How do I redo stripping?

The stripping is implemented in this package. Some of it can be modified from the DaVinci script, but other things require modifications in DaVinci. Therefore one has to checkout the right version of DaVinci, modify it and run the ntuple production with it.

What DaVinci version to use

For:

  • Data: Use the latest version. This is not relevant here because data is stored in MDST files which for which restripping is not possible.
  • Simulation: Use the version of DaVinci used for processing the corresponding data.

The latest stripping version for the data is here. So, if you are using 2017 data and the latest stripping available is 29r2p1 you check here and find that the DaVinci version used was v42r9p2.

Checking out DaVinci

Follow these instructions:

#disable default login
touch ~/.nogrouplogin

#set LHCb environment
source /cvmfs/lhcb.cern.ch/lib/LbEnv

#what platforms have a specific DaVinci version?
lb-sdb-query p DaVinci v42r9p2

#set platform
lb-set-platform x86_64-centos7-gcc62-opt

#pick a directory where all your DaVinci versions will be
cd /afs/cern.ch/work/a/acampove/RK/DaVinci/

#Create DaVinci skeleton
lb-dev DaVinci/v42r9p2
cd DaVinciDev_v42r9p2/

#Declare what packages you will be modifying
git lb-use Phys
git lb-use Stripping

each DaVinci version is associated with a specific version of the other packages. Here we need to know the version of Phys and Stripping, which can be found here.

#Place where PID cut on electrons are set, when building dielectron objects
git lb-checkout Phys/v23r10 Phys/CommonParticles
#Place where stripping lines are defined
git lb-checkout Stripping/v14r5p2 Phys/StrippingArchive

#build
make configure
make

#Edit code where PID is set for electrons
cd Phys/CommonParticles/python/CommonParticles/
vim StdLooseDiElectron.py

The change needed for StdLooseDiElectron.py is below:

   51#dieLL.Electron.Selection = ["RequiresDet='CALO' CombDLL(e-pi)>'-2'"]
   53dieLL.Electron.Selection = [""]

Then modify StdLooseDiMuon.py as indicated below:

   30#StdLooseDiMuon.Inputs = ["Phys/StdAllLooseMuons/Particles"]
   32StdLooseDiMuon.Inputs = ["Phys/StdAllNoPIDsMuons/Particles"]

Then implement changes in the file where the stripping line is defined:

cd Phys/StrippingArchive/python/StrippingArchive/Stripping29r2p1/StrippingRD/
vim StrippingB2LLXBDT.py

by changing:

  229#from StandardParticles import StdLooseANNPions, StdLooseANNKaons, StdLooseANNProtons
  231        from StandardParticles import StdAllNoPIDsPions, StdAllNoPIDsKaons, StdAllNoPIDsProtons
and
  232self.SelPions = self.createSubSel(
  234            OutputList=self.name + "SelPions",
  236            #InputList=StdLooseANNPions,
  238            InputList=StdAllNoPIDsPions,
  240            Cuts=config['PionCuts'])
  242
  244        self.SelKaons = self.createSubSel(
  246            OutputList=self.name + "SelKaons",
  248            #InputList=StdLooseANNKaons,
  250            InputList=StdAllNoPIDsKaons,
  252            Cuts=config['KaonCuts'])
  254
  256        self.SelProtons = self.createSubSel(
  258            OutputList=self.name + "SelProtons",
  260            #InputList=StdLooseANNProtons,
  262            InputList=StdAllNoPIDsProtons,
  264            Cuts=config['ProtonCuts'])

Adding or modifying tools

Stripping 42r9p2 is meant to be used with 2017 data and 44r10p5 with 2016 data. However TupleToolCorrectedMass does not exist in the former version but it does in the latter. Therefore you need to add that tool by doing:

git lb-use Analysis
git checkout Analysis/v18r9 Phys/DecayTreeTuple

Now from gitlab you should be able to see the missing tool in Analysis/v20r9p3 here. Download the .cpp and .h files and put them in the corresponding folder. Click on the History button and see the first commit here You see that not only the file with the class was added, but also the CMakeLists.txt was modified to link a library against the Physics library in ROOT. So, that modification is also needed. After that change do:

make configure
make

this should build the final version of the code.

