ecal::EcalRecProducer Class Reference

Performs basic ECal reconstruction. More...

#include <EcalRecProducer.h>

Public Member Functions
	EcalRecProducer (const std::string &name, framework::Process &process)
	Constructor.
virtual	~EcalRecProducer ()=default
	Destructor.
void	configure (framework::config::Parameters &) override
	Grabs configure parameters from the python config file.
void	produce (framework::Event &event) override
	Produce EcalHits and put them into the event bus using the EcalDigis as input.
void	onNewRun (const ldmx::RunHeader &) override
	Seed the phase smearing.
Public Member Functions inherited from framework::Producer
	Producer (const std::string &name, Process &process)
	Class constructor.
virtual void	process (Event &event) final
	Processing an event for a Producer is calling produce.
Public Member Functions inherited from framework::EventProcessor
	DECLARE_FACTORY (EventProcessor, EventProcessor *, const std::string &, Process &)
	declare that we have a factory for this class
	EventProcessor (const std::string &name, Process &process)
	Class constructor.
virtual	~EventProcessor ()=default
	Class destructor.
virtual void	beforeNewRun (ldmx::RunHeader &run_header)
	Callback for Producers to add parameters to the run header before conditions are initialized.
virtual void	onFileOpen (EventFile &event_file)
	Callback for the EventProcessor to take any necessary action when a new event input ROOT file is opened.
virtual void	onFileClose (EventFile &event_file)
	Callback for the EventProcessor to take any necessary action when a event input ROOT file is closed.
virtual void	onProcessStart ()
	Callback for the EventProcessor to take any necessary action when the processing of events starts, such as creating histograms.
virtual void	onProcessEnd ()
	Callback for the EventProcessor to take any necessary action when the processing of events finishes, such as calculating job-summary quantities.
template<class T >
const T &	getCondition (const std::string &condition_name)
	Access a conditions object for the current event.
TDirectory *	getHistoDirectory ()
	Access/create a directory in the histogram file for this event processor to create histograms and analysis tuples.
void	setStorageHint (framework::StorageControl::Hint hint)
	Mark the current event as having the given storage control hint from this module_.
void	setStorageHint (framework::StorageControl::Hint hint, const std::string &purposeString)
	Mark the current event as having the given storage control hint from this module and the given purpose string.
int	getLogFrequency () const
	Get the current logging frequency from the process.
int	getRunNumber () const
	Get the run number from the process.
std::string	getName () const
	Get the processor name.
void	createHistograms (const std::vector< framework::config::Parameters > &histos)
	Internal function which is used to create histograms passed from the python configuration @parma histos vector of Parameters that configure histograms to create.

Private Attributes
std::string	digi_coll_name_
	Digi Collection Name to use as input.
std::string	digi_pass_name_
	Digi Pass Name to use as input.
std::string	sim_hit_coll_name_
	simhit collection name
std::string	sim_hit_pass_name_
	simhit pass name
std::string	rec_hit_coll_name_
	output hit collection name
double	mip_si_energy_
	Energy [MeV] deposited by a MIP in Si 0.5mm thick.
double	clock_cycle_
	Length of clock cycle [ns].
double	charge_per_mip_
	Number of electrons generated by average MIP in Si 0.5mm thick.
std::vector< double >	layer_weights_
	Layer Weights to use for this reconstruction.
double	second_order_energy_correction_
	Second Order Energy Correction to use for this reconstruction.
std::string	amplitude_method_
	ADC mode amplitude: soi, max or fit.
int	fit_n_samples_
	number of samples in the fit, starting one before the max, <= 0 for all
std::unique_ptr< ldmx::HgcrocShape >	pulse_shape_
	HGCROC pulse template, only for amplitude_method fit.
bool	fit_fixed_t0_
	fit with t0 = event phase + offset, e.g. phase from HCal TOA
double	fit_t0_offset_
	offset [ns] between the event phase and the pulse t0
double	phase_resolution_
	Gaussian smearing [ns] of the event phase.
std::string	phase_name_
	event header float parameter holding the event phase [ns]
std::mt19937	rng_

Additional Inherited Members
Protected Member Functions inherited from framework::EventProcessor
void	abortEvent ()
	Abort the event immediately.
Protected Attributes inherited from framework::EventProcessor
HistogramPool	histograms_
	helper object for making and filling histograms
NtupleManager &	ntuple_ {NtupleManager::getInstance()}
	Manager for any ntuples.
logging::logger	the_log_
	The logger for this EventProcessor.

