5. Quick analysis¶
5.1 Analysis using qana.C¶
If you would like to perform a quick analysis on your simulated data, there is also a handy tool available which simplifies a ‘standard’ analysis to a single line of code. The basis for this is the combiner class PndSimpleCombiner, the corresponding task PndSimpleCombinerTask, and the steering macro qana.C.
Doing the analysis described above basically reduces to the command:
> root -l -b -q 'qana.C("signal_pid.root", 6.232,
"J/psi->mu+ mu-;pbarpSystem->;J/psi pi+ pi-",
"fit4c:fitvtx:mwin(J/psi)=0.8")'
The result is the output root file signal_pid_ana.root with the contents:
root [1] .ls
TFile** signal_pid_ana.root
TFile* signal_pid_ana.root
KEY: TFolder pnd;1 Main Folder
KEY: TList BranchList;1 Doubly linked list
KEY: TList TimeBasedBranchList;1 Doubly linked list
KEY: FairFileHeader FileHeader;1
KEY: TTree ntp0;1 J/psi->mu+ mu-
KEY: TTree ntp1;1 pbarpSystem->J/psi pi+ pi-
KEY: TTree pndsim;1 /pnd_0
where ntp0 and ntp1 are the ntuples containing the reconstructed J/psi and pbarpSystem candidates. When running the macro without parameters (or with just “H” as first parameter) a (long) help text about the usage is printed:
qana.C -- ROOT macro for quick and simple analysis of PANDA simulation or data.
USAGE:
qana.C( <input>, <mom>, <decay>, [parms], [nevt], [runnum], [mode], [runST] )
<input> : input file name with PndPidCandidates; "help"/"H" shows analysis parameter options
<mom> : pbar momentum; negative values are interpreted as -E_cm
<decay> : decay pattern to be reconstructed, e.g. "phi -> K+ K-; D_s+ -> phi pi+"; decay file: auto generate from decay file
[parms] : analysis options, e.g. "mwin=0.4:emin=0.1:pmin=0.1:qamc"; "auto": auto generate options; "qapart" runs particle QA
[nevt] : number of events to analyze (default = 0 -> all)
[runnum] : arbitrary integer run number (default = 0)
[mode] : arbitrary integer mode number (default = 0)
[runST] : run software trigger (default = false)
Example 1 - Analyze EvtGen events : root -l -b -q 'qana.C("signal_pid.root", 6.23, "J/psi -> mu+ mu-; pbp -> J/psi pi+ pi-", "fit4c:mwin=0.6")'
Example 2 - Analyze BOX generator events : root -l -b -q 'qana.C("box_pid.root", 15.0, "", "qapart")'
available options for parameter [parms]; use colon separated setting 'par1=value1:par2=value2' ...
- mwin : mass window for all composites
- mwin(X) : mass window for composite 'X'; X (or cc of X) has to appear in decay
- emin : minimum energy threshold for neutrals
- pmin : minimum momentum threshold for charged
- pid : PID criterion (al=All, vl=VeryLoose, lo=Loose, ti=Tight, vt=VeryTight, be=Best) for all species, e. g. pid=lo
- pidX : PID criterion for specific species X = (e, mu, pi, k, p), e.g. pide=vl, pidmu=ti,....
- algo : PID algorithm (e.g. "algo=ideal,emc;drc;disc" or "algo=comb") for all species
- algoX : PID algorithm for specific species X = (e, mu, pi, k, p)
- fit4c[<x] : perform 4C fit on last resonance; optional argument [<x] cuts on chi2 < x
- fit4cbest[<x] : perform 4C fit on last resonance and only keep best one; optional argument [<x] cuts on chi2 < x
- fitvtx[<x] : perform vertex fit on all resonances when possible (at least two daughters); optional argument [<x] cuts on chi2 < x
- ebrem : use the bremsstrahlung corrected electron/positron lists (ElectronPlusBrem/ElectronMinusBrem)
- qamc : stored MC information
- qaevtshape : store event shape information
- qaevs : '' (same as qaevtshape)
- qarec : condensed reco information
- qarecfull : comprehensive reco information
- nevt : write event wise n-tuple with event (shape) information
- !ntpX : skips dump of TTree for X-th resonance; e.g. in "phi -> K+ K-; D_s+ -> phi pi+", '!ntp0' would skip dump of TTree for phi->KK
Available PID algorithms (FullSim / FastSim):
- ideal = ideal particle id (PidAlgoIdealCharged / IdealPidProbability)
- comb = combination of all following algos
- mvd = Micro Vertex Detector (PidAlgoMvd / MvdPidProbability)
- mdt = Muon Detector (PidAlgoMdtHardCuts / ScMdtPidBarrelProbability;ScMdtPidForwardProbability)
- drc = DIRC detector (PidAlgoDrc / DrcBarrelProbability)
- disc = Disc DIRC (PidAlgoDisc / DrcDiscProbability)
- stt = Straw Tube Tracker (PidAlgoStt / SttPidProbability)
- emc = EM-Calorimeter (PidAlgoEmcBayes / ScEmcPidFSProbability;ScEmcPidFwCapProbability;ScEmcPidBarrelProbability;ScEmcPidBwCapProbability)
- rich = Forward RICH detector (PidAlgoRich / RichProbability)
- scit = SciTil detector (PidAlgoSciT / n.a.)
