataFormats/CMS.JetMatching

Running the code

The configuration file SWGuideHLTriggerOffline/BJet/test/hltBtagLifetimeAnalyzer.cfg will analyze the 10 1E32 btag HLT trigger paths, and make all jet plots (jets, tracks, with MC tructh matching, with offline matching) for each level, the vertex plots, and the rate plot.

The configuration file SWGuideHLTriggerOffline/BJet/test/hltBtagLifetimeAnalyzerExtra.cfg is aimed at SIM-HLT samples and will do a similar work, with two differences:

  • MC matching info are built from scratch
  • Offline matching is disables
Caveat: this has been tested only with CMSSW 2.0.2.

The configuration file SWGuideHLTriggerOffline/BJet/test/hltBtagLifetimeAnalyzerAOD.cfg will work on AOD-only samples. Currently, it will only make the event rate.

Running the code in CMSSW 1.6.11

Running with CMSSW 1.6.11 has not been tested, but is likely to work following the instructions for CMSSW 1.7.6.

Offline analysis

Further plots and more detailed analysis can be performed with the SWGuideHLTriggerOffline/BJet/test/plots.cc ROOT macro (still under development).

Validation Results

-- AndreaBocci - 23 Apr 2008

Responsible: JasonSlaunwhite

Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r8 - 2011-05-30 - MonicaVazquezAcosta
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    CMSPublic All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback