void plotDYUnfoldingMatrix(const TString input, int systematicsMode = DYTools::NORMAL, int randomSeed = 1, double reweightFsr = 1.0, double massLimit = -1.0)
//systematicsMode 0 (NORMAL) - no systematic calc
//1 (RESOLUTION_STUDY) - systematic due to smearing, 2 (FSR_STUDY) - systematics due to FSR, reweighting
//check mass spectra with reweightFsr = 0.95; 1.00; 1.05
//mass value until which do reweighting
{
// check whether it is a calculation
if (input.Contains("_DebugRun_")) {
std::cout << "plotDYUnfoldingMatrix: _DebugRun_ detected. Terminating the script\n";
return;
}
// normal calculation
gBenchmark->Start("plotDYUnfoldingMatrix");
if (systematicsMode==DYTools::NORMAL)
std::cout<<"Running script in the NORMAL mode"<<std::endl;
else if (systematicsMode==DYTools::RESOLUTION_STUDY)
std::cout<<"Running script in the RESOLUTION_STUDY mode"<<std::endl;
else if (systematicsMode==DYTools::FSR_STUDY)
std::cout<<"Running script in the FSR_STUDY mode"<<std::endl;
else if (systematicsMode==DYTools::ESCALE_RESIDUAL)
std::cout << "Running script in the ESCALE_RESIDUAL mode\n";
else {
std::cout<<"requested mode not recognized"<<std::endl;
assert(0);
}
//--------------------------------------------------------------------------------------------------------------
// Settings
//==============================================================================================================
Bool_t doSave = false; // save plots?
TString format = "png"; // output file format
int saveMadePlots = 1; // whether save produced canvases
vector<TString> fnamev; // file names
vector<TString> labelv; // legend label
vector<Int_t> colorv; // color in plots
vector<Int_t> linev; // line style
vector<Double_t> xsecv;
vector<Double_t> lumiv;
TString dirTag;
ifstream ifs;
ifs.open(input.Data());
if (!ifs.is_open()) std::cout << "failed to open a file <" << input.Data() << ">\n";
assert(ifs.is_open());
string line;
Int_t state=0;
const int inputFileMayContainEScaleDefinition = 1;
int expectEscaleLine=1;
ElectronEnergyScale escale((inputFileMayContainEScaleDefinition) ?
ElectronEnergyScale::UNDEFINED : // read electron energy scale from the file
ElectronEnergyScale::Date20120101_default);
while(getline(ifs,line)) {
if(line[0]=='#') continue;
if(state == 0){
dirTag = TString(line);
state++;
continue;
}else{
if (inputFileMayContainEScaleDefinition && expectEscaleLine) {
expectEscaleLine=0;
// try to determine whether the input file was updated
if (ElectronEnergyScale::DetermineCalibrationSet(line.c_str()) !=
ElectronEnergyScale::UNDEFINED) {
//std::cout << "got it ok: <" << line << ">" << std::endl;
escale.init(TString(line.c_str()));
if (!escale.isInitialized()) {
std::cout << "code error\n";
return;
}
// continue reading the file
getline(ifs,line);
}
else {
std::cout << "\n";
std::cout << "\n\tInput file does not contain electron energy scale. The expected file format:\n";
std::cout << "\tLine1: directory\n";
std::cout << "\tLine2: electron energy scale correction name (NEW from 2012 Jan 21)\n";
std::cout << "\tLine3: file_name.root xsect color linesty label\n";
std::cout << "using the default set\n\n";
escale.init("Date20120101_default");
if (!escale.isInitialized()) {
std::cout << "failed to correct the behavior\n";
return;
}
}
std::cout << "energy scale corrections: " << escale.calibrationSetName() << "\n";
}
string fname;
Int_t color, linesty;
stringstream ss(line);
Double_t xsec;
ss >> fname >> xsec >> color >> linesty;
string label = line.substr(line.find('@')+1);
fnamev.push_back(fname);
labelv.push_back(label);
colorv.push_back(color);
linev.push_back(linesty);
xsecv.push_back(xsec);
lumiv.push_back(0);
}
}
ifs.close();
const Double_t kGAP_LOW = 1.4442;
const Double_t kGAP_HIGH = 1.566;
//--------------------------------------------------------------------------------------------------------------
// Main analysis code
//==============================================================================================================
TRandom random;
// The random seeds are needed only if we are running this script in systematics mode
int seed = randomSeed;
random.SetSeed(seed);
gRandom->SetSeed(seed);
// In the case of systematic studies, generate an array of random offsets
TVectorD shift(escale._nEtaBins); // this vector is outdated by the new features in the escale obj.class
shift = 0;
if(systematicsMode==DYTools::RESOLUTION_STUDY) {
escale.randomizeSmearingWidth(seed);
for(int i=0; i<escale._nEtaBins; i++)
shift[i] = gRandom->Gaus(0,1);
}
// prepare tools for ESCALE_RESIDUAL
TH1F *shapeWeights=NULL;
if (systematicsMode==DYTools::ESCALE_RESIDUAL) {
TString shapeFName=TString("../root_files/yields/") + dirTag + TString("/shape_weights.root");
std::cout << "Obtaining shape_weights.root from <" << shapeFName << ">\n";
TFile fshape(shapeFName);
if (!fshape.IsOpen()) {
std::cout << "failed to open a file <" << shapeFName << ">\n";
throw 2;
}
shapeWeights = (TH1F*)fshape.Get("weights");
shapeWeights->SetDirectory(0);
dirTag += TString("_escale_residual");
std::cout << "changing dirTag to <" << dirTag << ">\n";
}
//
// Set up histograms
//
vector<TH1F*> hZMassv;//, hZMass2v, hZPtv, hZPt2v, hZyv, hZPhiv;
char hname[100];
for(UInt_t ifile = 0; ifile<fnamev.size(); ifile++) {
