ReweightUserHooks.h

Go to the documentation of this file.

#include "Pythia8/Pythia.h"
#include "TF1.h"

class PtHatReweightUserHook : public Pythia8::UserHooks {
public:
  PtHatReweightUserHook(double _pt = 15, double _power = 4.5) : pt(_pt), power(_power) {}
  ~PtHatReweightUserHook() override {}

  bool canBiasSelection() override { return true; }

  double biasSelectionBy(const Pythia8::SigmaProcess* sigmaProcessPtr,
                         const Pythia8::PhaseSpace* phaseSpacePtr,
                         bool inEvent) override {
    //the variable selBias of the base class should be used;
    if ((sigmaProcessPtr->nFinal() == 2)) {
      selBias = pow(phaseSpacePtr->pTHat() / pt, power);
      return selBias;
    }
    selBias = 1.;
    return selBias;
  }

private:
  double pt, power;
};

class PtHatEmpReweightUserHook : public Pythia8::UserHooks {
public:
  PtHatEmpReweightUserHook(const std::string& tuneName = "") {
    if (tuneName == "CP5")
      p = {7377.94700788, 8.38168461349, -4.70983112392, -0.0310148108446, -0.028798537937, 925.335472326};
    //Default reweighting - works good for tune CUEPT8M1
    else
      p = {5.3571961909810e+13,
           1.0907678218282e+01,
           -2.5898069229451e+00,
           -5.1575514014931e-01,
           5.5951279807561e-02,
           3.5e+02};
    sigma = [this](double x) -> double {
      return (p[0] * pow(x, p[2] + p[3] * log(0.01 * x) + p[4] * pow(log(0.01 * x), 2)) *
              pow(1 - 2 * x / (13000. + p[5]), p[1])) *
             x;
    };
  }
  ~PtHatEmpReweightUserHook() override {}

  bool canBiasSelection() override { return true; }

  double biasSelectionBy(const Pythia8::SigmaProcess* sigmaProcessPtr,
                         const Pythia8::PhaseSpace* phaseSpacePtr,
                         bool inEvent) override {
    //the variable selBias of the base class should be used;
    if ((sigmaProcessPtr->nFinal() == 2)) {
      selBias = 1.0 / sigma(phaseSpacePtr->pTHat());
      return selBias;
    }
    selBias = 1.;
    return selBias;
  }

private:
  std::vector<double> p;
  std::function<double(double)> sigma;
};

class RapReweightUserHook : public Pythia8::UserHooks {
public:
  RapReweightUserHook(const std::string& _yLabsigma_func,
                      double _yLab_power,
                      const std::string& _yCMsigma_func,
                      double _yCM_power,
                      double _pTHatMin,
                      double _pTHatMax)
      : yLabsigma_func(_yLabsigma_func),
        yCMsigma_func(_yCMsigma_func),
        yLab_power(_yLab_power),
        yCM_power(_yCM_power),
        pTHatMin(_pTHatMin),
        pTHatMax(_pTHatMax) {
    // empirical parametrizations defined in configuration file
    yLabsigma = TF1("yLabsigma", yLabsigma_func.c_str(), pTHatMin, pTHatMax);
    yCMsigma = TF1("yCMsigma", yLabsigma_func.c_str(), pTHatMin, pTHatMax);
  }
  ~RapReweightUserHook() override {}

  bool canBiasSelection() override { return true; }

  double biasSelectionBy(const Pythia8::SigmaProcess* sigmaProcessPtr,
                         const Pythia8::PhaseSpace* phaseSpacePtr,
                         bool inEvent) override {
    //the variable selBias of the base class should be used;
    if ((sigmaProcessPtr->nFinal() == 2)) {
      double x1 = phaseSpacePtr->x1();
      double x2 = phaseSpacePtr->x2();
      double yLab = 0.5 * log(x1 / x2);
      double yCM = 0.5 * log(phaseSpacePtr->tHat() / phaseSpacePtr->uHat());
      double pTHat = phaseSpacePtr->pTHat();
      double sigmaLab = yLabsigma.Eval(pTHat);
      double sigmaCM = yCMsigma.Eval(pTHat);
      // empirical reweighting function
      selBias = exp(pow(fabs(yLab), yLab_power) / (2 * sigmaLab * sigmaLab) +
                    pow(fabs(yCM), yCM_power) / (2 * sigmaCM * sigmaCM));
      return selBias;
    }
    selBias = 1.;
    return selBias;
  }

private:
  std::string yLabsigma_func, yCMsigma_func;
  double yLab_power, yCM_power, pTHatMin, pTHatMax;
  TF1 yLabsigma, yCMsigma;
};

class PtHatRapReweightUserHook : public Pythia8::UserHooks {
public:
  PtHatRapReweightUserHook(const std::string& _yLabsigma_func,
                           double _yLab_power,
                           const std::string& _yCMsigma_func,
                           double _yCM_power,
                           double _pTHatMin,
                           double _pTHatMax,
                           double _pt = 15,
                           double _power = 4.5)
      : yLabsigma_func(_yLabsigma_func),
        yCMsigma_func(_yCMsigma_func),
        yLab_power(_yLab_power),
        yCM_power(_yCM_power),
        pTHatMin(_pTHatMin),
        pTHatMax(_pTHatMax),
        pt(_pt),
        power(_power) {
    // empirical parametrizations defined in configuration file
    yLabsigma = TF1("yLabsigma", yLabsigma_func.c_str(), pTHatMin, pTHatMax);
    yCMsigma = TF1("yCMsigma", yLabsigma_func.c_str(), pTHatMin, pTHatMax);
  }
  ~PtHatRapReweightUserHook() override {}

  bool canBiasSelection() override { return true; }

  double biasSelectionBy(const Pythia8::SigmaProcess* sigmaProcessPtr,
                         const Pythia8::PhaseSpace* phaseSpacePtr,
                         bool inEvent) override {
    //the variable selBias of the base class should be used;
    if ((sigmaProcessPtr->nFinal() == 2)) {
      double x1 = phaseSpacePtr->x1();
      double x2 = phaseSpacePtr->x2();
      double yLab = 0.5 * log(x1 / x2);
      double yCM = 0.5 * log(phaseSpacePtr->tHat() / phaseSpacePtr->uHat());
      double pTHat = phaseSpacePtr->pTHat();
      double sigmaLab = yLabsigma.Eval(pTHat);
      double sigmaCM = yCMsigma.Eval(pTHat);
      // empirical reweighting function
      selBias = pow(pTHat / pt, power) * exp(pow(fabs(yLab), yLab_power) / (2 * sigmaLab * sigmaLab) +
                                             pow(fabs(yCM), yCM_power) / (2 * sigmaCM * sigmaCM));
      return selBias;
    }
    selBias = 1.;
    return selBias;
  }

private:
  std::string yLabsigma_func, yCMsigma_func;
  double yLab_power, yCM_power, pTHatMin, pTHatMax, pt, power;
  TF1 yLabsigma, yCMsigma;
};