//
// Subset of matheamtical functions not included in the TMath
//
-
-///////////////////////////////////////////////////////////////////////////
+//
+/////////////////////////////////////////////////////////////////////////
#include "TMath.h"
#include "Riostream.h"
#include "TH1F.h"
+#include "TH2F.h"
#include "TH3.h"
#include "TF1.h"
#include "TTree.h"
#include "TChain.h"
#include "TObjString.h"
#include "TLinearFitter.h"
-
+#include "TGraph2D.h"
+#include "TGraph.h"
+#include "TGraphErrors.h"
+#include "TMultiGraph.h"
+#include "TCanvas.h"
+#include "TLatex.h"
+#include "TCut.h"
//
// includes neccessary for test functions
//
#include "TStatToolkit.h"
+
+using std::cout;
+using std::cerr;
+using std::endl;
+
ClassImp(TStatToolkit) // Class implementation to enable ROOT I/O
}
Double_t norm = 1./Double_t(hh);
- Double_t norm2 = 1./Double_t(hh-1);
+ Double_t norm2 = (hh-1)>0 ? 1./Double_t(hh-1):1;
for (Int_t i=hh; i<nvectors; i++){
Double_t cmean = sumx*norm;
Double_t csigma = (sumx2 - hh*cmean*cmean)*norm2;
}
//___TStatToolkit__________________________________________________________________________
-void TStatToolkit::TruncatedMean(TH1F * his, TVectorD *param, Float_t down, Float_t up, Bool_t verbose){
+void TStatToolkit::TruncatedMean(const TH1 * his, TVectorD *param, Float_t down, Float_t up, Bool_t verbose){
//
//
//
if (verbose) printf("Mean\t%f\t Sigma2\t%f\n", mean,sigma2);
}
-void TStatToolkit::LTM(TH1F * his, TVectorD *param , Float_t fraction, Bool_t verbose){
+void TStatToolkit::LTM(TH1 * his, TVectorD *param , Float_t fraction, Bool_t verbose){
//
- // LTM
+ // LTM : Trimmed mean on histogram - Modified version for binned data
//
+ // Robust statistic to estimate properties of the distribution
+ // See http://en.wikipedia.org/w/index.php?title=Trimmed_estimator&oldid=582847999
+ //
+ // New faster version is under preparation
+ //
+ if (!param) return;
+ (*param)[0]=0;
+ (*param)[1]=0;
+ (*param)[2]=0;
Int_t nbins = his->GetNbinsX();
Int_t nentries = (Int_t)his->GetEntries();
+ if (nentries<=0) return;
Double_t *data = new Double_t[nentries];
Int_t npoints=0;
for (Int_t ibin=1;ibin<nbins; ibin++){
- Float_t entriesI = his->GetBinContent(ibin);
- Float_t xcenter= his->GetBinCenter(ibin);
+ Double_t entriesI = his->GetBinContent(ibin);
+ //Double_t xcenter= his->GetBinCenter(ibin);
+ Double_t x0 = his->GetXaxis()->GetBinLowEdge(ibin);
+ Double_t w = his->GetXaxis()->GetBinWidth(ibin);
for (Int_t ic=0; ic<entriesI; ic++){
if (npoints<nentries){
- data[npoints]= xcenter;
+ data[npoints]= x0+w*Double_t((ic+0.5)/entriesI);
npoints++;
}
}
}
- Double_t mean, sigma;
+ Double_t mean, sigma;
Int_t npoints2=TMath::Min(Int_t(fraction*Float_t(npoints)),npoints-1);
npoints2=TMath::Max(Int_t(0.5*Float_t(npoints)),npoints2);
TStatToolkit::EvaluateUni(npoints, data, mean,sigma,npoints2);
}
}
-Double_t TStatToolkit::FitGaus(TH1F* his, TVectorD *param, TMatrixD */*matrix*/, Float_t xmin, Float_t xmax, Bool_t verbose){
+
+void TStatToolkit::MedianFilter(TH1 * his1D, Int_t nmedian){
+ //
+ // Algorithm to filter histogram
+ // author: marian.ivanov@cern.ch
+ // Details of algorithm:
+ // http://en.wikipedia.org/w/index.php?title=Median_filter&oldid=582191524
+ // Input parameters:
+ // his1D - input histogam - to be modiefied by Medianfilter
+ // nmendian - number of bins in median filter
+ //
+ Int_t nbins = his1D->GetNbinsX();
+ TVectorD vectorH(nbins);
+ for (Int_t ibin=0; ibin<nbins; ibin++) vectorH[ibin]=his1D->GetBinContent(ibin+1);
