Using art in LArSoft

The art framework interface makes use of the following basic ideas - all the algorithms are stored in modules and the event information is accessed using objects living in the art namespace. The I/O uses ROOT, but it is not based on TObject.

For information on using art, please examine the art workbook. It may also be useful to issue the command art --help.

Basic Concepts

The art namespace contains, among other things, handles to information stored in an event. Objects that are stored in an event are collectively known as data products. They can either be added to an event using an art::EDProducer or art::EDFilter derived module, and they can be retrieved and operated on using an art::EDAnalyzer module. Once an object has been stored in the event, its data cannot be altered (a modified copy can be added to the event instead).

Other namespaces include:

  • fhicl - the namespace corresponding to the ParameterSet and Table information described below
  • cet - the namespace where exceptions and some other utilities live
  • mf - the namespace of the message logger facility
There is a link to the interface for each concept in the heading for that concept. The first three sections below, art::EDProducer, art::EDAnalyzer, art::EDFilter are the three main classes from which probably all module classes will derive. These are used to:
  1. add a data product to the event record, or
  2. read a data product from the event record and do analysis but not write to the event record and
  3. use the event record to determine if you want to keep this event at all to write to an output stream.

There are a number of features in art that can be found by issuing the command art --help. This will display options such as:

art --print-description <module_type>

There are also --print-available-modules and --print-available-services options available.


This is a type of module that makes data products and stores them in the art::Event. The module takes a fhicl::ParameterSet or art::EDProducer::Table in the constructor and uses that to configure that module.

The module can also implement a reconfigure method to allow for run time reconfiguration, for example while running the event display. This feature is not well supported in LArSoft, and you should not rely on it.

The user must supply the implementation for the art::EDProducer::produce() method to create and store data products.


This is a type of module that analyzes data products but cannot write them in an art::Event. The module takes a fhicl::ParameterSet or art::EDAnalyzer::Table in the constructor and uses that to configure that module.

The user must supply the implementation for the art::EDAnalyzer::analyze() method.
h3(count). art::EDFilter

The object allows the user to filter events based on information obtained about the event itself. Following modules can decide to process or ignore an event, according to the response of one or more filters.

art::EDProducer/Analyzer/Filter::beginJob, endJob, etc.

These are methods that do tasks that are needed only once a job starts, ends, or a run starts or ends, etc. For example, a beginJob method would likely contain definitions of histograms that might need to be filled during the operation of the module. A beginRun method could contain definitions of properties that can change from run to run (e.g. electron lifetime, temperature etc... ).


The art::Event is the primary way to access products made by art::EDProducer type modules.

It also provides the user with information about an event such as the run, event number, etc through methods like

// assume we have an art::Event &evt

// get the run number. RunNumber_t is a typedef to unsigned int
art::RunNumber_t run =;

// get the subrun number. SubRunNumber_t is a typedef to unsigned int
art::SubRunNumber_t subRun = evt.subRun();

// get the time stamp.  art::Timestamp::value() returns a TimeValue_t which is a typedef to unsigned long long.
// The conventional use is for the upper 32 bits to have the seconds since 1970 epoch and the lower 32 bits to be
// the number of nanoseconds within the current second.
art::Timestamp ts = evt.time();

// make it into a TTimeStamp
TTimeStamp tts(ts.timeHigh(), ts.timeLow());

// get the event number - this calls a reference to art::EventID().  EventNumber_t is a typedef to unsigned int
art::EventNumber_t event =;

The header files for the classes mentioned above are at:

The art::Event can also be used to access products, by asking it to return an art::Handle.

art::ValidHandle and art::Handle

An art::ValidHandle or an art::Handle is what is returned to an art module when a data product is requested. The request can either be from a art::EDProducer that is attempting to get objects stored in a previous reconstruction or analysis step, or it can be from a art::EDAnalyzer that is attempting to do some analysis task using the information in the object. For example, to get the data product mp::MyProd from the event, one should do

auto mplistHandle = evt.getValidHandle<std::vector<mp::MyProd>>("modulelabel");

where evt is an object of type art::Event.

You must specify the type of the data product: in this case, it is a collection of mp::MyProd, specifically a std::vector<mp::MyProd>. The "modulelabel" string is a shortcut: your program should use a art::InputTag instead, that is more specific and easy to identify as an input when reading the code. Usually, the label associated to the module that produced the data product you need is enough (e.g. "largeant" or "daq").

A art::ValidHandle always points to an existing and available data product, and if the data product is not available, an exception is thrown. Note that retrieved data products are always immutable and cannot be changed. An alternative way to read a data product is by getting a art::Handle via art::Event::getByLabel() call, but then you have to check if it found it (return value true) and if not, do the right thing.

You can also get all the data product of a given type out of the event using the art::Event::getByType method:

std::vector<art::Handle<std::vector<mp::MyProd>>> mpHandles;

Notice that no module label is used because we are getting all objects of the specified type out of the event.

art::Handle looks like a pointer in the code in that the data members of the object being handled are accessed using the -> operator (it's a "smart pointer"). For example, to get the size of the vector of mp::MyProd one can do


To use the objects in the handle collection, you can

  • Use the art::ValidHandle directly,
    // in the code below, ev is an art::Event...
    auto hnd = evt.getValidHandle<std::vector<mp::MyProd>>("..."); // use the appropriate input tag, not "..." 
    for (mp::MyProd const& prod: *hnd) {
      // do something with prod
  • Get the collection from the art::ValidHandle,
    // in the code below, ev is an art::Event...
    auto hnd = evt.getValidHandle<std::vector<mp::MyProd>>("..."); // use the appropriate input tag, not "..." 
    std::vector<mp::MyProd> const& prodvec(*hnd);
    for (mp::MyProd const& prod: prodvec) {
      // do something with prod

Occasionally, it is necessary to sort such a collection, in which case, a sorted container of pointers to the original collection can be used (e.g.):

std::vector<mp::MyProd const*> ptrColl;
for (mp::MyProd const& prod: prodvec)

or, in a more C++-aware style using functions from the algorithm or larcorealg/CoreUtils/operations.h C++ headers1:

std::vector<mp::MyProd const*> ptrColl(prodvec.size());
std::transform(cbegin(prodvec), cend(prodvec), begin(ptrColl), [](auto& p) { return &p; }); 

// or using the LArSoft utility:
std::transform(cbegin(prodvec), cend(prodvec), begin(ptrColl), util::takeAddress);
(util::takeAddress() is defined in larcorealg:source:larcorealg/CoreUtils/operations.h).

Also remember that as long as the pointers in your collection are from the data product, you can track back the index of the item in the data product (for example, prodPtr - &prodvec.front() for your vector of pointers, and prodIter - prodvec.begin(), or std::distance(prodvec.begin(), prodIter), for iterator vectors).

In all cases, you won't be able to change the content of a mp::MyProd object in the data product.

NB Only art::PtrVector<mp::MyProd> and std::vector<art::Ptr<mp::MyProd>> correctly save the references to objects made in other modules, so if you want to save the collection, it must be one of them:

art::PtrVector<mp::MyProd> ptrColl;
for (size_t i = 0; i < hnd->size(); ++i)
  ptrColl.push_back( { hnd, i } );

(the braces construct in place the argument that art::PtrVector::push_back expects, that is a art::Ptr<mp::MyProd>, out of the handle and the index of the element in the data product that handle points to).

1 Note that the right cbegin(), cend() and begin() functions are picked by the compiler from the "proper namespace" depending on the type of prod (usually prod is a std::vector, so the proper namespace is std): this is called argument-dependent lookup.

Upcast on read

The upcast on read functionality can be used to read back objects written into a file that follow a simple inheritance scheme, i.e. reading in objects of a derived type using the base class type.
For instance, imagine that recob::Track inherited from a recob::Prong. A collection of tracks, data product of type std::vector<recob::Track>, can be retrieved from the art::Event by passing a std::vector<recob::Prong const*> to the art::Event::getView() method:

// declare the std::vector
std::vector<const recob::Prong*> prongs;

// Read in the list of recob::Tracks we made in fTrackInputTag as recob::Prongs.
evt.getView(fTrackInputTag, prongs)

If you write an algorithm (a function) that uses prongs in this way:

void doSomething(std::vector<recob::Prong const*>& prongs);

you can use it with any data product derived from recob::Prong.

LArSoft strongly discourages the use of polymorphism (that is, inheritance with virtual functions) in data products. Occurrences of inheritance in data products at all are rare (if any), as the preferred solution is containment: the "derived" class contains a "base" class as a data member.

Note that the vector is of constant pointers of the requested type. You shouldn't save these vectors anywhere (i.e. as collections in other data objects).


This is a utility that will search a list of predefined directories for a relative location of a file. For instance, if you want to use the geometry definition for a detector, mydet.gdml, as the input for the Geometry service, you would pass the geometry service the value "mydet.gdml" as a std::string. The cet::search_path object then searches through the previously defined $FW_SEARCH_PATH variable to see if it can locate the specified file. It will then allow access to information about the file, including its full path. This is helpful when writing code that needs data files that could be on a different path depending on the execution environment (e.g. a BlueArc area when running locally, but dCache when running on the grid), or for generic code that can't know where customisation points are (e.g. a experiment-specific data area), or for versioned data whose path includes a version number and therefore can change (e.g. the deployed UPS products). The necessary variables are set when a user sets up the environment.

An example of using cet::search_path to find a file your code needs can be found in LArSoft's ShowerCalibrationGaloreByPID example.
Here is another:

// this is the name of the GDML file, from configuration
std::string GDMLname = pset.get<std::string>("GDML");

cet::search_path sp("FW_SEARCH_PATH");

// taking a parameter from the parameter set passed into the geometry as the first argument,
// the second argument is the local variable where to store the full path - both are std::string objects
// return value is true unless the file could not be found
if (!sp.find_file(GDMLname, fGDMLFile) {
  // failed to resolve the file name
  throw art::Exception(art::errors::Configuration)
    << "geometry file '" << GDMLname << "' not found!";

FHiCL configuration: fhicl::ParameterSet and fhicl::Table

Either objects keep track of the configuration passed to the current art module or service.
The configuration is read from files in a specific language, FHiCL, and art distributes to each module and service the chunk of configuration that pertains that module and service. For examples on using FHiCL, please read/listen to the information available at:

The interpretation of that information can follow one of two possible approaches:
  • fhicl::ParameterSet is the old way to access configuration; access can be done in one line, does not impose any structure, and it is documented in fhicl::ParameterSet. Being the first one, this is what most of LArSoft still uses.
  • fhicl::Table is the new way of accessing configuration: it provides validation of the configuration (that is, it will warn if there is unexpected additional configuration, that is usually just a typo) and can be used to provide users with run-time documentation about the configuration. It requires some structure and is in general somehow less flexible than the above. It is the future, and many of the newest modules in LArSoft adopt it. The documentation is split in two areas:

All the examples in larexamples use the self-validating configuration.

In both approaches, each element of the configuration is translated into a C++ data type, whether a trivial type or a class. Some examples:
type fhicl::ParameterSet approach fhicl::Table approach
bool pset.get<bool>("MyBool") fhicl::Atom<bool>{ fhicl::Name{"MyBool"} };
int pset.get<int>("MyInt") fhicl::Atom<int>{ fhicl::Name{"MyInt"} };
unsigned int pset.get<unsigned int>("MyUInt", 18U) fhicl::Atom<unsigned int>{ fhicl::Name{"MyUInt"}, 18U }; (with default value 18)
double pset.get<double>("MyReal") fhicl::Atom<double>{ fhicl::Name{"MyReal"} };
std::vector<int> pset.get<std::vector<int>>("MyVInt") fhicl::Sequence<int>{ fhicl::Name{"MyVInt"} };
std::string pset.get<std::string>("MyString") fhicl::Atom<std::string>{ fhicl::Name{"MyString"} };
std::vector<std::string> pset.get<std::vector<std::string>>("MyVString") fhicl::Sequence<std::string>{ fhicl::Name{"MyVString"} };
art::InputTag pset.get<art::InputTag>("inputLabel") fhicl::Atom<art::InputTag>{ fhicl::Name{"inputLabel"} };
parameter set pset.get<fhicl::ParameterSet>("subConfig") fhicl::Table<SubConfigStruct>{ fhicl::Name{"subConfig"} };

The art::InputTag is the recommended way to read an input data product label from configuration.
The latter example allows to read a chunk of configuration enclosed in nested braces. This is useful to isolate and organise parameters belonging to different algorithms. The configuration validation approach requires the specific structure of that parameters subset to be defined in a C data structure.