Detailed Description

Performs basic ECal reconstruction.

Reconstruction is done from the EcalDigi samples. Some hard-coded parameters are used for position and energy calculation.

Definition at line 30 of file EcalRecProducer.h.

Constructor & Destructor Documentation

EcalRecProducer()

ecal::EcalRecProducer::EcalRecProducer	(	const std::string &	name,
		framework::Process &	process )

Constructor.

Definition at line 23 of file EcalRecProducer.cxx.

    : Producer(name, process) {}

Member Function Documentation

configure()

void ecal::EcalRecProducer::configure ( framework::config::Parameters & ps )

overridevirtual

Grabs configure parameters from the python config file.

Reimplemented from framework::EventProcessor.

Definition at line 27 of file EcalRecProducer.cxx.

                                                             {
  // collection names
  digi_coll_name_ = ps.get<std::string>("digi_coll_name");
  digi_pass_name_ = ps.get<std::string>("digi_pass_name");
  sim_hit_coll_name_ = ps.get<std::string>("sim_hit_coll_name");
  sim_hit_pass_name_ = ps.get<std::string>("sim_hit_pass_name");
  rec_hit_coll_name_ = ps.get<std::string>("rec_hit_coll_name");
 
  layer_weights_ = ps.get<std::vector<double>>("layer_weights");
  second_order_energy_correction_ =
      ps.get<double>("second_order_energy_correction");
 
  mip_si_energy_ = ps.get<double>("mip_si_energy");
  clock_cycle_ = ps.get<double>("clock_cycle");
  charge_per_mip_ = ps.get<double>("charge_per_mip");
 
  amplitude_method_ = ps.get<std::string>("amplitude_method", "soi");
  fit_n_samples_ = ps.get<int>("fit_n_samples", 0);
  fit_fixed_t0_ = ps.get<bool>("fit_fixed_t0", false);
  fit_t0_offset_ = ps.get<double>("fit_t0_offset", 0.);
  phase_resolution_ = ps.get<double>("phase_resolution", 0.);
  phase_name_ = ps.get<std::string>("phase_name", "ecal_digi_phase");
  if (amplitude_method_ == "fit") {
    auto shape = ps.get<std::vector<double>>("pulse_shape");
    if (shape.size() != 5) {
      EXCEPTION_RAISE("InvalidConfig", "pulse_shape needs 5 parameters.");
    }
    pulse_shape_ = std::make_unique<ldmx::HgcrocShape>(
        shape[0], shape[1], shape[2], shape[3], shape[4]);
  } else if (amplitude_method_ != "soi" and amplitude_method_ != "max") {
    EXCEPTION_RAISE("InvalidConfig", "Unknown amplitude_method '" +
                                         amplitude_method_ +
                                         "', use soi, max or fit.");
  }
}

References amplitude_method_, charge_per_mip_, clock_cycle_, digi_coll_name_, digi_pass_name_, fit_fixed_t0_, fit_n_samples_, fit_t0_offset_, framework::config::Parameters::get(), layer_weights_, mip_si_energy_, phase_name_, phase_resolution_, pulse_shape_, rec_hit_coll_name_, second_order_energy_correction_, sim_hit_coll_name_, and sim_hit_pass_name_.

onNewRun()

void ecal::EcalRecProducer::onNewRun ( const ldmx::RunHeader & )

overridevirtual

Seed the phase smearing.