+ for (Int_t ibin=0; ibin<nbins; ibin++) {
+ Int_t index0=ibin-nmedian;
+ Int_t index1=ibin+nmedian;
+ if (index0<0) {index1+=-index0; index0=0;}
+ if (index1>=nbins) {index0-=index1-nbins+1; index1=nbins-1;}
+ Double_t value= TMath::Median(index1-index0,&(vectorH.GetMatrixArray()[index0]));
+ his1D->SetBinContent(ibin+1, value);
+ }
+}
+
+Bool_t TStatToolkit::LTMHisto(TH1 *his1D, TVectorD ¶ms , Float_t fraction){
+ //
+ // LTM : Trimmed mean on histogram - Modified version for binned data
+ //
+ // Robust statistic to estimate properties of the distribution
+ // To handle binning error special treatment
+ // for definition of unbinned data see:
+ // http://en.wikipedia.org/w/index.php?title=Trimmed_estimator&oldid=582847999
+ //
+ // Function parameters:
+ // his1D - input histogram
+ // params - vector with parameters
+ // - 0 - area
+ // - 1 - mean
+ // - 2 - rms
+ // - 3 - error estimate of mean
+ // - 4 - error estimate of RMS
+ // - 5 - first accepted bin position
+ // - 6 - last accepted bin position
+ //
+ Int_t nbins = his1D->GetNbinsX();
+ Int_t nentries = (Int_t)his1D->GetEntries();
+ const Double_t kEpsilon=0.0000000001;
+
+ if (nentries<=0) return 0;
+ if (fraction>1) fraction=0;
+ if (fraction<0) return 0;
+ TVectorD vectorX(nbins);
+ TVectorD vectorMean(nbins);
+ TVectorD vectorRMS(nbins);
+ Double_t sumCont=0;
+ for (Int_t ibin0=1; ibin0<=nbins; ibin0++) sumCont+=his1D->GetBinContent(ibin0);
+ //
+ Double_t minRMS=his1D->GetRMS()*10000;
+ Int_t maxBin=0;
+ //
+ for (Int_t ibin0=1; ibin0<nbins; ibin0++){
+ Double_t sum0=0, sum1=0, sum2=0;
+ Int_t ibin1=ibin0;
+ for ( ibin1=ibin0; ibin1<nbins; ibin1++){
+ Double_t cont=his1D->GetBinContent(ibin1);
+ Double_t x= his1D->GetBinCenter(ibin1);
+ sum0+=cont;
+ sum1+=cont*x;
+ sum2+=cont*x*x;
+ if ( (ibin0!=ibin1) && sum0>=fraction*sumCont) break;
+ }
+ vectorX[ibin0]=his1D->GetBinCenter(ibin0);
+ if (sum0<fraction*sumCont) continue;
+ //
+ // substract fractions of bin0 and bin1 to keep sum0=fration*sumCont
+ //
+ Double_t diff = sum0-fraction*sumCont;
+ Double_t mean = sum1/sum0;
+ //
+ Double_t x0=his1D->GetBinCenter(ibin0);
+ Double_t x1=his1D->GetBinCenter(ibin1);
+ Double_t y0=his1D->GetBinContent(ibin0);
+ Double_t y1=his1D->GetBinContent(ibin1);
+ //
+ Double_t d = y0+y1-diff; //enties to keep
+ Double_t w0=0,w1=0;
+ if (y0<=kEpsilon&&y1>kEpsilon){
+ w1=d/y1;
+ }
+ if (y1<=kEpsilon&&y0>kEpsilon){
+ w0=d/y0;
+ }
+ if (y0>kEpsilon && y1>kEpsilon && x1>x0 ){
+ w0 = (d*(x1-mean))/((x1-x0)*y0);
+ w1 = (d-y0*w0)/y1;
+ //
+ if (w0>1) {w1+=(w0-1)*y0/y1; w0=1;}
+ if (w1>1) {w0+=(w1-1)*y1/y0; w1=1;}
+ }
+ if ( (x1>x0) &&TMath::Abs(y0*w0+y1*w1-d)>kEpsilon*sum0){
+ printf(" TStatToolkit::LTMHisto error\n");
+ }
+ sum0-=y0+y1;
+ sum1-=y0*x0;
+ sum1-=y1*x1;
+ sum2-=y0*x0*x0;
+ sum2-=y1*x1*x1;
+ //
+ Double_t xx0=his1D->GetXaxis()->GetBinUpEdge(ibin0)-0.5*w0*his1D->GetBinWidth(ibin0);
+ Double_t xx1=his1D->GetXaxis()->GetBinLowEdge(ibin1)+0.5*w1*his1D->GetBinWidth(ibin1);
+ sum0+=y0*w0+y1*w1;
+ sum1+=y0*w0*xx0;
+ sum1+=y1*w1*xx1;
+ sum2+=y0*w0*xx0*xx0;
+ sum2+=y1*w1*xx1*xx1;
+
+ //
+ // choose the bin with smallest rms
+ //
+ if (sum0>0){
+ vectorMean[ibin0]=sum1/sum0;
+ vectorRMS[ibin0]=TMath::Sqrt(TMath::Abs(sum2/sum0-vectorMean[ibin0]*vectorMean[ibin0]));
+ if (vectorRMS[ibin0]<minRMS){
+ minRMS=vectorRMS[ibin0];
+ params[0]=sum0;
+ params[1]=vectorMean[ibin0];
+ params[2]=vectorRMS[ibin0];
+ params[3]=vectorRMS[ibin0]/TMath::Sqrt(sumCont*fraction);
+ params[4]=0; // what is the formula for error of RMS???