The art::ServiceHandle is a handle with the semantics of a smart pointer, just like the art::ValidHandle above. The resource it handles is not data, but code.
In particular, any instance of this class points to the same art service object (that is a singleton).

A art service can be used within any art module. Services can be configured using the job configuration file. A typical example of the use of a service is the detector geometry. The Geometry service is needed in most modules, but you don't want to keep making new instances of the geometry. Additionally, the different detectors may have to set different parameter values that should be handled in the job configuration.

LArSoft services pursue the factorization of the services into a framework interface (art service described above) and a service provider independent of any framework.
For this reason you will see calls to LArSoft's lar::providerFrom() in place of the art::ServiceHandle instance you would expect. For example:

auto const* geom = lar::providerFrom<geo::Geometry>();

replaces art::ServiceHandle<geo::Geometry> geom;. The auto type is effective type of the service provider, that is not geo::Geometry.
We encourage to call lar::providerFrom() no more than once per event (and suggesting no less), and to save it for further use as in the example above.


This is a specialized service that connects up to the file where histograms made by modules are to be stored. It provides a mechanism for making TObject to be stored in that file and managing the memory for those objects. For example

    // get the geometry to figure out something for the histogram
    auto const* geom = lar::providerFrom<geo::Geometry>();

    // get access to the TFile service
    art::ServiceHandle<art::TFileService> tfs;

    // make the histogram - fHist is assumed to have been declared in the .h file
    fHist    = tfs->make<TH1D>("channels",  ";channel;# PE",  geo->Nchannels(), 0, geo->Nchannels());

    // note that ROOT does not automatically register TGraphs into a TFile;
    // so we tell TFileService to register it for us;
    // the register function demands name and title of the new object,
    // followed by the object constructor arguments; here we use the default constructor
    // (therefore, no additional constructor arguments) and will set number of points and values later
    fGraph   = tfs->makeAndRegister<TGraph>("channelGraph", "Graph title;channel;# PE");

Message Facility and MessageLogger

The Message Levels

The MessageLogger provides several levels of messages that can be used to print information to an output log. The decision of which level to use is your first way to determine how frequently your message will be printed. If you choose the INFO level, then typically all such messages are printed; if instead you choose the DEBUG level, then typically those messages are not printed unless a specific environmental flag is set or the message service is configured properly.

The levels most likely to be useful are LOG_DEBUG, mf::LogInfo, mf::LogVerbatim, mf::LogWarning and mf::LogError. Note there is no mf:: in front of LOG_DEBUG because it is a macro that will include the file and line numbers of the code producing the output. These are listed in order of severity and have the following behaviors:

  • LogInfo, LogWarning, and LogError represent three levels of "severity" of the message. It is possible (see MessageLogger Parameters ) to configure the job to ignore all LogInfo messages, or all messages of severity less than LogError.
  • LogVerbatim is identical in all ways to LogInfo, except that absolutely no message formatting is performed, and no context, module, or other information is appended to the message. This is appropriate, for example, if the program has prepared a nicely-formatted table for output.
  • The category should specify what this message is about. The category will generally appear as the first part of the message, but it also plays two other roles:
    1. It is possible to set limits on how many times some specific type of message will be sent to the outputs of the logger. By "type" we mean any messages with some specific category. See MessageLogger Parameters for details.
    2. When message statistics are provided, the counts of how many times messages of a given type were issued are keyed to category, module label, and (if provided) subroutine.
    Normally a category name should be up to 20 characters long, without special characters other than underscore.
  • It is unnecessary to explicitly specify the module label or the run/event number; these are automatically provided by the logger.
  • An arbitrary number of additional objects << a << b << ... << z can be appended to the message. These can be strings, ints, doubles, or any object that has an operator<< to an ostream. (Or the message can be issued with no additional objects at all.)
  • There is no need to add spaces at the beginning or end of text items sent to the message, or to add text separators between numerical items. This spacing is taken care of by the logger. However, if any item appended to a message ends in a space, then it is assumed that the user is handling spacing explicitly, and no further automatic spaces are inserted for that message.
  • There is no need to affix any sort of endl; when the statement ends the message will be dispatched.
  • Newline characters can be included in the objects appended to the message. These will be used in formatting the message. But they are generally not necessary: Line breaks are automatically inserted if the next appended object would cause the line to exceed 80 characters.
  • LOG_DEBUG is identical to the others, except:
    • LOG_DEBUG affixes the FILE and LINE number to the message.
    • LOG_DEBUG messages are considered to be lower in severity than LogInfo messages.
    • By default, LOG_DEBUG messages will be rapidly discarded with a minimum of overhead. The user must specify in the .fcl file LOG_DEBUG messages from various modules that are to be enabled
    • Because it must get FILE and LINE from the spot issuing the message, LOG_DEBUG is implemented as a macro rather than a free function.
    • Because LOG_DEBUG is a macro, it is not prepended with the mf:: namespace designation.

