xlinear.internals

Class ColtSparseMatrix

java.lang.Object

xlinear.internals.ColtSparseMatrix

All Implemented Interfaces:

Matrix, SparseMatrix

@Data
public class ColtSparseMatrix
extends java.lang.Object
implements SparseMatrix

Design decision: for first version, use Colt instead of Math Commons sparse matrices, because Math Commons has the artificial restriction that nRows * nCols has to be smaller than Integer.MAX_VALUE (no matter how sparse it is). Colt can hold up to Long.LONG_VALUE (but this is poorly documented), which should be more than enough; for more would need more than int's for rows and cols.

Constructor Summary

Constructors

Constructor and Description
ColtSparseMatrix(cern.colt.matrix.tdouble.DoubleMatrix2D implementation)

Method Summary

Modifier and Type	Method and Description
CholeskyDecomposition	cholesky() Default behavior for views: convert to concrete implementation and then compute the Cholesky
ColtSparseMatrix	createEmpty(int nRows, int nCols)
void	editNonZerosInPlace(MatrixVisitorEditInPlace visitor)
boolean	equals(java.lang.Object obj)
double	get(int row, int col)
cern.colt.matrix.tdouble.DoubleMatrix2D	getImplementation()
int	hashCode()
LUDecomposition	lu() Default behavior for views: convert to concrete implementation and then compute the LU
ColtSparseMatrix	mul(SparseMatrix another)
int	nCols()
int	nRows()
void	set(int row, int col, double v)
SparseMatrix	slice(int row0Incl, int row1Excl, int col0Incl, int col1Excl, boolean readOnly)
java.lang.String	toString()
void	visitNonZeros(MatrixVisitorViewOnly visitor) Efficient traversal of non zero entries.

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface xlinear.SparseMatrix

add, add, addInPlace, addInPlace, col, div, inverse, mul, mul, mulInPlace, nonZeroEntries, operator_divide, operator_minus, operator_minus, operator_minus, operator_multiply, operator_multiply, operator_multiply, operator_multiply, operator_plus, operator_plus, operator_plus, readOnlyView, row, slice, sub, sub, subInPlace, subInPlace, transpose

Methods inherited from interface xlinear.Matrix

add, addInPlace, divInPlace, get, isVector, mul, nEntries, operator_add, operator_add, operator_add, operator_divideAssign, operator_multiplyAssign, operator_remove, operator_remove, operator_remove, set, sub, subInPlace

Constructor Detail

ColtSparseMatrix

public ColtSparseMatrix(cern.colt.matrix.tdouble.DoubleMatrix2D implementation)

Method Detail

visitNonZeros

public void visitNonZeros(MatrixVisitorViewOnly visitor)

Description copied from interface: SparseMatrix

Efficient traversal of non zero entries. Specific order at which these entries are visited is up to the implementation. We assume the visits are done sequentially (refactor at some point with the Stream framework to give choice in future?)

Specified by:

visitNonZeros in interface SparseMatrix

editNonZerosInPlace

public void editNonZerosInPlace(MatrixVisitorEditInPlace visitor)

Specified by:

editNonZerosInPlace in interface SparseMatrix

createEmpty

public ColtSparseMatrix createEmpty(int nRows,
                                    int nCols)

Specified by:

createEmpty in interface Matrix

Specified by:

createEmpty in interface SparseMatrix

slice

public SparseMatrix slice(int row0Incl,
                          int row1Excl,
                          int col0Incl,
                          int col1Excl,
                          boolean readOnly)

Specified by:

slice in interface Matrix

Specified by:

slice in interface SparseMatrix

mul

public ColtSparseMatrix mul(SparseMatrix another)

Specified by:

mul in interface Matrix

Specified by:

mul in interface SparseMatrix

cholesky

public CholeskyDecomposition cholesky()

Description copied from interface: SparseMatrix

Default behavior for views: convert to concrete implementation and then compute the Cholesky

Specified by:

cholesky in interface Matrix

Specified by:

cholesky in interface SparseMatrix

lu

public LUDecomposition lu()

Description copied from interface: SparseMatrix

Default behavior for views: convert to concrete implementation and then compute the LU

Specified by:

lu in interface Matrix

Specified by:

lu in interface SparseMatrix

nRows

public int nRows()

Specified by:

nRows in interface Matrix

nCols

public int nCols()

Specified by:

nCols in interface Matrix

get

public double get(int row,
                  int col)

Specified by:

get in interface Matrix

set

public void set(int row,
                int col,
                double v)

Specified by:

set in interface Matrix

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

hashCode

@Pure
public int hashCode()

Overrides:

hashCode in class java.lang.Object

equals

@Pure
public boolean equals(java.lang.Object obj)

Overrides:

equals in class java.lang.Object

getImplementation

@Pure
public cern.colt.matrix.tdouble.DoubleMatrix2D getImplementation()