CMS 3D CMS Logo

 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Properties Friends Macros Pages
LRHelpFunctions.cc
Go to the documentation of this file.
2 #include "TopQuarkAnalysis/TopTools/test/tdrstyle.C"
3 
4 // constructors
6  constructPurity = false;
7  setTDRStyle();
8  gStyle->SetCanvasDefW(900);
9 }
10 
11 
12 LRHelpFunctions::LRHelpFunctions(const std::vector<int>& obsNr, int nrBins, const std::vector<double>& obsMin, const std::vector<double>& obsMax, const std::vector<const char*>& functions) {
13  obsNumbers = obsNr;
14  constructPurity = false;
15  setTDRStyle();
16  gStyle->SetCanvasDefW(900);
17  for(size_t o=0; o<obsNr.size(); o++){
18  // create Signal, background and s/(s+b) histogram
19  TString htS = "Obs"; htS += obsNr[o]; htS += "_S";
20  TString htB = "Obs"; htB += obsNr[o]; htB += "_B";
21  TString htSB = "Obs"; htSB += obsNr[o]; htSB += "_SoverSplusB";
22  hObsS.push_back( new TH1F(htS,htS,nrBins,obsMin[o],obsMax[o]) );
23  hObsB.push_back( new TH1F(htB,htB,nrBins,obsMin[o],obsMax[o]) );
24  hObsSoverSplusB.push_back( new TH1F(htSB,htSB,nrBins,obsMin[o],obsMax[o]) );
25 
26  // create the correlation 2D plots among the observables (for signal)
27  for (size_t o2 = o+1; o2 < obsNr.size(); o2++) {
28  TString hcorr = "Corr_Obs"; hcorr += obsNr[o]; hcorr += "_Obs"; hcorr += obsNr[o2];
29  hObsCorr.push_back( new TH2F(hcorr,hcorr,nrBins,obsMin[o],obsMax[o],nrBins,obsMin[o2],obsMax[o2]) );
30  }
31 
32  // create fit functions
33  TString ftSB = "F_Obs"; ftSB += obsNr[o]; ftSB += "_SoverSplusB";
34  fObsSoverSplusB.push_back( new TF1(ftSB,functions[o],hObsS[o]->GetXaxis()->GetXmin(),hObsS[o]->GetXaxis()->GetXmax()) );
35  }
36 }
37 
38 void LRHelpFunctions::recreateFitFct(const std::vector<int>& obsNr, const std::vector<const char*>& functions) {
39  if (!fObsSoverSplusB.empty()) fObsSoverSplusB.clear();
40  for(size_t o=0; o<obsNr.size(); o++){
41  // create fit functions
42  TString ftSB = "F_Obs"; ftSB += obsNr[o]; ftSB += "_SoverSplusB";
43  fObsSoverSplusB.push_back( new TF1(ftSB,functions[o],hObsS[o]->GetXaxis()->GetXmin(),hObsS[o]->GetXaxis()->GetXmax()) );
44  }
45 }
46 
47 LRHelpFunctions::LRHelpFunctions(int nrLRbins, double LRmin, double LRmax, const char* LRfunction) {
48  constructPurity = true;
49  setTDRStyle();
50  gStyle->SetCanvasDefW(900);
51 
52  // create LR histograms
53  hLRtotS = new TH1F("hLRtotS","hLRtotS",nrLRbins,LRmin,LRmax);
54  hLRtotS->GetXaxis()->SetTitle("Combined LR");
55  hLRtotB = new TH1F("hLRtotB","hLRtotB",nrLRbins,LRmin,LRmax);
56  hLRtotB->GetXaxis()->SetTitle("Combined LR");
57  hLRtotSoverSplusB = new TH1F("hLRtotSoverSplusB","hLRtotSoverSplusB",nrLRbins,LRmin,LRmax);
58 
59  // create LR fit function
60  fLRtotSoverSplusB = new TF1("fLRtotSoverSplusB",LRfunction,LRmin,LRmax);
61 }
62 
63 
64 // destructor
66 
67 
68 // member function to initialize the LR hists and fits
69 void LRHelpFunctions::initLRHistsAndFits(int nrLRbins, double LRmin, double LRmax, const char* LRfunction){
70  constructPurity = true;
71  // create LR histograms
72  hLRtotS = new TH1F("hLRtotS","hLRtotS",nrLRbins,LRmin,LRmax);