Using the Message Service in Code

In order to issue messages, the module must include the MessageLogger header:

#include "messagefacility/MessageLogger/MessageLogger.h" 

Having included the necessary MessageLogger header, when code wishes to issue a message, one of these functions can be used:

mf::LogError    ("category") << a << b << ... << z;
mf::LogWarning  ("category") << a << b << ... << z;
mf::LogInfo     ("category") << a << b << ... << z;
mf::LogVerbatim ("category") << a << b << ... << z;
LOG_DEBUG       ("category") << a << b << ... << z;

The easiest way to produce output that is formatted to your specifications is to emply the mf::LogVerbatim level. This level has absolutely no extra formatting from the message service tacked on to it and most closely resembles what one would expect from std::cout.

You can use std formatting functions in all message service levels. For example

mf::LogVerbatim("MyMessageStream") << "define the width of the following field to be 5, " 
                                   << std::setw(5) << 999 
                                   << " and now define the precision of the following field to be 1, " 
                                   << std::setprecision(2) << 1.00012;

The above example uses mf::LogVerbatim, but the same would work in LOG_DEBUG, mf::LogInfo, mf::LogWarning, and mf::LogError.

Configuring the Message Service

Here is a tutorial for configuring the message facility. Below are some helpful tips for configuring message output.

The .fcl file controlling the job should contain the line

message:      @local::standard_xxx

in the services block of the file, where xxx indicates the message level severity. The various predefined configurations are defined in the source:trunk/Utilities/messageservice.fcl file. Those options are


NB you can only use one of the above in a job .fcl file.

The MessageLogger can be configured to set the number of messages printed and to send each class of message to a different output stream. For example, see the standard_warning configuration of the source:trunk/Utilities/messageservice.fcl file, which is repeated here

standard_warning: { 

 destinations:  #defines the behavior of the configuration
  warningmsg:   #conveniently defined name for a destination, this is not a FHICL key word
   type:      "cout"      #tells the message service to output this destination to cout
   threshold: "WARNING"   #tells the message service that this destination applies to WARNING and higher level messages
   append:    true        #says to append all messages to the output
   categories:            #these correspond to the strings in the mf::LogWarning("") calls in the code
     default:             #defines a default behavior, this is a FHICL key word
      limit:       1000   #limits the total number of messages in this category to 1000
      timespan:    60     #in seconds, tells the message service it can output 1000 messages from this category every minute

     RecoBaseDefaultCtor: #an example of a loud category in LArSoft mf::LogWarning messages
      limit: 10           #limit the total number printed to 10
      timespan: 600       #only print this category every 5 minutes

   } # end categories
  } # end warningmsg
 } # end destinations
} # end standard_warning

If one wanted to add another category in the job control .fcl file, it can be done with these lines after the services block:

services.message.destinations.warningmsg.categories.YourStringHere: { limit: 100 timespan: 60 }

If instead, you want to add some message printing for info level messages you can add the following to the fcl file after the services block:

services.message.destinations.infomsg: {
   type:      "cout"      #tells the message service to output this destination to cout
   threshold: "INFO"      #tells the message service that this destination applies to INFO and higher level messages
   append:    true        #says to append all messages to the output
     default:        {limit: 0}  #don't print anything at the infomsg level except the explicitly named categories
     YourStringHere: {limit: 100 timespan: 60 }
     YourOtherString:{limit: 1   timespan: 1000} 


art::Exception is defined in canvas (./canvas/Utilities/Exception.h).
Doxygen source file

