H1D.cpp

Go to the documentation of this file.

/***********************************************************************************\
* (c) Copyright 1998-2025 CERN for the benefit of the LHCb and ATLAS collaborations *
*                                                                                   *
* This software is distributed under the terms of the Apache version 2 licence,     *
* copied verbatim in the file "LICENSE".                                            *
*                                                                                   *
* In applying this licence, CERN does not waive the privileges and immunities       *
* granted to it by virtue of its status as an Intergovernmental Organization        *
* or submit itself to any jurisdiction.                                             *
\***********************************************************************************/
#ifdef __ICC
// disable icc remark #2259: non-pointer conversion from "X" to "Y" may lose significant bits
//   TODO: To be removed, since it comes from ROOT TMathBase.h
#  pragma warning( disable : 2259 )
#endif
 
#include "GaudiPI.h"
#include <Gaudi/MonitoringHub.h>
#include <GaudiCommonSvc/H1D.h>
#include <GaudiCommonSvc/HistogramUtility.h>
#include <GaudiKernel/ObjectFactory.h>
#include <GaudiKernel/StreamBuffer.h>
 
#include <cmath>
 
std::pair<DataObject*, AIDA::IHistogram1D*> Gaudi::createH1D( ISvcLocator* svcLocator, const std::string& path,
                                                              const std::string& title, int nBins, double xlow,
                                                              double xup ) {
  auto p = new Histogram1D( new TH1D( title.c_str(), title.c_str(), nBins, xlow, xup ) );
  svcLocator->monitoringHub().registerEntity( "", path, "histogram:Histogram:double", *p );
  return { p, p };
}
 
std::pair<DataObject*, AIDA::IHistogram1D*> Gaudi::createH1D( ISvcLocator* svcLocator, const std::string& path,
                                                              const std::string& title, const Edges& e ) {
  auto p = new Histogram1D( new TH1D( title.c_str(), title.c_str(), e.size() - 1, &e.front() ) );
  svcLocator->monitoringHub().registerEntity( "", path, "histogram:Histogram:double", *p );
  return { p, p };
}
 
std::pair<DataObject*, AIDA::IHistogram1D*> Gaudi::createH1D( ISvcLocator* svcLocator, const std::string& path,
                                                              const AIDA::IHistogram1D& hist ) {
  TH1D* h = getRepresentation<AIDA::IHistogram1D, TH1D>( hist );
  auto  n = ( h ? new Histogram1D( new TH1D( *h ) ) : nullptr );
  if ( n ) { svcLocator->monitoringHub().registerEntity( "", path, "histogram:Histogram:double", *n ); }
  return { n, n };
}
 
namespace Gaudi {
 
  template <>
  void* Generic1D<AIDA::IHistogram1D, TH1D>::cast( const std::string& className ) const {
    if ( className == "AIDA::IHistogram1D" ) return const_cast<AIDA::IHistogram1D*>( (AIDA::IHistogram1D*)this );
    if ( className == "AIDA::IHistogram" ) return const_cast<AIDA::IHistogram*>( (AIDA::IHistogram*)this );
    return nullptr;
  }
 
  template <>
  int Generic1D<AIDA::IHistogram1D, TH1D>::binEntries( int index ) const {
    if ( binHeight( index ) <= 0 ) return 0;
    double xx = binHeight( index ) / binError( index );
    return int( xx * xx + 0.5 );
  }
 
  template <>
  void Generic1D<AIDA::IHistogram1D, TH1D>::adoptRepresentation( TObject* rep ) {
    TH1D* imp = dynamic_cast<TH1D*>( rep );
    if ( !imp ) throw std::runtime_error( "Cannot adopt native histogram representation." );
    m_rep.reset( imp );
  }
} // namespace Gaudi
 
