void validateInput(const char* filename, int level=0)
{
TFile* file = new TFile(filename, "update");
TIter nextDirectory(file->GetListOfKeys());
TKey* idir;
while((idir = (TKey*)nextDirectory())){
file->cd();
if( idir->IsFolder() ){
if( level>-1 ){ std::cout << "Found directory: " << idir->GetName() << std::endl; }
file->cd(idir->GetName());
validateFolder(file, idir->GetName(), level);
}
else{
if( level> 0 ){ std::cout << "Found histogram: " << idir->GetName() << std::endl; }
if( level>-1 ){
TH1F* h = (TH1F*)file->Get(idir->GetName());
if(h->Integral() == 0){
std::cout << "----- E R R O R ----- : histogram has 0 integral please fix this: --> " << idir->GetName() << std::endl;
}
}
}
}
file->Close();
return;
}
/** The LES may have changed. Repoint to the current directory if it still exists,
Otherwise, go to the next one.
*/
bool BookDirIterator::resync (LedgerEntrySet& les)
{
if (mIndex.isZero ())
mIndex = mBase;
else if (mIndex != mBase)
--mIndex;
return nextDirectory (les);
}
bool BookDirIterator::firstDirectory (LedgerEntrySet& les)
{
WriteLog (lsTRACE, Ledger) << "BookDirIterator(" << mBase.GetHex() << ") firstDirectory";
/** Jump to the beginning
*/
mIndex = mBase;
return nextDirectory (les);
}
void rescaleSignal(bool armed, double scale, const char* filename, const char* pattern="", unsigned int debug=0)
{
std::vector<std::string> histnames; histnames.clear();
if( debug>0 ){
std::cout << "file = " << filename << std::endl;
std::cout << "scale = " << scale << std::endl;
std::cout << "armed = " << armed << std::endl;
}
TFile* file = new TFile(filename, "update");
TIter nextDirectory(file->GetListOfKeys());
std::vector<std::string> buffer;
TKey* idir;
while((idir = (TKey*)nextDirectory())){
buffer.clear();
if( idir->IsFolder() ){
file->cd(); // make sure to start in directory head
if( debug>0 ){ std::cout << "Found directory: " << idir->GetName() << std::endl; }
if( file->GetDirectory(idir->GetName()) ){
file->cd(idir->GetName()); // change to sub-directory
buffer = signalList(idir->GetName(), pattern, debug);
}
// append to the vector of histograms to be rescaled
for(std::vector<std::string>::const_iterator elem=buffer.begin(); elem!=buffer.end(); ++elem){
histnames.push_back(*elem);
}
if(debug>1){
std::cout << "added " << buffer.size() << " elements to histnames [" << histnames.size() << "] for directory " << idir->GetName() << std::endl;
}
}
}
// pick up files which are not kept in an extra folder
file->cd(); buffer.clear();
buffer = signalList("", pattern, debug);
// append to the vector of histograms to be rescaled
for(std::vector<std::string>::const_iterator elem=buffer.begin(); elem!=buffer.end(); ++elem){
histnames.push_back(*elem);
}
if(debug>1){
std::cout << "added " << buffer.size() << " elements to histnames [" << histnames.size() << "] for file head" << std::endl;
}
for(std::vector<std::string>::const_iterator hist=histnames.begin(); hist!=histnames.end(); ++hist){
file->cd();
TH1F* h = (TH1F*)file->Get(hist->c_str());
std::string histName;
if(hist->find("/")!=std::string::npos){
histName = hist->substr(hist->find("/")+1);
}
else{
histName = *hist;
}
TH1F* hout = (TH1F*)h->Clone(histName.c_str());
if(debug>1){
std::cout << "...folder : " << hist->substr(0, hist->find("/")).c_str() << std::endl;
std::cout << "...histogram : " << hout->GetName () << " / " << hist->c_str() << std::endl;
std::cout << "...old scale : " << hout->Integral() << std::endl;
}
hout->Scale(scale);
if(match(pattern, "data")){
//make sure to have an integer integral when rescaling data yields
hout->Scale(int(hout->Integral())/hout->Integral());
}
if(debug>1){
std::cout << "...new scale : " << hout->Integral() << std::endl;
}
if(armed){
if(hist->find("/")!=std::string::npos){
file->cd(hist->substr(0, hist->find("/")).c_str());
}
else{
file->cd();
}
std::cout << "writing to file: " << hout->GetName() << std::endl;
hout->Write(hist->substr(hist->find("/")+1).c_str(), TObject::kOverwrite);
}
}
file->Close();
return;
}
void rescale2SM4(bool armed, const char* filename, double ecms=7., double mass=-1)
{
unsigned int debug = 1;
