Main MRPT website > C++ reference for MRPT 1.4.0
List of all members | Classes | Private Member Functions | Private Attributes
mrpt::math::CSparseMatrix Class Reference

Detailed Description

A sparse matrix structure, wrapping T.

Davis' CSparse library (part of suitesparse) The type of the matrix entries is fixed to "double".

There are two formats for the non-zero entries in this matrix:

The latter is the "normal" format, which is expected by all mathematical operations defined in this class. There're three ways of initializing and populating a sparse matrix:

  1. As a triplet (empty), then add entries, then compress:
    CSparseMatrix SM(100,100);
    SM.insert_entry(i,j, val); // or
    SM.insert_submatrix(i,j, MAT); // ...
    SM.compressFromTriplet();
    CSparseMatrix(const size_t nRows=0, const size_t nCols=0)
    Create an initially empty sparse matrix, in the "triplet" form.
  2. As a triplet from a CSparseMatrixTemplate<double>:
    CSparseMatrixTemplate<double> data;
    data(row,col) = val;
    ...
    CSparseMatrix SM(data);
  3. From an existing dense matrix:

    CMatrixDouble data(100,100); // or
    CMatrixFloat data(100,100); // or
    CMatrixFixedNumeric<double,4,6> data; // etc...
    CSparseMatrix SM(data);
    CMatrixTemplateNumeric< float > CMatrixFloat
    Declares a matrix of float numbers (non serializable).
    CMatrixTemplateNumeric< double > CMatrixDouble
    Declares a matrix of double numbers (non serializable).

Due to its practical utility, there is a special inner class CSparseMatrix::CholeskyDecomp to handle Cholesky-related methods and data.

Note
This class was initially adapted from "robotvision", by Hauke Strasdat, Steven Lovegrove and Andrew J. Davison. See http://www.openslam.org/robotvision.html
CSparse is maintained by Timothy Davis: http://people.sc.fsu.edu/~jburkardt/c_src/csparse/csparse.html .
See also his book "Direct methods for sparse linear systems". http://books.google.es/books?id=TvwiyF8vy3EC&pg=PA12&lpg=PA12&dq=cs_compress&source=bl&ots=od9uGJ793j&sig=Wa-fBk4sZkZv3Y0Op8FNH8PvCUs&hl=es&ei=UjA0TJf-EoSmsQay3aXPAw&sa=X&oi=book_result&ct=result&resnum=8&ved=0CEQQ6AEwBw#v=onepage&q&f=false
See also
mrpt::math::MatrixBlockSparseCols, mrpt::math::CMatrixFixedNumeric, mrpt::math::CMatrixTemplateNumeric, etc.

Definition at line 92 of file CSparseMatrix.h.

#include <mrpt/math/CSparseMatrix.h>

Classes

class  CholeskyDecomp
 Auxiliary class to hold the results of a Cholesky factorization of a sparse matrix. More...
 

Public Member Functions

Constructors, destructor & copy operations


 CSparseMatrix (const size_t nRows=0, const size_t nCols=0)
 Create an initially empty sparse matrix, in the "triplet" form. More...
 
template<typename T >
 CSparseMatrix (const CSparseMatrixTemplate< T > &data)
 A good way to initialize a sparse matrix from a list of non NULL elements. More...
 
template<typename T , size_t N, size_t M>
 CSparseMatrix (const CMatrixFixedNumeric< T, N, M > &MAT)
 Constructor from a dense matrix of any kind existing in MRPT, creating a "column-compressed" sparse matrix. More...
 
template<typename T >
 CSparseMatrix (const CMatrixTemplateNumeric< T > &MAT)
 Constructor from a dense matrix of any kind existing in MRPT, creating a "column-compressed" sparse matrix. More...
 
 CSparseMatrix (const CSparseMatrix &other)
 Copy constructor. More...
 
 CSparseMatrix (const cs *const sm)
 Copy constructor from an existing "cs" CSparse data structure. More...
 
virtual ~CSparseMatrix ()
 Destructor. More...
 
void operator= (const CSparseMatrix &other)
 Copy operator from another existing object. More...
 
void swap (CSparseMatrix &other)
 Fast swap contents with another sparse matrix. More...
 
void clear (const size_t nRows=1, const size_t nCols=1)
 Erase all previous contents and leave the matrix as a "triplet" ROWS x COLS matrix without any nonzero entry. More...
 
Math operations (the interesting stuff...)
void add_AB (const CSparseMatrix &A, const CSparseMatrix &B)
 this = A+B More...
 
void multiply_AB (const CSparseMatrix &A, const CSparseMatrix &B)
 this = A*B More...
 
void multiply_Ab (const mrpt::math::CVectorDouble &b, mrpt::math::CVectorDouble &out_res) const
 out_res = this * b More...
 