Gaudi::Histogram1D::Histogram1D() : Base( new TH1D() ) { init( "", false ); }
 
Gaudi::Histogram1D::Histogram1D( TH1D* rep ) {
  m_rep.reset( rep );
  init( m_rep->GetTitle() );
  initSums();
}
 
void Gaudi::Histogram1D::init( const std::string& title, bool initialize_axis ) {
  m_classType = "IHistogram1D";
  if ( initialize_axis ) { m_axis.initialize( m_rep->GetXaxis(), false ); }
  const TArrayD* a = m_rep->GetSumw2();
  if ( !a || ( a && a->GetSize() == 0 ) ) m_rep->Sumw2();
  setTitle( title );
  m_rep->SetDirectory( nullptr );
  m_sumEntries = 0;
  m_sumwx      = 0;
}
 
void Gaudi::Histogram1D::initSums() {
  m_sumwx      = 0;
  m_sumEntries = 0;
  for ( int i = 1, n = m_rep->GetNbinsX(); i <= n; ++i ) {
    m_sumwx += m_rep->GetBinContent( i ) * m_rep->GetBinCenter( i );
    m_sumEntries += m_rep->GetBinContent( i );
  }
}
 
bool Gaudi::Histogram1D::reset() {
  m_sumwx      = 0;
  m_sumEntries = 0;
  return Base::reset();
}

void Gaudi::Histogram1D::adoptRepresentation( TObject* rep ) {
  Gaudi::Generic1D<AIDA::IHistogram1D, TH1D>::adoptRepresentation( rep );
  if ( m_rep ) {
    init( m_rep->GetTitle() );
    initSums();
  }
}
 
bool Gaudi::Histogram1D::setBinContents( int i, int entries, double height, double error, double centre ) {
  m_rep->SetBinContent( rIndex( i ), height );
  m_rep->SetBinError( rIndex( i ), error );
  // accumulate sumwx for in range bins
  if ( i != AIDA::IAxis::UNDERFLOW_BIN && i != AIDA::IAxis::OVERFLOW_BIN ) m_sumwx += centre * height;
  m_sumEntries += entries;
  return true;
}
 
bool Gaudi::Histogram1D::setRms( double rms ) {
  m_rep->SetEntries( m_sumEntries );
  std::vector<double> stat( 11 );
  // sum weights
  stat[0] = sumBinHeights();
  stat[1] = 0;
  if ( std::abs( equivalentBinEntries() ) > std::numeric_limits<double>::epsilon() )
    stat[1] = ( sumBinHeights() * sumBinHeights() ) / equivalentBinEntries();
  stat[2]     = m_sumwx;
  double mean = 0;
  if ( std::abs( sumBinHeights() ) > std::numeric_limits<double>::epsilon() ) mean = m_sumwx / sumBinHeights();
  stat[3] = ( mean * mean + rms * rms ) * sumBinHeights();
  m_rep->PutStats( &stat.front() );
  return true;
}
 
// set histogram statistics
bool Gaudi::Histogram1D::setStatistics( int allEntries, double eqBinEntries, double mean, double rms ) {
  m_rep->SetEntries( allEntries );
  // fill statistcal vector for Root
  std::vector<double> stat( 11 );
  // sum weights
  stat[0] = sumBinHeights();
  // sum weights **2
  stat[1] = 0;
  if ( std::abs( eqBinEntries ) > std::numeric_limits<double>::epsilon() )
    stat[1] = ( sumBinHeights() * sumBinHeights() ) / eqBinEntries;
  // sum weights * x
  stat[2] = mean * sumBinHeights();
  // sum weight * x **2
  stat[3] = ( mean * mean + rms * rms ) * sumBinHeights();
  m_rep->PutStats( &stat.front() );
  return true;
}
 
bool Gaudi::Histogram1D::fill( double x, double weight ) {
  // avoid race conditions when filling the histogram
  auto guard = std::scoped_lock{ m_fillSerialization };
  m_rep->Fill( x, weight );
  return true;
}
 
void Gaudi::Histogram1D::copyFromAida( const AIDA::IHistogram1D& h ) {
  // implement here the copy
  std::string title = h.title() + "Copy";
  if ( h.axis().isFixedBinning() ) {
    m_rep.reset(
        new TH1D( title.c_str(), title.c_str(), h.axis().bins(), h.axis().lowerEdge(), h.axis().upperEdge() ) );
  } else {
    Edges e;
    for ( int i = 0; i < h.axis().bins(); ++i ) { e.push_back( h.axis().binLowerEdge( i ) ); }
    // add also upperedges at the end
    e.push_back( h.axis().upperEdge() );
    m_rep.reset( new TH1D( title.c_str(), title.c_str(), e.size() - 1, &e.front() ) );
  }
  m_axis.initialize( m_rep->GetXaxis(), false );
  m_rep->Sumw2();
  m_sumEntries = 0;
  m_sumwx      = 0;
  // sumw
  double sumw = h.sumBinHeights();
  // sumw2
  double sumw2 = 0;
  if ( std::abs( h.equivalentBinEntries() ) > std::numeric_limits<double>::epsilon() )
    sumw2 = ( sumw * sumw ) / h.equivalentBinEntries();
 
