gtsam  4.0.0
gtsam::VariableIndex Class Reference

Detailed Description

The VariableIndex class computes and stores the block column structure of a factor graph.

The factor graph stores a collection of factors, each of which involves a set of variables. In contrast, the VariableIndex is built from a factor graph prior to elimination, and stores the list of factors that involve each variable. This information is stored as a deque of lists of factor indices.

Standard Constructors

 VariableIndex ()
 Default constructor, creates an empty VariableIndex.
template<class FG >
 VariableIndex (const FG &factorGraph)
 Create a VariableIndex that computes and stores the block column structure of a factor graph.

Standard Interface

size_t size () const
 The number of variable entries. This is equal to the number of unique variable Keys.
size_t nFactors () const
 The number of factors in the original factor graph.
size_t nEntries () const
 The number of nonzero blocks, i.e. the number of variable-factor entries.
const Factors & operator[] (Key variable) const
 Access a list of factors by variable.


bool equals (const VariableIndex &other, double tol=0.0) const
 Test for equality (for unit tests and debug assertions).
void print (const std::string &str="VariableIndex: ", const KeyFormatter &keyFormatter=DefaultKeyFormatter) const
 Print the variable index (for unit tests and debugging).
void outputMetisFormat (std::ostream &os) const
 Output dual hypergraph to Metis file format for use with hmetis In the dual graph, variables are hyperedges, factors are nodes.

Advanced Interface

template<class FG >
void augment (const FG &factors, boost::optional< const FactorIndices & > newFactorIndices=boost::none)
 Augment the variable index with new factors. More...
template<typename ITERATOR , class FG >
void remove (ITERATOR firstFactor, ITERATOR lastFactor, const FG &factors)
 Remove entries corresponding to the specified factors. More...
template<typename ITERATOR >
void removeUnusedVariables (ITERATOR firstKey, ITERATOR lastKey)
 Remove unused empty variables (in debug mode verifies they are empty).
const_iterator begin () const
 Iterator to the first variable entry.
const_iterator end () const
 Iterator to the first variable entry.
const_iterator find (Key key) const
 Find the iterator for the requested variable entry.
Factor_iterator factorsBegin (Key variable)
Factor_iterator factorsEnd (Key variable)
Factor_const_iterator factorsBegin (Key variable) const
Factor_const_iterator factorsEnd (Key variable) const
const Factors & internalAt (Key variable) const
 Internal version of 'at' that asserts existence.
Factors & internalAt (Key variable)
 Internal version of 'at' that asserts existence.

Public Types

typedef boost::shared_ptr< VariableIndexshared_ptr
typedef FactorIndices Factors
typedef Factors::iterator Factor_iterator
typedef Factors::const_iterator Factor_const_iterator
typedef KeyMap::const_iterator const_iterator
typedef KeyMap::const_iterator iterator
typedef KeyMap::value_type value_type

Protected Types

typedef FastMap< Key, Factors > KeyMap

Protected Attributes

KeyMap index_
size_t nFactors_
size_t nEntries_

Member Function Documentation

◆ augment()

template<class FG >
void gtsam::VariableIndex::augment ( const FG &  factors,
boost::optional< const FactorIndices & >  newFactorIndices = boost::none 

Augment the variable index with new factors.

This can be used when solving problems incrementally.

◆ remove()

template<typename ITERATOR , class FG >
void gtsam::VariableIndex::remove ( ITERATOR  firstFactor,
ITERATOR  lastFactor,
const FG &  factors 

Remove entries corresponding to the specified factors.

NOTE: We intentionally do not decrement nFactors_ because the factor indices need to remain consistent. Removing factors from a factor graph does not shift the indices of other factors. Also, we keep nFactors_ one greater than the highest-numbered factor referenced in a VariableIndex.

indicesThe indices of the factors to remove, which must match factors
factorsThe factors being removed, which must symbolically correspond exactly to the factors with the specified indices that were added.

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