sprintf(hname,"hZMass_%i",ifile); hZMassv.push_back(new TH1F(hname,"",500,0,1500)); hZMassv[ifile]->Sumw2();
}
TH1F *hMassDiff = new TH1F("hMassDiff","", 100, -30, 30);
TH1F *hMassDiffBB = new TH1F("hMassDiffBB","", 100, -30, 30);
TH1F *hMassDiffEB = new TH1F("hMassDiffEB","", 100, -30, 30);
TH1F *hMassDiffEE = new TH1F("hMassDiffEE","", 100, -30, 30);
TH1F *hMassDiffV[DYTools::nMassBins];
for(int i=0; i<DYTools::nMassBins; i++){
sprintf(hname,"hMassDiffV_%d",i);
hMassDiffV[i] = new TH1F(hname,"",100,-50,50);
}
// MC spectra for storage in ROOT file
// The FSR of RECO means that this is the spectrum of generator-level
// post-FSR mass for events that were actually reconstructed (i.e. includes
// efficiency and acceptances losses)
TVectorD yieldsMcFsrOfRec (DYTools::nMassBins);
TVectorD yieldsMcFsrOfRecErr (DYTools::nMassBins);
TVectorD yieldsMcRec (DYTools::nMassBins);
TVectorD yieldsMcRecErr (DYTools::nMassBins);
yieldsMcFsrOfRec = 0;
yieldsMcFsrOfRecErr = 0;
yieldsMcRec = 0;
yieldsMcRecErr = 0;
// The yields at generator level with mass bins defined by post-FSR di-leptons
TVectorD yieldsMcFsr (DYTools::nMassBins);
yieldsMcFsr = 0;
// Vectors for bin to bin corrections
TVectorD DetCorrFactorNumerator (DYTools::nMassBins);
TVectorD DetCorrFactorDenominator(DYTools::nMassBins);
TVectorD DetCorrFactor (DYTools::nMassBins);
TVectorD DetCorrFactorErrPos (DYTools::nMassBins);
TVectorD DetCorrFactorErrNeg (DYTools::nMassBins);
DetCorrFactorNumerator = 0;
DetCorrFactorDenominator = 0;
DetCorrFactor = 0;
DetCorrFactorErrPos = 0;
DetCorrFactorErrNeg = 0;
// Matrices for unfolding
TMatrixD DetMigration(DYTools::nMassBins, DYTools::nMassBins);
TMatrixD DetResponse(DYTools::nMassBins, DYTools::nMassBins);
TMatrixD DetResponseErrPos(DYTools::nMassBins, DYTools::nMassBins);
TMatrixD DetResponseErrNeg(DYTools::nMassBins, DYTools::nMassBins);
for(int i=0; i<DYTools::nMassBins; i++){
for(int j=0; j<DYTools::nMassBins; j++){
DetMigration(i,j) = 0;
DetResponse(i,j) = 0;
DetResponseErrPos(i,j) = 0;
DetResponseErrNeg(i,j) = 0;
}
}
//
// Access samples and fill histograms
//
TFile *infile=0;
TTree *eventTree=0;
// Data structures to store info from TTrees
mithep::TEventInfo *info = new mithep::TEventInfo();
mithep::TGenInfo *gen = new mithep::TGenInfo();
TClonesArray *dielectronArr = new TClonesArray("mithep::TDielectron");
// loop over samples
for(UInt_t ifile=0; ifile<fnamev.size(); ifile++) {
// Read input file
cout << "Processing " << fnamev[ifile] << "..." << endl;
infile = new TFile(fnamev[ifile]);
assert(infile);
// Get the TTrees
eventTree = (TTree*)infile->Get("Events"); assert(eventTree);
// Find weight for events for this file
// The first file in the list comes with weight 1,
// all subsequent ones are normalized to xsection and luminosity
double xsec=xsecv[ifile];
AdjustXSectionForSkim(infile,xsec,eventTree->GetEntries(),1);
lumiv[ifile] = eventTree->GetEntries()/xsec;
double scale = lumiv[0]/lumiv[ifile];
cout << " -> sample weight is " << scale << endl;
// Set branch address to structures that will store the info
eventTree->SetBranchAddress("Info",&info); TBranch *infoBr = eventTree->GetBranch("Info");
eventTree->SetBranchAddress("Gen",&gen); TBranch *genBr = eventTree->GetBranch("Gen");
eventTree->SetBranchAddress("Dielectron",&dielectronArr); TBranch *dielectronBr = eventTree->GetBranch("Dielectron");
// loop over events
for(UInt_t ientry=0; ientry<eventTree->GetEntries(); ientry++) {
//if (ientry>100000) break;
genBr->GetEntry(ientry);
infoBr->GetEntry(ientry);
double reweight;
if (systematicsMode!=DYTools::FSR_STUDY) reweight=1.0;
else if (((gen->mass)-(gen->vmass))>massLimit) reweight=1.0;
else reweight=reweightFsr;
if (ientry<20) std::cout<<"reweight="<<reweight<<std::endl;
// Use post-FSR generator level mass in unfolding
int ibinFsr = DYTools::findMassBin(gen->mass);
if(ibinFsr != -1 && ibinFsr < yieldsMcFsr.GetNoElements()){
yieldsMcFsr[ibinFsr] += reweight * scale * gen->weight;
}
/*
// For EPS2011 for both data and MC (starting from Summer11 production)
// we use an OR of the twi triggers below. Both are unpresecaled.
ULong_t eventTriggerBit = kHLT_Ele17_CaloIdL_CaloIsoVL_Ele8_CaloIdL_CaloIsoVL
| kHLT_Ele17_CaloIdT_TrkIdVL_CaloIsoVL_TrkIsoVL_Ele8_CaloIdT_TrkIdVL_CaloIsoVL_TrkIsoVL;
ULong_t leadingTriggerObjectBit = kHLT_Ele17_CaloIdL_CaloIsoVL_Ele8_CaloIdL_CaloIsoVL_Ele1Obj
| kHLT_Ele17_CaloIdT_TrkIdVL_CaloIsoVL_TrkIsoVL_Ele8_CaloIdT_TrkIdVL_CaloIsoVL_TrkIsoVL_Ele1Obj;
ULong_t trailingTriggerObjectBit = kHLT_Ele17_CaloIdL_CaloIsoVL_Ele8_CaloIdL_CaloIsoVL_Ele2Obj
| kHLT_Ele17_CaloIdT_TrkIdVL_CaloIsoVL_TrkIsoVL_Ele8_CaloIdT_TrkIdVL_CaloIsoVL_TrkIsoVL_Ele2Obj;
*/
const bool isData=kFALSE;
TriggerConstantSet constantsSet = Full2011DatasetTriggers; // Enum from TriggerSelection.hh
TriggerSelection requiredTriggers(constantsSet, isData, info->runNum);
ULong_t eventTriggerBit = requiredTriggers.getEventTriggerBit(info->runNum);
ULong_t leadingTriggerObjectBit = requiredTriggers.getLeadingTriggerObjectBit(info->runNum);
ULong_t trailingTriggerObjectBit = requiredTriggers.getTrailingTriggerObjectBit(info->runNum);
if(!(info->triggerBits & eventTriggerBit)) continue; // no trigger accept? Skip to next event...
