org.ejml.simple.SimpleBase Maven / Gradle / Ivy
/*
* Copyright (c) 2022, Peter Abeles. All Rights Reserved.
*
* This file is part of Efficient Java Matrix Library (EJML).
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.ejml.simple;
import org.ejml.UtilEjml;
import org.ejml.data.*;
import org.ejml.dense.row.CommonOps_DDRM;
import org.ejml.dense.row.NormOps_DDRM;
import org.ejml.equation.Equation;
import org.ejml.ops.ConvertMatrixType;
import org.ejml.ops.DConvertMatrixStruct;
import org.ejml.ops.FConvertMatrixStruct;
import org.ejml.ops.MatrixIO;
import org.ejml.simple.ops.*;
import org.jetbrains.annotations.Nullable;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.PrintStream;
import java.io.Serializable;
import java.lang.reflect.InvocationTargetException;
import java.lang.reflect.Method;
/**
* Parent of {@link SimpleMatrix} implements all the standard matrix operations and uses
* generics to allow the returned matrix type to be changed. This class should be extended
* instead of SimpleMatrix.
*
* @author Peter Abeles
*/
@SuppressWarnings({"unchecked", "NullAway.Init"})
public abstract class SimpleBase> implements Serializable {
static final long serialVersionUID = 2342556642L;
/**
* Internal matrix which this is a wrapper around.
*/
protected Matrix mat;
protected SimpleOperations ops;
protected transient AutomaticSimpleMatrixConvert convertType = new AutomaticSimpleMatrixConvert();
protected SimpleBase( int numRows, int numCols ) {
setMatrix(new DMatrixRMaj(numRows, numCols));
}
protected SimpleBase() {}
private void readObject( java.io.ObjectInputStream in )
throws IOException, ClassNotFoundException {
in.defaultReadObject();
convertType = new AutomaticSimpleMatrixConvert();
}
/**
* Used internally for creating new instances of SimpleMatrix. If SimpleMatrix is extended
* by another class this function should be overridden so that the returned matrices are
* of the correct type.
*
* @param numRows number of rows in the new matrix.
* @param numCols number of columns in the new matrix.
* @param type Type of matrix it should create
* @return A new matrix.
*/
protected abstract T createMatrix( int numRows, int numCols, MatrixType type );
protected abstract T wrapMatrix( Matrix m );
/**
*
* Returns a reference to the matrix that it uses internally. This is useful
* when an operation is needed that is not provided by this class.
*
*
* @return Reference to the internal DMatrixRMaj.
*/
public InnerType getMatrix() {
return (InnerType)mat;
}
public DMatrixRMaj getDDRM() {
return (mat.getType() == MatrixType.DDRM) ? (DMatrixRMaj)mat : (DMatrixRMaj)ConvertMatrixType.convert(mat, MatrixType.DDRM);
}
public FMatrixRMaj getFDRM() {
return (mat.getType() == MatrixType.FDRM) ? (FMatrixRMaj)mat : (FMatrixRMaj)ConvertMatrixType.convert(mat, MatrixType.FDRM);
}
public ZMatrixRMaj getZDRM() {
return (mat.getType() == MatrixType.ZDRM) ? (ZMatrixRMaj)mat : (ZMatrixRMaj)ConvertMatrixType.convert(mat, MatrixType.ZDRM);
}
public CMatrixRMaj getCDRM() {
return (mat.getType() == MatrixType.CDRM) ? (CMatrixRMaj)mat : (CMatrixRMaj)ConvertMatrixType.convert(mat, MatrixType.CDRM);
}
public DMatrixSparseCSC getDSCC() {
return (mat.getType() == MatrixType.DSCC) ? (DMatrixSparseCSC)mat : (DMatrixSparseCSC)ConvertMatrixType.convert(mat, MatrixType.DSCC);
}
public FMatrixSparseCSC getFSCC() {
return (mat.getType() == MatrixType.FSCC) ? (FMatrixSparseCSC)mat : (FMatrixSparseCSC)ConvertMatrixType.convert(mat, MatrixType.FSCC);
}
protected static SimpleOperations lookupOps( MatrixType type ) {
switch (type) {
case DDRM:
return new SimpleOperations_DDRM();
case FDRM:
return new SimpleOperations_FDRM();
case ZDRM:
return new SimpleOperations_ZDRM();
case CDRM:
return new SimpleOperations_CDRM();
case DSCC:
return new SimpleOperations_DSCC();
case FSCC:
return new SimpleOperations_FSCC();
default:
throw new RuntimeException("Unknown Matrix Type. " + type);
}
}
/**
*
* Returns the transpose of this matrix.