Modifying the DaVinci options

A working example of an option file that works with the software set up above is here. The relevant lines are:

   27# Imports the stripping builder and the configuration dictionary from the right stripping version and stripping line. 
   29from StrippingArchive.Stripping29r2p1.StrippingRD.StrippingB2LLXBDT import B2LLXBDTConf   as builder
   31from StrippingArchive.Stripping29r2p1.StrippingRD.StrippingB2LLXBDT import default_config as config
   33
   35#Picks the dictionary with the stripping cuts
   37conf= config['CONFIG']
   39
   41print(conf)
   43print("---------------")
   45print("---------------")
   47
   49#redefine stripping cuts
   51if RESTRIP and LINE == "bdt":
   53    conf['DiElectronCuts']="""
   55                               (HASVERTEX) & (VFASPF(VCHI2)<16) & (MM<5.0*GeV)
   57                               & (INTREE( (ID=='e+') & (PT>200*MeV) & (MIPCHI2DV(PRIMARY)>1.) & (TRGHOSTPROB<0.5) ))
   59                               & (INTREE( (ID=='e-') & (PT>200*MeV) & (MIPCHI2DV(PRIMARY)>1.) & (TRGHOSTPROB<0.5) ))
   61                           """
   63    conf['PionCuts'  ]    ="(PROBNNpi> -1000) & (PT>250*MeV) & (TRGHOSTPROB<0.4)"
   65    conf['KaonCuts'  ]    ="(PROBNNk > -1000) & (PT>300*MeV) & (TRGHOSTPROB<0.4)"
   67    conf['ProtonCuts']    ="(PROBNNp > -1000) & (PT>300*MeV) & (TRGHOSTPROB<0.4)"
   69elif RESTRIP and LINE == "cut":
   71    conf['PIDe']       = -1000
   73
   75print(conf)
   77print("---------------")
   79print("---------------")
   81
   83#Create a new stripping
   85from StrippingConf.Configuration import StrippingConf, StrippingStream
   87MyStream = StrippingStream("MyStream")
   89
   91if LINE == "bdt":
   93    lb = builder('B2LLXBDT', conf)
   95else:
   97    lb = builder('Bu2LLK'  , conf)
   99
  101#pick up the old line with redefined selection
  103for line in lb.lines(): 
  105    if line.name() == 'Stripping' + STR_LINE:
  107        print("Appending {}".format(STR_LINE))
  109        MyStream.appendLines([line])
  111
  113# Configure Stripping
  115from Configurables import ProcStatusCheck
  117filterBadEvents = ProcStatusCheck()
  119
  121sc = StrippingConf( Streams = [ MyStream ], MaxCandidates = 2000, AcceptBadEvents = False, BadEventSelection = filterBadEvents )

and at the end of the file:

  372from Configurables import DaVinci, CheckPV, GaudiSequencer, LoKi__HDRFilter
  374#Need to kill old stripping
  376from Configurables import EventNodeKiller
  378
  380eventNodeKiller = EventNodeKiller('Stripkiller')
  382eventNodeKiller.Nodes = [ '/Event/AllStreams', '/Event/Strip' ]
  384
  386DaVinci()
  388DaVinci().EvtMax = -1 
  390DaVinci().PrintFreq = 1000
  392DaVinci().HistogramFile = 'ROOT.root'
  394DaVinci().Simulation = True
  396DaVinci().Lumi = False
  398#Add killing stripping algorithm
  400DaVinci().appendToMainSequence( [ eventNodeKiller ] )
  402#Add new stripping
  404DaVinci().appendToMainSequence( [ sc.sequence() ] )
  406DaVinci().UserAlgorithms = [ tupleee ]
  408DaVinci().VerboseMessages = True
  410DaVinci().DataType = "2017"

Producing the ntuple

After this, to test everything cd to the DaVinciDev_v42r9p2 directory and:

./run gaudirun.py DV_B2Kee_2017.py

this should produce the ntuple.

-- AngelCampoverde - 2020-04-10

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r2 - 2020-04-10 - AngelCampoverde
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Sandbox All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback