org.nuiton.math.matrix

Class MatrixFactory

java.lang.Object

org.nuiton.math.matrix.MatrixFactory

public class MatrixFactory
extends Object

Cette classe permet de creer des matrices, toutes les creations de matrice doivent etre faite a travers cette classe. Cette classe permet de modifier la representation interne des matrices de facon simple.

Created: 11 octobre 2005 20:15:20 CEST

Version:

$Revision: 503 $ Last update: $Date: 2014-07-11 13:20:52 +0200 (Fri, 11 Jul 2014) $ by : $Author: bpoussin $

Author:

Benjamin POUSSIN

Field Summary

Fields

Modifier and Type	Field and Description
protected static SemanticMapper	defaultSemanticMapper
protected static Class<?>	defaultSparseVectorClass
protected static int	defaultThresholdSparse if we try to allocate matrix with more that this threshold, sparse matrix is used
protected static boolean	defaultUseLazyVector If true, createVector return all time LazyVector to prevent memory allocation when not necessary.
protected static Class<?>	defaultVectorClass Valeur par defaut si aucun type de Vector n'est donné
protected Class<?>	sparseVectorClass
protected int	thresholdSparse
protected boolean	useLazyVector
protected Class<?>	vectorClass

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	MatrixFactory(Class<?> vectorClass, Class<?> sparseVectorClass, int thresholdSparse, boolean useLazyVector)

Method Summary

Methods

Modifier and Type	Method and Description
MatrixND	create(double[] values, int[] dim) Convert a double array into matrix.
MatrixND	create(int[] dim)
MatrixND	create(List<?>[] semantics)
MatrixND	create(MatrixND matrix) Create new matrix by copying matrix in argument
MatrixND	create(MatrixND matrix, Vector data) Create new matrix by copying matrix in argument, and force backend to be data.
MatrixND	create(String name, int[] dim)
MatrixND	create(String name, int[] dim, String[] dimNames)
MatrixND	create(String name, int[] dim, String[] dimNames, Vector data) Create new matrix and force backend to be data.
MatrixND	create(String name, List<?>[] semantics)
MatrixND	create(String name, List<?>[] semantics, String[] dimNames)
MatrixND	create(String name, List<?>[] semantics, String[] dimNames, Vector data) Create new matrix and force backend to be data.
MatrixProxy	createProxy(List<?>[] semantics, MatrixProvider matrixProvider)
MatrixProxy	createProxy(String name, int[] dim, MatrixProvider matrixProvider)
MatrixProxy	createProxy(String name, int[] dim, String[] dimNames, MatrixProvider matrixProvider)
MatrixProxy	createProxy(String name, List<?>[] semantics, MatrixProvider matrixProvider)
MatrixProxy	createProxy(String name, List<?>[] semantics, String[] dimNames, MatrixProvider matrixProvider)
protected Vector	createVector(int length) cree un nouveau Vector.
static Class<?>	getDefaultVectorClass()
static MatrixFactory	getInstance() Utilise par defaut defaultVectorClass.
static MatrixFactory	getInstance(Class<?> vectorClass) Retourne une factory utilisant vectorClass comme classe de base a l'implantation des matrices, aussi bien pour les matrices dense que les matrices creuse.
static MatrixFactory	getInstance(Class<?> vectorClass, Class<?> sparseVectorClass, int thresholdSparse) Retourne une factory utilisant, par defaut on utilise les LazyVector
static MatrixFactory	getInstance(Class<?> vectorClass, Class<?> sparseVectorClass, int thresholdSparse, boolean useLazyVector) Retourne une factory utilisant vectorClass comme classe de base a l'implantation des matrices.
static SemanticMapper	getSemanticMapper()
Class<?>	getSparseVectorClass()
int	getThresholdSparse()
Class<?>	getVectorClass()
static void	initMatrixFactoryThreadLocal(Class<?> vectorClass, Class<?> sparseVectorClass, int thresholdSparse, boolean useLazyVector) Initialise une factory pour le thread courant avec le vectorClass passe en parametre.
boolean	isUseLazyVector()
MatrixND	matrixId(int size) Crée une nouvelle matrice identité.
static void	removeMatrixFactoryThreadLocal() Retire la factory pour le thread courant
static void	setDefaultSparseVectorClass(Class<?> defaultSparseVectorClass)
static void	setDefaultThresholdSparse(int defaultThresholdSparse)
static void	setDefaultUseLazyVector(boolean useLazyVector)
static void	setDefaultVectorClass(Class<?> vectorClass)
static void	setSemanticMapper(SemanticMapper semanticMapper)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

defaultUseLazyVector

protected static boolean defaultUseLazyVector

If true, createVector return all time LazyVector to prevent memory allocation when not necessary. LazyVector init real vector only when necessary

defaultVectorClass

protected static Class<?> defaultVectorClass

Valeur par defaut si aucun type de Vector n'est donné

defaultSparseVectorClass

protected static Class<?> defaultSparseVectorClass

defaultThresholdSparse

protected static int defaultThresholdSparse

if we try to allocate matrix with more that this threshold, sparse matrix is used

vectorClass

protected Class<?> vectorClass

sparseVectorClass

protected Class<?> sparseVectorClass

thresholdSparse

protected int thresholdSparse

useLazyVector

protected boolean useLazyVector

defaultSemanticMapper

protected static SemanticMapper defaultSemanticMapper

Constructor Detail

MatrixFactory

protected MatrixFactory(Class<?> vectorClass,
             Class<?> sparseVectorClass,
             int thresholdSparse,
             boolean useLazyVector)

Method Detail

getVectorClass

public Class<?> getVectorClass()

getSparseVectorClass

public Class<?> getSparseVectorClass()

getThresholdSparse

public int getThresholdSparse()

isUseLazyVector

public boolean isUseLazyVector()

setDefaultVectorClass

public static void setDefaultVectorClass(Class<?> vectorClass)

setDefaultSparseVectorClass

public static void setDefaultSparseVectorClass(Class<?> defaultSparseVectorClass)

setDefaultThresholdSparse

public static void setDefaultThresholdSparse(int defaultThresholdSparse)

setDefaultUseLazyVector

public static void setDefaultUseLazyVector(boolean useLazyVector)

getDefaultVectorClass

public static Class<?> getDefaultVectorClass()

setSemanticMapper

public static void setSemanticMapper(SemanticMapper semanticMapper)

getSemanticMapper

public static SemanticMapper getSemanticMapper()

getInstance

public static MatrixFactory getInstance(Class<?> vectorClass)

Retourne une factory utilisant vectorClass comme classe de base a l'implantation des matrices, aussi bien pour les matrices dense que les matrices creuse. Par defaut on utilise les LazyVector

Parameters:

vectorClass - vector class implementation

Returns:

factory

getInstance

public static MatrixFactory getInstance(Class<?> vectorClass,
                        Class<?> sparseVectorClass,
                        int thresholdSparse)

Retourne une factory utilisant, par defaut on utilise les LazyVector

Parameters:

vectorClass - vector class implementation for dense matrix

sparseVectorClass - vector class implementation for sparse matrix

thresholdSparse - threshold to determine usage of dense or sparse matrix

Returns:

factory

getInstance

public static MatrixFactory getInstance(Class<?> vectorClass,
                        Class<?> sparseVectorClass,
                        int thresholdSparse,
                        boolean useLazyVector)

Retourne une factory utilisant vectorClass comme classe de base a l'implantation des matrices.

Parameters:

vectorClass - vector class implementation for dense matrix

sparseVectorClass - vector class implementation for sparse matrix

thresholdSparse - threshold to determine usage of dense or sparse matrix

useLazyVector - if true use LazyVector

Returns:

factory

getInstance

public static MatrixFactory getInstance()

Utilise par defaut defaultVectorClass. Si une factory a ete initialisee pour le thread local, alors celle-ci sera reutilise, sinon une nouvelle est cree avec le backend par defaut

Returns:

factory

initMatrixFactoryThreadLocal

public static void initMatrixFactoryThreadLocal(Class<?> vectorClass,
                                Class<?> sparseVectorClass,
                                int thresholdSparse,
                                boolean useLazyVector)

Initialise une factory pour le thread courant avec le vectorClass passe en parametre.

Parameters:

vectorClass - vector class implementation for dense matrix

sparseVectorClass - vector class implementation for sparse matrix

thresholdSparse - threshold to determine usage of dense or sparse matrix

useLazyVector - if true use LazyVector

removeMatrixFactoryThreadLocal

public static void removeMatrixFactoryThreadLocal()

Retire la factory pour le thread courant

create

public MatrixND create(int[] dim)

create

public MatrixND create(double[] values,
              int[] dim)

Convert a double array into matrix.

Parameters:

values - The values to fill the matrix

dim - An array representing the dimensions of the matrix

Returns:

a 2D matrix filled with the values, null if the dimension is more than 2

create

public MatrixND create(List<?>[] semantics)

create

public MatrixND create(String name,
              int[] dim)

create

public MatrixND create(String name,
              int[] dim,
              String[] dimNames)

create

public MatrixND create(String name,
              int[] dim,
              String[] dimNames,
              Vector data)

Create new matrix and force backend to be data.

Parameters:

name - matrix's name (can be null)

dim - matrix's dimension (must be provided)

dimNames - dimensions' name (can be null)

data - backend used to new matrix

Returns:

new Matrix with specified backend

create

public MatrixND create(String name,
              List<?>[] semantics)

create

public MatrixND create(String name,
              List<?>[] semantics,
              String[] dimNames)

create

public MatrixND create(String name,
              List<?>[] semantics,
              String[] dimNames,
              Vector data)

Create new matrix and force backend to be data.

Parameters:

name - matrix's name (can be null)

semantics - semantiques (must be provided)

dimNames - dimensions' name (can be null)

data - backend used to new matrix

Returns:

new Matrix with specified backend

create

public MatrixND create(MatrixND matrix)

Create new matrix by copying matrix in argument

Parameters:

matrix -

Returns:

create

public MatrixND create(MatrixND matrix,
              Vector data)

Create new matrix by copying matrix in argument, and force backend to be data.

Parameters:

matrix - matrix to copied

data - backend used to new matrix

matrixId

public MatrixND matrixId(int size)

Crée une nouvelle matrice identité. Une matrice identité est une matrice à 2 dimensions dont tous les éléments de la diagonal vaut 1

Parameters:

size - la taille de la matrice

Returns:

une nouvelle matrice identité

createVector

protected Vector createVector(int length)

cree un nouveau Vector. Si la taille demandee est inferieur au seuil alors on cree une matrice plaine, sinon on cree une matrice creuse.

Parameters:

length -

Returns:

createProxy

public MatrixProxy createProxy(List<?>[] semantics,
                      MatrixProvider matrixProvider)

createProxy

public MatrixProxy createProxy(String name,
                      int[] dim,
                      MatrixProvider matrixProvider)

createProxy

public MatrixProxy createProxy(String name,
                      int[] dim,
                      String[] dimNames,
                      MatrixProvider matrixProvider)

createProxy

public MatrixProxy createProxy(String name,
                      List<?>[] semantics,
                      MatrixProvider matrixProvider)

createProxy

public MatrixProxy createProxy(String name,
                      List<?>[] semantics,
                      String[] dimNames,
                      MatrixProvider matrixProvider)