* aT
*
*
* @return A matrix that is n by m.
* @see CommonOps_DDRM#transpose(DMatrixRMaj, DMatrixRMaj)
*/
public T transpose() {
T ret = createMatrix(mat.getNumCols(), mat.getNumRows(), mat.getType());
ops.transpose(mat, ret.mat);
return ret;
}
/**
*
* Returns a matrix which is the result of matrix multiplication:
*
* c = a * b
*
* where c is the returned matrix, a is this matrix, and b is the passed in matrix.
*
*
* @param B A matrix that is n by bn. Not modified.
* @return The results of this operation.
* @see CommonOps_DDRM#mult(DMatrix1Row, DMatrix1Row, DMatrix1Row)
*/
public T mult( T B ) {
convertType.specify(this, B);
// Look to see if there is a special function for handling this case
if (this.mat.getType() != B.getType()) {
Method m = findAlternative("mult", mat, B.mat, convertType.commonType.getClassType());
if (m != null) {
T ret = wrapMatrix(convertType.commonType.create(1, 1));
invoke(m, this.mat, B.mat, ret.mat);
return ret;
}
}
// Otherwise convert into a common matrix type if necessary
T A = convertType.convert(this);
B = convertType.convert(B);
T ret = A.createMatrix(mat.getNumRows(), B.getMatrix().getNumCols(), A.getType());
A.ops.mult(A.mat, B.mat, ret.mat);
return ret;
}
/**
*
* Computes the Kronecker product between this matrix and the provided B matrix:
*
* C = kron(A,B)
*
*
* @param B The right matrix in the operation. Not modified.
* @return Kronecker product between this matrix and B.
* @see CommonOps_DDRM#kron(DMatrixRMaj, DMatrixRMaj, DMatrixRMaj)
*/
public T kron( T B ) {
convertType.specify(this, B);
T A = convertType.convert(this);
B = convertType.convert(B);
T ret = A.createMatrix(mat.getNumRows()*B.numRows(), mat.getNumCols()*B.numCols(), A.getType());
A.ops.kron(A.mat, B.mat, ret.mat);
return ret;
}
/**
*
* Returns the result of matrix addition:
*
* c = a + b
*
* where c is the returned matrix, a is this matrix, and b is the passed in matrix.
*
*
* @param B m by n matrix. Not modified.
* @return The results of this operation.
* @see CommonOps_DDRM#mult(DMatrix1Row, DMatrix1Row, DMatrix1Row)
*/
public T plus( T B ) {
convertType.specify(this, B);
T A = convertType.convert(this);
B = convertType.convert(B);
T ret = A.createMatrix(mat.getNumRows(), mat.getNumCols(), A.getType());
A.ops.plus(A.mat, B.mat, ret.mat);
return ret;
}
/**
*
* Returns the result of matrix subtraction:
*
* c = a - b
*
* where c is the returned matrix, a is this matrix, and b is the passed in matrix.
*
*
* @param B m by n matrix. Not modified.
* @return The results of this operation.
* @see CommonOps_DDRM#subtract(DMatrixD1, DMatrixD1, DMatrixD1)
*/
public T minus( T B ) {
convertType.specify(this, B);
T A = convertType.convert(this);
B = convertType.convert(B);
T ret = A.createLike();
A.ops.minus(A.mat, B.mat, ret.mat);
return ret;
}
/**
*
* Returns the result of matrix-double subtraction:
*
* c = a - b
*
* where c is the returned matrix, a is this matrix, and b is the passed in double.
*
*
* @param b Value subtracted from each element
* @return The results of this operation.
* @see CommonOps_DDRM#subtract(DMatrixD1, double, DMatrixD1)
*/
public T minus( double b ) {
T ret = createLike();
ops.minus(mat, b, ret.mat);
return ret;
}
/**
*
* Returns the result of scalar addition:
*
* c = a + b
*
* where c is the returned matrix, a is this matrix, and b is the passed in double.
*
*
* @param b Value added to each element
* @return A matrix that contains the results.
* @see CommonOps_DDRM#add(DMatrixD1, double, DMatrixD1)
*/
public T plus( double b ) {
T ret = createLike();
ops.plus(mat, b, ret.mat);
return ret;
}
/**
*
* Performs a matrix addition and scale operation.
*
* c = a + β*b
*
* where c is the returned matrix, a is this matrix, and b is the passed in matrix.
*
*
* @param B m by n matrix. Not modified.
* @return A matrix that contains the results.
* @see CommonOps_DDRM#add(DMatrixD1, double, DMatrixD1, DMatrixD1)
*/
public T plus( double beta, T B ) {
convertType.specify(this, B);
T A = convertType.convert(this);
B = convertType.convert(B);
T ret = A.createLike();
A.ops.plus(A.mat, beta, B.mat, ret.mat);
return ret;
}
/**
* Computes the dot product (a.k.a. inner product) between this vector and vector 'v'.
*
* @param v The second vector in the dot product. Not modified.
* @return dot product
*/
public double dot( T v ) {
convertType.specify(this, v);
T A = convertType.convert(this);
v = convertType.convert(v);
if (!isVector()) {
throw new IllegalArgumentException("'this' matrix is not a vector.");
} else if (!v.isVector()) {
throw new IllegalArgumentException("'v' matrix is not a vector.");
}
return A.ops.dot(A.mat, v.getMatrix());
}
/**
* Returns true if this matrix is a vector. A vector is defined as a matrix
* that has either one row or column.
*
* @return Returns true for vectors and false otherwise.
*/
public boolean isVector() {
return mat.getNumRows() == 1 || mat.getNumCols() == 1;
}
/**
*
* Returns the result of scaling each element by 'val':
* bi,j = val*ai,j
*
*
* @param val The multiplication factor.
* @return The scaled matrix.
* @see CommonOps_DDRM#scale(double, DMatrixD1)
*/
public T scale( double val ) {
T ret = createLike();
ops.scale(mat, val, ret.getMatrix());
return ret;
}
/**
*
* Returns the result of dividing each element by 'val':
* bi,j = ai,j/val
*
*
* @param val Divisor.
* @return Matrix with its elements divided by the specified value.
* @see CommonOps_DDRM#divide(DMatrixD1, double)
*/
public T divide( double val ) {
T ret = createLike();
ops.divide(mat, val, ret.getMatrix());
return ret;
}
/**
*
* Returns the inverse of this matrix.
*
* b = a-1
*
*
*
* If the matrix could not be inverted then SingularMatrixException is thrown. Even
* if no exception is thrown the matrix could still be singular or nearly singular.
*
*
* @return The inverse of this matrix.
* @see CommonOps_DDRM#invert(DMatrixRMaj, DMatrixRMaj)
*/
public T invert() {
T ret = createLike();
if (!ops.invert(mat, ret.mat))
throw new SingularMatrixException();
if (ops.hasUncountable(ret.mat))
throw new SingularMatrixException("Solution contains uncountable numbers");
return ret;
}
/**
*
* Computes the Moore-Penrose pseudo-inverse
*
*
* @return inverse computed using the pseudo inverse.
*/
public T pseudoInverse() {
T ret = createLike();
ops.pseudoInverse(mat, ret.mat);
return ret;
}
/**
*
* Solves for X in the following equation:
*
* x = a-1b
*
* where 'a' is this matrix and 'b' is an n by p matrix.