+ params[5]=ibin0;
+ params[6]=ibin1;
+ params[7]=his1D->GetBinCenter(ibin0);
+ params[8]=his1D->GetBinCenter(ibin1);
+ maxBin=ibin0;
+ }
+ }else{
+ break;
+ }
+ }
+ return kTRUE;
+}
+
+
+Double_t TStatToolkit::FitGaus(TH1* his, TVectorD *param, TMatrixD */*matrix*/, Float_t xmin, Float_t xmax, Bool_t verbose){
//
// Fit histogram with gaussian function
//
fitter.ClearPoints();
TVectorD par(3);
TVectorD sigma(3);
- TMatrixD A(3,3);
+ TMatrixD matA(3,3);
TMatrixD b(3,1);
Float_t rms = TMath::RMS(nBins,arr);
Float_t max = TMath::MaxElement(nBins,arr);
Float_t val = TMath::Log(Float_t(entriesI));
fitter.AddPoint(&xcenter,val,error);
if (npoints<3){
- A(npoints,0)=1;
- A(npoints,1)=xcenter;
- A(npoints,2)=xcenter*xcenter;
+ matA(npoints,0)=1;
+ matA(npoints,1)=xcenter;
+ matA(npoints,2)=xcenter*xcenter;
b(npoints,0)=val;
meanCOG+=xcenter*entriesI;
rms2COG +=xcenter*entriesI*xcenter;
if (npoints>=3){
if ( npoints == 3 ){
//analytic calculation of the parameters for three points
- A.Invert();
+ matA.Invert();
TMatrixD res(1,3);
- res.Mult(A,b);
+ res.Mult(matA,b);
par[0]=res(0,0);
par[1]=res(0,1);
par[2]=res(0,2);
}
-Float_t TStatToolkit::GetCOG(Short_t *arr, Int_t nBins, Float_t xMin, Float_t xMax, Float_t *rms, Float_t *sum)
+Float_t TStatToolkit::GetCOG(const Short_t *arr, Int_t nBins, Float_t xMin, Float_t xMax, Float_t *rms, Float_t *sum)
{
//
// calculate center of gravity rms and sum for array 'arr' with nBins an a x range xMin to xMax
// ROOT gauss fit
// nhistos - number of histograms to be used for test
//
- TTreeSRedirector *pcstream = new TTreeSRedirector("fitdebug.root");
+ TTreeSRedirector *pcstream = new TTreeSRedirector("fitdebug.root","recreate");
Float_t *xTrue = new Float_t[nhistos];
Float_t *sTrue = new Float_t[nhistos];
h1f[i]->FillRandom("myg");
}
- TStopwatch s;
+ TStopwatch s;
s.Start();
//standard gaus fit
for (Int_t i=0; i<nhistos; i++){
if (type==0) stat = projection->GetMean();
if (type==1) stat = projection->GetRMS();
if (type==2 || type==3){
- TVectorD vec(3);
+ TVectorD vec(10);
TStatToolkit::LTM((TH1F*)projection,&vec,0.7);
if (type==2) stat= vec[1];
if (type==3) stat= vec[0];
return graph;
}
-TGraph * TStatToolkit::MakeStat1D(TH3 * his, Int_t delta1, Int_t type){
- //
+TGraphErrors * TStatToolkit::MakeStat1D(TH2 * his, Int_t deltaBin, Double_t fraction, Int_t returnType, Int_t markerStyle, Int_t markerColor){
//
+ // function to retrieve the "mean and RMS estimate" of 2D histograms
+ //
+ // Robust statistic to estimate properties of the distribution
+ // See http://en.wikipedia.org/wiki/Trimmed_estimator
//
- // delta - number of bins to integrate
- // type - 0 - mean value
-
+ // deltaBin - number of bins to integrate (bin+-deltaBin)
+ // fraction - fraction of values for the LTM and for the gauss fit
+ // returnType -
+ // 0 - mean value
+ // 1 - RMS
+ // 2 - LTM mean
+ // 3 - LTM sigma
+ // 4 - Gaus fit mean - on LTM range
+ // 5 - Gaus fit sigma - on LTM range
+ //
TAxis * xaxis = his->GetXaxis();
- TAxis * yaxis = his->GetYaxis();
- // TAxis * zaxis = his->GetZaxis();
Int_t nbinx = xaxis->GetNbins();
- Int_t nbiny = yaxis->GetNbins();
char name[1000];
Int_t icount=0;
- TGraph *graph = new TGraph(nbinx);
+ //
+ TVectorD vecX(nbinx);
+ TVectorD vecXErr(nbinx);
+ TVectorD vecY(nbinx);
+ TVectorD vecYErr(nbinx);
+ //
TF1 f1("f1","gaus");
- for (Int_t ix=0; ix<nbinx;ix++){
- Float_t xcenter = xaxis->GetBinCenter(ix);
- // Float_t ycenter = yaxis->GetBinCenter(iy);
+ TVectorD vecLTM(10);
+
+ for (Int_t jx=1; jx<=nbinx;jx++){
+ Int_t ix=jx-1;
+ Float_t xcenter = xaxis->GetBinCenter(jx);
snprintf(name,1000,"%s_%d",his->GetName(), ix);
- TH1 *projection = his->ProjectionZ(name,ix-delta1,ix+delta1,0,nbiny);
- Float_t stat= 0;
- if (type==0) stat = projection->GetMean();
- if (type==1) stat = projection->GetRMS();
- if (type==2 || type==3){
- TVectorD vec(3);
- TStatToolkit::LTM((TH1F*)projection,&vec,0.7);
- if (type==2) stat= vec[1];
- if (type==3) stat= vec[0];
+ TH1 *projection = his->ProjectionY(name,TMath::Max(jx-deltaBin,1),TMath::Min(jx+deltaBin,nbinx));
+ Double_t stat= 0;
+ Double_t err =0;
+ TStatToolkit::LTMHisto((TH1F*)projection,vecLTM,fraction);
+ //
+ if (returnType==0) {
+ stat = projection->GetMean();
+ err = projection->GetMeanError();
}
- if (type==4|| type==5){
- projection->Fit(&f1);
- if (type==4) stat= f1.GetParameter(1);
- if (type==5) stat= f1.GetParameter(2);
+ if (returnType==1) {
+ stat = projection->GetRMS();
+ err = projection->GetRMSError();
}
- //printf("%d\t%f\t%f\t%f\n", icount,xcenter, ycenter, stat);
- graph->SetPoint(icount,xcenter, stat);
+ if (returnType==2 || returnType==3){
+ if (returnType==2) {stat= vecLTM[1]; err =projection->GetRMSError();}
+ if (returnType==3) {stat= vecLTM[2]; err =projection->GetRMSError();}
+ }