// loop through dielectrons
dielectronArr->Clear();
dielectronBr->GetEntry(ientry);
for(Int_t i=0; i<dielectronArr->GetEntriesFast(); i++) {
const mithep::TDielectron *dielectron = (mithep::TDielectron*)((*dielectronArr)[i]);
// Apply selection
// Eta cuts
if((fabs(dielectron->scEta_1)>kGAP_LOW) && (fabs(dielectron->scEta_1)<kGAP_HIGH)) continue;
if((fabs(dielectron->scEta_2)>kGAP_LOW) && (fabs(dielectron->scEta_2)<kGAP_HIGH)) continue;
if((fabs(dielectron->scEta_1) > 2.5) || (fabs(dielectron->scEta_2) > 2.5)) continue; // outside eta range? Skip to next event...
// Asymmetric SC Et cuts
if( ! ( ( dielectron->scEt_1 > DYTools::etMinLead && dielectron->scEt_2 > DYTools::etMinTrail)
|| ( dielectron->scEt_1 > DYTools::etMinTrail && dielectron->scEt_2 > DYTools::etMinLead) )) continue;
// Both electrons must match trigger objects. At least one ordering
// must match
if( ! (
(dielectron->hltMatchBits_1 & leadingTriggerObjectBit &&
dielectron->hltMatchBits_2 & trailingTriggerObjectBit )
||
(dielectron->hltMatchBits_1 & trailingTriggerObjectBit &&
dielectron->hltMatchBits_2 & leadingTriggerObjectBit ) ) ) continue;
// The Smurf electron ID package is the same as used in HWW analysis
// and contains cuts like VBTF WP80 for pt>20, VBTF WP70 for pt<10
// with some customization, plus impact parameter cuts dz and dxy
if(!passSmurf(dielectron)) continue;
// We have a Z candidate! HURRAY!
// Apply extra smearing to MC reconstructed dielectron mass
// to better resemble the data
/*
outdated lines. kept for reference
double smear1 = escale.extraSmearingSigma(dielectron->scEta_1);
double smear2 = escale::extraSmearingSigma(dielectron->scEta_2);
// In systematics mode, overwrite the smear values with
// shifted ones.
if(systematicsMode==DYTools::RESOLUTION_STUDY){
smear1 = escale::extraSmearingSigmaShifted(dielectron->scEta_1,shift);
smear2 = escale::extraSmearingSigmaShifted(dielectron->scEta_2,shift);
}
double smearTotal = sqrt(smear1*smear1 + smear2*smear2);
double massResmeared = dielectron->mass + random.Gaus(0.0,smearTotal);
*/
/* lines based on new features -- begin */
double massSmear = (systematicsMode == DYTools::RESOLUTION_STUDY) ?
escale.generateMCSmearRandomized(dielectron->scEta_1,dielectron->scEta_2) :
escale.generateMCSmear(dielectron->scEta_1,dielectron->scEta_2);
double massResmeared = dielectron->mass + massSmear;
/* lines based on new features -- end */
hZMassv[ifile]->Fill(massResmeared,scale * gen->weight);
// Fill structures for response matrix and bin by bin corrections
if(ibinFsr != -1 && ibinFsr < yieldsMcFsrOfRec.GetNoElements()){
yieldsMcFsrOfRec[ibinFsr] += reweight * scale * gen->weight;
DetCorrFactorDenominator(ibinFsr) += reweight * scale * gen->weight;
}
else if(ibinFsr >= yieldsMcFsrOfRec.GetNoElements())
cout << "ERROR: binning problem" << endl;
int ibinRec = DYTools::findMassBin(massResmeared);
double shape_weight = (shapeWeights && (ibinRec!=-1)) ?
shapeWeights->GetBinContent(ibinRec+1) : 1.0;
if(ibinRec != -1 && ibinRec < yieldsMcRec.GetNoElements()){
yieldsMcRec[ibinRec] += reweight * scale * gen->weight;
DetCorrFactorNumerator(ibinRec) += reweight * scale * gen->weight * shape_weight;
}
else if(ibinRec >= yieldsMcRec.GetNoElements())
cout << "ERROR: binning problem" << endl;
if( ibinRec != -1 && ibinRec < yieldsMcRec.GetNoElements()
&& ibinFsr != -1 && ibinFsr < yieldsMcRec.GetNoElements() ){
DetMigration(ibinFsr,ibinRec) += reweight * scale * gen->weight * shape_weight;
}
Bool_t isB1 = (fabs(dielectron->scEta_1)<kGAP_LOW);
Bool_t isB2 = (fabs(dielectron->scEta_2)<kGAP_LOW);
hMassDiff->Fill(massResmeared - gen->mass);
if( isB1 && isB2 )
hMassDiffBB->Fill(massResmeared - gen->mass);
if( (isB1 && !isB2) || (!isB1 && isB2) )
hMassDiffEB->Fill(massResmeared - gen->mass);
if( !isB1 && !isB2 )
hMassDiffEE->Fill(massResmeared - gen->mass);
if(ibinFsr != -1){
hMassDiffV[ibinFsr]->Fill(massResmeared - gen->mass);
}
// if(ibinRec == -1)
// printf("Missed bin: M_fsr=%f M_reco=%f\n", gen->mass, massResmeared);
} // end loop over dielectrons
} // end loop over events
delete infile;
infile=0, eventTree=0;
} // end loop over files
delete gen;
//finish reweighting of mass spectra
// Find bin by bin corrections and the errors
double tCentral, tErrNeg, tErrPos;
for(int i=0; i<DYTools::nMassBins; i++){
if ( DetCorrFactorDenominator(i) != 0 ){
// This method does not take into account correlation between numerator
// and denominator in calculation of errors. This is a flaw to be corrected
// in the future.
