LDMX Software
tracking::reco::PedestalCalculator Class Reference

Compute per-channel, per-sample pedestal mean and noise from a baseline run. More...

#include <PedestalCalculator.h>

Classes

struct  Accumulator
 

Public Member Functions

 PedestalCalculator (const std::string &name, framework::Process &process)
 
void configure (framework::config::Parameters &ps) override
 Callback for the EventProcessor to configure itself from the given set of parameters.
 
void analyze (const framework::Event &event) override
 Process the event and make histograms or summaries.
 
void onProcessEnd () override
 Callback for the EventProcessor to take any necessary action when the processing of events finishes, such as calculating job-summary quantities.
 
- Public Member Functions inherited from framework::Analyzer
 Analyzer (const std::string &name, Process &process)
 Class constructor.
 
virtual void process (Event &event) final
 Processing an event for an Analyzer is calling analyze.
 
virtual void beforeNewRun (ldmx::RunHeader &run_header) final
 Don't allow Analyzers to add parameters to the run header.
 
- 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 onNewRun (const ldmx::RunHeader &run_header)
 Callback for the EventProcessor to take any necessary action when the run being processed changes.
 
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.
 
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 Member Functions

void writePedestalsJson ()
 Write the accumulated pedestals to a JSON file (output_file_).
 

Private Attributes

std::string input_collection_ {"TrackerRawData"}
 
std::string input_pass_name_ {""}
 
std::string output_file_ {"pedestals.json"}
 
std::string output_format_ {"json"}
 Storage backend for the pedestals ("json"; future: "sqlite", ...).
 
std::unordered_map< std::string, Accumulatoraccumulators_
 
long n_events_ {0}
 

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
 
NtupleManagerntuple_ {NtupleManager::getInstance()}
 Manager for any ntuples.
 
logging::logger the_log_
 The logger for this EventProcessor.
 

Detailed Description

Compute per-channel, per-sample pedestal mean and noise from a baseline run.

Accumulates ADC values from every RawSiStripHit across all events, then on process end writes a JSON file containing the mean and RMS noise for each of the three APV25 samples. Feed the resulting JSON to PedestalSubtractor when running physics reconstruction.

Definition at line 21 of file PedestalCalculator.h.

Constructor & Destructor Documentation

◆ PedestalCalculator()

tracking::reco::PedestalCalculator::PedestalCalculator ( const std::string & name,
framework::Process & process )
inline

Definition at line 23 of file PedestalCalculator.h.

Base class for a module which does not produce a data product.
virtual void process(Event &event) final
Processing an event for an Analyzer is calling analyze.

Member Function Documentation

◆ analyze()

void tracking::reco::PedestalCalculator::analyze ( const framework::Event & event)
overridevirtual

Process the event and make histograms or summaries.

Parameters
eventThe Event to analyze

Implements framework::Analyzer.

Definition at line 20 of file PedestalCalculator.cxx.

20 {
21 const auto& hits = event.getCollection<ldmx::RawSiStripHit>(input_collection_,
22 input_pass_name_);
23
24 for (const auto& hit : hits) {
25 const auto key = channelmap::channelKey(hit.getFebId(), hit.getHybridId(),
26 hit.getApvId(), hit.getChannel());
27 auto& acc = accumulators_[key];
28 acc.n_++;
29 const auto& samples = hit.getSamples();
30 for (int s = 0; s < channelmap::K_SAMPLES_PER_APV_TRIGGER &&
31 s < static_cast<int>(samples.size());
32 ++s) {
33 // Welford's online algorithm: numerically stable incremental
34 // mean+variance.
35 double v = samples[s];
36 double delta = v - acc.mean_[s];
37 acc.mean_[s] += delta / acc.n_;
38 acc.m2_[s] += delta * (v - acc.mean_[s]);
39 }
40 }
41 ++n_events_;
42}
Implementation of a raw digitized hit from a silicon strip detector.
std::string channelKey(uint8_t feb, uint8_t hybrid, uint8_t apv, uint8_t channel)
Build the per-channel pedestal-map key "feb:hybrid:apv:channel".
constexpr int K_SAMPLES_PER_APV_TRIGGER
Number of ADC samples the APV25 reads out per channel per trigger.

References tracking::reco::channelmap::channelKey(), and tracking::reco::channelmap::K_SAMPLES_PER_APV_TRIGGER.

◆ configure()

void tracking::reco::PedestalCalculator::configure ( framework::config::Parameters & parameters)
overridevirtual

Callback for the EventProcessor to configure itself from the given set of parameters.