The use of the art::Exception can be configured to skip a module, or skip to the next event, run, etc. Different exception classes can be set to do different things. Since LArSoft is built upon art, the art::Exception is preferred to the cet::exception.

if(x > 2) throw art::Exception("SomeUsefulDescription") << "x = " << x << " is too big";

The default behavior of the framework upon catching an exception is to rethrow (except for some system exceptions, which skip the event). However, one can specify the behavior for exceptions with the following configuration fragment:

services.scheduler.XXXX: [ "SomeUsefulDescription", "SomeOtherUsefulDescription", ... ]
where XXX can be:
  • Rethrow;
  • IgnoreCompletely;
  • SkipEvent;
  • FailModule (behave as if the module returned a failure value for this event);
  • FailPath (behave as if the path rejected this event).

Note that the FailModule setting does not imply a path rejection if the module throwing the exception so configured is a filter set to "VETO" or "IGNORE."

art::Ptr<T> and art::PtrVector<T>

The art::Ptr<T> is a template class that acts like a ROOT TRef. It provides a linkage between objects written into different areas of the event (and output file). For example, the rb::Cluster object holds an art::PtrVector<recob::Hit> pointing to the hits contained in the recob::Cluster. The art::Ptr<T> behaves like a pointer (ie you access the methods of the T using the "->"). It also has functionality to return the actual pointer to the object in question using art::Ptr<T>::get() or to check if the art::Ptr<T> is pointing to a legitimate object using art::Ptr<T>::isNull().

An art::Ptr<T> can only be made from an object that has been stored in the event record and is being fetched from the event record. Put another way, you can only make an art::Ptr<T> if you have an art::Handle to a collection of objects of type T. art::Ptr<T> cannot be instantiated like an object of type T,

// The following line will NOT work
art::Ptr<T> myt = new T(); 

or similar will not work because the object you are interested in for that code has not been first stored in the event record.

Please see this note about art::Ptrs as well.

An art::PtrVector<T> is a vector of art::Ptr<T> objects. It provides the basic functionality you would expect from a std::vector, including iterators, size(), begin() and end() methods.


art::Assns are a way to associate (Assns is a contraction for associations) objects of one type with another. For example, you may want to associate a recob::Cluster with the recob::Hit objects comprising the recob::Cluster. The art:Assns object allows you to navigate these associations bidirectionally, that is you can ask a recob::Cluster which recob::Hits it contains, as well as as a recob::Hit to which recob::Cluster it belongs. The art::Assns also allow us to avoid storing these connections in the higher level object that is being associated. Instead the associations are stored in the event record.

A set of utility functions to perform the necessary steps to associate objects for storage in the art::Event are found in
source:trunk/Utilities/AssociationUtil.h There are also methods to retrieve a collection of objects of one type that are not associated with objects of another type. Those methods allow one to retrieve, for example, all recob::Hits from an event that are not associated with recob::Tracks. The comments in that file describe how to use the functions.

To make use of the associations and retrieve objects from the file, one would do

// below trackListHandle is of type art::Handle< std::vector<recob::Track> >, 
// evt is an art::Event, and trackCreatorModule is an std::string holding the 
// label of the module making the tracks and the associations of hits to tracks
art::FindMany<recob::Hit> fmh(trackListHandle, evt, trackCreatorModule);

// loop over all tracks in the handle and get their hits
for(size_t t = 0; t < trackListHandle->size(); ++t){
   std::vector<const recob::Hit*> hits =;

// can also get a collection of art::Ptrs instead of const pointers
art::FindManyP<recob::Hit> fmh(trackListHandle, evt, trackCreatorModule);

// loop over all tracks in the handle and get their hits
for(size_t t = 0; t < trackListHandle->size(); ++t){
   std::vector<art::Ptr<recob::Hit> > hits =;

One can also use the art::FindOne and art::FindOneP, see the detailed description on how to use art::Assns is here. The art::FindOne returns a cet::maybe_ref, whose interface is defined here. The cet::maybe_ref can be tested for validity, allowing a user to be sure a valid association was created.