SimpleDivide( DetCorrFactorNumerator(i), DetCorrFactorDenominator(i), tCentral, tErrNeg, tErrPos);
DetCorrFactor(i) = tCentral;
DetCorrFactorErrPos(i) = tErrPos;
DetCorrFactorErrNeg(i) = tErrNeg;
} else {
DetCorrFactor(i) = 0;
DetCorrFactorErrPos(i) = 0;
DetCorrFactorErrNeg(i) = 0;
}
}
// Find response matrix, which is simply the normalized migration matrix
for(int ifsr = 0; ifsr < DetMigration.GetNrows(); ifsr++){
double nEventsInFsrMassBin = 0;
for(int ireco = 0; ireco < DetMigration.GetNcols(); ireco++)
nEventsInFsrMassBin += DetMigration(ifsr,ireco);
// Now normalize each element and find errors
for(int ireco = 0; ireco < DetMigration.GetNcols(); ireco++){
tCentral = 0;
tErrPos = 0;
tErrNeg = 0;
// Note: the event counts supposedly are dominated with events with weight "1"
// coming from the primary MC sample, so the error is assumed Poissonian in
// the call for efficiency-calculating function below.
if( nEventsInFsrMassBin != 0 ){
EfficiencyDivide(DetMigration(ifsr,ireco), nEventsInFsrMassBin, tCentral, tErrNeg, tErrPos);
// BayesEfficiency does not seem to be working in newer ROOT versions,
// so it is replaced by simler method
// BayesEfficiency( DetMigration(ifsr,ireco), nEventsInFsrMassBin, tCentral, tErrNeg, tErrPos);
}
DetResponse (ifsr,ireco) = tCentral;
DetResponseErrPos(ifsr,ireco) = tErrPos;
DetResponseErrNeg(ifsr,ireco) = tErrNeg;
}
}
// Find inverted response matrix
TMatrixD DetInvertedResponse = DetResponse;
Double_t det;
DetInvertedResponse.Invert(&det);
TMatrixD DetInvertedResponseErr(DetInvertedResponse.GetNrows(), DetInvertedResponse.GetNcols());
calculateInvertedMatrixErrors(DetResponse, DetResponseErrPos, DetResponseErrNeg, DetInvertedResponseErr);
// Package bin limits into TVectorD for storing in a file
TVectorD BinLimitsArray(DYTools::nMassBins+1);
for(int i=0; i<DYTools::nMassBins+1; i++)
BinLimitsArray(i) = DYTools::massBinLimits[i];
// Store constants in the file
TString outputDir(TString("../root_files/constants/")+dirTag);
if((systematicsMode==DYTools::RESOLUTION_STUDY) || (systematicsMode==DYTools::FSR_STUDY))
outputDir = TString("../root_files/systematics/")+dirTag;
gSystem->mkdir(outputDir,kTRUE);
TString unfoldingConstFileName(outputDir+TString("/unfolding_constants.root"));
if(systematicsMode==DYTools::RESOLUTION_STUDY){
unfoldingConstFileName = outputDir+TString("/unfolding_constants_seed_");
unfoldingConstFileName += seed;
unfoldingConstFileName += ".root";
}
if(systematicsMode==DYTools::FSR_STUDY){
unfoldingConstFileName = outputDir+TString("/unfolding_constants_reweight_");
unfoldingConstFileName += int(100*reweightFsr);
unfoldingConstFileName += ".root";
}
TFile fConst(unfoldingConstFileName, "recreate" );
DetResponse .Write("DetResponse");
DetInvertedResponse .Write("DetInvertedResponse");
DetInvertedResponseErr .Write("DetInvertedResponseErr");
BinLimitsArray .Write("BinLimitsArray");
DetCorrFactor .Write("DetCorrFactor");
DetCorrFactorErrPos .Write("DetCorrFactorErrPos");
DetCorrFactorErrNeg .Write("DetCorrFactorErrNeg");
fConst.Close();
// Store reference MC arrays in a file
TString refFileName(outputDir+TString("/yields_MC_unfolding_reference.root"));
TFile fRef(refFileName, "recreate" );
BinLimitsArray .Write("BinLimitsArray");
yieldsMcFsr .Write("yieldsMcFsr");
yieldsMcFsrOfRec .Write("yieldsMcFsrOfRec");
yieldsMcRec .Write("yieldsMcRec");
fRef.Close();
//--------------------------------------------------------------------------------------------------------------
// Make plots
//==============================================================================================================
std::vector<TCanvas*> canvasV; // holder for a quick saver
canvasV.reserve(10);
TCanvas *c = MakeCanvas("zMass1","zMass1",800,600);
canvasV.push_back(c);
// string buffers
char ylabel[50]; // y-axis label
// Z mass
sprintf(ylabel,"a.u. / %.1f GeV/c^{2}",hZMassv[0]->GetBinWidth(1));
CPlot plotZMass1("zmass1","","m(ee) [GeV/c^{2}]",ylabel);
if (fnamev.size()) hZMassv[0]->GetXaxis()->SetLimits(10.,2000.);
for(UInt_t i=0; i<fnamev.size(); i++) {
plotZMass1.AddHist1D(hZMassv[i],labelv[i],"hist",colorv[i],linev[i]);
}
plotZMass1.SetLogy();
plotZMass1.SetLogx();
plotZMass1.TransLegend(0.02, 0.0);
plotZMass1.Draw(c,doSave,format);
// Create plots of how reco mass looks and how unfolded mass should look like
TVectorD massBinCentral (DYTools::nMassBins);
TVectorD massBinHalfWidth (DYTools::nMassBins);
TVectorD yieldMcFsrOfRecErr (DYTools::nMassBins);
TVectorD yieldMcRecErr (DYTools::nMassBins);
for(int i=0; i < DYTools::nMassBins; i++){
massBinCentral (i) = (DYTools::massBinLimits[i+1] + DYTools::massBinLimits[i])/2.0;
massBinHalfWidth(i) = (DYTools::massBinLimits[i+1] - DYTools::massBinLimits[i])/2.0;
yieldsMcFsrOfRecErr(i) = sqrt(yieldsMcFsrOfRec[i]);
yieldsMcRecErr (i) = sqrt(yieldsMcRec[i]);
}
