Clusterizer< ACellHolder, ACluster > Class Template Reference

Implementation of the clustering Clusters elements of a given collection using the relations presented by a neighorhood. More...

#include <Clusterizer.h>

Inheritance diagram for Clusterizer< ACellHolder, ACluster >:

Collaboration diagram for Clusterizer< ACellHolder, ACluster >:

Public Types
using	IOTypes = std::tuple< AIOTypes... >
	Types that should be served to apply on invokation.
using	IOVectors = std::tuple< std::vector< AIOTypes >... >
	Vector types that should be served to apply on invokation.

Public Member Functions
void	apply (std::vector< ACellHolder * > const &cellHolders, std::vector< WeightedRelation< ACellHolder > > const &cellHolderRelations, std::vector< ACluster > &clusters) override
	Creates the clusters. More...
virtual std::string	getDescription ()
	Brief description of the purpose of the concret findlet.
virtual void	exposeParameters (ModuleParamList *moduleParamList, const std::string &prefix)
	Forward prefixed parameters of this findlet to the module parameter list.
virtual void	apply (ToVector< AIOTypes > &... ioVectors)=0
	Main function executing the algorithm.
void	initialize () override
	Receive and dispatch signal before the start of the event processing.
void	beginRun () override
	Receive and dispatch signal for the beginning of a new run.
void	beginEvent () override
	Receive and dispatch signal for the start of a new event.
void	endRun () override
	Receive and dispatch signal for the end of the run.
void	terminate () override
	Receive and dispatch Signal for termination of the event processing.

Protected Types
using	ToVector = typename ToVectorImpl< T >::Type
	Short hand for ToRangeImpl.

Protected Member Functions
void	addProcessingSignalListener (ProcessingSignalListener *psl)
	Register a processing signal listener to be notified.
int	getNProcessingSignalListener ()
	Get the number of currently registered listeners.

Private Types
using	Super = ProcessingSignalListener
	Type of the base class.

Private Member Functions
void	expandCluster (std::vector< WeightedRelation< ACellHolder >> const &cellHolderRelations, std::vector< ACellHolder * > &cluster) const
	Helper function. Starting a new cluster and iterativelly expands it.
void	setCellState (ACellHolder *cellHolder, Weight cellState) const
	Setter for the cell state of a pointed object that holds an AutomatonCell.
Weight	getCellState (ACellHolder *cellHolder) const
	Getter for the cell state of a pointed object that holds an AutomatonCell.
void	setCellWeight (ACellHolder *cellHolder, Weight cellWeight) const
	Setter for the cell weight of a pointed object that holds an AutomatonCell.

Private Attributes
std::vector< ProcessingSignalListener * >	m_subordinaryProcessingSignalListeners
	References to subordinary signal processing listener contained in this findlet.
bool	m_initialized = false
	Flag to keep track whether initialization happend before.
bool	m_terminated = false
	Flag to keep track whether termination happend before.
std::string	m_initializedAs
	Name of the type during initialisation.

Detailed Description

template<class ACellHolder, class ACluster = std::vector<ACellHolder*>>
class Belle2::TrackFindingCDC::Clusterizer< ACellHolder, ACluster >

Implementation of the clustering Clusters elements of a given collection using the relations presented by a neighorhood.

A cluster is essentially a connected subset of all cells that can reach each other by one more relations in a neighborhood. The algorithm is essentially an iterative expansion of the neighborhood relations keeping track of the already used cells by using the CellState of the AutomatonCell.

• ACellHolder must therefore provide an AutomatonCell accessable by a getAutomatonCell() method. In case the objects you what to cluster do not contain an automaton cell already you may adopt it by using the WithAutomatonCell mixin.
• ACluster can be anything that is default constructable and supports .insert(end(), ACellHolder*).

Definition at line 40 of file Clusterizer.h.

Member Function Documentation

apply()

void apply ( std::vector< ACellHolder * > const &  cellHolders, std::vector< WeightedRelation< ACellHolder > > const &  cellHolderRelations, std::vector< ACluster > &  clusters  )

inlineoverride

Creates the clusters.

Repeatly expands a neighborhood of referenced objects that have an AutomatonCell. The CellState after the clusterization is the index of the generated cluster. The CellWeight is set to the total number of neighbors of each cell.

Parameters

	cellHolders	Pointers to objects that should be clustered.
	cellHolderRelations	Relations between the objects that should be clustered
[out]	clusters	Groups of connected objects in the neighborhood.

Definition at line 54 of file Clusterizer.h.

      {
        // Expect the relations to be sorted for lookup
        assert(std::is_sorted(cellHolderRelations.begin(),
                              cellHolderRelations.end()));
 
        // Prepare some output clusters
        clusters.reserve(30);
 
        // Prepare states
        for (ACellHolder* cellHolder : cellHolders) {
          setCellState(cellHolder, -1);
        }
 
        // Work horse cluster
        std::vector<ACellHolder*> cluster;
 
        // Go through each cell holder and start a cluster on each that is unassigned yet
        int iCluster = -1;
        for (ACellHolder* cellHolder : cellHolders) {
          if (getCellState(cellHolder) != -1) continue;
 
          cluster.clear();
 
          ++iCluster;
          setCellState(cellHolder, iCluster);
          cluster.push_back(cellHolder);
 
          expandCluster(cellHolderRelations, cluster);
 
          clusters.emplace_back(std::move(cluster));
          cluster.clear();
        }
      }

---

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

• tracking/trackFindingCDC/ca/include/Clusterizer.h