Macro to compare cluster and hits coordinates

[u/mrichter/AliRoot.git] / ITS / UPGRADE / readmeKMC

Here are some instructions for running:
Once the KMCDetector object is initialized (similar to FER, see "Detector, tracking initialization"
part of the testDetKMC.C macro) one should define for each kinematic bin to test a TObjArray with
KMCTrackSummary objects (done in the testDetKMC.C by the CreateTrackConditions routine):
this object is basically a set of standard histos (with DCA, pt_rec - pt_gen distibutions etc + stat. of generated
and found tracks) which are filled provided the reco track satisfies to user defined conditions, like the number of
hits (one can specify correct or fake hits, specific layers pattern etc). For instance:

sm = new KMCTrackSummary("corr4hdGoodPatt","corr4hdGoodPatt",nlr); // create a summary object with this name

sm->SetMinMaxClITS(4);             // require at least 4 cluster in the track (the upper limit here is default 999 - no limit)

sm->SetMinMaxClITSFake(0,0); // set min number of fake clusters required (used for fakes studies) and their max number tolerated

sm->SetNamePrefix(nm);          // this histo names will be prefixed by nm

sm->AddPatternITSCorr( sm->Bits(1,1));    // require that the innermost 2 double layers must contribute at least 1 correct cluster

sm->AddPatternITSCorr( sm->Bits(0,0,1));  // require the 3d layer cluster to be correct

sm->AddPatternITSCorr( sm->Bits(0,0,0,1,1)); // measurement in the middle: ask at leas 1 correct cluster in 2 middle layers

sm->AddPatternITSCorr( sm->Bits(0,0,0, 0,0, 1,1)); // same for outer 2 layers

One can define and put in the TObjArray many such objects with different conditions, they will be filled independently.
The simulation for single kinem. bin is done via call:

its.SolveSingleTrack(mass, pt, eta, trSum, nevR);
This will generate nevR tracks of given mass, pt, eta, propagate them, reconstruct under imposed bg, sensor err. conditions
and fill the KMCTrackSummary objects provided via trSum TObjArray (i.e. histos of each object are filled if given reco track
satisfies the criteria defined for this object).

In the testDetKMC trSum arrays for each pt bin are added to another TObjArray (e.g. summary) which is then stored in
the sumOut file.

I also attach the PrepSummaryKMC.C macro to fill/return from this output file the TObjArray of TGraphErrors corresponding
to given id of KMCTrackSummary. The call
PrepSummaryKMC(const char* sumf, int cls, const char* pref=0, const char* outF=0);
will open sumf output of the testDetKMC, and  for all pt bins (indexing the entries of the final "summary" TObjArray
of testDetKMC) it will extract the array of KMCTrackSummary objects defined in the CreateTrackConditions
for this bin, then extract from it the cls-th KMCTrackSummary object and put the performance values for this pt bin
(resolutions, eff ) as an entry in the graph.

prepPlotsKMC1.C is just an example of preparing the graphs from the testDetKMC.C output objects.