std::vector<std::string> histnames; histnames.clear();
if( debug>0 ){
std::cout << "file = " << filename << std::setw(10);
std::cout << "mass = " << mass << std::setw(10);
std::cout << "armed = " << armed << std::endl;
}
TFile* file = new TFile(filename, "update");
TIter nextDirectory(file->GetListOfKeys());
std::vector<std::string> buffer; TKey* idir;
while((idir = (TKey*)nextDirectory())){
buffer.clear();
if( idir->IsFolder() ){
file->cd(); // make sure to start in directory head
if( debug>1 ){ std::cout << "Found directory: " << idir->GetName() << std::endl; }
if( file->GetDirectory(idir->GetName()) ){
file->cd(idir->GetName()); // change to sub-directory
buffer = signalList(idir->GetName(), "", debug);
}
// append to the vector of histograms to be rescaled
for(std::vector<std::string>::const_iterator elem=buffer.begin(); elem!=buffer.end(); ++elem){
histnames.push_back(*elem);
}
if(debug>1){
std::cout << "added " << buffer.size() << " elements to histnames [" << histnames.size() << "] for directory " << idir->GetName() << std::endl;
}
}
}
// pick up histograms which are not kept in an extra folder
file->cd(); buffer.clear();
buffer = signalList("", "", debug);
// append to the vector of histograms to be rescaled
for(std::vector<std::string>::const_iterator elem=buffer.begin(); elem!=buffer.end(); ++elem){
histnames.push_back(*elem);
}
if(debug>1){
std::cout << "added " << buffer.size() << " elements to histnames [" << histnames.size() << "] for file head" << std::endl;
}
HiggsCSandWidth smx; HiggsCSandWidthSM4 sm4;
for(std::vector<std::string>::const_iterator hist=histnames.begin(); hist!=histnames.end(); ++hist){
int type = 0;
// determine mass from hostogram name
std::string strippedName = (hist->find("/")!=std::string::npos ? hist->substr(hist->find("/")+1) : *hist);
std::string massName;
if(strippedName.find("ggH")!=std::string::npos) {
massName = strippedName.substr(3, 3); type = 1;
}
if(strippedName.find("qqH")!=std::string::npos) {
massName = strippedName.substr(3, 3); type = 2;
}
if(strippedName.find("VH" )!=std::string::npos) {
massName = strippedName.substr(2, 3); type = 3;
}
if( type==0 ) {
std::cout << "not supported process" << std::endl;
continue;
}
else {
file->cd();
float mdx = atof(massName.c_str());
TH1F* h = (TH1F*)file->Get(hist->c_str());
float smxXS = type==1 ? smx.HiggsCS(type, mdx, ecms, true) : 0.; float smxBR = smx.HiggsBR(2, mdx, true);
float sm4XS = type==1 ? sm4.HiggsCS(type, mdx, ecms, true) : 0.; float sm4BR = sm4.HiggsBR(2, mdx, true);
if( debug>1 ){
std::cout << " -- hist = " << std::setw(10) << h->GetName() << std::endl
<< " -- type = " << std::setw(10) << type << std::endl
<< " -- mass = " << std::setw(10) << mdx << std::endl
<< " -- SM = " << std::setw(10) << smxXS*smxBR << " (BR = " << smxBR << ")" << std::endl
<< " -- SM4 = " << std::setw(10) << sm4XS*sm4BR << " (BR = " << sm4BR << ")" << std::endl
<< " -- scale = " << std::setw(10) << (type==1 ? sm4XS*sm4BR/(smxXS*smxBR) : 0) << std::endl
<< std::endl;
}
if( type==1 ){ h->Scale(sm4XS*sm4BR/(smxXS*smxBR)); }
//scaling old style
//if( type==2 ){ h->Scale(sm4BR/smxBR); }
//if( type==3 ){ h->Scale(sm4BR/smxBR); }
// scaling new style
if( type==2 ){ h->Scale(0.); }
if( type==3 ){ h->Scale(0.); }
if(armed){
if(hist->find("/")!=std::string::npos){
file->cd(hist->substr(0, hist->find("/")).c_str());
}
else{
file->cd();
}
h->Write(strippedName.c_str(), TObject::kOverwrite);
}
}
}
file->Close();
return;
}
void changeAxis(bool armed, TString iExprMatch,float iMin, const char* filename, const char* pattern="", unsigned int debug=0)
{
std::vector<std::string> histnames; histnames.clear();
if( debug>0 ){
std::cout << "file = " << filename << std::endl;
std::cout << "old = " << iExprMatch.Data() << std::endl;
std::cout << "armed = " << armed << std::endl;
}
TFile* file = new TFile(filename, "update");
TIter nextDirectory(file->GetListOfKeys());
std::vector<std::string> buffer;
TKey* idir;
while((idir = (TKey*)nextDirectory())){
buffer.clear();
if( idir->IsFolder() ){
file->cd(); // make sure to start in directory head
