/*
* Copyright (C) 2008-2015 by Holger Arndt
*
* This file is part of the Universal Java Matrix Package (UJMP).
* See the NOTICE file distributed with this work for additional
* information regarding copyright ownership and licensing.
*
* UJMP is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* UJMP is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with UJMP; if not, write to the
* Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
* Boston, MA 02110-1301 USA
*/
package org.ujmp.colt;
import java.util.HashSet;
import java.util.Iterator;
import java.util.Set;
import org.ujmp.core.Coordinates;
import org.ujmp.core.Matrix;
import org.ujmp.core.doublematrix.stub.AbstractSparseDoubleMatrix2D;
import org.ujmp.core.interfaces.Wrapper;
import org.ujmp.core.mapmatrix.MapMatrix;
import org.ujmp.core.util.CoordinateSetToLongWrapper;
import org.ujmp.core.util.MathUtil;
import cern.colt.function.IntIntDoubleFunction;
import cern.colt.matrix.DoubleFactory2D;
import cern.colt.matrix.DoubleMatrix2D;
import cern.colt.matrix.impl.DenseDoubleMatrix2D;
import cern.colt.matrix.impl.SparseDoubleMatrix2D;
import cern.colt.matrix.linalg.Algebra;
import cern.colt.matrix.linalg.CholeskyDecomposition;
import cern.colt.matrix.linalg.EigenvalueDecomposition;
import cern.colt.matrix.linalg.LUDecomposition;
import cern.colt.matrix.linalg.QRDecomposition;
import cern.colt.matrix.linalg.SingularValueDecomposition;
import cern.jet.math.Functions;
public class ColtSparseDoubleMatrix2D extends AbstractSparseDoubleMatrix2D implements Wrapper<SparseDoubleMatrix2D> {
private static final long serialVersionUID = -3223474248020842822L;
public static final ColtSparseDoubleMatrix2DFactory Factory = new ColtSparseDoubleMatrix2DFactory();
private final SparseDoubleMatrix2D matrix;
public ColtSparseDoubleMatrix2D(int rows, int columns) {
super(rows, columns);
this.matrix = new SparseDoubleMatrix2D(rows, columns);
}
public ColtSparseDoubleMatrix2D(SparseDoubleMatrix2D m) {
super(m.rows(), m.columns());
this.matrix = m;
}
public ColtSparseDoubleMatrix2D(Matrix source) {
super(source.getRowCount(), source.getColumnCount());
this.matrix = new SparseDoubleMatrix2D((int) source.getRowCount(), (int) source.getColumnCount());
for (long[] c : source.availableCoordinates()) {
setDouble(source.getAsDouble(c), c);
}
if (source.getMetaData() != null) {
setMetaData(source.getMetaData().clone());
}
}
public final void clear() {
matrix.forEachNonZero(new IntIntDoubleFunction() {
public double apply(int arg0, int arg1, double arg2) {
return 0;
}
});
}
public double getDouble(long row, long column) {
return matrix.getQuick(MathUtil.longToInt(row), MathUtil.longToInt(column));
}
public double getDouble(int row, int column) {
return matrix.getQuick(row, column);
}
// must be synchronized for sparse matrix implementation
public synchronized void setDouble(double value, long row, long column) {
matrix.setQuick(MathUtil.longToInt(row), MathUtil.longToInt(column), value);
}
// must be synchronized for sparse matrix implementation
public synchronized void setDouble(double value, int row, int column) {
matrix.setQuick(row, column, value);
}
public SparseDoubleMatrix2D getWrappedObject() {
return matrix;
}
public Matrix inv() {
return new ColtDenseDoubleMatrix2D((DenseDoubleMatrix2D) new Algebra().inverse(matrix));
}
public Iterable<long[]> availableCoordinates() {
return new AvailableCoordinateIterable();
}
class AvailableCoordinateIterable implements Iterable<long[]> {
// TODO: optimize
public Iterator<long[]> iterator() {
Set<Coordinates> cset = new HashSet<Coordinates>();
for (long r = getRowCount() - 1; r >= 0; r--) {
for (long c = getColumnCount() - 1; c >= 0; c--) {
if (getDouble(r, c) != 0.0) {
cset.add(Coordinates.wrap(r, c).clone());
}
}
}
return new CoordinateSetToLongWrapper(cset).iterator();
}
}
public final boolean containsCoordinates(long... coordinates) {
return getAsDouble(coordinates) != 0.0;
}
public Matrix transpose() {
return new ColtSparseDoubleMatrix2D((SparseDoubleMatrix2D) matrix.viewDice().copy());
}
public Matrix plus(double value) {
Matrix result = new ColtSparseDoubleMatrix2D((SparseDoubleMatrix2D) matrix.copy().assign(Functions.plus(value)));