Reimplemented from framework::EventProcessor.

Definition at line 63 of file EcalRecProducer.cxx.

                                                   {
  if (fit_fixed_t0_) {
    const auto& rseed = getCondition<framework::RandomNumberSeedService>(
        framework::RandomNumberSeedService::CONDITIONS_OBJECT_NAME);
    rng_.seed(rseed.getSeed(getName() + "::phase"));
  }
}

References framework::RandomNumberSeedService::CONDITIONS_OBJECT_NAME, fit_fixed_t0_, framework::EventProcessor::getCondition(), and framework::EventProcessor::getName().

produce()

void ecal::EcalRecProducer::produce ( framework::Event & event )

overridevirtual

Produce EcalHits and put them into the event bus using the EcalDigis as input.

This function unfolds the digi samples taken by the HGC ROC and reconstructs their energy using knowledge of how the chip operates and the position using EcalGeometry.

Non-positive Electron (charge) count This reconstruction error occurs when the ADC value is at or below the ADC pedestal for that channel. In the normal running mode, this will never happen because our front-end (the digi emulator or the digitizer itself) will suppress any signals that are below the readout threshold. Nevertheless, in some running modes, we don't have this zero suppression, so we need to check that the reconstruction charge (count of electrons) is positive. A channel sitting exactly at its pedestal carries no energy, so it is dropped as well.

Implements framework::Producer.

Definition at line 71 of file EcalRecProducer.cxx.

                                                   {
  // Get the Ecal Geometry
  const auto& geometry = getCondition<ldmx::EcalGeometry>(
      ldmx::EcalGeometry::CONDITIONS_OBJECT_NAME);
 
  // Safety check: ensure layer_weights_ covers all geometry layers.
  // A mismatch here typically means an outdated geometry configuration
  // (e.g. v14) is being used with layer weights for a different geometry
  // version (e.g. v15).
  if (geometry.getNumLayers() > static_cast<int>(layer_weights_.size())) {
    EXCEPTION_RAISE(
        "InvalidConfig",
        "The number of layers in the Ecal geometry (" +
            std::to_string(geometry.getNumLayers()) +
            ") exceeds the number of configured layer_weights (" +
            std::to_string(layer_weights_.size()) +
            "). This is likely caused by a mismatch between the geometry "
            "version and the reconstruction configuration. Please ensure "
            "that the correct geometry is being used.");
  }
 
  // Get the reconstruction parameters
  EcalReconConditions the_conditions(
      getCondition<conditions::DoubleTableCondition>(
          EcalReconConditions::CONDITIONS_NAME));
 
  std::vector<ldmx::EcalHit> ecal_rec_hits;
  auto ecal_digis = event.getObject<ldmx::HgcrocDigiCollection>(
      digi_coll_name_, digi_pass_name_);
  const int i_soi = ecal_digis.getSampleOfInterestIndex();
 
  // one measured phase per event, shared by all channels
  double t0_fixed{0.};
  if (fit_fixed_t0_) {
    t0_fixed =
        event.getEventHeader().getFloatParameter(phase_name_) + fit_t0_offset_;
    if (phase_resolution_ > 0.) {
      t0_fixed += std::normal_distribution<double>(0., phase_resolution_)(rng_);
    }
  }
  // loop through digis
  for (auto digi : ecal_digis) {
    // ID from first digi sample
    //  assuming rest of samples have same ID
    ldmx::EcalID id(digi.id());
 
    // ID to real space position
    auto [x_, y_, z_] = geometry.getPosition(id);
 
    // the TOT measurement is in the sample where the pulse fell back below
    // threshold, which is the SOI only for in-time hits (iss #1944)
    int i_tot = digi.totSampleIndex();
 
    // get the estimated charge deposited from digi samples
    double charge(0.);
    // TOA with respect to the 25ns clock window that measured it
    double hit_time(0.);
 
    if (i_tot >= 0) {
      // TOT - number of clock ticks that pulse was over threshold
      //  this is related to the amplitude of the pulse approximately through a
      //  linear drain rate the amplitude of the pulse is related to the energy
      //  deposited
      auto tot_sample = digi.at(i_tot);
 