if( debug>0 ){ std::cout << "Found directory: " << idir->GetName() << std::endl; }
if( file->GetDirectory(idir->GetName()) ){
file->cd(idir->GetName()); // change to sub-directory
buffer = signalList(idir->GetName(), pattern,iExprMatch.Data(), debug);
}
// append to the vector of histograms to be rescaled
for(std::vector<std::string>::const_iterator elem=buffer.begin(); elem!=buffer.end(); ++elem){
histnames.push_back(*elem);
}
if(debug>1){
std::cout << "added " << buffer.size() << " elements to histnames [" << histnames.size() << "] for directory " << idir->GetName() << std::endl;
}
}
}
// pick up files which are not kept in an extra folder
file->cd(); buffer.clear();
buffer = signalList("", pattern,iExprMatch.Data(), debug);
// append to the vector of histograms to be rescaled
for(std::vector<std::string>::const_iterator elem=buffer.begin(); elem!=buffer.end(); ++elem){
histnames.push_back(*elem);
}
if(debug>1){
std::cout << "added " << buffer.size() << " elements to histnames [" << histnames.size() << "] for file head" << std::endl;
}
for(std::vector<std::string>::const_iterator hist=histnames.begin(); hist!=histnames.end(); ++hist){
file->cd();
TH1F* h = (TH1F*)file->Get(hist->c_str());
std::string histName;
if(hist->find("/")!=std::string::npos){
histName = hist->substr(hist->find("/")+1);
}
else{
histName = *hist;
}
TH1F* hout = (TH1F*)h->Clone(histName.c_str());
if(debug>1){
std::cout << "...folder : " << hist->substr(0, hist->find("/")).c_str() << std::endl;
std::cout << "...histogram : " << hout->GetName () << " / " << hist->c_str() << std::endl;
}
for(int i0 = 0; i0 < hout->GetNbinsX()+1; i0++) if(hout->GetXaxis()->GetBinCenter(i0) < iMin) hout->SetBinContent(i0,0);
if(debug>1){
std::cout << "...new name : " << hout->GetName() << std::endl;
std::cout << "...new title : " << hout->GetTitle() << std::endl;
}
if(armed){
if(hist->find("/")!=std::string::npos){
file->cd(hist->substr(0, hist->find("/")).c_str());
}
else{
file->cd();
}
std::cout << "writing to file: " << hout->GetName() << " -- " << hist->substr(hist->find("/")+1).c_str() << std::endl;
hout->Write(hist->substr(hist->find("/")+1).c_str(), TObject::kOverwrite);
}
}
file->Close();
return;
}
void blindData(const char* filename, const char* background_patterns="Fakes, EWK, ttbar, Ztt", const char* signal_patterns="ggH125, qqH125, VH125", const char* directory_patterns="*", bool armed=false, int rnd=-1, float signal_scale=1., const char* outputLabel="", unsigned int debug=1)
{
/// prepare input parameters
std::vector<std::string> signals;
string2Vector(cleanupWhitespaces(signal_patterns), signals);
std::vector<std::string> samples;
string2Vector(cleanupWhitespaces(background_patterns), samples);
samples.insert(samples.end(), signals.begin(), signals.end());
// check if mssm or sm input file
bool sm=true;
if(std::string(filename).find("mssm") != std::string::npos){
sm=false;
std::cerr << "INFO : MSSM File " << std::endl;
}
// in case data_obs is supposed to be written to an extra output file
// open the file, otherwise the data_obs in the input file will be
// overwritten
string used_filename=string(filename);
TFile* outputFile = 0;
if(!std::string(outputLabel).empty()){
std::string out = std::string(used_filename);
if(std::string(outputLabel).find("MSSM") != std::string::npos){
outputFile = new TFile((out.substr(0, out.rfind("."))+".root").c_str(), "update");
}
// make sure data_obs in correct file for MSSM injection is changed
// this is the old file which is overwritten
else{
outputFile = new TFile((out.substr(0, out.rfind("."))+"_"+outputLabel+".root").c_str(), "update");
}
}
TKey* idir;
TH1F* buffer = 0;
TH1F* blind_data_obs = 0;
TFile* file = new TFile(used_filename.c_str(), "update");
TIter nextDirectory(file->GetListOfKeys());
while((idir = (TKey*)nextDirectory())){
if( idir->IsFolder() ){
file->cd(); // make sure to start in directory head
if( debug>0 ){ std::cerr << "Found directory: " << idir->GetName() << std::endl; }
// check if we want to muck w/ this directory. For the vhtt case, we
// have different background types in the same root file, so we have
// to run blindData twice.
