TestBeamHitProducer.cxx

Go to the documentation of this file.

 
#include "TrigScint/TestBeamHitProducer.h"
 
namespace trigscint {
 
TestBeamHitProducer::TestBeamHitProducer(const std::string& name,
                                         framework::Process& process)
    : Producer(name, process) {}
 
void TestBeamHitProducer::configure(framework::config::Parameters& parameters) {
  inputCol_ = parameters.getParameter<std::string>("inputCollection");
  outputCollection_ = parameters.getParameter<std::string>("outputCollection");
  inputPassName_ = parameters.getParameter<std::string>("inputPassName");
  MIPresponse_ = parameters.getParameter<std::vector<double> >("MIPresponse");
  peds_ = parameters.getParameter<std::vector<double> >("pedestals");
  gain_ =
      parameters.getParameter<std::vector<double> >("gain");  // to do: vector
  startSample_ = parameters.getParameter<int>("startSample");
  pulseWidth_ = parameters.getParameter<int>("pulseWidth");
  pulseWidthLYSO_ = parameters.getParameter<int>("pulseWidthLYSO");
  nInstrumentedChannels_ =
      parameters.getParameter<int>("nInstrumentedChannels");
  doCleanHits_ = parameters.getParameter<bool>("doCleanHits");
 
  std::cout << " [ TestBeamHitProducer ] In configure(), got parameters "
            << "\n\t inputCollection = " << inputCol_
            << "\n\t inputPassName = " << inputPassName_
            << "\n\t outputCollection = " << outputCollection_
            << "\n\t startSample = " << startSample_
            << "\n\t pulseWidth = " << pulseWidth_
            << "\n\t pulseWidthLYSO = " << pulseWidthLYSO_
            << "\n\t gain[0] = " << gain_[0]
            << "\n\t nInstrumentedChannels = " << nInstrumentedChannels_
            << "\n\t doCleanHits = " << doCleanHits_
            << "\n\t pedestals[0] = " << peds_[0]
            << "\n\t MIPresponse[0] = " << MIPresponse_[0] << "\t."
            << std::endl;
 
  return;
}
 
void TestBeamHitProducer::produce(framework::Event& event) {
  /*
    hit producer.
    sum up all charge within a certain time window. could work two ways:
        - find a time sample above a threshold or where TDC is 0-50 (0-2)
                 * this could allow for finding several pulses per event, just
    keep sliding along in time
                     -- in that case, record nPulses, startsample, pulsewidth
    for each hit
            - use a fixed start sample, and keep summing from there until
    nSamples is reached or all time samples used
 
    specifics:
            - each channel has its own pedestal to subtract. also try using the
    first 5 samples to establish a threshold in the event. store both
            - store hit Q and hit PE conversion
            - for now use the same reconstruction paradigm for LYSO and plastic.
    two notes though
              1. LYSO pulse looks wider, and might need wider window. plastic is
    fine with 5. start by 8 for LYSO (even if for large pulses, sometimes 10
    seem to be needed). could probe this by correlating plastic and LYSO, and
    checking if at some amplitude (in plastic), we start cutting off charge in
    LYSO
                  2. there is a time offset between fibers. need to have a
    parameter fiber2offset to use for elecID >= 8
                  -- added this as a variable in EventReadout, along with fiber
    number. so this class doesn't need any detailed knowledge
 
    variables to write out:
        - start sample
            - pulse width
            - n samples above threshold
            - nPulses
            - early pedestal (from first 5)
            - assumed/average channel pedestal
            - total Q
            - ped subtracted total Q
            - PE value
            - max amplitude in pulse
            - store material assumption? isLYSO?
   */
 
  float mevPerMip = 0.3;
  float pePerMip = 100;
 
  const auto channels{
      event.getCollection<trigscint::EventReadout>(inputCol_, inputPassName_)};
 