73  hLRtotB = new TH1F("hLRtotB","hLRtotB",nrLRbins,LRmin,LRmax);
74  hLRtotSoverSplusB = new TH1F("hLRtotSoverSplusB","hLRtotSoverSplusB",nrLRbins,LRmin,LRmax);
75  // create LR fit function
76  fLRtotSoverSplusB = new TF1("fLRtotSoverSplusB",LRfunction,LRmin,LRmax);
77 }
78 
79 
80 
81 // member function to set initial values to the observable fit function
82 void LRHelpFunctions::setObsFitParameters(int obs,const std::vector<double>& fitPars){
83  for(size_t fit=0; fit<fObsSoverSplusB.size(); fit++){
84  TString fn = "_Obs"; fn += obs;
85  if(((TString)fObsSoverSplusB[fit]->GetName()).Contains(fn)){
86  for(size_t p=0; p<fitPars.size(); p++){
87  fObsSoverSplusB[fit]->SetParameter(p,fitPars[p]);
88  }
89  }
90  }
91 }
92 
93 
94 
95 
96 
97 // member function to add observable values to the signal histograms
98 void LRHelpFunctions::fillToSignalHists(const std::vector<double>& obsVals, double weight){
99  int hIndex = 0;
100  for(size_t o=0; o<obsVals.size(); o++) {
101  hObsS[o]->Fill(obsVals[o], weight);
102  for(size_t o2=o+1; o2<obsVals.size(); o2++) {
103  hObsCorr[hIndex] -> Fill(obsVals[o],obsVals[o2], weight);
104  ++hIndex;
105  }
106  }
107 }
108 
109 // member function to add observable values to the signal histograms
110 void LRHelpFunctions::fillToSignalHists(int obsNbr, double obsVal, double weight){
111  TString obs = "Obs"; obs += obsNbr; obs += "_";
112  for(size_t f = 0; f<hObsS.size(); f++){
113  if(((TString)(hObsS[f]->GetName())).Contains(obs)) {
114  hObsS[f]->Fill(obsVal, weight);
115  return;
116  }
117  }
118 }
119 
120 // member function to add observable values to the signal histograms
121 void LRHelpFunctions::fillToSignalCorrelation(int obsNbr1, double obsVal1, int obsNbr2,
122  double obsVal2, double weight){
123  TString hcorr = "Corr_Obs"; hcorr += std::min(obsNbr1,obsNbr2) ; hcorr += "_Obs"; hcorr += std::max(obsNbr1, obsNbr2);
124  for(size_t i=0; i<hObsCorr.size(); i++) {
125  if(((TString)(hObsCorr[i]->GetName())) == hcorr) {
126  if (obsNbr1 < obsNbr2) {
127  hObsCorr[i] -> Fill(obsVal1,obsVal2, weight);
128  } else {
129  hObsCorr[i] -> Fill(obsVal2,obsVal1, weight);
130  }
131  return;
132  }
133  }
134 }
135 
136 
137 // member function to add observable values to the background histograms
138 void LRHelpFunctions::fillToBackgroundHists(const std::vector<double>& obsVals, double weight){
139  for(size_t o=0; o<obsVals.size(); o++) hObsB[o]->Fill(obsVals[o], weight);
140 }
141 
142 // member function to add observable values to the signal histograms
143 void LRHelpFunctions::fillToBackgroundHists(int obsNbr, double obsVal, double weight){
144  TString obs = "Obs"; obs += obsNbr; obs += "_";
145  for(size_t f = 0; f<hObsB.size(); f++){
146  if(((TString)(hObsB[f]->GetName())).Contains(obs)) {
147  hObsB[f]->Fill(obsVal, weight);
148  return;
149  }
150  }
151 }
152 
153 // member function to normalize the S and B spectra
155  for(size_t o=0; o<hObsS.size(); o++){
156  // count entries in each histo. Do it this way instead of GetEntries method,
157  // since the latter does not account for the weights.
