/*
* File: SparseRowMatrix.java
* Authors: Kevin R. Dixon
* Company: Sandia National Laboratories
* Project: Cognitive Foundry
*
* Copyright February 21, 2006, Sandia Corporation. Under the terms of Contract
* DE-AC04-94AL85000, there is a non-exclusive license for use of this work by
* or on behalf of the U.S. Government. Export of this program may require a
* license from the United States Government. See CopyrightHistory.txt for
* complete details.
*
*/
package gov.sandia.cognition.math.matrix.mtj;
import gov.sandia.cognition.annotation.CodeReview;
import gov.sandia.cognition.math.matrix.Matrix;
import gov.sandia.cognition.math.matrix.MatrixEntry;
import gov.sandia.cognition.math.matrix.MatrixFactory;
import java.io.IOException;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import no.uib.cipr.matrix.sparse.FlexCompRowMatrix;
/**
* A sparse matrix, represented as a collection of sparse row vectors.
* Generally, this is the fastest sparse matrix for premultiplying against
* a vector.
*
* @author Kevin R. Dixon
* @since 1.0
*/
@CodeReview(
reviewer="Justin Basilico",
date="2006-07-27",
changesNeeded=false,
comments="Looks good."
)
public class SparseRowMatrix
extends AbstractSparseMatrix
{
/**
* Creates a new empty instance of SparseRowMatrix.
*
* @param numRows Number of rows in the matrix.
* @param numColumns Number of columns in the matrix.
*/
protected SparseRowMatrix(
int numRows,
int numColumns )
{
this( new no.uib.cipr.matrix.sparse.FlexCompRowMatrix(
numRows, numColumns ) );
if (numRows < 0)
{
throw new IllegalArgumentException( "Num rows must be >= 0" );
}
if (numColumns < 0)
{
throw new IllegalArgumentException( "Num columns must be >= 0" );
}
}
/**
* Copy constructor for SparseRowMatrix matrices
*
* @param matrix Matrix from which to copy the internal MTJ matrix.
*/
protected SparseRowMatrix(
SparseRowMatrix matrix )
{
this( (FlexCompRowMatrix) matrix.getInternalMatrix().copy() );
}
/**
* Copy constructor for general matrices, copies over nonzero values.
*
* @param matrix Matrix from which to copy the nonzero elements into this.
*/
protected SparseRowMatrix(
Matrix matrix )
{
this( matrix.getNumRows(), matrix.getNumColumns() );
for (MatrixEntry e : matrix)
{
double value = e.getValue();
if (value != 0.0)
{
this.setElement( e.getRowIndex(), e.getColumnIndex(), value );
}
}
}
/**
* Creates a SparseRowMatrix based on the appropriate MTJ matrix,
* does NOT create a copy of internalMatrix.
*
* @param internalMatrix New internal matrix for this, no copy made.
*/
protected SparseRowMatrix(
FlexCompRowMatrix internalMatrix )
{
super( internalMatrix );
}
@Override
public FlexCompRowMatrix getInternalMatrix()
{
return (FlexCompRowMatrix) super.getInternalMatrix();
}
public SparseColumnMatrix transpose()
{
SparseColumnMatrix result = new SparseColumnMatrix(
this.getNumColumns(), this.getNumRows() );
this.transposeInto( result );
return result;
}
public SparseRowMatrix times(
final AbstractMTJMatrix matrix )
{
int returnRows = this.getNumRows();
int returnColumns = matrix.getNumColumns();
SparseRowMatrix result = new SparseRowMatrix(
returnRows, returnColumns );
this.timesInto( matrix, result );
return result;
}
public SparseRowMatrix getSubMatrix(
int minRow,
int maxRow,
int minColumn,
int maxColumn )
{
if (minRow > maxRow)
{
throw new IllegalArgumentException( "minRow > maxRow" );
}
if (minColumn > maxColumn)
{
throw new IllegalArgumentException( "minColumn > maxColumn" );
}
int numRows = maxRow - minRow + 1;
int numColumns = maxColumn - minColumn + 1;
SparseRowMatrix submatrix = new SparseRowMatrix( numRows, numColumns );
this.getSubMatrixInto(
minRow, maxRow, minColumn, maxColumn, submatrix );
return submatrix;
}
/**
* Gets the specified row of the matrix, using MTJ's internal routine
* to speed things up.
*
* @param rowIndex Zero-based row index.
* @return SparseVector with the same dimensions as this has columns.
*/
@Override
public SparseVector getRow(
int rowIndex )
{
return new SparseVector( this.getInternalMatrix().getRow( rowIndex ) );
}
/**
* Sets the specified row of the matrix using rowVector, using MTJ's
* internal routine to speed things up.
*
* @param rowIndex Zero-based row index.
* @param rowVector SparseVector containing the elements to replace in
* this.
*/
public void setRow(
int rowIndex,
SparseVector rowVector )
{
this.getInternalMatrix().setRow(
rowIndex, rowVector.getInternalVector() );
}
public void compact()
{
this.getInternalMatrix().compact();
}
@Override
public int getEntryCount()
{
final FlexCompRowMatrix m = this.getInternalMatrix();
final int rowCount = this.getNumRows();
int result = 0;
for (int i = 0; i < rowCount; i++)
{
result += m.getRow(i).getUsed();
}
return result;
}
/**
* Custom deserialization is needed.
*
* @param in The ObjectInputStream to deserialize from.
* @throws java.io.IOException If there is an error with the stream.
* @throws java.lang.ClassNotFoundException If a class used by this one
* cannot be found.
*/
private void readObject(
ObjectInputStream in )
throws IOException, ClassNotFoundException
{
in.defaultReadObject();
int numRows = in.readInt();
int numCols = in.readInt();
this.setInternalMatrix( new FlexCompRowMatrix( numRows, numCols ) );
for (int i = 0; i < numRows; i++)
{
SparseVector row = (SparseVector) in.readObject();
this.setRow( i, row );
}
this.compact();
}
/**
* Custom serialization is needed.
*
* @param out The ObjectOutputStream to write this object to.
* @throws java.io.IOException If there is an error writing to the stream.
*/
private void writeObject(
ObjectOutputStream out )
throws IOException
{
this.compact();
out.defaultWriteObject();
int numRows = this.getNumRows();
int numCols = this.getNumColumns();
out.writeInt( numRows );
out.writeInt( numCols );
for (int i = 0; i < numRows; i++)
{
out.writeObject( this.getRow( i ) );
}
}
@Override
public MatrixFactory<?> getMatrixFactory()
{
return SparseMatrixFactoryMTJ.INSTANCE;
}
}