+ if (returnType==4|| returnType==5){
+ projection->Fit(&f1,"QN","QN", vecLTM[7], vecLTM[8]);
+ if (returnType==4) {
+ stat= f1.GetParameter(1);
+ err=f1.GetParError(1);
+ }
+ if (returnType==5) {
+ stat= f1.GetParameter(2);
+ err=f1.GetParError(2);
+ }
+ }
+ vecX[icount]=xcenter;
+ vecY[icount]=stat;
+ vecYErr[icount]=err;
icount++;
+ delete projection;
}
+ TGraphErrors *graph = new TGraphErrors(icount,vecX.GetMatrixArray(), vecY.GetMatrixArray(),0, vecYErr.GetMatrixArray());
+ graph->SetMarkerStyle(markerStyle);
+ graph->SetMarkerColor(markerColor);
return graph;
}
TObjArray* valTokens = strVal.Tokenize(":");
drawStr = valTokens->At(0)->GetName();
ferr = valTokens->At(1)->GetName();
+ delete valTokens;
}
fitter->ClearPoints();
Int_t entries = tree->Draw(drawStr.Data(), cutStr.Data(), "goff", stop-start, start);
- if (entries == -1) return new TString("An ERROR has occured during fitting!");
- Double_t **values = new Double_t*[dim+1] ;
+ if (entries == -1) {
+ delete formulaTokens;
+ return new TString("An ERROR has occured during fitting!");
+ }
+ Double_t **values = new Double_t*[dim+1] ;
+ for (Int_t i=0; i<dim+1; i++) values[i]=NULL;
//
entries = tree->Draw(ferr.Data(), cutStr.Data(), "goff", stop-start, start);
if (entries == -1) {
+ delete formulaTokens;
delete []values;
return new TString("An ERROR has occured during fitting!");
}
TObjArray* valTokens = strVal.Tokenize(":");
drawStr = valTokens->At(0)->GetName();
ferr = valTokens->At(1)->GetName();
+ delete valTokens;
}
fitter->ClearPoints();
Int_t entries = tree->Draw(drawStr.Data(), cutStr.Data(), "goff", stop-start, start);
- if (entries == -1) return new TString("An ERROR has occured during fitting!");
+ if (entries == -1) {
+ delete formulaTokens;
+ return new TString("An ERROR has occured during fitting!");
+ }
Double_t **values = new Double_t*[dim+1] ;
+ for (Int_t i=0; i<dim+1; i++) values[i]=NULL;
//
entries = tree->Draw(ferr.Data(), cutStr.Data(), "goff", stop-start, start);
if (entries == -1) {
+ delete formulaTokens;
delete [] values;
return new TString("An ERROR has occured during fitting!");
}
if (entries != centries) {
delete []errors;
delete []values;
+ delete formulaTokens;
return new TString("An ERROR has occured during fitting!");
}
values[i] = new Double_t[entries];
if (strVal.Contains(":")){
TObjArray* valTokens = strVal.Tokenize(":");
drawStr = valTokens->At(0)->GetName();
- ferr = valTokens->At(1)->GetName();
+ ferr = valTokens->At(1)->GetName();
+ delete valTokens;
}
fitter->ClearPoints();
Int_t entries = tree->Draw(drawStr.Data(), cutStr.Data(), "goff", stop-start, start);
- if (entries == -1) return new TString("An ERROR has occured during fitting!");
+ if (entries == -1) {
+ delete formulaTokens;
+ return new TString("An ERROR has occured during fitting!");
+ }
Double_t **values = new Double_t*[dim+1] ;
+ for (Int_t i=0; i<dim+1; i++) values[i]=NULL;
//
entries = tree->Draw(ferr.Data(), cutStr.Data(), "goff", stop-start, start);
if (entries == -1) {
delete []values;
+ delete formulaTokens;
return new TString("An ERROR has occured during fitting!");
}
Double_t *errors = new Double_t[entries];
if (entries != centries) {
delete []errors;
delete []values;
+ delete formulaTokens;
return new TString("An ERROR has occured during fitting!");
}
values[i] = new Double_t[entries];
-Int_t TStatToolkit::GetFitIndex(TString fString, TString subString){
+Int_t TStatToolkit::GetFitIndex(const TString fString, const TString subString){
//
// fitString - ++ separated list of fits
// substring - ++ separated list of the requiered substrings
}
if (isOK) index=i;
}
+ delete arrFit;
+ delete arrSub;
return index;
}
-TString TStatToolkit::FilterFit(TString &input, TString filter, TVectorD ¶m, TMatrixD & covar){
+TString TStatToolkit::FilterFit(const TString &input, const TString filter, TVectorD ¶m, TMatrixD & covar){
//
// Filter fit expression make sub-fit
//
}
}
result+="-0.)";
+ delete array0;
+ delete array1;
return result;
}
-void TStatToolkit::Constrain1D(TString &input, TString filter, TVectorD ¶m, TMatrixD & covar, Double_t mean, Double_t sigma){
+void TStatToolkit::Constrain1D(const TString &input, const TString filter, TVectorD ¶m, TMatrixD & covar, Double_t mean, Double_t sigma){
//
// constrain linear fit
// input - string description of fit function
// param,covar - parameters and covariance matrix of the fit
// mean,sigma - new measurement uning which the fit is updated
//
+
TObjArray *array0= input.Tokenize("++");
TObjArray *array1= filter.Tokenize("++");
TMatrixD paramM(param.GetNrows(),1);
for (Int_t i=0; i<=array0->GetEntries(); i++){paramM(i,0)=param(i);}
- for (Int_t i=0; i<array0->GetEntries(); i++){
- Bool_t isOK=kTRUE;
- TString str(array0->At(i)->GetName());
- for (Int_t j=0; j<array1->GetEntries(); j++){