NB The art::FindMany(P) and art::FindOne(P) are smart query objects and should only be instantiated once for a given collection. If they are instantiated once for each item in a art::Handle, art::PtrVector, art::View or std::vector< art::Ptr<T> > then a heavy performance price will be paid as a lookup table is made multiple times.

art::RandomNumberGenerator Service

A nice description of how to use this service can be found here. Note that this write up is for the older fw implementation of art, but there are only minor differences.

To store the state of the random number engines for each event one must add the RandomNumberSaver module to the list of physics producers to be run by the job. The necessary line to add to the fcl file is

   rns: { module_type: "RandomNumberSaver" }


A more detailed description of how to use the service in your code and store the state of the generator in the event record and retrieve it later is located here.

Making Objects to Store in the Output

Making objects to store in the art::Event is a straightforward operation. The first step is to declare a container for the objects,

std::unique_ptr<std::vector<mp::MyProd> > mpCollection(new std::vector<mp::MyProd>);

Here we used the std::auto_ptr because it handles the cleanup of the memory for the collection for us.

The mpCollection now behaves just like a std::vector, except one accesses the std::vector methods using the "->" operator. Once the collection has been filled (and it can be a collection of just one object), it is written to the event by doing


Now the ownership of the collection belongs to the event and the user cannot modify the collection or the objects in the collection any more.

Schema Evolution for Data Products

Data products may change over time and we will want to do everything we can to ensure backwards compatibility between old and new versions. Please see this page on the ART wiki for information on how to evolve a data product while still maintaining backwards compatibility.

Making a Module

Below are examples of how to make both an EDProducer module and an EDAnalyzer module. The examples show a basic .cc file for each.

Important Module names may not contain underscores. Underscores are special characters used by the ART system for storing data products in an event to label the products according to what module, instance and process created them. Using underscores in module names will result in your job not running.

The artmod command is useful for generating these files required for plugging a new module (Analyzer, Producer, or Filter) into the framework. The command documents itself:

% artmod --help

Suppose I'm working in a package called "MyPackage" which uses the namespace "mp" and I want to create a producer module called "MyProducer". Moreover, assume that I want to include certain methods from the art::EDProducer base class, like art::EDProducer::beginJob() and art::EDProducer::reconfigure():

% artmod --header-loc MyPackage -e beginJob -e reconfigure producer mp::MyProducer

I can also pass it methods specific to my module using the -e tag as long as the interface is completely defined, ie

% artmod --header-loc MyPackage -e beginJob -e reconfigure -e "bool CheckSomethingOut(double d)" producer mp::MyProducer

Similarly I could pass data members like

% artmod --header-loc MyPackage -e beginJob -e reconfigure -e "bool CheckSomethingOut(double d)" -e "double fMyDouble" producer mp::MyProducer

Read the help screen for more details.

NB Objects are stored in the art::Event as collections of references, not pointers. You need to get them out of the event as collections of references, not pointers.

A word about Algorithm objects

Before describing how to use the ART module objects, it is worth discussing a complementary concept used in LArSoft.

In LArSoft we have the concept of an Algorithm object. Its basic purpose is to take data products in and return either new or modified data products back to the user. In almost all cases, the user will be a EDProducer or EDFilter type module. The reason to have this additional type of object, which is not defined in the ART framework, is to facilitate use of algorithms in multiple modules. That is, a user may write a module that makes use of both Algorithm objects A&B while another may only want to use Algorithm object A. If the algorithm code is written outside of the modules, then the code can be reused in many modules.

The Algorithm objects that are used in modules should have constructors that take a fhicl::ParameterSet const& p as an argument, and they must also define reconfigure(fhicl::ParameterSet const& p) methods. The constructor should call the reconfigure method in order to configure any parameters needed by the Algorithm. The modules owning the Algorithm object should call the Algorithm's reconfigure method from within the module reconfigure method.

Checking the Configuration of a ROOT file

To dump the information regarding all the ParameterSet stored in a given art data file, then the program config_dumper is useful:

$ config_dumper -P file_name.root

Use config_dumper --help to find the available options.