      // convert the time over threshold into a total energy deposited in the
      // silicon
      //  (time over threshold [ns] - pedestal) * gain
      charge = (tot_sample.tot() - the_conditions.totPedestal(id)) *
               the_conditions.totGain(id);
 
      // TOA is measured from the start of its own sample
      hit_time = (i_tot - i_soi) * clock_cycle_ +
                 tot_sample.toa() * (clock_cycle_ / 1024);
 
      ldmx_log(trace) << "Recon { TOA: " << hit_time << " ns } ";
      ldmx_log(trace) << "TOT Mode (sample " << i_tot << ") -> "
                      << tot_sample.tot() << "TDC -> " << charge << " fC";
    } else {
      hit_time = digi.soi().toa() * (clock_cycle_ / 1024);  // ns
      ldmx_log(trace) << "Recon { TOA: " << hit_time << " ns } ";
 
      // ADC mode of readout
      // ADC - voltage measurement at a specific time of the pulse
      // Pulse Shape:
      //  p[0]/(1.0+exp(p[1](t-p[2]+p[3]-p[4])))/(1.0+exp(p[5]*(t-p[6]+p[3]-p[4])))
      //  p[0] = amplitude to be fit (TBD)
      //  p[1] = -0.345 shape parameter - rate of up slope
      //  p[2] = 70.6547 shape parameter - time of up slope relative to shape
      //  fit p[3] = 77.732 shape parameter - time of peak relative to shape fit
      //  p[4] = peak time to be fit (TBD)
      //  p[5] = 0.140068 shape parameter - rate of down slope
      //  p[6] = 87.7649 shape paramter - time of down slope relative to shape
      //  fit
      // These measurements can be used to fit the pulse shape if TOT is not
      // available. For now, we simply take the measurement of the SOI as the
      // peak amplitude.
 
      const double pedestal = the_conditions.adcPedestal(id);
      if (amplitude_method_ == "soi") {
        charge = (digi.soi().adcT() - pedestal) * the_conditions.adcGain(id);
      } else {
        std::vector<short> adc;
        for (unsigned int i{0}; i < digi.size(); i++)
          adc.push_back(static_cast<short>(digi.at(i).adcT()));
        int i_max = std::max_element(adc.begin(), adc.end()) - adc.begin();
        if (amplitude_method_ == "max") {
          charge = (adc[i_max] - pedestal) * the_conditions.adcGain(id);
          // TOA is measured from the start of its own sample
          hit_time = (i_max - i_soi) * clock_cycle_ +
                     digi.at(i_max).toa() * (clock_cycle_ / 1024);
        } else {
          // window starts one sample before the peak to catch the rise
          const int n_adc = static_cast<int>(adc.size());
          const int n =
              fit_n_samples_ > 0 ? std::min(fit_n_samples_, n_adc) : n_adc;
          int i_peak = i_max;
          if (fit_fixed_t0_) {
            i_peak =
                i_soi + static_cast<int>(std::lround(t0_fixed / clock_cycle_));
          }
          const int first = std::clamp(i_peak - 1, 0, n_adc - n);
          std::vector<short> window(adc.begin() + first,
                                    adc.begin() + first + n);
          // peak within half a clock of the window, else noise tails fake
          // pulses
          ldmx::PulseFitter fitter(
              *pulse_shape_, (first - i_soi) * clock_cycle_, clock_cycle_,
              pedestal, 1., (first - i_soi - 0.5) * clock_cycle_,
              (first + n - 1 - i_soi + 0.5) * clock_cycle_, 0.25);
          auto r = fit_fixed_t0_ ? fitter.fitFixed(window, t0_fixed)
                                 : fitter.fit(window);
          charge = r.converged ? r.amplitude * the_conditions.adcGain(id) : 0.;
          hit_time = r.t0;
        }
      }
 
      ldmx_log(trace) << "ADC Mode -> " << charge << " fC";
    }
 
    if (charge <= 0) continue;
 
    double num_mips_equivalent = charge / charge_per_mip_;
    double energy_deposited_in_si = num_mips_equivalent * mip_si_energy_;
 
    ldmx_log(trace) << " -> " << num_mips_equivalent << " equiv MIPs -> "
                    << energy_deposited_in_si << " MeV";
 