TGraphErrors *grFsrOfRec = new TGraphErrors(massBinCentral, yieldsMcFsrOfRec,
massBinHalfWidth, yieldsMcFsrOfRecErr);
TGraphErrors *grRec = new TGraphErrors(massBinCentral, yieldsMcRec,
massBinHalfWidth, yieldsMcRecErr);
TCanvas *d = MakeCanvas("zMass2","zMass2",800,600);
canvasV.push_back(d);
CPlot plotZMass2("zmass2","","m(ee) [GeV/c^{2}]","events");
plotZMass2.SetLogx(1);
plotZMass2.AddGraph(grFsrOfRec,"no detector effects","PE",kBlack);
plotZMass2.AddGraph(grRec,"reconstructed","PE",kBlue);
plotZMass2.Draw(d);
// double xsize = 600;
// double ysize = 600;
int xsize = 600;
int ysize = 400;
// Create the plot of the response matrix
TH2F *hResponse = new TH2F("hResponse","",DYTools::nMassBins, DYTools::massBinLimits,
DYTools::nMassBins, DYTools::massBinLimits);
for(int i=1; i<=DYTools::nMassBins; i++)
for(int j=1; j<=DYTools::nMassBins; j++)
hResponse->SetBinContent(i,j,DetResponse(i-1,j-1));
TCanvas *e = MakeCanvas("response","response",xsize,ysize);
canvasV.push_back(e);
CPlot plotResponse("response","",
"generator post-FSR m(ee) [GeV/c^{2}]",
"reconstructed m(ee) [GeV/c^{2}]");
plotResponse.AddHist2D(hResponse,"COLZ");
e->cd();
plotResponse.SetLogx();
plotResponse.SetLogy();
gPad->SetRightMargin(0.15);
gStyle->SetPalette(1);
hResponse->GetXaxis()->SetMoreLogLabels();
hResponse->GetXaxis()->SetNoExponent();
hResponse->GetYaxis()->SetNoExponent();
plotResponse.Draw(e);
// Create the plot of the inverted response matrix
TH2F *hInvResponse = new TH2F("hInvResponse","",DYTools::nMassBins, DYTools::massBinLimits,
DYTools::nMassBins, DYTools::massBinLimits);
for(int i=1; i<=DYTools::nMassBins; i++)
for(int j=1; j<=DYTools::nMassBins; j++)
hInvResponse->SetBinContent(i,j,DetInvertedResponse(i-1,j-1));
TCanvas *f = MakeCanvas("invResponse","invResponse",xsize,ysize);
canvasV.push_back(f);
CPlot plotInvResponse("inverted response","",
"reconstructed m(ee) [GeV/c^{2}]",
"generator post-FSR m(ee) [GeV/c^{2}]");
plotInvResponse.AddHist2D(hInvResponse,"COLZ");
f->cd();
plotInvResponse.SetLogx();
plotInvResponse.SetLogy();
gPad->SetRightMargin(0.15);
gStyle->SetPalette(1);
hInvResponse->GetXaxis()->SetMoreLogLabels();
hInvResponse->GetXaxis()->SetNoExponent();
hInvResponse->GetYaxis()->SetNoExponent();
plotInvResponse.Draw(f);
// Create a plot of detector resolution without mass binning
TCanvas *g = MakeCanvas("massDiff","massDiff",600,600);
canvasV.push_back(g);
CPlot plotMassDiff("massDiff","","reco mass - gen post-FSR mass [GeV/c^{2}]","a.u.");
hMassDiffBB->Scale(1.0/hMassDiffBB->GetSumOfWeights());
hMassDiffEB->Scale(1.0/hMassDiffEB->GetSumOfWeights());
hMassDiffEE->Scale(1.0/hMassDiffEE->GetSumOfWeights());
plotMassDiff.AddHist1D(hMassDiffBB,"EB-EB","hist",kBlack);
plotMassDiff.AddHist1D(hMassDiffEB,"EE-EB","hist",kBlue);
plotMassDiff.AddHist1D(hMassDiffEE,"EE-EE","hist",kRed);
plotMassDiff.TransLegend(0.1, 0.0);
plotMassDiff.Draw(g);
// Create a plot of reco - gen post-FSR mass difference for several mass bins
TCanvas *h = MakeCanvas("massDiffV","massDiffV",600,600);
canvasV.push_back(h);
CPlot plotMassDiffV("massDiffV","","reco mass - gen post-FSR mass [GeV/c^{2}]","a.u.");
hMassDiffV[7]->Scale(1.0/hMassDiffV[7]->GetSumOfWeights());
hMassDiffV[6]->Scale(1.0/hMassDiffV[6]->GetSumOfWeights());
hMassDiffV[5]->Scale(1.0/hMassDiffV[5]->GetSumOfWeights());
hMassDiffV[4]->Scale(1.0/hMassDiffV[4]->GetSumOfWeights());
plotMassDiffV.AddHist1D(hMassDiffV[4],"50-60 GeV/c^{2}","hist",kBlack);
plotMassDiffV.AddHist1D(hMassDiffV[5],"60-76 GeV/c^{2}","hist",kBlue);
plotMassDiffV.AddHist1D(hMassDiffV[6],"76-86 GeV/c^{2}","hist",kRed);
plotMassDiffV.AddHist1D(hMassDiffV[7],"86-96 GeV/c^{2}","hist",kMagenta);
plotMassDiffV.SetXRange(-20,50);
plotMassDiffV.Draw(h);
// Create graph of bin-to-bin corrections
TGraphAsymmErrors *grCorrFactor
= new TGraphAsymmErrors(massBinCentral, DetCorrFactor,
massBinHalfWidth, massBinHalfWidth,
DetCorrFactorErrNeg, DetCorrFactorErrPos);
TCanvas *c11 = MakeCanvas("corrFactor","corrFactor",800,600);
canvasV.push_back(c11);
CPlot plotCorrFactor("corrFactor","","m(ee) [GeV/c^{2}]","correction factor");
plotCorrFactor.AddGraph(grCorrFactor,"PE",kBlue);
plotCorrFactor.SetLogx();
plotCorrFactor.SetYRange(0,2.0);
plotCorrFactor.Draw(c11);
if (saveMadePlots) {
TString plotPath=TString("../root_files/plots/") + dirTag;
TString name=plotPath + TString("/unfolding_plots.root");
gSystem->mkdir(plotPath,true);
TFile* fplots = TFile::Open(name,"RECREATE");
for (unsigned int i=0; i<canvasV.size(); ++i) {
canvasV[i]->Write();
}
fplots->Close();
}
//--------------------------------------------------------------------------------------------------------------
// Summary print out
//==============================================================================================================
cout << endl;
cout << "*" << endl;
cout << "* SUMMARY" << endl;
cout << "*--------------------------------------------------" << endl;