MapMatrix<String, Object> a = getMetaData();
if (a != null) {
result.setMetaData(a.clone());
}
return result;
}
public Matrix[] qr() {
if (getColumnCount() > getRowCount()) {
throw new RuntimeException("matrix size must be m>=n");
}
QRDecomposition qr = new QRDecomposition(matrix);
Matrix q = new ColtDenseDoubleMatrix2D(qr.getQ());
Matrix r = new ColtDenseDoubleMatrix2D(qr.getR());
return new Matrix[] { q, r };
}
public Matrix[] svd() {
if (getColumnCount() > getRowCount()) {
SingularValueDecomposition svd = new SingularValueDecomposition(matrix.viewDice());
Matrix u = new ColtDenseDoubleMatrix2D(svd.getV());
Matrix s = new ColtDenseDoubleMatrix2D(svd.getS());
Matrix v = new ColtDenseDoubleMatrix2D(svd.getU());
return new Matrix[] { u, s, v };
} else {
SingularValueDecomposition svd = new SingularValueDecomposition(matrix);
Matrix u = new ColtDenseDoubleMatrix2D(svd.getU());
Matrix s = new ColtDenseDoubleMatrix2D(svd.getS());
Matrix v = new ColtDenseDoubleMatrix2D(svd.getV());
return new Matrix[] { u, s, v };
}
}
public Matrix times(double value) {
Matrix result = new ColtSparseDoubleMatrix2D((SparseDoubleMatrix2D) matrix.copy().assign(Functions.mult(value)));
MapMatrix<String, Object> a = getMetaData();
if (a != null) {
result.setMetaData(a.clone());
}
return result;
}
public Matrix copy() {
Matrix m = new ColtSparseDoubleMatrix2D((SparseDoubleMatrix2D) matrix.copy());
if (getMetaData() != null) {
m.setMetaData(getMetaData().clone());
}
return m;
}
public Matrix chol() {
CholeskyDecomposition chol = new CholeskyDecomposition(matrix);
Matrix r = new ColtDenseDoubleMatrix2D(chol.getL());
return r;
}
public Matrix divide(double value) {
Matrix result = new ColtSparseDoubleMatrix2D((SparseDoubleMatrix2D) matrix.copy().assign(Functions.div(value)));
MapMatrix<String, Object> a = getMetaData();
if (a != null) {
result.setMetaData(a.clone());
}
return result;
}
public Matrix minus(double value) {
Matrix result = new ColtSparseDoubleMatrix2D((SparseDoubleMatrix2D) matrix.copy()
.assign(Functions.minus(value)));
MapMatrix<String, Object> a = getMetaData();
if (a != null) {
result.setMetaData(a.clone());
}
return result;
}
public Matrix solve(Matrix b) {
if (b instanceof ColtDenseDoubleMatrix2D) {
ColtDenseDoubleMatrix2D b2 = (ColtDenseDoubleMatrix2D) b;
if (isSquare()) {
DoubleMatrix2D ret = new LUDecomposition(matrix).solve(b2.getWrappedObject());
return new ColtDenseDoubleMatrix2D(ret);
} else {
DoubleMatrix2D ret = new QRDecomposition(matrix).solve(b2.getWrappedObject());
return new ColtDenseDoubleMatrix2D(ret);
}
} else {
return super.solve(b);
}
}
public Matrix solveSPD(Matrix b) {
if (b instanceof ColtDenseDoubleMatrix2D) {
ColtDenseDoubleMatrix2D b2 = (ColtDenseDoubleMatrix2D) b;
DoubleMatrix2D ret = new CholeskyDecomposition(matrix).solve(b2.getWrappedObject());
return new ColtDenseDoubleMatrix2D(ret);
} else {
return super.solve(b);
}
}
public Matrix invSPD() {
DoubleMatrix2D ret = new CholeskyDecomposition(matrix).solve(DoubleFactory2D.dense.identity(matrix.rows()));
return new ColtDenseDoubleMatrix2D(ret);
}
public Matrix[] eig() {
EigenvalueDecomposition eig = new EigenvalueDecomposition(matrix);
Matrix v = new ColtDenseDoubleMatrix2D(eig.getV());
Matrix d = new ColtDenseDoubleMatrix2D(eig.getD());
return new Matrix[] { v, d };
}
public Matrix mtimes(Matrix m) {
if (m instanceof ColtSparseDoubleMatrix2D) {
SparseDoubleMatrix2D ret = new SparseDoubleMatrix2D((int) getRowCount(), (int) m.getColumnCount());
matrix.zMult(((ColtSparseDoubleMatrix2D) m).matrix, ret);
return new ColtSparseDoubleMatrix2D(ret);
} else {
return super.mtimes(m);
}
}
public Matrix[] lu() {
if (getColumnCount() > getRowCount()) {
throw new RuntimeException("only supported for m>=n");
}
LUDecomposition lu = new LUDecomposition(matrix);
Matrix l = new ColtDenseDoubleMatrix2D(lu.getL());
Matrix u = new ColtDenseDoubleMatrix2D(lu.getU().viewPart(0, 0, (int) getColumnCount(), (int) getColumnCount()));
int[] piv = lu.getPivot();
int m = (int) getRowCount();
Matrix p = new ColtDenseDoubleMatrix2D(m, m);
for (int i = 0; i < m; i++) {
p.setAsDouble(1, i, piv[i]);
}
return new Matrix[] { l, u, p };
}
public ColtSparseDoubleMatrix2DFactory getFactory() {
return Factory;
}
}