    // incorporate layer_ weights
    double reconstructed_energy =
        (num_mips_equivalent *
             layer_weights_.at(
                 id.layer())       // energy lost in non-sensitive layers
         + energy_deposited_in_si  // energy deposited in Si itself
         ) *
        second_order_energy_correction_;
 
    // copy over information to rec hit structure in new collection
    ldmx::EcalHit rec_hit;
    rec_hit.setID(id.raw());
    rec_hit.setXPos(x_);
    rec_hit.setYPos(y_);
    rec_hit.setZPos(z_);
    rec_hit.setAmplitude(energy_deposited_in_si);
    rec_hit.setEnergy(reconstructed_energy);
    rec_hit.setTime(hit_time);
 
    ecal_rec_hits.push_back(rec_hit);
  }
 
  if (event.exists(sim_hit_coll_name_, sim_hit_pass_name_)) {
    // ecal sim hits_ exist ==> label which hits_ are real and which are pure
    // noise
    auto ecal_sim_hits{event.getCollection<ldmx::SimCalorimeterHit>(
        sim_hit_coll_name_, sim_hit_pass_name_)};
    std::set<int> real_hits;
    for (auto const& sim_hit : ecal_sim_hits) real_hits.insert(sim_hit.getID());
    for (auto& hit : ecal_rec_hits)
      hit.setNoise(real_hits.find(hit.getID()) == real_hits.end());
  }
 
  // add collection to event bus
  event.add(rec_hit_coll_name_, ecal_rec_hits);
}

References ecal::EcalReconConditions::adcGain(), ecal::EcalReconConditions::adcPedestal(), amplitude_method_, charge_per_mip_, clock_cycle_, ecal::EcalReconConditions::CONDITIONS_NAME, ldmx::PulseFitter::FitResult::converged, digi_coll_name_, digi_pass_name_, framework::Event::exists(), ldmx::PulseFitter::fit(), fit_fixed_t0_, fit_n_samples_, fit_t0_offset_, ldmx::PulseFitter::fitFixed(), framework::EventProcessor::getCondition(), ldmx::HgcrocDigiCollection::getSampleOfInterestIndex(), layer_weights_, mip_si_energy_, phase_name_, phase_resolution_, pulse_shape_, rec_hit_coll_name_, second_order_energy_correction_, ldmx::CalorimeterHit::setAmplitude(), ldmx::CalorimeterHit::setEnergy(), ldmx::CalorimeterHit::setID(), ldmx::CalorimeterHit::setTime(), ldmx::CalorimeterHit::setXPos(), ldmx::CalorimeterHit::setYPos(), ldmx::CalorimeterHit::setZPos(), sim_hit_coll_name_, sim_hit_pass_name_, ecal::EcalReconConditions::totGain(), and ecal::EcalReconConditions::totPedestal().

Member Data Documentation

amplitude_method_

std::string ecal::EcalRecProducer::amplitude_method_

private

ADC mode amplitude: soi, max or fit.

Definition at line 106 of file EcalRecProducer.h.

Referenced by configure(), and produce().

charge_per_mip_

double ecal::EcalRecProducer::charge_per_mip_

private

Number of electrons generated by average MIP in Si 0.5mm thick.

Definition at line 83 of file EcalRecProducer.h.

Referenced by configure(), and produce().

clock_cycle_

double ecal::EcalRecProducer::clock_cycle_

private

Length of clock cycle [ns].

Definition at line 80 of file EcalRecProducer.h.