cout << endl;
// Printout of all constants, uncomment if needed
//printf("DetMigration:\n"); DetMigration.Print();
//printf("DetResponse:\n"); DetResponse.Print();
//printf("DetInvertedResponse:\n"); DetInvertedResponse.Print();
// DetCorrFactor.Print();
// DetResponse.Print();
// printf("Detector corr factor numerator:\n");
// DetCorrFactorNumerator.Print();
// printf("yieldsMcRec:\n");
// yieldsMcRec.Print();
// printf("Detector corr factor denominator:\n");
// DetCorrFactorDenominator.Print();
// printf("yieldsMcFsrOfRec:\n");
// yieldsMcFsrOfRec.Print();
gBenchmark->Show("plotDYUnfoldingMatrix");
}
int main( int argc, char** argv )
{
CCP4program prog( "cecalc", "0.1", "$Date: 2004/07/01" );
// defaults
clipper::String ipfile = "NONE";
clipper::String ipcolfo = "NONE";
clipper::String opfile = "ecalc.mtz";
clipper::String opcol = "ecalc";
const int nprm = 12;
// command input
CommandInput args( argc, argv, true );
int arg = 0;
while ( ++arg < args.size() ) {
if ( args[arg] == "-mtzin" ) {
if ( ++arg < args.size() ) ipfile = args[arg];
} else if ( args[arg] == "-mtzout" ) {
if ( ++arg < args.size() ) opfile = args[arg];
} else if ( args[arg] == "-colin-fo" ) {
if ( ++arg < args.size() ) ipcolfo = args[arg];
} else if ( args[arg] == "-colout" ) {
if ( ++arg < args.size() ) opcol = args[arg];
} else {
std::cout << "Unrecognized:\t" << args[arg] << "\n";
args.clear();
}
}
if ( args.size() <= 1 ) {
std::cout << "Usage: cecalc\n\t-mtzin <filename>\n\t-mtzout <filename>\n\t-colin-fo <colpath>\n\t-colout <colpath>\nCalculate E's from F's\n";
exit(1);
}
// make data objects
clipper::CCP4MTZfile mtzin, mtzout;
clipper::HKL_info hkls;
typedef clipper::HKL_data_base::HKL_reference_index HRI;
// open file
mtzin.open_read( ipfile );
mtzin.import_hkl_info( hkls );
clipper::HKL_data<clipper::data32::F_sigF> fsig( hkls );
mtzin.import_hkl_data( fsig, ipcolfo );
if ( opcol[0] != '/' ) opcol = mtzin.assigned_paths()[0].notail()+"/"+opcol;
mtzin.close_read();
// create initial E
clipper::HKL_data<clipper::data32::E_sigE> esig( hkls );
esig.compute( fsig, clipper::data32::Compute_EsigE_from_FsigF() );
// calc E-scaling
std::vector<double> params_init( nprm, 1.0 );
clipper::BasisFn_spline basis_fo( esig, nprm, 2.0 );
clipper::TargetFn_scaleEsq<clipper::data32::E_sigE> target_fo( esig );
clipper::ResolutionFn escale( hkls, basis_fo, target_fo, params_init );
// apply E-scaling
for ( HRI ih = esig.first(); !ih.last(); ih.next() )
if ( !esig[ih].missing() ) esig[ih].scale( sqrt( escale.f(ih) ) );
// output data
mtzout.open_append( ipfile, opfile );
mtzout.export_hkl_data( esig, opcol );
mtzout.close_append();
}
void calcEscaleSystematics(TString lumiTag="DY_j22_19712pb",
int saveTexTable=0){
int nUnfoldingBins = DYTools::getTotalNumberOfBins();
TVectorD observedYields(nUnfoldingBins);
TVectorD observedYieldsErr(nUnfoldingBins);
TVectorD dummyArr(nUnfoldingBins);
TVectorD unfoldedYields(nUnfoldingBins);
//
// Indicators (and also counters) whether the data was present
//
int countEScaleSyst=0;
int countResidualShape=0;
int countEtaSyst=0;
int countFittingShape=0;
std::vector<TString> usedFiles;
ElectronEnergyScale escale("Date20120802_default");
//
// Calculate error associated with statistical
// uncertainty on the energy scale factors
//
TVectorD unfoldedYieldsMean(nUnfoldingBins);
TVectorD unfoldedYieldsRMS(nUnfoldingBins);
TVectorD unfoldedYieldsSquaredMean(nUnfoldingBins);
unfoldedYieldsMean = 0;
unfoldedYieldsRMS = 0;
unfoldedYieldsSquaredMean = 0;
TString matrixFileName = TString("../root_files/constants/") + lumiTag +
TString("/detResponse_unfolding_constants") + DYTools::analysisTag + TString("_PU.root");
const int nFiles1 = 20; // expected number of files
if (1)
for(int ifile=0; ifile<nFiles1; ifile++){
int seed = 1001+ifile;
escale.randomizeEnergyScaleCorrections(seed);
TString fname = TString("../root_files/yields/") + lumiTag +
TString("_escale_randomized/yields_bg-subtracted") + DYTools::analysisTag +
TString("__") + escale.calibrationSetShortName();
fname += ".root";
TFile file(fname);
if (!file.IsOpen()) {
std::cout << "failed to open a file <" << fname << ">\n";
}
else {
file.Close();
// register
usedFiles.push_back(fname);
// work with data
int res= (readData(fname, observedYields,observedYieldsErr,dummyArr) == 1)
&& (applyUnfoldingLocal(observedYields,unfoldedYields,matrixFileName) == 1);
if (res==1) {
countEScaleSyst++;
//std::cout << "unfoldedYields for seed=" << seed << ": "; unfoldedYields.Print();
// Accumulate mean and RMS
for(int idx = 0; idx < nUnfoldingBins; idx++){
unfoldedYieldsMean[idx] += unfoldedYields[idx];
unfoldedYieldsSquaredMean[idx] += unfoldedYields[idx]*unfoldedYields[idx];
}
}
}
}
// Final calculation of the mean and RMS for Smearing
TVectorD escaleRandomizedSystRelative(nUnfoldingBins);
if (countEScaleSyst) {