158  double nrSignEntries = 0, nrBackEntries = 0;
159  for (int i=0; i<=hObsS[o]->GetNbinsX()+1; i++){
160  nrSignEntries += hObsS[o]->GetBinContent(i);
161  nrBackEntries += hObsB[o]->GetBinContent(i);
162  }
163  for(int b=0; b <= hObsS[o]->GetNbinsX()+1; b++){
164  hObsS[o]->SetBinContent(b,hObsS[o]->GetBinContent(b)/(nrSignEntries));
165  hObsB[o]->SetBinContent(b,hObsB[o]->GetBinContent(b)/(nrBackEntries));
166  hObsS[o]->SetBinError(b,hObsS[o]->GetBinError(b)/(nrSignEntries));
167  hObsB[o]->SetBinError(b,hObsB[o]->GetBinError(b)/(nrBackEntries));
168  }
169  //std::cout<<"Integral for obs"<<o<<" S: "<<hObsS[o]->Integral(0,10000)<<" & obs"<<o<<" B: "<<hObsB[o]->Integral(0,10000)<<std::endl;
170  }
171 }
172 
173 
174 
175 // member function to produce and fit the S/(S+B) histograms
177  for(size_t o=0; o<hObsS.size(); o++){
178  for(int b=0; b <= hObsS[o]->GetNbinsX()+1; b++){
179  if ((hObsS[o]->GetBinContent(b)+ hObsB[o]->GetBinContent(b)) > 0) {
180  hObsSoverSplusB[o]->SetBinContent(b, hObsS[o]->GetBinContent(b) / (hObsS[o]->GetBinContent(b) + hObsB[o]->GetBinContent(b)));
181  double error = sqrt( pow(hObsS[o]->GetBinError(b) * hObsB[o]->GetBinContent(b)/pow(hObsS[o]->GetBinContent(b) + hObsB[o]->GetBinContent(b),2),2)
182  + pow(hObsB[o]->GetBinError(b) * hObsS[o]->GetBinContent(b)/pow(hObsS[o]->GetBinContent(b) + hObsB[o]->GetBinContent(b),2),2) );
183  hObsSoverSplusB[o]->SetBinError(b,error);
184  }
185  }
186  hObsSoverSplusB[o]->Fit(fObsSoverSplusB[o]->GetName(),"RQ");
187  }
188 }
189 
190 
191 // member function to read the observable hists & fits from a root-file
192 void LRHelpFunctions::readObsHistsAndFits(const TString& fileName, const std::vector<int>& observables, bool readLRplots){
193  hObsS.clear();
194  hObsB.clear();
195  hObsSoverSplusB.clear();
196  fObsSoverSplusB.clear();
197  TFile *fitFile = new TFile(fileName, "READ");
198  if(observables[0] == -1){
199  std::cout<<" ... will read hists and fit for all available observables in file "<<fileName<<std::endl;
200  TList *list = fitFile->GetListOfKeys();
201  TIter next(list);
202  TKey *el;
203  while ((el = (TKey*)next())) {
204  TString keyName = el->GetName();
205  if(keyName.Contains("F_") && keyName.Contains("_SoverSplusB")){
206  fObsSoverSplusB.push_back( new TF1(*((TF1*) el -> ReadObj())));
207  }
208  else if(keyName.Contains("_SoverSplusB")){
209  hObsSoverSplusB.push_back( new TH1F(*((TH1F*) el -> ReadObj())));
210  }
211  else if(keyName.Contains("_S")){
212  hObsS.push_back( new TH1F(*((TH1F*) el -> ReadObj())));
213  }
214  else if(keyName.Contains("_B")){
215  hObsB.push_back( new TH1F(*((TH1F*) el -> ReadObj())));
216  }
217  else if(keyName.Contains("Corr")){
218  hObsCorr.push_back( new TH2F(*((TH2F*) el -> ReadObj())));
219  }
220  }
221  }
222  else{
223  obsNumbers = observables;
224  for(unsigned int obs = 0; obs < observables.size(); obs++){
225  std::cout<<" ... will read hists and fit for obs "<<observables[obs]<<" from file "<<fileName<<std::endl;
226  TString hStitle = "Obs"; hStitle += observables[obs]; hStitle += "_S";
227  hObsS.push_back( new TH1F(*((TH1F*)fitFile->GetKey(hStitle)->ReadObj())));