- if (str.Contains(array1->At(j)->GetName())==0) isOK=kFALSE;
- }
- if (isOK) {
- TStatToolkit::Update1D(mean, sigma, i+1, paramM, covar);//
+ if (filter.Length()==0){
+ TStatToolkit::Update1D(mean, sigma, 0, paramM, covar);//
+ }else{
+ for (Int_t i=0; i<array0->GetEntries(); i++){
+ Bool_t isOK=kTRUE;
+ TString str(array0->At(i)->GetName());
+ for (Int_t j=0; j<array1->GetEntries(); j++){
+ if (str.Contains(array1->At(j)->GetName())==0) isOK=kFALSE;
+ }
+ if (isOK) {
+ TStatToolkit::Update1D(mean, sigma, i+1, paramM, covar);//
+ }
}
}
for (Int_t i=0; i<=array0->GetEntries(); i++){
param(i)=paramM(i,0);
}
+ delete array0;
+ delete array1;
}
-TString TStatToolkit::MakeFitString(TString &input, TVectorD ¶m, TMatrixD & covar){
+TString TStatToolkit::MakeFitString(const TString &input, const TVectorD ¶m, const TMatrixD & covar, Bool_t verbose){
//
//
//
TObjArray *array0= input.Tokenize("++");
- TString result="(0.0";
+ TString result=Form("(%f",param[0]);
+ printf("%f\t%f\t\n", param[0], TMath::Sqrt(covar(0,0)));
for (Int_t i=0; i<array0->GetEntries(); i++){
TString str(array0->At(i)->GetName());
result+="+"+str;
result+=Form("*(%f)",param[i+1]);
- printf("%f\t%f\t%s\n", param[i+1], TMath::Sqrt(covar(i+1,i+1)),str.Data());
+ if (verbose) printf("%f\t%f\t%s\n", param[i+1], TMath::Sqrt(covar(i+1,i+1)),str.Data());
}
result+="-0.)";
+ delete array0;
return result;
}
+TGraphErrors * TStatToolkit::MakeGraphErrors(TTree * tree, const char * expr, const char * cut, Int_t mstyle, Int_t mcolor, Float_t msize, Float_t offset){
+ //
+ // Query a graph errors
+ // return TGraphErrors specified by expr and cut
+ // Example usage TStatToolkit::MakeGraphError(tree,"Y:X:ErrY","X>0", 25,2,0.4)
+ // tree - tree with variable
+ // expr - examp
+ const Int_t entries = tree->Draw(expr,cut,"goff");
+ if (entries<=0) {
+ TStatToolkit t;
+ t.Error("TStatToolkit::MakeGraphError",Form("Empty or Not valid expression (%s) or cut *%s)", expr,cut));
+ return 0;
+ }
+ if ( tree->GetV2()==0){
+ TStatToolkit t;
+ t.Error("TStatToolkit::MakeGraphError",Form("Not valid expression (%s) ", expr));
+ return 0;
+ }
+ TGraphErrors * graph=0;
+ if ( tree->GetV3()!=0){
+ graph = new TGraphErrors (entries, tree->GetV2(),tree->GetV1(),0,tree->GetV3());
+ }else{
+ graph = new TGraphErrors (entries, tree->GetV2(),tree->GetV1(),0,0);
+ }
+ graph->SetMarkerStyle(mstyle);
+ graph->SetMarkerColor(mcolor);
+ graph->SetLineColor(mcolor);
+ graph->SetTitle(expr);
+ TString chstring(expr);
+ TObjArray *charray = chstring.Tokenize(":");
+ graph->GetXaxis()->SetTitle(charray->At(1)->GetName());
+ graph->GetYaxis()->SetTitle(charray->At(0)->GetName());
+ delete charray;
+ if (msize>0) graph->SetMarkerSize(msize);
+ for(Int_t i=0;i<graph->GetN();i++) graph->GetX()[i]+=offset;
+ return graph;
+
+}
+
-TGraph * TStatToolkit::MakeGraphSparse(TTree * tree, const char * expr, const char * cut){
+TGraph * TStatToolkit::MakeGraphSparse(TTree * tree, const char * expr, const char * cut, Int_t mstyle, Int_t mcolor, Float_t msize, Float_t offset){
//
// Make a sparse draw of the variables
+ // Format of expr : Var:Run or Var:Run:ErrorY or Var:Run:ErrorY:ErrorX
+ // offset : points can slightly be shifted in x for better visibility with more graphs
//
- const Int_t entries = tree->Draw(expr,cut,"goff");
- // TGraph * graph = (TGraph*)gPad->GetPrimitive("Graph"); // 2D
- TGraph * graph = new TGraph (entries, tree->GetV2(),tree->GetV1());
+ // Written by Weilin.Yu
+ // updated & merged with QA-code by Patrick Reichelt
//
- Int_t *index = new Int_t[entries];
- TMath::Sort(entries,graph->GetX(),index,kFALSE);
-
- Double_t *VV = new Double_t[entries];
+ const Int_t entries = tree->Draw(expr,cut,"goff");
+ if (entries<=0) {
+ TStatToolkit t;
+ t.Error("TStatToolkit::MakeGraphSparse",Form("Empty or Not valid expression (%s) or cut (%s)", expr, cut));
+ return 0;
+ }
+ // TGraph * graph = (TGraph*)gPad->GetPrimitive("Graph"); // 2D
+ Double_t *graphY, *graphX;
+ graphY = tree->GetV1();
+ graphX = tree->GetV2();
+
+ // sort according to run number
+ Int_t *index = new Int_t[entries*4];
+ TMath::Sort(entries,graphX,index,kFALSE);
+
+ // define arrays for the new graph
+ Double_t *unsortedX = new Double_t[entries];
+ Int_t *runNumber = new Int_t[entries];
Double_t count = 0.5;
- vector<Int_t> vrun;
- VV[index[0]] = count;
- vrun.push_back(graph->GetX()[index[0]]);
- for(Int_t i=1;i<entries;i++){
- if(graph->GetX()[index[i]]==graph->GetX()[index[i-1]])
- VV[index[i]] = count;
- else if(graph->GetX()[index[i]]!=graph->GetX()[index[i-1]]){
+
+ // evaluate arrays for the new graph according to the run-number
+ Int_t icount=0;
+ //first entry
+ unsortedX[index[0]] = count;
+ runNumber[0] = graphX[index[0]];
+ // loop the rest of entries
+ for(Int_t i=1;i<entries;i++)
+ {
+ if(graphX[index[i]]==graphX[index[i-1]])