for(int idx = 0; idx < nUnfoldingBins; idx++){
unfoldedYieldsMean[idx] = unfoldedYieldsMean[idx]/double(countEScaleSyst);
unfoldedYieldsSquaredMean[idx] = unfoldedYieldsSquaredMean[idx]/double(countEScaleSyst);
unfoldedYieldsRMS[idx] = sqrt(unfoldedYieldsSquaredMean[idx] -
unfoldedYieldsMean[idx]*unfoldedYieldsMean[idx]);
escaleRandomizedSystRelative[idx] = unfoldedYieldsRMS[idx]/unfoldedYieldsMean[idx];
}
//std::cout << "unfoldedYieldsMean=" << ": "; unfoldedYieldsMean.Print();
//std::cout << "escaleRandomizedSystRelative: "; escaleRandomizedSystRelative.Print();
}
else {
std::cout << "escaleRandomizomized files were not found\n";
}
//
// Calculate error related to the difference between
// data and MC mass distribution shape after all corrections
// had been applied to data and MC. "Residual differences"
//
TVectorD unfoldedYieldsVariation(nUnfoldingBins);
TVectorD escaleResidualDiffSystRelative(nUnfoldingBins);
TString fname = TString("../root_files/yields/") + lumiTag +
TString("/yields_bg-subtracted") + DYTools::analysisTag + TString(".root");
int res=1;
{
TFile file(fname);
if (file.IsOpen()) {
file.Close();
if (readData(fname,observedYields,observedYieldsErr,dummyArr) != 1) {
std::cout << "failed to get data from file <" << fname << ">\n";
throw 2;
}
//
matrixFileName = TString("../root_files/constants/") + lumiTag +
TString("/detResponse_unfolding_constants") + DYTools::analysisTag + TString("_PU.root");
res=applyUnfoldingLocal(observedYields, unfoldedYields, matrixFileName);
if (res==1) usedFiles.push_back(matrixFileName);
//
if (res==1) {
matrixFileName = TString("../root_files/constants/") + lumiTag +
TString("_escale_residual/unfolding_constants") + DYTools::analysisTag +
TString("_escaleResidual.root");
res=applyUnfoldingLocal(observedYields, unfoldedYieldsVariation, matrixFileName);
if (res==1) {
countResidualShape++;
usedFiles.push_back(matrixFileName);
}
}
}
if (res==1) {
for(int idx=0; idx<nUnfoldingBins; idx++){
escaleResidualDiffSystRelative[idx] =
fabs(unfoldedYields[idx] - unfoldedYieldsVariation[idx])
/(unfoldedYields[idx] + unfoldedYieldsVariation[idx]);
}
}
}
//
// Calculate error related to extra smearing function shape
//
std::vector<TString> shapeNames;
shapeNames.push_back("_6binNegs_BreitWigner_20120802");
shapeNames.push_back("_6binNegs_Voigtian_20120802");
std::vector<TVectorD*> unfoldedYieldsShape;
if (1)
for (unsigned int i=0; i<shapeNames.size(); ++i) {
TVectorD observedYieldsShape(nUnfoldingBins);
TVectorD observedYieldsShapeErr(nUnfoldingBins);
TString shapeFName=TString("../root_files/yields/") + lumiTag +
TString("_escale_shape/yields_bg-subtracted") + DYTools::analysisTag +
TString("_") + shapeNames[i] + TString(".root");
TFile fShape(shapeFName);
if (fShape.IsOpen()) {
fShape.Close();
usedFiles.push_back(shapeFName);
if ( readData(shapeFName,observedYieldsShape,observedYieldsShapeErr,dummyArr) !=1) {
std::cout << "failed to load data from <" << shapeFName << ">\n";
throw 2;
}
TString matrixFileNameShape = TString("../root_files/constants/") + lumiTag +
TString("_escale_shape/unfolding_constants") + DYTools::analysisTag +
TString("_") + shapeNames[i] + TString(".root");
TVectorD *shapeYields=new TVectorD(nUnfoldingBins);
{
TFile ftmp(TString("tmp_") + shapeNames[i] + TString(".root"),"recreate");
observedYieldsShape.Write("YieldsSignal");
TVectorD tmp(nUnfoldingBins);
tmp=0;
tmp.Write("YieldsSignalErr");
tmp.Write("YieldsSignalSystErr");
ftmp.Close();
}
res=applyUnfoldingLocal(observedYieldsShape, *shapeYields, matrixFileNameShape);
if (res) {
countFittingShape++;
usedFiles.push_back(matrixFileNameShape);
unfoldedYieldsShape.push_back(shapeYields);
}
else delete shapeYields;
}
}
//
TVectorD escaleFitShapeSystRelative(nUnfoldingBins);
for(int idx=0; idx<nUnfoldingBins; idx++){
double min = unfoldedYields[idx];
double max = unfoldedYields[idx];
for (unsigned int iShape=0; iShape<unfoldedYieldsShape.size(); ++iShape) {
min = TMath::Min( min, (*unfoldedYieldsShape[iShape])[idx] );
max = TMath::Max( max, (*unfoldedYieldsShape[iShape])[idx] );
}
escaleFitShapeSystRelative[idx] = fabs(max-min)/(max+min);
}
//
// Calculate error related to eta binning
//
std::vector<TString> etaBinNames;
etaBinNames.push_back("_2binNegs_Gauss_20120802");
etaBinNames.push_back("_3EB3EENegs_Gauss_20120802");
etaBinNames.push_back("_4binNegs_Gauss_20120802");
etaBinNames.push_back("_4EB3EENegs_Gauss_20120802");
etaBinNames.push_back("_5binNegs_Gauss_20120802");
etaBinNames.push_back("_6binNegs_Gauss_20120802"); // default
etaBinNames.push_back("_6bins_Gauss_20120802");
std::vector<TVectorD*> unfoldedYieldsEta;
unfoldedYieldsEta.reserve(etaBinNames.size());
if (1)
for (unsigned int i=0; i<etaBinNames.size(); ++i) {
TString fEtaFileName=TString("../root_files/yields/") + lumiTag +
TString("_escale_eta/yields_bg-subtracted") + DYTools::analysisTag +