CSparseMatrix operator+ (const CSparseMatrix &other) const
 
CSparseMatrix operator* (const CSparseMatrix &other) const
 
mrpt::math::CVectorDouble operator* (const mrpt::math::CVectorDouble &other) const
 
void operator+= (const CSparseMatrix &other)
 
void operator*= (const CSparseMatrix &other)
 
CSparseMatrix transpose () const
 

Private Member Functions

template<class MATRIX >
void construct_from_mrpt_mat (const MATRIX &C)
 Initialization from a dense matrix of any kind existing in MRPT. More...
 
void construct_from_triplet (const cs &triplet)
 Initialization from a triplet "cs", which is first compressed. More...
 
void construct_from_existing_cs (const cs &sm)
 To be called by constructors only, assume previous pointers are trash and overwrite them. More...
 
void internal_free_mem ()
 free buffers (deallocate the memory of the i,p,x buffers) More...
 
void copy (const cs *const sm)
 Copy the data from an existing "cs" CSparse data structure. More...
 
void copy_fast (cs *const sm)
 Fast copy the data from an existing "cs" CSparse data structure, copying the pointers and leaving NULLs in the source structure. More...
 

Private Attributes

cs sparse_matrix
 
void insert_entry (const size_t row, const size_t col, const double val)
 @ Access the matrix, get, set elements, etc. More...
 
void insert_entry_fast (const size_t row, const size_t col, const double val)
 This was an optimized version, but is now equivalent to insert_entry() due to the need to be compatible with unmodified CSparse system libraries. More...
 
template<class MATRIX >
void insert_submatrix (const size_t row, const size_t col, const MATRIX &M)
 ONLY for TRIPLET matrices: insert a given matrix (in any of the MRPT formats) at a given location of the sparse matrix. More...
 
void compressFromTriplet ()
 ONLY for TRIPLET matrices: convert the matrix in a column-compressed form. More...
 
void get_dense (CMatrixDouble &outMat) const
 Return a dense representation of the sparse matrix. More...
 
bool saveToTextFile_dense (const std::string &filName)
 save as a dense matrix to a text file More...
 
bool saveToTextFile_sparse (const std::string &filName)
 Save sparse structure to a text file loadable from MATLAB (can be called on triplet or CCS matrices). More...
 
size_t getRowCount () const
 
size_t getColCount () const
 
void setRowCount (const size_t nRows)
 Change the number of rows in the matrix (can't be lower than current size) More...
 
void setColCount (const size_t nCols)
 
bool isTriplet () const
 Returns true if this sparse matrix is in "triplet" form. More...
 
bool isColumnCompressed () const
 Returns true if this sparse matrix is in "column compressed" form. More...
 
static void cs2dense (const cs &SM, CMatrixDouble &outMat)
 Static method to convert a "cs" structure into a dense representation of the sparse matrix. More...
 

Constructor & Destructor Documentation

◆ CSparseMatrix() [1/6]

mrpt::math::CSparseMatrix::CSparseMatrix ( const size_t  nRows = 0,
const size_t  nCols = 0 
)

Create an initially empty sparse matrix, in the "triplet" form.

Notice that you must call "compressFromTriplet" after populating the matrix and before using the math operatons on this matrix. The initial size can be later on extended with insert_entry() or setRowCount() & setColCount().

See also
insert_entry, setRowCount, setColCount

◆ CSparseMatrix() [2/6]

template<typename T >
mrpt::math::CSparseMatrix::CSparseMatrix ( const CSparseMatrixTemplate< T > &  data)
inline

A good way to initialize a sparse matrix from a list of non NULL elements.

This constructor takes all the non-zero entries in "data" and builds a column-compressed sparse representation.

Definition at line 162 of file CSparseMatrix.h.

References ASSERTMSG_, mrpt::math::CSparseMatrixTemplate< T >::begin(), mrpt::math::CSparseMatrixTemplate< T >::empty(), mrpt::math::CSparseMatrixTemplate< T >::end(), mrpt::math::CSparseMatrixTemplate< T >::getColCount(), mrpt::math::CSparseMatrixTemplate< T >::getRowCount(), insert_entry_fast(), and sparse_matrix.

◆ CSparseMatrix() [3/6]

template<typename T , size_t N, size_t M>
mrpt::math::CSparseMatrix::CSparseMatrix ( const CMatrixFixedNumeric< T, N, M > &  MAT)
inlineexplicit

Constructor from a dense matrix of any kind existing in MRPT, creating a "column-compressed" sparse matrix.

Definition at line 183 of file CSparseMatrix.h.

◆ CSparseMatrix() [4/6]

template<typename T >
mrpt::math::CSparseMatrix::CSparseMatrix ( const CMatrixTemplateNumeric< T > &  MAT)
inlineexplicit

Constructor from a dense matrix of any kind existing in MRPT, creating a "column-compressed" sparse matrix.