  int evNb = event.getEventNumber();
  std::vector<trigscint::TestBeamHit> hits;
  for (auto chan : channels) {
    trigscint::TestBeamHit hit;
    int bar = chan.getChanID();
    if (bar >=
        nInstrumentedChannels_)  // don't run hit reconstruction on junk signal
      continue;
    int width = pulseWidth_;
    if (bar % 2 == 0) {         // LYSO channel: allow for wider pulses
      width = pulseWidthLYSO_;  // avoid hardwiring
    }
    hit.setPulseWidth(width);
    hit.setStartSample(startSample_);
    float ped = peds_.at(bar);  // chan.getPedestal() ;
    float earlyPed = chan.getEarlyPedestal();
    hit.setPedestal(ped);
    hit.setEarlyPedestal(earlyPed);
    int isClean = 0;              // false;
    float threshold = fabs(ped);  // 2*fabs(peds_[ bar ]); // or sth
    if (doCleanHits_) threshold = 7 * fabs(ped);  // stricter cut
 
    int startT = startSample_ + chan.getTimeOffset();
    float maxQ = -999.;
    int nSampAbovePed = 0;
    int nSampAboveThr = 0;
    float totSubtrQ = 0;
    std::vector<float> q = chan.getQ();
    // go to the start sample defined for this channel.
    for (int iT = startT; iT < q.size(); iT++) {
      ldmx_log(debug) << "in event " << evNb << "; channel " << bar
                      << ", got charge[" << iT << "] = " << q.at(iT);
      // for the defined number of samples, subtract pedestal. if > 0, increment
      // sample counter.
      float subQ =
          q.at(iT) - ped;  // this might be addition, if ped is negative. should
                           // see this as channel dependence in nSampAbove
      // once beyond nSamples, want to see how long positive threshold
      // subtracted tail is --> increment sample counter in any case.
      if (subQ > 0) nSampAbovePed++;
      if (subQ > threshold) nSampAboveThr++;
      if (iT - startT < width) {  // we're in the pulse integration window
        if (subQ > maxQ)  // keep track of max Q. this is the pulse amplitude
          maxQ = subQ;    // q.at(iT);
        if (subQ > 0)
          totSubtrQ +=
              subQ;  // add positive subtracted Q to total pulse charge.
      } else if (subQ < 0 ||
                 q.at(iT) < 0)  // if after the full pulse width, subQ <
                                // pedestal, break. special case to break at q=0
                                // for cases where ped < 0
        break;
      // done
    }  // over time samples
 
    // first check that hit passes any quality cuts
    uint flag = chan.getQualityFlag();
    if (doCleanHits_) {
      //        int isLongPulse=(nSampAboveThr < 2 || nSampAboveThr >
      // width + 2);
      int isLongPulse =
          (nSampAboveThr >
           width);  // the short ones we'll have to single out otherwise (like
                    // spike flag or low Q) or live with
      flag += 4 * isLongPulse;
      if (maxQ > 2e5 && nSampAboveThr < 3 &&
          flag % 2 == 0)  // this was not flagged as a spike but is eerily
                          // narrow and with high Q; flag it as a spike.
        flag += 1;
      if (flag == 0) isClean = 1;
    }
 
    float PE = totSubtrQ * 6250. / gain_[bar];
    if (PE > 20)  // dont't want to intercalibrate the shot noise
      PE *= MIPresponse_[bar];
 
    // set pulse properties like PE and amplitude
    hit.setSampAbovePed(nSampAbovePed);
    hit.setSampAboveThr(nSampAboveThr);
    hit.setQ(totSubtrQ);
    hit.setAmplitude(maxQ);
    hit.setPE(PE);
    hit.setHitQuality(isClean);
    hit.setQualityFlag(flag);
    // set bar id. set moduleNB = 0
    hit.setBarID(bar);
    hit.setModuleID(0);  // test beam
    // the rest are a little ill-defined for now (PE-energy conversion not
    // known/different between LYSO and plastic)
    hit.setTime(-999);  // maybe?
    hit.setBeamEfrac(-1.);
    hit.setEnergy(hit.getPE() * mevPerMip / pePerMip);
 
    // add hit
    hits.push_back(hit);
  }  // over channels
 
  // at end of event, write the collection of trigger scintillator hits.
  event.add(outputCollection_, hits);
 
  return;
}
 
}  // namespace trigscint
 
DECLARE_PRODUCER_NS(trigscint, TestBeamHitProducer)