- ftof = Forward Time Of Flight (PidAlgoFtof / n.a.)
The two most important parameters are the decay tree definition decay, and the steering parameter string parms, which are discussed a bit more in detail below.
5.1.1 The decay configuration parameter¶
Although the syntax of this parameter is quite intuitive, there are some rules which have to be followed for proper function.
The particle names used are those from EvtGen.
The decay chain has to be build up from generic particles: e+, e-, mu+, mu+, pi+, pi-, K+, K-, p+, anti-p-, gamma.
The sub-decays are separated by a semicolon.
The decay chain must be build bottom up, i.e. a particle used in a decay definition has to be defined before
correct:
"pi0 -> gamma gamma; omega -> pi+ pi- pi0"wrong:
"omega -> pi+ pi- pi0; pi0 -> gamma gamma"As default, charge conjugation is applied automatically. If this is not wanted, the key word nocc can be added at the end of each decay. This can be useful, if particle and anti-particle go to different decay channels.
"D0 -> K- pi+; pbarpSystem -> D0 anti-D0"reconstructs both D0 -> K- pi+ and anti-D0 -> K+ pi- and combines them to a pbar-p-system. Only one n-tuple is stored forD0andanti-D0, distinguishable by the branchxpdg
"D0 -> K- pi- nocc; pbarpSystem -> D0 anti-D0"fails to run, since the listanti-D0is missing.
"D0 -> K- pi+ nocc; anti-D0 -> pi+ pi- nocc; pbarpSystem -> D0 anti-D0"reconstructpbarpSystemfromD0andanti-D0from different decay channels.For each sub-decay an n-tuple is stored by default. These are named ntp0, ntp1, ntp2,… with the index according to the position of the sub-decay in the decay configuration string. This behaviour can be modified in the
parmssetting.
5.1.2 The steering parameter setting¶
There is a default behavior how the analysis is carried out, which basically is restricted to pure combinatorics. Many aspects like fitting, PID, mass windows, etc, can be adjusted with the parms parameter. Some of the settings are interpreted by the macro quickana.C, some by the combiner class PndSimpleCombiner, and the others by the task class PndSimpleCombinerTask. The following table lists all available options. The individual parameters are separated by a colon (:).
Parameter |
Meaning |
fit4c[<x] |
Performs 4-constraint-fitting of the last resonance in the decay list, applying an optional cut <x on the chi². |
fit4cbest[<x] |
Performs 4-constraint-fitting of the last resonance in the decay list, applying an optional cut <x on the chi². Only keeps candidate with best chi². |
fitvtx[<x] |
Performs vertex fit for all resonances with at least two tracks, applying an optional cut <x on the chi². |
mwin=x |
Applies mass window (mrec- mPDG) < x/2 GeV/c² (window width x) for all resonances. Is overridden by following particle specific definitions. |
mwin(R)=x |
Applies mass window (mrec- mPDG) < x/2 GeV/c² for resonance type R, e.g. mwin(pi0)=0.05. Overrides any previous settings, e.g. by mwin. |
mwin(R)=x|y |
Applies mass window x < mrec< y GeV/c² for resonance R. Overrides any previous setting, e.g. by mwin. |
emin=x |
Minimum energy cut E>x GeV for photon candidates. |
pmin=x |
Minimum momentum cut p>x GeV/c for tracks. |
pid=x |
Common PID criterion for all particle species. x ∈ [All, VeryLoose, Loose, Tight, VeryTight, Best]. Default is All. |
pidY=x |
PID criterion for species Y. x ∈ [All, VeryLoose, Loose, Tight, VeryTight, Best], Y ∈ [e, mu, pi, k, p]. Examples: pidmu=VeryTight, pidk=Loose. |
algo=x |
Common PID algorithm configuration x, e.g algo PidAlgoEmcBayes;PidAlgoDrc. Default for all species is PidAlgoEmcBayes;PidAlgoDrc;PidAlgoDisc;PidAlgoStt;PidAlgoMdtHardCuts;PidAlgoSciT;PidAlgoRich |
algoY=x |
PID algorithm setting for species Y, Y ∈ [e, mu, pi, k, p]. Example: algoe=PidAlgoEmcBayes:algok=PidAlgoDrc;PidAlgoDisc;PidAlgoStt. |
gtrk=x |
GoodTrackSelector mode. x ∈ [All, VeryLoose, Loose, Tight, VeryTight] |
gphot=x |
GoodPhotonSelector mode. x ∈ [All, VeryLoose, Loose, Tight, VeryTight] |
ebrem |
Use the electron lists with bremsstrahlung correction (ElectronBrem). |