Definition at line 186 of file CSparseMatrix.h.

◆ CSparseMatrix() [5/6]

mrpt::math::CSparseMatrix::CSparseMatrix ( const CSparseMatrix other)

Copy constructor.

◆ CSparseMatrix() [6/6]

mrpt::math::CSparseMatrix::CSparseMatrix ( const cs *const  sm)
explicit

Copy constructor from an existing "cs" CSparse data structure.

◆ ~CSparseMatrix()

virtual mrpt::math::CSparseMatrix::~CSparseMatrix ( )
virtual

Destructor.

Member Function Documentation

◆ add_AB()

void mrpt::math::CSparseMatrix::add_AB ( const CSparseMatrix A,
const CSparseMatrix B 
)

this = A+B

Referenced by operator+().

◆ clear()

void mrpt::math::CSparseMatrix::clear ( const size_t  nRows = 1,
const size_t  nCols = 1 
)

Erase all previous contents and leave the matrix as a "triplet" ROWS x COLS matrix without any nonzero entry.

◆ compressFromTriplet()

void mrpt::math::CSparseMatrix::compressFromTriplet ( )

ONLY for TRIPLET matrices: convert the matrix in a column-compressed form.

See also
insert_entry

Referenced by mrpt::graphslam::optimize_graph_spa_levmarq().

◆ construct_from_existing_cs()

void mrpt::math::CSparseMatrix::construct_from_existing_cs ( const cs &  sm)
private

To be called by constructors only, assume previous pointers are trash and overwrite them.

◆ construct_from_mrpt_mat()

template<class MATRIX >
void mrpt::math::CSparseMatrix::construct_from_mrpt_mat ( const MATRIX &  C)
inlineprivate

Initialization from a dense matrix of any kind existing in MRPT.

Definition at line 99 of file CSparseMatrix.h.

References mrpt::system::os::memcpy().

◆ construct_from_triplet()

void mrpt::math::CSparseMatrix::construct_from_triplet ( const cs &  triplet)
private

Initialization from a triplet "cs", which is first compressed.

◆ copy()

void mrpt::math::CSparseMatrix::copy ( const cs *const  sm)
private

Copy the data from an existing "cs" CSparse data structure.

◆ copy_fast()

void mrpt::math::CSparseMatrix::copy_fast ( cs *const  sm)
private

Fast copy the data from an existing "cs" CSparse data structure, copying the pointers and leaving NULLs in the source structure.

◆ cs2dense()

static void mrpt::math::CSparseMatrix::cs2dense ( const cs &  SM,
CMatrixDouble outMat 
)
static

Static method to convert a "cs" structure into a dense representation of the sparse matrix.

◆ get_dense()

void mrpt::math::CSparseMatrix::get_dense ( CMatrixDouble outMat) const

Return a dense representation of the sparse matrix.

See also
saveToTextFile_dense

◆ getColCount()

size_t mrpt::math::CSparseMatrix::getColCount ( ) const
inline

Definition at line 322 of file CSparseMatrix.h.

◆ getRowCount()

size_t mrpt::math::CSparseMatrix::getRowCount ( ) const
inline

Definition at line 321 of file CSparseMatrix.h.

◆ insert_entry()

void mrpt::math::CSparseMatrix::insert_entry ( const size_t  row,
const size_t  col,
const double  val 
)

@ Access the matrix, get, set elements, etc.

ONLY for TRIPLET matrices: insert a new non-zero entry in the matrix. This method cannot be used once the matrix is in column-compressed form. The size of the matrix will be automatically extended if the indices are out of the current limits.

See also
isTriplet, compressFromTriplet

◆ insert_entry_fast()

void mrpt::math::CSparseMatrix::insert_entry_fast ( const size_t  row,
const size_t  col,
const double  val 
)
inline

This was an optimized version, but is now equivalent to insert_entry() due to the need to be compatible with unmodified CSparse system libraries.

Definition at line 254 of file CSparseMatrix.h.

Referenced by CSparseMatrix(), and mrpt::graphslam::optimize_graph_spa_levmarq().

◆ insert_submatrix()

template<class MATRIX >
void mrpt::math::CSparseMatrix::insert_submatrix ( const size_t  row,
const size_t  col,
const MATRIX &  M 
)
inline

ONLY for TRIPLET matrices: insert a given matrix (in any of the MRPT formats) at a given location of the sparse matrix.

This method cannot be used once the matrix is in column-compressed form. The size of the matrix will be automatically extended if the indices are out of the current limits. Since this is inline, it can be very efficient for fixed-size MRPT matrices.