The parameters a processor has access to are the member variables of the python class in the sequence that has class_name equal to the EventProcessor class name.

For an example, look at MyProcessor.

Parameters
parametersParameters for configuration.

Reimplemented from framework::EventProcessor.

Definition at line 12 of file PedestalCalculator.cxx.

12 {
13 input_collection_ =
14 ps.get<std::string>("input_collection", input_collection_);
15 input_pass_name_ = ps.get<std::string>("input_pass_name", input_pass_name_);
16 output_file_ = ps.get<std::string>("output_file", output_file_);
17 output_format_ = ps.get<std::string>("output_format", output_format_);
18}
std::string output_format_
Storage backend for the pedestals ("json"; future: "sqlite", ...).

References framework::config::Parameters::get(), and output_format_.

◆ onProcessEnd()

void tracking::reco::PedestalCalculator::onProcessEnd ( )
overridevirtual

Callback for the EventProcessor to take any necessary action when the processing of events finishes, such as calculating job-summary quantities.

Reimplemented from framework::EventProcessor.

Definition at line 44 of file PedestalCalculator.cxx.

44 {
45 // Dispatch on the configured storage backend. Only JSON is supported today;
46 // additional backends (e.g. SQLite) can be added as new write*() methods.
47 if (output_format_ == "json") {
49 } else {
50 ldmx_log(error) << "Unknown output_format '" << output_format_
51 << "' (supported: 'json')";
52 }
53}
void writePedestalsJson()
Write the accumulated pedestals to a JSON file (output_file_).

References output_format_, and writePedestalsJson().

◆ writePedestalsJson()

void tracking::reco::PedestalCalculator::writePedestalsJson ( )
private

Write the accumulated pedestals to a JSON file (output_file_).

Definition at line 55 of file PedestalCalculator.cxx.

55 {
56 nlohmann::json channels = nlohmann::json::object();
57 for (const auto& [key, acc] : accumulators_) {
58 std::array<double, channelmap::K_SAMPLES_PER_APV_TRIGGER> noise{};
59 for (int s = 0; s < channelmap::K_SAMPLES_PER_APV_TRIGGER; ++s) {
60 // Welford: M2/(n-1) is the sample variance; M2/n is population variance.
61 // Use population variance (divide by n) since we want noise of the
62 // distribution.
63 noise[s] = acc.n_ > 1 ? std::sqrt(acc.m2_[s] / acc.n_) : 0.0;
64 }
65 channels[key] = {{"mean", acc.mean_}, {"noise", noise}};
66 }
67
68 nlohmann::json doc = {{"n_events", n_events_}, {"channels", channels}};
69
70 std::ofstream out(output_file_);
71 if (!out) {
72 ldmx_log(error) << "Cannot write to '" << output_file_ << "'";
73 return;
74 }
75 out << doc.dump(2) << std::endl;
76
77 ldmx_log(info) << "Wrote " << accumulators_.size() << " channels from "
78 << n_events_ << " events to '" << output_file_ << "'";
79}

References tracking::reco::channelmap::K_SAMPLES_PER_APV_TRIGGER.

Referenced by onProcessEnd().

Member Data Documentation

◆ accumulators_

std::unordered_map<std::string, Accumulator> tracking::reco::PedestalCalculator::accumulators_
private

Definition at line 49 of file PedestalCalculator.h.

◆ input_collection_

std::string tracking::reco::PedestalCalculator::input_collection_ {"TrackerRawData"}
private

Definition at line 43 of file PedestalCalculator.h.

43{"TrackerRawData"};

◆ input_pass_name_

std::string tracking::reco::PedestalCalculator::input_pass_name_ {""}
private

Definition at line 44 of file PedestalCalculator.h.

44{""};

◆ n_events_

long tracking::reco::PedestalCalculator::n_events_ {0}
private

Definition at line 50 of file PedestalCalculator.h.

50{0};

◆ output_file_

std::string tracking::reco::PedestalCalculator::output_file_ {"pedestals.json"}
private

Definition at line 45 of file PedestalCalculator.h.

45{"pedestals.json"};

◆ output_format_

std::string tracking::reco::PedestalCalculator::output_format_ {"json"}
private

Storage backend for the pedestals ("json"; future: "sqlite", ...).

Definition at line 47 of file PedestalCalculator.h.

47{"json"};

Referenced by configure(), and onProcessEnd().


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