+ unsortedX[index[i]] = count;
+ else if(graphX[index[i]]!=graphX[index[i-1]]){
count++;
- VV[index[i]] = count;
- vrun.push_back(graph->GetX()[index[i]]);
+ icount++;
+ unsortedX[index[i]] = count;
+ runNumber[icount]=graphX[index[i]];
}
}
-
+
+ // count the number of xbins (run-wise) for the new graph
const Int_t newNbins = int(count+0.5);
Double_t *newBins = new Double_t[newNbins+1];
for(Int_t i=0; i<=count+1;i++){
newBins[i] = i;
}
-
- TGraph *graphNew = new TGraph(entries,VV,graph->GetY());
+
+ // define and fill the new graph
+ TGraph *graphNew = 0;
+ if (tree->GetV3()) {
+ if (tree->GetV4()) {
+ graphNew = new TGraphErrors(entries,unsortedX,graphY,tree->GetV4(),tree->GetV3());
+ }
+ else { graphNew = new TGraphErrors(entries,unsortedX,graphY,0,tree->GetV3()); }
+ }
+ else { graphNew = new TGraphErrors(entries,unsortedX,graphY,0,0); }
+ // with "Set(...)", the x-axis is being sorted
graphNew->GetXaxis()->Set(newNbins,newBins);
-
+
+ // set the bins for the x-axis, apply shifting of points
Char_t xName[50];
- Double_t bin_unit = graphNew->GetXaxis()->GetNbins()/count;
for(Int_t i=0;i<count;i++){
- snprintf(xName,50,"%d",vrun.at(i));
+ snprintf(xName,50,"%d",runNumber[i]);
graphNew->GetXaxis()->SetBinLabel(i+1,xName);
+ graphNew->GetX()[i]+=offset;
}
+
graphNew->GetHistogram()->SetTitle("");
-
- delete [] VV;
+ graphNew->SetMarkerStyle(mstyle);
+ graphNew->SetMarkerColor(mcolor);
+ if (msize>0) graphNew->SetMarkerSize(msize);
+ delete [] unsortedX;
+ delete [] runNumber;
delete [] index;
delete [] newBins;
+ //
+ graphNew->SetTitle(expr);
+ TString chstring(expr);
+ TObjArray *charray = chstring.Tokenize(":");
+ graphNew->GetXaxis()->SetTitle(charray->At(1)->GetName());
+ graphNew->GetYaxis()->SetTitle(charray->At(0)->GetName());
+ delete charray;
return graphNew;
}
+
+
+//
+// functions used for the trending
+//
+
+Int_t TStatToolkit::MakeStatAlias(TTree * tree, const char * expr, const char * cut, const char * alias)
+{
+ //
+ // Add alias using statistical values of a given variable.
+ // (by MI, Patrick Reichelt)
+ //
+ // tree - input tree
+ // expr - variable expression
+ // cut - selection criteria
+ // Output - return number of entries used to define variable
+ // In addition mean, rms, median, and robust mean and rms (choosing fraction of data with smallest RMS)
+ //
+ /* Example usage:
+ 1.) create the robust estimators for variable expr="QA.TPC.CPass1.meanTPCncl" and create a corresponding
+ aliases with the prefix alias[0]="ncl", calculated using fraction alias[1]="0.90"
+
+ TStatToolkit::MakeStatAlias(tree,"QA.TPC.CPass1.meanTPCncl","QA.TPC.CPass1.status>0","ncl:0.9");
+ root [4] tree->GetListOfAliases().Print()
+ OBJ: TNamed ncl_Median (130.964333+0)
+ OBJ: TNamed ncl_Mean (122.120387+0)
+ OBJ: TNamed ncl_RMS (33.509623+0)
+ OBJ: TNamed ncl_Mean90 (131.503862+0)
+ OBJ: TNamed ncl_RMS90 (3.738260+0)
+ */
+ //
+ Int_t entries = tree->Draw(expr,cut,"goff");
+ if (entries<=1){
+ printf("Expression or cut not valid:\t%s\t%s\n", expr, cut);
+ return 0;
+ }
+ //
+ TObjArray* oaAlias = TString(alias).Tokenize(":");
+ if (oaAlias->GetEntries()<2) return 0;
+ Float_t entryFraction = atof( oaAlias->At(1)->GetName() );
+ //
+ Double_t median = TMath::Median(entries,tree->GetV1());
+ Double_t mean = TMath::Mean(entries,tree->GetV1());
+ Double_t rms = TMath::RMS(entries,tree->GetV1());
+ Double_t meanEF=0, rmsEF=0;
+ TStatToolkit::EvaluateUni(entries, tree->GetV1(), meanEF, rmsEF, entries*entryFraction);
+ //
+ tree->SetAlias(Form("%s_Median",oaAlias->At(0)->GetName()), Form("(%f+0)",median));
+ tree->SetAlias(Form("%s_Mean",oaAlias->At(0)->GetName()), Form("(%f+0)",mean));
+ tree->SetAlias(Form("%s_RMS",oaAlias->At(0)->GetName()), Form("(%f+0)",rms));
+ tree->SetAlias(Form("%s_Mean%d",oaAlias->At(0)->GetName(),Int_t(entryFraction*100)), Form("(%f+0)",meanEF));
+ tree->SetAlias(Form("%s_RMS%d",oaAlias->At(0)->GetName(),Int_t(entryFraction*100)), Form("(%f+0)",rmsEF));
+ delete oaAlias;
+ return entries;
+}
+
+Int_t TStatToolkit::SetStatusAlias(TTree * tree, const char * expr, const char * cut, const char * alias)
+{
+ //
+ // Add alias to trending tree using statistical values of a given variable.
+ // (by MI, Patrick Reichelt)
+ //
+ // format of expr : varname (e.g. meanTPCncl)
+ // format of cut : char like in TCut
+ // format of alias: alias:query:entryFraction(EF) (fraction of entries used for uniformity evaluation)
+ // e.g.: varname_Out:(abs(varname-meanEF)>6.*rmsEF):0.8
+ // available internal variables are: 'varname, Median, Mean, MeanEF, RMS, RMSEF'
+ // in the alias, 'varname' will be replaced by its content, and 'EF' by the percentage (e.g. MeanEF -> Mean80)
+ //
+ /* Example usage:
+ 1.) Define robust mean (possible, but easier done with TStatToolkit::MakeStatAlias(...))