qamc |
Stores the MC truth list in the n-tuple nmc. |
qaevtshape |
Stores event shape variables in each n-tuple. |
nevt |
Stores additional n-tuple with event (shape) variables; one entry per event. |
qapart |
Runs the task PndParticleQATask, useful for single particle studies. Creates n-tuples ntp (charged), ntpn (neutrals), nmc (MC truth list) |
!ntpx |
Prevents storing the n-tuple from the x-th decay definition. Example: !ntp0:!ntp1 suppresses storage of the two first decay definitions in decay chain. |
!mc |
Prevents PndParticleQATask to store the MC information in n-tuple nmc. |
!neut |
Prevents PndParticleQATask to store the single neutral information in n-tuple ntpn. |
!chrg |
Prevents PndParticleQATask to store the charged track information in n-tuple ntp. |
5.2 On-the-fly Fast Simulation and Analysis using qfsana.C¶
In case your want to take a really quick look on how a particular decay channel reconstructs in terms of acceptance or kinematic distributions, you can use make use of Fast Simulations together with the above mentioned simplified combinatorics by using the macro qfsana.C. Instead of simulated data, you just need a EvtGen decay file, or, for background studies not even that. The complete simulation and analysis described above for signal and background can be performed with two lines of code:
> root -l -b -q 'qfsana.C("signal", // prefix
"pp_jpsi2pi_jpsi_mumu.dec", // decay file
6.232, // pbar momentum
"J/psi->mu+ mu-;pbarpSystem->J/psi pi+ pi-", // decay pattern
1000, // number of events
"fit4c:fitvtx:mwin=0.8")' // config parameters
> root -l -b -q 'qfsana.C("bkg",
"DPM", // other generator
6.232,
"J/psi->mu+ mu-;pbarpSystem->J/psi pi+ pi-",
10000,
"fit4c:fitvtx:mwin=0.8")'
The resultant files contains the same n-tuples as shown above. Documentation about the names of the branches can be found under the description of PndRhoTupleQA (https://panda-wiki.gsi.de/foswiki/bin/view/Computing/PandaRootAnalysisJuly13#PndRhoTupleQA).
The parameters for this macro are slightly different than those for qana.C, however, the parameters for decay and parms are the same as above:
qfsana.C -- ROOT macro for quick and simple fast simulation and analysis for PANDA
USAGE:
qfsana.C( <pref>, <decfile>, [mom], [nevt], [decay], [parms], [runST], [runnum], [mode] )
<pref> : output file names prefix; "help"/"H" shows analysis parameter options
<decfile> : EvtGen decfile "xxx.dec" or "xxx.dec:iniResonance"; DPM/FTF/BOX uses DPM, FTF or BOX generator
- DPM settings: DPM = inelastic only, DPM1 = inel. + elastic, DPM2 = elastic only
- FTF settings: FTF = inelastic only, FTF1 = inel. + elastic
- BOX settings: "BOX:type[pdgcode,mult]"; optional range/value with "p/tht/ctht/phi[min,max]/[value]" separated by ":", e.g. "BOX:type[211,2]:p[1,5]:tht[45]:phi[90,210]"
[mom] : pbar momentum (negative values are interpreted as -E_cm) for EvtGen/DPM/FTF; BOX generator: maximum particle momentum (if not specified above) (default = 15.0)
[nevt] : number of events to simulate and analyze (default = 1000)
[decay] : decay pattern to be reconstructed, e.g. "phi -> K+ K-; D_s+ -> phi pi+"; "auto"/decfile: generate from decay file; "": fastsim only w/o reco (default = auto)
[parms] : analysis options, e.g. "mwin=0.4:emin=0.1:pmin=0.1:qamc"; "auto": auto generate options; "qapart" runs particle QA; "persist" saves PndPidCandidates (default = auto)
[runnum] : arbitrary integer run number (default = 0)
[mode] : arbitrary integer mode number (default = 0)
[runST] : run software trigger (default = false)
Example 1 - EvtGen sim/analysis : root -l -b -q 'qfsana.C("jpsi2pi", "pp_jpsi2pi_jpsi_mumu.dec:pbarpSystem", 6.23, 1000, "J/psi -> mu+ mu-; pbp -> J/psi pi+ pi-", "fit4c:mwin=0.6")'
Example 2 - EvtGen auto mode : root -l -b -q 'qfsana.C("jpsi2pi_auto", "pp_jpsi2pi_jpsi_mumu.dec", 6.23)'
Example 3 - Particle QA BOX gen : root -l -b -q 'qfsana.C("single_kplus", "BOX:type[321,1]", 8.0, 1000, "", "qapart")'
Example 4 - Simulation output DPM : root -l -b -q 'qfsana.C("bkg", "DPM", 6.23, 1000, "", "")'