228  TString hBtitle = "Obs"; hBtitle += observables[obs]; hBtitle += "_B";
229  hObsB.push_back( new TH1F(*((TH1F*)fitFile->GetKey(hBtitle)->ReadObj())));
230  TString hSBtitle = "Obs"; hSBtitle += observables[obs]; hSBtitle += "_SoverSplusB";
231  TString fSBtitle = "F_"; fSBtitle += hSBtitle;
232  hObsSoverSplusB.push_back( new TH1F(*((TH1F*)fitFile->GetKey(hSBtitle)->ReadObj())));
233  fObsSoverSplusB.push_back( new TF1(*((TF1*)fitFile->GetKey(fSBtitle)->ReadObj())));
234  for(unsigned int obs2 = obs+1; obs2 < observables.size(); obs2++){
235  TString hCorrtitle = "Corr_Obs"; hCorrtitle += observables[obs];
236  hCorrtitle += "_Obs"; hCorrtitle += observables[obs2];
237  hObsCorr.push_back( new TH2F(*((TH2F*)fitFile->GetKey(hCorrtitle)->ReadObj())));
238  }
239  }
240  }
241 
242  if(readLRplots){
243  constructPurity = true;
244  std::cout<<" ... will LR s and B histos from file "<<fileName<<std::endl;
245  hLRtotS = new TH1F(*((TH1F*)fitFile->GetKey("hLRtotS")->ReadObj()));
246  hLRtotB = new TH1F(*((TH1F*)fitFile->GetKey("hLRtotB")->ReadObj()));
247  hLRtotSoverSplusB = new TH1F(*((TH1F*)fitFile->GetKey("hLRtotSoverSplusB")->ReadObj()));
248  fLRtotSoverSplusB = new TF1(*((TF1*)fitFile->GetKey("fLRtotSoverSplusB") -> ReadObj()));
249  }
250 }
251 
252 
253 
254 
255 // member function to store all observable plots and fits to a ROOT-file
257  TFile fOut(fname,"RECREATE");
258  fOut.cd();
259  for(size_t o=0; o<hObsS.size(); o++){
260  hObsS[o] -> Write();
261  hObsB[o] -> Write();
262  hObsSoverSplusB[o] -> Write();
263  fObsSoverSplusB[o] -> Write();
264  }
265  int hIndex = 0;
266  for(size_t o=0; o<hObsS.size(); o++) {
267  for(size_t o2=o+1; o2<hObsS.size(); o2++) {
268  hObsCorr[hIndex] -> Write();
269  ++ hIndex;
270  }
271  }
272  if(constructPurity){
273  hLRtotS -> Write();
274  hLRtotB -> Write();
275  hLRtotSoverSplusB -> Write();
276  fLRtotSoverSplusB -> Write();
277  hEffvsPur -> Write();
278  hLRValvsPur -> Write();
279  hLRValvsEff -> Write();
280  }
281  fOut.cd();
282  fOut.Write();
283  fOut.Close();
284 }
285 
286 
287 
288 
289 
290 // member function to make some simple control plots and store them in a ps-file
292  TCanvas c("dummy","",1);
293  c.Print(fname + "[","landscape");
294  for(size_t o=0; o<hObsS.size(); o++) {
295  TCanvas c2("c2","",1);
296  c2.Divide(2,1);
297  c2.cd(1);
298  hObsS[o] -> SetLineColor(2);
299  if(hObsB[o]->GetMaximum() > hObsS[o]->GetMaximum()){
300  hObsB[o] -> Draw("hist");
301  hObsS[o] -> Draw("histsame");
302  }
303  else
304  {
305  hObsS[o] -> Draw("hist");
306  hObsB[o] -> Draw("histsame");
307  }
308  c2.cd(2);
309  hObsSoverSplusB[o] -> Draw();
310  fObsSoverSplusB[o] -> Draw("same");
311  c2.Print(fname,"Landscape");
312  }
313 
314  int hIndex = 0;
315  for(size_t o=0; o<hObsS.size(); o++) {
316  for(size_t o2=o+1; o2<hObsS.size(); o2++) {
317  TCanvas cc("cc","",1);
318  hObsCorr[hIndex] -> Draw("box");
319  TPaveText pt(0.5,0.87,0.98,0.93,"blNDC");
320  pt.SetFillStyle(1);
321  pt.SetFillColor(0);
322  pt.SetBorderSize(0);
323  TString tcorr = "Corr. of "; tcorr += (int)(100.*hObsCorr[hIndex] -> GetCorrelationFactor()); tcorr += " %";
324  //TText *text = pt.AddText(tcorr);