  double sumwx  = h.mean() * h.sumBinHeights();
  double sumwx2 = ( h.mean() * h.mean() + h.rms() * h.rms() ) * h.sumBinHeights();
 
  // copy the contents in
  for ( int i = -2; i < axis().bins(); ++i ) {
    // root binning starts from one !
    m_rep->SetBinContent( rIndex( i ), h.binHeight( i ) );
    m_rep->SetBinError( rIndex( i ), h.binError( i ) );
  }
  // need to do set entries after setting contents otherwise root will recalulate them
  // taking into account how many time  SetBinContents() has been called
  m_rep->SetEntries( h.allEntries() );
  // stat vector
  std::vector<double> stat( 11 );
  stat[0] = sumw;
  stat[1] = sumw2;
  stat[2] = sumwx;
  stat[3] = sumwx2;
  m_rep->PutStats( &stat.front() );
}
 
StreamBuffer& Gaudi::Histogram1D::serialize( StreamBuffer& s ) {
  // DataObject::serialize(s);
  std::string title;
  int         size;
  s >> size;
  for ( int j = 0; j < size; j++ ) {
    std::string key, value;
    s >> key >> value;
    if ( !annotation().addItem( key, value ) ) { annotation().setValue( key, value ); };
    if ( "Title" == key ) { title = value; }
  }
  double lowerEdge, upperEdge, binHeight, binError;
  int    isFixedBinning, bins;
  s >> isFixedBinning >> bins;
 
  if ( isFixedBinning ) {
    s >> lowerEdge >> upperEdge;
    m_rep.reset( new TH1D( title.c_str(), title.c_str(), bins, lowerEdge, upperEdge ) );
  } else {
    Edges edges;
    edges.resize( bins );
    for ( int i = 0; i <= bins; ++i ) s >> edges[i];
    m_rep.reset( new TH1D( title.c_str(), title.c_str(), edges.size() - 1, &edges.front() ) );
  }
  m_axis.initialize( m_rep->GetXaxis(), true );
  m_rep->Sumw2();
  m_sumEntries = 0;
  m_sumwx      = 0;
 
  for ( int i = 0; i <= bins + 1; ++i ) {
    s >> binHeight >> binError;
    m_rep->SetBinContent( i, binHeight );
    m_rep->SetBinError( i, binError );
  }
  Stat_t allEntries;
  s >> allEntries;
  m_rep->SetEntries( allEntries );
  Stat_t stats[4]; // stats array
  s >> stats[0] >> stats[1] >> stats[2] >> stats[3];
  m_rep->PutStats( stats );
  return s;
}
 
StreamBuffer& Gaudi::Histogram1D::serialize( StreamBuffer& s ) const {
  // DataObject::serialize(s);
  s << static_cast<int>( annotation().size() );
  for ( int i = 0; i < annotation().size(); i++ ) {
    s << annotation().key( i );
    s << annotation().value( i );
  }
  const AIDA::IAxis& axis( this->axis() );
  const int          isFixedBinning = axis.isFixedBinning();
  const int          bins           = axis.bins();
  s << isFixedBinning << bins;
  if ( isFixedBinning ) {
    s << axis.lowerEdge();
  } else {
    for ( int i = 0; i < bins; ++i ) s << axis.binLowerEdge( i );
  }
  s << axis.upperEdge();
  for ( int i = 0; i <= bins + 1; ++i ) s << m_rep->GetBinContent( i ) << m_rep->GetBinError( i );
 
  s << m_rep->GetEntries();
  Stat_t stats[4]; // stats array
  m_rep->GetStats( stats );
  s << stats[0] << stats[1] << stats[2] << stats[3];
  return s;
}