TString("_") + etaBinNames[i] + TString(".root");
TFile fEta(fEtaFileName);
if (fEta.IsOpen()) {
fEta.Close();
usedFiles.push_back(fEtaFileName);
TVectorD observedYieldsEta(nUnfoldingBins);
TVectorD observedYieldsEtaErr(nUnfoldingBins);
if (readData(fEtaFileName,observedYieldsEta,observedYieldsEtaErr,dummyArr)!=1) {
std::cout << "failed to get info from <" << fEtaFileName << ">\n";
throw 2;
}
matrixFileName = TString("../root_files/constants/") + lumiTag +
TString("_escale_eta/unfolding_constants") + DYTools::analysisTag +
TString("_") + etaBinNames[i] + TString(".root");
TVectorD *unfYields = new TVectorD(nUnfoldingBins);
res=applyUnfoldingLocal(observedYieldsEta, *unfYields, matrixFileName);
if (res==1) {
countEtaSyst++;
usedFiles.push_back(matrixFileName);
unfoldedYieldsEta.push_back(unfYields);
}
else {
delete unfYields;
}
}
}
//
TVectorD escaleEtaBinSystRelative(nUnfoldingBins);
for(int idx=0; idx<nUnfoldingBins; idx++){
double min = unfoldedYields[idx];
double max = unfoldedYields[idx];
for (unsigned int iShape=0; iShape<unfoldedYieldsEta.size(); ++iShape) {
min = TMath::Min( min, (*unfoldedYieldsEta[iShape])[idx] );
max = TMath::Max( max, (*unfoldedYieldsEta[iShape])[idx] );
}
escaleEtaBinSystRelative[idx] = fabs(max-min)/(max+min);
}
//
// Put all errors together
//
TVectorD escaleSystPercent(nUnfoldingBins);
for(int idx = 0; idx < nUnfoldingBins; idx++){
escaleSystPercent[idx] = 100.0 *
sqrt(
escaleRandomizedSystRelative[idx]*escaleRandomizedSystRelative[idx]
+ escaleResidualDiffSystRelative[idx]*escaleResidualDiffSystRelative[idx]
+ escaleFitShapeSystRelative[idx]*escaleFitShapeSystRelative[idx]
+ escaleEtaBinSystRelative[idx]*escaleEtaBinSystRelative[idx]
);
}
// Don't write TObject part of the objects
TDescriptiveInfo_t::Class()->IgnoreTObjectStreamer();
TDescriptiveInfo_t *info= new TDescriptiveInfo_t();
std::vector<std::string> *lines = (info) ? (& info->_info) : NULL;
const int bufsize=300;
char buf[bufsize];
std::string sbuf;
snprintf(buf,bufsize," Summary of successful loads:\n");
printf(buf);
if (lines) lines->push_back(buf);
snprintf(buf,bufsize," escale systematics file count= %2d\n", countEScaleSyst);
printf(buf);
if (lines) lines->push_back(buf);
snprintf(buf,bufsize," residual shape syst. file count= %2d\n", countResidualShape);
printf(buf);
if (lines) lines->push_back(buf);
snprintf(buf,bufsize," eta systematics file count= %2d\n", countEtaSyst);
printf(buf);
if (lines) lines->push_back(buf);
snprintf(buf,bufsize," fitting shape file count= %2d\n", countFittingShape);
printf(buf);
if (lines) { lines->push_back(buf); lines->push_back("\n"); }
std::cout << "\n";
int listFilesOnScreen=1;
snprintf(buf,bufsize," Loaded files:\n");
if (listFilesOnScreen) printf(buf);
if (lines) lines->push_back(buf);
for (unsigned int i=0; i<usedFiles.size(); ++i) {
snprintf(buf,bufsize," %s\n",usedFiles[i].Data());
if (listFilesOnScreen) printf(buf);
if (lines) lines->push_back(buf);
}
std::cout << std::endl;
if (saveTexTable) {
std::vector<TString> headers;
std::vector<int> padding;
std::vector<TVectorD*> data;
std::vector<double> factors;
headers.push_back("mass range");
headers.push_back("rapidity range");
headers.push_back("\\multicolumn{1}{c|}{statistical, \\%}"); data.push_back(&escaleRandomizedSystRelative); factors.push_back(100.); padding.push_back(6);
headers.push_back("\\multicolumn{1}{c|}{shape, \\%}"); data.push_back(&escaleFitShapeSystRelative); factors.push_back(100.); padding.push_back(4);
headers.push_back("\\multicolumn{1}{c|}{residual, \\%}"); data.push_back(&escaleResidualDiffSystRelative); factors.push_back(100.); padding.push_back(5);
headers.push_back("\\multicolumn{1}{c|}{$\\eta$ binning, \\%}"); data.push_back(&escaleEtaBinSystRelative); factors.push_back(100.); padding.push_back(6);
headers.push_back("\\multicolumn{1}{c|}{total, \\%}"); data.push_back(&escaleSystPercent); factors.push_back(1.); padding.push_back(3);
TString texFName=TString("../root_files/systematics/") + lumiTag +
TString("/escale_systematics") + DYTools::analysisTag + TString("_tmp.tex");
printTexTable(texFName,headers,padding,data,factors);
}
TString finalFName=TString("../root_files/systematics/") + lumiTag +
TString("/escale_systematics") + DYTools::analysisTag + TString("_tmp.root");
TFile fout(finalFName,"recreate");
unfolding::writeBinningArrays(fout);
escaleRandomizedSystRelative.Write("escaleRandomizedSystRelativeFI");
escaleFitShapeSystRelative.Write("escaleFitShapeSystRelativeFI");
escaleResidualDiffSystRelative.Write("escaleResidualDiffSystRelativeFI");
escaleEtaBinSystRelative.Write("escaleEtaBinSystRelativeFI");
escaleSystPercent.Write("escaleSystPercentFI");
if (info) {
TTree *infoTree = new TTree("Description","description");
infoTree->Branch("description","TDescriptiveInfo_t",&info);
infoTree->Fill();
infoTree->Write();
}
fout.Close();
return;
}