325  pt.Draw("same");
326  ++hIndex;
327  cc.Print(fname,"Landscape");
328  }
329  }
330 
331  if(constructPurity){
332  TCanvas c3("c3","",1);
333  c3.Divide(2,1);
334  c3.cd(1);
335  hLRtotB -> Draw();
336  hLRtotS -> SetLineColor(2);
337  hLRtotS -> Draw("same");
338  c3.cd(2);
339  hLRtotSoverSplusB -> Draw();
340  c3.Print(fname,"Landscape");
341 
342  TCanvas c4("c4","",1);
343  hEffvsPur -> Draw("AL*");
344  c4.Print(fname,"Landscape");
345 
346  TCanvas c5("c5","",1);
347  hLRValvsPur -> Draw("AL*");
348  c5.Print(fname,"Landscape");
349 
350  TCanvas c6("c6","",1);
351  hLRValvsEff -> Draw("AL*");
352  c6.Print(fname,"Landscape");
353  }
354  c.Print(fname + "]","landscape");
355 }
356 
357 
358 
359 
360 // member function to calculate the LR value, using the S/N definition
361 double LRHelpFunctions::calcLRval(const std::vector<double>& vals){
362  double logLR = 0.;
363  for(size_t o=0; o<fObsSoverSplusB.size(); o++){
364  double SoverSplusN = fObsSoverSplusB[o]->Eval(vals[o]);
365  double SoverN = 1./((1./SoverSplusN) - 1.);
366  logLR += log(SoverN);
367  }
368  return logLR;
369 }
370 
371 
372 // member function to calculate the LR value, using the definition that was
373 // used in the P-TDR: S/(S+N)
374 
375 double LRHelpFunctions::calcPtdrLRval(const std::vector<double>& vals, bool useCorrelation) {
376  double logLR = 1.;
377  for(size_t o=0; o<fObsSoverSplusB.size(); o++){
378  double SoverSplusN = fObsSoverSplusB[o]->Eval(vals[o]);
379  if (SoverSplusN<0.0001) SoverSplusN=0.0001;
380  if (useCorrelation){
381  double corr = 0;
382  for(size_t j=0; j<fObsSoverSplusB.size(); j++){
383  if (o==j) ++corr;
384  else {
385  TString hcorr = "Corr_Obs"; hcorr += std::min(obsNumbers[o],obsNumbers[j]) ; hcorr += "_Obs"; hcorr += std::max(obsNumbers[o],obsNumbers[j]);
386  for(size_t i=0; i<hObsCorr.size(); i++) {
387  if(((TString)(hObsCorr[i]->GetName())) == hcorr)
388  corr += fabs(hObsCorr[i]->GetCorrelationFactor());
389  }
390  }
391  }
392  logLR *= pow(SoverSplusN/fObsSoverSplusB[o]->GetMaximum(), 1./corr);
393  }
394  else logLR *= SoverSplusN/fObsSoverSplusB[o]->GetMaximum();
395  }
396  //std::cout <<logLR<<std::endl;
397  return logLR;
398 }
399 
400 
401 // member function to fill a signal contribution to the LR histogram
402 void LRHelpFunctions::fillLRSignalHist(double val, double weight){
403  // std::cout<< "@@@===> LRHelpf Signal LRval = "<< val<<std::endl;
404  hLRtotS -> Fill(val, weight);
405 }
406 
407 
408 
409 // member function to fill a background contribution to the LR histogram
411  // std::cout<< "@@@===> LRHelpf Backgroud LRval = "<< val<<std::endl;
412  hLRtotB -> Fill(val, weight);
413 }
414 
415 
416 
417 // member function to make and fit the purity vs. LRval histogram, and the purity vs. efficiency graph
419 
420  for(int b=0; b<=hLRtotS->GetNbinsX(); b++) {
421  float Sint = hLRtotS->Integral(b, hLRtotS->GetNbinsX()+1);
422  float Bint = hLRtotB->Integral(b, hLRtotB->GetNbinsX()+1);
423  if (Sint + Bint > 0) {
424  hLRtotSoverSplusB->SetBinContent(b,1. * Sint / (Sint + Bint));
425  hLRtotSoverSplusB->SetBinError(b,sqrt((pow(Sint*sqrt(Bint),2)+pow(Bint*sqrt(Sint),2)))/pow((Sint+Bint),2));