Referenced by configure(), and produce().

digi_coll_name_

std::string ecal::EcalRecProducer::digi_coll_name_

private

Digi Collection Name to use as input.

Definition at line 62 of file EcalRecProducer.h.

Referenced by configure(), and produce().

digi_pass_name_

std::string ecal::EcalRecProducer::digi_pass_name_

private

Digi Pass Name to use as input.

Definition at line 65 of file EcalRecProducer.h.

Referenced by configure(), and produce().

fit_fixed_t0_

bool ecal::EcalRecProducer::fit_fixed_t0_

private

fit with t0 = event phase + offset, e.g. phase from HCal TOA

Definition at line 115 of file EcalRecProducer.h.

Referenced by configure(), onNewRun(), and produce().

fit_n_samples_

int ecal::EcalRecProducer::fit_n_samples_

private

number of samples in the fit, starting one before the max, <= 0 for all

Definition at line 109 of file EcalRecProducer.h.

Referenced by configure(), and produce().

fit_t0_offset_

double ecal::EcalRecProducer::fit_t0_offset_

private

offset [ns] between the event phase and the pulse t0

Definition at line 118 of file EcalRecProducer.h.

Referenced by configure(), and produce().

layer_weights_

std::vector<double> ecal::EcalRecProducer::layer_weights_

private

Layer Weights to use for this reconstruction.

Layer weights account for the energy lost in the absorber directly in front of the Silicon layer where the measured energy was deposited. These are determined by calculating the average amount of energy lost by a MIP passing through the extra material between sensitive layers.

Definition at line 93 of file EcalRecProducer.h.

Referenced by configure(), and produce().

mip_si_energy_

double ecal::EcalRecProducer::mip_si_energy_

private

Energy [MeV] deposited by a MIP in Si 0.5mm thick.

Definition at line 77 of file EcalRecProducer.h.

Referenced by configure(), and produce().

phase_name_

std::string ecal::EcalRecProducer::phase_name_

private

event header float parameter holding the event phase [ns]

Definition at line 124 of file EcalRecProducer.h.

Referenced by configure(), and produce().

phase_resolution_

double ecal::EcalRecProducer::phase_resolution_

private

Gaussian smearing [ns] of the event phase.

Definition at line 121 of file EcalRecProducer.h.

Referenced by configure(), and produce().

pulse_shape_

std::unique_ptr<ldmx::HgcrocShape> ecal::EcalRecProducer::pulse_shape_

private

HGCROC pulse template, only for amplitude_method fit.

Definition at line 112 of file EcalRecProducer.h.

Referenced by configure(), and produce().

rec_hit_coll_name_

std::string ecal::EcalRecProducer::rec_hit_coll_name_

private

output hit collection name

Definition at line 74 of file EcalRecProducer.h.

Referenced by configure(), and produce().

rng_

std::mt19937 ecal::EcalRecProducer::rng_

private

Definition at line 126 of file EcalRecProducer.h.

second_order_energy_correction_

double ecal::EcalRecProducer::second_order_energy_correction_

private

Second Order Energy Correction to use for this reconstruction.

This is a shift applied to all of the energies in order to have the mean of the total energy deposited in the ECal be accurate. This is less physically motivated than the layer weights and is more of a calibration number.

Definition at line 103 of file EcalRecProducer.h.

Referenced by configure(), and produce().

sim_hit_coll_name_

std::string ecal::EcalRecProducer::sim_hit_coll_name_

private

simhit collection name

Definition at line 68 of file EcalRecProducer.h.

Referenced by configure(), and produce().

sim_hit_pass_name_

std::string ecal::EcalRecProducer::sim_hit_pass_name_

private

simhit pass name

Definition at line 71 of file EcalRecProducer.h.

Referenced by configure(), and produce().

---

The documentation for this class was generated from the following files:

• Ecal/include/Ecal/EcalRecProducer.h
• Ecal/src/Ecal/EcalRecProducer.cxx