if (!inPatterns(std::string(idir->GetName()), directory_patterns)) {
if( debug>0 ){
std::cerr << "WARNING: Skipping directory: " << idir->GetName() << std::endl;
std::cerr << " No match found in pattern: " << directory_patterns << std::endl;
}
continue;
}
if( file->GetDirectory(idir->GetName()) ){
file->cd(idir->GetName()); // change to sub-directory
buffer = (TH1F*)file->Get((std::string(idir->GetName())+"/data_obs").c_str());
if(!buffer){
std::cout << "WARNING: Did not find histogram data_obs in directory: " << idir->GetName() << std::endl;
std::cout << " Will skip directory: " << std::endl;
continue;
}
blind_data_obs = (TH1F*)buffer->Clone("data_obs");
if(!samples.empty()){
blind_data_obs->Reset();
std::cout << "INFO : Blinding datacads now." << std::endl;
for(std::vector<std::string>::const_iterator sample = samples.begin(); sample!=samples.end(); ++sample){
if( debug>0 ){ std::cerr << "Looking for histogram: " << (std::string(idir->GetName())+"/"+(*sample)) << std::endl; }
// add special treatment for et/mt ZLL,ZJ,ZL here. You can run the
// macro with ZLL, ZL, ZJ in the background samples. Those samples,
// which do not apply for one or the other event category are
// skipped here.
if(sm==true){
if(std::string(idir->GetName()).find("vbf") != std::string::npos && (*sample == std::string("ZL") || *sample == std::string("ZJ"))){
continue;
}
else if(std::string(idir->GetName()).find("vbf") == std::string::npos && *sample == std::string("ZLL")){
continue;
}
}
else{
if(std::string(idir->GetName()).find("nobtag") == std::string::npos && (*sample == std::string("ZL") || *sample == std::string("ZJ"))){
continue;
}
else if(std::string(idir->GetName()).find("nobtag") != std::string::npos && *sample == std::string("ZLL")){
continue;
}
}
buffer = (TH1F*)file->Get((std::string(idir->GetName())+"/"+(*sample)).c_str());
if (!buffer) {
std::cerr << "ERROR : Could not get histogram from: " << std::string(idir->GetName())+"/"+(*sample) << std::endl;
std::cerr << " Histogram will be skipped : " << std::string(idir->GetName())+"/"+(*sample) << std::endl;
continue;
}
if(inPatterns(*sample, signal_patterns)) {
if( debug>1 ){
std::cerr << "INFO : Scale signal sample " << *sample << " by scale " << signal_scale << std::endl;
}
buffer->Scale(signal_scale);
}
blind_data_obs->Add(buffer);
if (debug > 1){
std::cerr << "INFO : Adding: " << buffer->GetName() << " -- " << buffer->Integral() << " --> New value: " << blind_data_obs->Integral() << std::endl;
}
}
}
else{
std::cout << "INFO : Data are not blinded." << std::endl;
}
if(rnd>=0){
// randomize histogram; this will automatically have integer integral
std::cerr << "-- R A N D O M I Z I N G --" << std::endl;
randomize(blind_data_obs, rnd, debug);
}
else{
// use expected mean with signal injected
blind_data_obs->Scale(TMath::Nint(blind_data_obs->Integral())/blind_data_obs->Integral());
// adjust uncertaintie
adjustUncerts(blind_data_obs);
}
std::cout << "INFO : New data_obs yield: " << idir->GetName() << " " << TMath::Nint(blind_data_obs->Integral()) << std::endl;
if(armed){
if (debug > 1){
std::cerr << "INFO : Writing to file: " << blind_data_obs->GetName() << std::endl;
}
if(outputFile){
// write to a dedicated new file with name output in case output has been specified
// make sure data_obs in correct file for MSSM injection is changed
// this is the old file which is overwritten
if(std::string(outputLabel).find("MSSM") != std::string::npos){
outputFile->cd(idir->GetName());
blind_data_obs->Write("data_obs", TObject::kOverwrite);
}
else{
outputFile->mkdir(idir->GetName()); outputFile->cd(idir->GetName());
blind_data_obs->Write("data_obs");
}
}
else{
// override old data_obs in the inputfile otherwise
file->cd(idir->GetName());
blind_data_obs->Write("data_obs", TObject::kOverwrite);
}
}
}
}
}
file->Close();
if(outputFile){
outputFile->Close();
}
return;
}