+ TStatToolkit::SetStatusAlias(tree, "meanTPCnclF", "meanTPCnclF>0", "meanTPCnclF_MeanEF:MeanEF:0.80") ;
+ root [10] tree->GetListOfAliases()->Print()
+ Collection name='TList', class='TList', size=1
+ OBJ: TNamed meanTPCnclF_Mean80 0.899308
+ 2.) create alias outlyers - 6 sigma cut
+ TStatToolkit::SetStatusAlias(tree, "meanTPCnclF", "meanTPCnclF>0", "meanTPCnclF_Out:(abs(meanTPCnclF-MeanEF)>6.*RMSEF):0.8")
+ meanTPCnclF_Out ==> (abs(meanTPCnclF-0.899308)>6.*0.016590)
+ 3.) the same functionality as in 2.)
+ TStatToolkit::SetStatusAlias(tree, "meanTPCnclF", "meanTPCnclF>0", "varname_Out2:(abs(varname-MeanEF)>6.*RMSEF):0.8")
+ meanTPCnclF_Out2 ==> (abs(meanTPCnclF-0.899308)>6.*0.016590)
+ */
+ //
+ Int_t entries = tree->Draw(expr,cut,"goff");
+ if (entries<1){
+ printf("Expression or cut not valid:\t%s\t%s\n", expr, cut);
+ return 0;
+ }
+ //
+ TObjArray* oaVar = TString(expr).Tokenize(":");
+ char varname[50];
+ snprintf(varname,50,"%s", oaVar->At(0)->GetName());
+ //
+ TObjArray* oaAlias = TString(alias).Tokenize(":");
+ if (oaAlias->GetEntries()<3) return 0;
+ Float_t entryFraction = atof( oaAlias->At(2)->GetName() );
+ //
+ Double_t median = TMath::Median(entries,tree->GetV1());
+ Double_t mean = TMath::Mean(entries,tree->GetV1());
+ Double_t rms = TMath::RMS(entries,tree->GetV1());
+ Double_t meanEF=0, rmsEF=0;
+ TStatToolkit::EvaluateUni(entries, tree->GetV1(), meanEF, rmsEF, entries*entryFraction);
+ //
+ TString sAlias( oaAlias->At(0)->GetName() );
+ sAlias.ReplaceAll("varname",varname);
+ sAlias.ReplaceAll("MeanEF", Form("Mean%1.0f",entryFraction*100) );
+ sAlias.ReplaceAll("RMSEF", Form("RMS%1.0f",entryFraction*100) );
+ TString sQuery( oaAlias->At(1)->GetName() );
+ sQuery.ReplaceAll("varname",varname);
+ sQuery.ReplaceAll("MeanEF", Form("%f",meanEF) );
+ sQuery.ReplaceAll("RMSEF", Form("%f",rmsEF) ); //make sure to replace 'RMSEF' before 'RMS'...
+ sQuery.ReplaceAll("Median", Form("%f",median) );
+ sQuery.ReplaceAll("Mean", Form("%f",mean) );
+ sQuery.ReplaceAll("RMS", Form("%f",rms) );
+ printf("define alias:\t%s = %s\n", sAlias.Data(), sQuery.Data());
+ //
+ char query[200];
+ char aname[200];
+ snprintf(query,200,"%s", sQuery.Data());
+ snprintf(aname,200,"%s", sAlias.Data());
+ tree->SetAlias(aname, query);
+ delete oaVar;
+ delete oaAlias;
+ return entries;
+}
+
+TMultiGraph* TStatToolkit::MakeStatusMultGr(TTree * tree, const char * expr, const char * cut, const char * alias, Int_t igr)
+{
+ //
+ // Compute a trending multigraph that shows for which runs a variable has outliers.
+ // (by MI, Patrick Reichelt)
+ //
+ // format of expr : varname:xaxis (e.g. meanTPCncl:run)
+ // format of cut : char like in TCut
+ // format of alias: (1):(varname_Out==0):(varname_Out)[:(varname_Warning):...]
+ // in the alias, 'varname' will be replaced by its content (e.g. varname_Out -> meanTPCncl_Out)
+ // note: the aliases 'varname_Out' etc have to be defined by function TStatToolkit::SetStatusAlias(...)
+ // counter igr is used to shift the multigraph in y when filling a TObjArray.
+ //
+ TObjArray* oaVar = TString(expr).Tokenize(":");
+ if (oaVar->GetEntries()<2) return 0;
+ char varname[50];
+ char var_x[50];
+ snprintf(varname,50,"%s", oaVar->At(0)->GetName());
+ snprintf(var_x ,50,"%s", oaVar->At(1)->GetName());
+ //
+ TString sAlias(alias);
+ sAlias.ReplaceAll("varname",varname);
+ TObjArray* oaAlias = TString(sAlias.Data()).Tokenize(":");
+ if (oaAlias->GetEntries()<3) return 0;
+ //
+ char query[200];
+ TMultiGraph* multGr = new TMultiGraph();
+ Int_t marArr[6] = {24+igr%2, 20+igr%2, 20+igr%2, 20+igr%2, 22, 23};
+ Int_t colArr[6] = {kBlack, kBlack, kRed, kOrange, kMagenta, kViolet};
+ Double_t sizArr[6] = {1.2, 1.1, 1.0, 1.0, 1, 1};
+ const Int_t ngr = oaAlias->GetEntriesFast();
+ for (Int_t i=0; i<ngr; i++){
+ if (i==2) continue; // the Fatal(Out) graph will be added in the end to be plotted on top!
+ snprintf(query,200, "%f*(%s-0.5):%s", 1.+igr, oaAlias->At(i)->GetName(), var_x);
+ multGr->Add( (TGraphErrors*) TStatToolkit::MakeGraphSparse(tree,query,cut,marArr[i],colArr[i],sizArr[i]) );
+ }
+ snprintf(query,200, "%f*(%s-0.5):%s", 1.+igr, oaAlias->At(2)->GetName(), var_x);
+ multGr->Add( (TGraphErrors*) TStatToolkit::MakeGraphSparse(tree,query,cut,marArr[2],colArr[2],sizArr[2]) );
+ //
+ multGr->SetName(varname);
+ multGr->SetTitle(varname); // used for y-axis labels. // details to be included!