See also
isTriplet, compressFromTriplet, insert_entry

Definition at line 263 of file CSparseMatrix.h.

References THROW_EXCEPTION.

Referenced by mrpt::graphslam::optimize_graph_spa_levmarq().

◆ internal_free_mem()

void mrpt::math::CSparseMatrix::internal_free_mem ( )
private

free buffers (deallocate the memory of the i,p,x buffers)

◆ isColumnCompressed()

bool mrpt::math::CSparseMatrix::isColumnCompressed ( ) const
inline

Returns true if this sparse matrix is in "column compressed" form.

See also
isTriplet

Definition at line 332 of file CSparseMatrix.h.

◆ isTriplet()

bool mrpt::math::CSparseMatrix::isTriplet ( ) const
inline

Returns true if this sparse matrix is in "triplet" form.

See also
isColumnCompressed

Definition at line 329 of file CSparseMatrix.h.

◆ multiply_AB()

void mrpt::math::CSparseMatrix::multiply_AB ( const CSparseMatrix A,
const CSparseMatrix B 
)

this = A*B

Referenced by operator*().

◆ multiply_Ab()

void mrpt::math::CSparseMatrix::multiply_Ab ( const mrpt::math::CVectorDouble b,
mrpt::math::CVectorDouble out_res 
) const

out_res = this * b

◆ operator*() [1/2]

CSparseMatrix mrpt::math::CSparseMatrix::operator* ( const CSparseMatrix other) const
inline

Definition at line 221 of file CSparseMatrix.h.

References multiply_AB().

◆ operator*() [2/2]

mrpt::math::CVectorDouble mrpt::math::CSparseMatrix::operator* ( const mrpt::math::CVectorDouble other) const
inline

Definition at line 227 of file CSparseMatrix.h.

References multiply_Ab().

◆ operator*=()

void mrpt::math::CSparseMatrix::operator*= ( const CSparseMatrix other)
inline

Definition at line 235 of file CSparseMatrix.h.

References multiply_AB().

◆ operator+()

CSparseMatrix mrpt::math::CSparseMatrix::operator+ ( const CSparseMatrix other) const
inline

Definition at line 215 of file CSparseMatrix.h.

References add_AB().

◆ operator+=()

void mrpt::math::CSparseMatrix::operator+= ( const CSparseMatrix other)
inline

Definition at line 232 of file CSparseMatrix.h.

◆ operator=()

void mrpt::math::CSparseMatrix::operator= ( const CSparseMatrix other)

Copy operator from another existing object.

◆ saveToTextFile_dense()

bool mrpt::math::CSparseMatrix::saveToTextFile_dense ( const std::string &  filName)

save as a dense matrix to a text file

Returns
False on any error.

◆ saveToTextFile_sparse()

bool mrpt::math::CSparseMatrix::saveToTextFile_sparse ( const std::string &  filName)

Save sparse structure to a text file loadable from MATLAB (can be called on triplet or CCS matrices).

The format of the text file is:

NUM_ROWS NUM_COLS NUM_NON_ZERO_MAX
row_1 col_1 value_1
row_2 col_2 value_2
...

Instructions for loading from MATLAB in triplet form will be automatically writen to the output file as comments in th first lines:

D=load('file.txt');
SM=spconvert(D(2:end,:));
or, to always preserve the actual matrix size m x n:
m=D(1,1); n=D(1,2); nzmax=D(1,3);
Di=D(2:end,1); Dj=D(2:end,2); Ds=D(2:end,3);
M=sparse(Di,Dj,Ds, m,n, nzmax);
EIGEN_STRONG_INLINE iterator end()
Definition: eigen_plugins.h:27
size_t size(const MATRIXLIKE &m, int dim)
Definition: bits.h:38
Returns
False on any error.

◆ setColCount()

void mrpt::math::CSparseMatrix::setColCount ( const size_t  nCols)
inline

Definition at line 326 of file CSparseMatrix.h.

References ASSERT_.

◆ setRowCount()

void mrpt::math::CSparseMatrix::setRowCount ( const size_t  nRows)
inline

Change the number of rows in the matrix (can't be lower than current size)

Definition at line 325 of file CSparseMatrix.h.

References ASSERT_.

◆ swap()

void mrpt::math::CSparseMatrix::swap ( CSparseMatrix other)

Fast swap contents with another sparse matrix.

◆ transpose()

CSparseMatrix mrpt::math::CSparseMatrix::transpose ( ) const

Member Data Documentation

◆ sparse_matrix

cs mrpt::math::CSparseMatrix::sparse_matrix
private

Definition at line 95 of file CSparseMatrix.h.

Referenced by CSparseMatrix().




Page generated by Doxygen 1.9.2 for MRPT 1.4.0 SVN: at Mon Sep 20 00:47:55 UTC 2021