426  }
427  }
428 
429  hLRtotS->GetXaxis()->SetTitle("Combined LR ratio");
430  hLRtotB->GetXaxis()->SetTitle("Combined LR ratio");
431  hLRtotSoverSplusB->GetXaxis()->SetTitle("Cut on Combined LR");
432  hLRtotSoverSplusB->GetYaxis()->SetTitle("Purity");
433 
434  hLRtotSoverSplusB -> Fit(fLRtotSoverSplusB->GetName(),"RQ");
435  double totSignal = hLRtotS->Integral(0,hLRtotS->GetNbinsX()+1);
436  double Eff[200], Pur[200], LRVal[200];
437  if (hLRtotS->GetNbinsX()>200) {
438  std::cout << "Number of bins of LR histograms can not execeed 200!"<<std::endl;
439  return;
440  }
441  for(int cut=0; (cut<=hLRtotS->GetNbinsX())&&(cut<200) ; cut ++){
442  double LRcutVal = hLRtotS->GetBinLowEdge(cut);
443  Eff[cut] = hLRtotS->Integral(cut,hLRtotS->GetNbinsX()+1)/totSignal;
444  Pur[cut] = fLRtotSoverSplusB->Eval(LRcutVal);
445  LRVal[cut] = LRcutVal;
446  }
447  hEffvsPur = new TGraph(hLRtotS->GetNbinsX(),Eff,Pur);
448  hEffvsPur -> SetName("hEffvsPur");
449  hEffvsPur -> SetTitle("");
450  hEffvsPur -> GetXaxis() -> SetTitle((TString)("Efficiency of cut on log combined LR"));
451  hEffvsPur -> GetYaxis() -> SetTitle((TString)("Purity"));
452  hEffvsPur -> GetYaxis() -> SetRangeUser(0,1.1);
453 
454  hLRValvsPur = new TGraph(hLRtotS->GetNbinsX(),LRVal,Pur);
455  hLRValvsPur -> SetName("hLRValvsPur");
456  hLRValvsPur -> SetTitle("");
457  hLRValvsPur -> GetXaxis() -> SetTitle((TString)("Cut on the log combined LR value"));
458  hLRValvsPur -> GetYaxis() -> SetTitle((TString)("Purity"));
459  hLRValvsPur -> GetYaxis() -> SetRangeUser(0,1.1);
460 
461  hLRValvsEff = new TGraph(hLRtotS->GetNbinsX(),LRVal,Eff);
462  hLRValvsEff -> SetName("hLRValvsEff");
463  hLRValvsEff -> SetTitle("");
464  hLRValvsEff -> GetXaxis() -> SetTitle((TString)("Cut on the log combined LR value"));
465  hLRValvsEff -> GetYaxis() -> SetTitle((TString)("Efficiency of cut on log combined LR"));
466  hLRValvsEff -> GetYaxis() -> SetRangeUser(0,1.1);
467 
468 }
469 
470 
471 
472 
473 // member function to calculate the probability for signal given a logLR value
474 double LRHelpFunctions::calcProb(double logLR){
475  return fLRtotSoverSplusB -> Eval(logLR);
476 }
477 
478 
479 // member to check if a certain S/S+B fit function is read from the root-file
481  bool included = false;
482  TString obs = "_Obs"; obs += o; obs += "_";
483  for(size_t f = 0; f<fObsSoverSplusB.size(); f++){
484  if(((TString)(fObsSoverSplusB[f]->GetName())).Contains(obs)) included = true;
485  }
486  return included;
487 }
488 
489 void LRHelpFunctions::setXlabels(const std::vector<std::string> & xLabels)
490 {
491  if (hObsS.size() != xLabels.size()) {
492  std::cout << "LRHelpFunctions::setXlabels: Number of labels ("
493  << xLabels.size()<< ") does not match number of obervables("
494  << hObsS.size()<<").\n";
495  return;
496  }
497  for(size_t i = 0; i<hObsS.size(); ++i){
498  hObsS[i]->GetXaxis()->SetTitle(TString(xLabels[i]));
499  hObsB[i]->GetXaxis()->SetTitle(TString(xLabels[i]));
500  hObsSoverSplusB[i]->GetXaxis()->SetTitle(TString(xLabels[i]));
501  hObsSoverSplusB[i]->GetYaxis()->SetTitle(TString("S/(S+B)"));
502  }
503 }
504 
505 void LRHelpFunctions::setYlabels(const std::vector<std::string> & yLabels)