+ delete oaVar;
+ delete oaAlias;
+ return multGr;
+}
+
+
+void TStatToolkit::AddStatusPad(TCanvas* c1, Float_t padratio, Float_t bottommargin)
+{
+ //
+ // add pad to bottom of canvas for Status graphs (by Patrick Reichelt)
+ // call function "DrawStatusGraphs(...)" afterwards
+ //
+ TCanvas* c1_clone = (TCanvas*) c1->Clone("c1_clone");
+ c1->Clear();
+ // produce new pads
+ c1->cd();
+ TPad* pad1 = new TPad("pad1", "pad1", 0., padratio, 1., 1.);
+ pad1->Draw();
+ pad1->SetNumber(1); // so it can be called via "c1->cd(1);"
+ c1->cd();
+ TPad* pad2 = new TPad("pad2", "pad2", 0., 0., 1., padratio);
+ pad2->Draw();
+ pad2->SetNumber(2);
+ // draw original canvas into first pad
+ c1->cd(1);
+ c1_clone->DrawClonePad();
+ pad1->SetBottomMargin(0.001);
+ pad1->SetRightMargin(0.01);
+ // set up second pad
+ c1->cd(2);
+ pad2->SetGrid(3);
+ pad2->SetTopMargin(0);
+ pad2->SetBottomMargin(bottommargin); // for the long x-axis labels (runnumbers)
+ pad2->SetRightMargin(0.01);
+}
+
+
+void TStatToolkit::DrawStatusGraphs(TObjArray* oaMultGr)
+{
+ //
+ // draw Status graphs into active pad of canvas (by MI, Patrick Reichelt)
+ // ...into bottom pad, if called after "AddStatusPad(...)"
+ //
+ const Int_t nvars = oaMultGr->GetEntriesFast();
+ TGraph* grAxis = (TGraph*) ((TMultiGraph*) oaMultGr->At(0))->GetListOfGraphs()->At(0);
+ grAxis->SetMaximum(0.5*nvars+0.5);
+ grAxis->SetMinimum(0);
+ grAxis->GetYaxis()->SetLabelSize(0);
+ Int_t entries = grAxis->GetN();
+ printf("entries (via GetN()) = %d\n",entries);
+ grAxis->GetXaxis()->SetLabelSize(5.7*TMath::Min(TMath::Max(5./entries,0.01),0.03));
+ grAxis->GetXaxis()->LabelsOption("v");
+ grAxis->Draw("ap");
+ //
+ // draw multigraphs & names of status variables on the y axis
+ for (Int_t i=0; i<nvars; i++){
+ ((TMultiGraph*) oaMultGr->At(i))->Draw("p");
+ TLatex* ylabel = new TLatex(-0.1, 0.5*i+0.5, ((TMultiGraph*) oaMultGr->At(i))->GetTitle());
+ ylabel->SetTextAlign(32); //hor:right & vert:centered
+ ylabel->SetTextSize(0.025/gPad->GetHNDC());
+ ylabel->Draw();
+ }
+}
+
+
+TH1* TStatToolkit::DrawHistogram(TTree * tree, const char* drawCommand, const char* cuts, const char* histoname, const char* histotitle, Int_t nsigma, Float_t fraction )
+{
+ //
+ // Draw histogram from TTree with robust range
+ // Only for 1D so far!
+ //
+ // Parameters:
+ // - histoname: name of histogram
+ // - histotitle: title of histgram
+ // - fraction: fraction of data to define the robust mean
+ // - nsigma: nsigma value for range
+ //
+
+ TString drawStr(drawCommand);
+ TString cutStr(cuts);
+ Int_t dim = 1;
+
+ if(!tree) {
+ cerr<<" Tree pointer is NULL!"<<endl;
+ return 0;
+ }
+
+ // get entries
+ Int_t entries = tree->Draw(drawStr.Data(), cutStr.Data(), "goff");
+ if (entries == -1) {
+ cerr<<"TTree draw returns -1"<<endl;
+ return 0;
+ }
+
+ // get dimension
+ if(tree->GetV1()) dim = 1;
+ if(tree->GetV2()) dim = 2;
+ if(tree->GetV3()) dim = 3;
+ if(dim > 2){
+ cerr<<"TTree has more than 2 dimensions (not yet supported)"<<endl;
+ return 0;
+ }
+
+ // draw robust
+ Double_t meanX, rmsX=0;
+ Double_t meanY, rmsY=0;
+ TStatToolkit::EvaluateUni(entries, tree->GetV1(),meanX,rmsX, fraction*entries);
+ if(dim==2){
+ TStatToolkit::EvaluateUni(entries, tree->GetV1(),meanY,rmsY, fraction*entries);
+ TStatToolkit::EvaluateUni(entries, tree->GetV2(),meanX,rmsX, fraction*entries);
+ }
+ TH1* hOut;
+ if(dim==1){
+ hOut = new TH1F(histoname, histotitle, 200, meanX-nsigma*rmsX, meanX+nsigma*rmsX);
+ for (Int_t i=0; i<entries; i++) hOut->Fill(tree->GetV1()[i]);
+ hOut->GetXaxis()->SetTitle(tree->GetHistogram()->GetXaxis()->GetTitle());
+ hOut->Draw();
+ }
+ else if(dim==2){
+ hOut = new TH2F(histoname, histotitle, 200, meanX-nsigma*rmsX, meanX+nsigma*rmsX,200, meanY-nsigma*rmsY, meanY+nsigma*rmsY);
+ for (Int_t i=0; i<entries; i++) hOut->Fill(tree->GetV2()[i],tree->GetV1()[i]);
+ hOut->GetXaxis()->SetTitle(tree->GetHistogram()->GetXaxis()->GetTitle());
+ hOut->GetYaxis()->SetTitle(tree->GetHistogram()->GetYaxis()->GetTitle());
+ hOut->Draw("colz");
+ }
+ return hOut;
+}