PositiveSideGhostTrackFitter.cc

Go to the documentation of this file.

#include <cmath>
#include <vector>

#include "DataFormats/GeometryVector/interface/GlobalVector.h" 
#include "DataFormats/GeometryVector/interface/GlobalPoint.h" 

#include "RecoVertex/GhostTrackFitter/interface/GhostTrackState.h"
#include "RecoVertex/GhostTrackFitter/interface/GhostTrackPrediction.h"

#include "RecoVertex/GhostTrackFitter/interface/PositiveSideGhostTrackFitter.h"

using namespace reco;

GhostTrackPrediction PositiveSideGhostTrackFitter::fit(
            const GhostTrackFitter::PredictionUpdater &updater,
            const GhostTrackPrediction &prior,
            std::vector<GhostTrackState> &states,
            double &ndof, double &chi2)
{
    double rho = prior.rho();
    for(unsigned int i = 0; i < states.size(); i++) {
        GhostTrackState &state = states[i];
        state.linearize(prior, true, .5 / rho);
    }

    GhostTrackPrediction pred =
            actualFitter_->fit(updater, prior, states, ndof, chi2);

    double origin = pred.lambda(origin_);
    bool done = true;
    for(unsigned int i = 0; i < states.size(); i++) {
        GhostTrackState &state = states[i];
        double lambda = state.lambda();
        if (lambda < origin && (origin - lambda) < 3.5) {
            GhostTrackState testState = state;
            testState.linearize(pred, 2. * origin - lambda);
            double ndof, chi2;

            updater.contribution(prior, testState, ndof, chi2, true);
            if (ndof > 0. && chi2 < 10.) {
                state = testState;
                if (state.weight() != 1.)
                    state.setWeight(3.);
                done = false;
            }
        }
    }

    if (!done) {
        for(unsigned int i = 0; i < states.size(); i++) {
            GhostTrackState &state = states[i];
            double lambda = state.lambda();
            if (state.weight() != 1. && lambda < origin) {
                double weight =
                    std::exp(10. * (origin - lambda) - .1);
                state.setWeight(
                    std::min(state.weight(), weight));
            }
        }

        pred = actualFitter_->fit(updater, prior, states, ndof, chi2);
    }

    return pred;
}