506 {
507  if (hObsS.size() != yLabels.size()) {
508  std::cout << "LRHelpFunctions::setYlabels: Number of labels ("
509  << yLabels.size()<< ") does not match number of obervables("
510  << hObsS.size()<<").\n";
511  return;
512  }
513  for(size_t i = 0; i<hObsS.size(); ++i){
514  hObsS[i]->GetYaxis()->SetTitle(TString(yLabels[i]));
515  hObsB[i]->GetYaxis()->SetTitle(TString(yLabels[i]));
516  }
517 }
518 
519 void LRHelpFunctions::singlePlot(const TString& fname, int obsNbr, const TString& extension) {
520  if (!obsFitIncluded(obsNbr)) return;
521 
522  TStyle *tdrStyle = gROOT->GetStyle("tdrStyle");
523  tdrStyle->SetOptFit(0);
524  tdrStyle->SetOptStat(0);
525  tdrStyle->SetOptTitle(0);
526 // tdrStyle->SetPadTopMargin(0.01);
527 // tdrStyle->SetPadBottomMargin(0.01);
528 // tdrStyle->SetPadLeftMargin(0.01);
529 // tdrStyle->SetPadRightMargin(0.01);
530 
531  TCanvas c2("c2","",600,300);
532  c2.SetTopMargin(0.01);
533  c2.SetBottomMargin(0.01);
534  c2.SetLeftMargin(0.01);
535  c2.SetRightMargin(0.01);
536  std::cout <<fname<<std::endl;
537  c2.Divide(2,1);
538 
539  TString obs = "Obs"; obs += obsNbr; obs += "_";
540  for(size_t o = 0; o<hObsB.size(); ++o){
541  if(((TString)(hObsB[o]->GetName())).Contains(obs)) {
542  c2.cd(1);
543  hObsS[o] -> SetLineColor(2);
544  if(hObsB[o]->GetMaximum() > hObsS[o]->GetMaximum()){
545  hObsB[o] -> Draw("hist");
546  hObsS[o] -> Draw("histsame");
547  }
548  else
549  {
550  hObsS[o] -> Draw("hist");
551  hObsB[o] -> Draw("histsame");
552  }
553  c2.cd(2);
554 
555  hObsSoverSplusB[o] -> Draw();
556  fObsSoverSplusB[o] -> Draw("same");
557  }
558  }
559  std::cout << fname+"."+extension<<std::endl;
560  c2.Print(fname+"."+extension);
561 }
562 
563 void LRHelpFunctions::purityPlot(const TString& fname,const TString& extension)
564 {
565  TStyle *tdrStyle = gROOT->GetStyle("tdrStyle");
566  tdrStyle->SetOptFit(0);
567  tdrStyle->SetOptStat(0);
568  tdrStyle->SetOptTitle(0);
569 // tdrStyle->SetPadTopMargin(0.01);
570 // tdrStyle->SetPadBottomMargin(0.01);
571 // tdrStyle->SetPadLeftMargin(0.01);
572 // tdrStyle->SetPadRightMargin(0.01);
573 
574  TCanvas c2("c2","",600,300);
575  c2.SetTopMargin(0.01);
576  c2.SetBottomMargin(0.01);
577  c2.SetLeftMargin(0.01);
578  c2.SetRightMargin(0.01);
579  std::cout <<fname<<std::endl;
580  c2.Divide(2,1);
581 
582  c2.cd(1);
583  hLRValvsPur->Draw("AP");
584  c2.cd(2);
585  hEffvsPur->Draw("AP");
586  c2.Print(fname+"Purity."+extension);
587 
588  hLRtotS->GetXaxis()->SetNdivisions(505);
589  hLRtotB->GetXaxis()->SetNdivisions(505);
590  TCanvas c3("c2","",300,300);
591 // c3.SetTopMargin(0.01);
592 // c3.SetBottomMargin(0.01);
593 // c3.SetLeftMargin(0.01);
594 // c3.SetRightMargin(0.01);
595  hLRtotS->GetXaxis()->SetTitle("Combined LR");
596  hLRtotB->GetXaxis()->SetTitle("Combined LR");
597 
598  hLRtotB -> Draw();
599  hLRtotS -> SetLineColor(2);
600  hLRtotS -> Draw("same");
601  c3.Print(fname+"Dist."+extension);
602 
603 
604  TCanvas c5("c3","",900,600);
605  c5.Divide(2,2);
606  c5.cd(1);
607  hLRtotB -> Draw();
608  hLRtotS -> SetLineColor(2);
609  hLRtotS -> Draw("same");
610  c5.cd(3);
611  hLRValvsEff -> Draw("AP");
612  c5.cd(2);
613  hEffvsPur -> Draw("AP");
614  c5.cd(4);
615  hLRtotSoverSplusB -> Draw();
616  c5.Print(fname+"all."+extension);
617 }
void storeControlPlots(const TString &)
double calcLRval(const std::vector< double > &)
std::vector< TH1F * > hObsSoverSplusB
int i
Definition: DBlmapReader.cc:9
void setXlabels(const std::vector< std::string > &xLabels)
void recreateFitFct(const std::vector< int > &obsNr, const std::vector< const char * > &functions)
def setTDRStyle
Definition: plotscripts.py:87
std::vector< TH1F * > hObsS
double calcPtdrLRval(const std::vector< double > &vals, bool useCorrelation=false)
void makeAndFitPurityHists()
double calcProb(double)
std::vector< TH1F * > hObsB
void setYlabels(const std::vector< std::string > &yLabels)
TGraph * hLRValvsEff
bool obsFitIncluded(int)
std::vector< TH2F * > hObsCorr
TGraph * hLRValvsPur
void fillToBackgroundHists(const std::vector< double > &, double weight=1.0)
void fillToSignalCorrelation(int obsNbr1, double obsVal1, int obsNbr2, double obsVal2, double weight)
void makeAndFitSoverSplusBHists()
Definition: Fit.h:34
void Fill(HcalDetId &id, double val, std::vector< TH2F > &depth)
T sqrt(T t)
Definition: SSEVec.h:18
int j
Definition: DBlmapReader.cc:9
double f[11][100]
void storeToROOTfile(const TString &)
T min(T a, T b)
Definition: MathUtil.h:58
JetCorrectorParameters corr
Definition: classes.h:5
std::vector< int > obsNumbers
void fillLRSignalHist(double, double weight=1.0)
void singlePlot(const TString &fname, int obsNbr, const TString &extension=TString("eps"))
void readObsHistsAndFits(const TString &, const std::vector< int > &, bool)
void fillToSignalHists(const std::vector< double > &, double weight=1.0)
void setObsFitParameters(int obs, const std::vector< double > &)
double b
Definition: hdecay.h:120
string fname
main script
std::vector< TF1 * > fObsSoverSplusB
tuple cout
Definition: gather_cfg.py:145
void initLRHistsAndFits(int, double, double, const char *)
void fillLRBackgroundHist(double, double weight=1.0)
int weight
Definition: histoStyle.py:50
TH1F * hLRtotSoverSplusB
Power< A, B >::type pow(const A &a, const B &b)
Definition: Power.h:40
How EventSelector::AcceptEvent() decides whether to accept an event for output otherwise it is excluding the probing of A single or multiple positive and the trigger will pass if any such matching triggers are PASS or EXCEPTION[A criterion thatmatches no triggers at all is detected and causes a throw.] A single negative with an expectation of appropriate bit checking in the decision and the trigger will pass if any such matching triggers are FAIL or EXCEPTION A wildcarded negative criterion that matches more than one trigger in the trigger list("!*","!HLTx*"if it matches 2 triggers or more) will accept the event if all the matching triggers are FAIL.It will reject the event if any of the triggers are PASS or EXCEPTION(this matches the behavior of"!*"before the partial wildcard feature was incorporated).Triggers which are in the READY state are completely ignored.(READY should never be returned since the trigger paths have been run
void purityPlot(const TString &fname, const TString &extension=TString("eps"))