*
*
*
* If the system could not be solved then SingularMatrixException is thrown. Even
* if no exception is thrown 'a' could still be singular or nearly singular.
*
*
* @param B n by p matrix. Not modified.
* @return The solution for 'x' that is n by p.
* @see CommonOps_DDRM#solve(DMatrixRMaj, DMatrixRMaj, DMatrixRMaj)
*/
public T solve( T B ) {
convertType.specify(this, B);
// Look to see if there is a special function for handling this case
if (this.mat.getType() != B.getType()) {
Method m = findAlternative("solve", mat, B.mat, convertType.commonType.getClassType());
if (m != null) {
T ret = wrapMatrix(convertType.commonType.create(1, 1));
invoke(m, this.mat, B.mat, ret.mat); // TODO handle boolean return from solve
return ret;
}
}
T A = convertType.convert(this);
B = convertType.convert(B);
T x = A.createMatrix(mat.getNumCols(), B.getMatrix().getNumCols(), A.getType());
if (!A.ops.solve(A.mat, x.mat, B.mat))
throw new SingularMatrixException();
if (A.ops.hasUncountable(x.mat))
throw new SingularMatrixException("Solution contains uncountable numbers");
return x;
}
/**
* Sets the elements in this matrix to be equal to the elements in the passed in matrix.
* Both matrix must have the same dimension.
*
* @param a The matrix whose value this matrix is being set to.
*/
public void setTo( T a ) {
if (a.getType() == getType())
mat.setTo(a.getMatrix());
else {
setMatrix(a.mat.copy());
}
}
/**
*
* Sets all the elements in this matrix equal to the specified value.
*
* aij = val
*
*
* @param val The value each element is set to.
* @see CommonOps_DDRM#fill(DMatrixD1, double)
*/
public void fill( double val ) {
try {
ops.fill(mat, val);
} catch (ConvertToDenseException e) {
convertToDense();
fill(val);
}
}
/**
* Sets all the elements in the matrix equal to zero.
*
* @see CommonOps_DDRM#fill(DMatrixD1, double)
*/
public void zero() {
fill(0);
}
/**
*
* Computes the Frobenius normal of the matrix:
*
* normF = Sqrt{ ∑i=1:m ∑j=1:n { aij2} }
*
*
* @return The matrix's Frobenius normal.
* @see NormOps_DDRM#normF(DMatrixD1)
*/
public double normF() {
return ops.normF(mat);
}
/**
*
* The condition p = 2 number of a matrix is used to measure the sensitivity of the linear
* system Ax=b. A value near one indicates that it is a well conditioned matrix.
*
*
* @return The condition number.
* @see NormOps_DDRM#conditionP2(DMatrixRMaj)
*/
public double conditionP2() {
return ops.conditionP2(mat);
}
/**
* Computes the determinant of the matrix.
*
* @return The determinant.
* @see CommonOps_DDRM#det(DMatrixRMaj)
*/
public double determinant() {
double ret = ops.determinant(mat);
if (UtilEjml.isUncountable(ret))
return 0;
return ret;
}
/**
*
* Computes the trace of the matrix.
*
*
* @return The trace of the matrix.
* @see CommonOps_DDRM#trace(DMatrix1Row)
*/
public double trace() {
return ops.trace(mat);
}
/**
*
* Reshapes the matrix to the specified number of rows and columns. If the total number of elements
* is ≤ number of elements it had before the data is saved. Otherwise a new internal array is
* declared and the old data lost.
*
*
*
* This is equivalent to calling A.getMatrix().reshape(numRows,numCols,false).
*
*
* @param numRows The new number of rows in the matrix.
* @param numCols The new number of columns in the matrix.
* @see DMatrixRMaj#reshape(int, int, boolean)
*/
public void reshape( int numRows, int numCols ) {
if (mat.getType().isFixed()) {
throw new IllegalArgumentException("Can't reshape a fixed sized matrix");
} else {
((ReshapeMatrix)mat).reshape(numRows, numCols);
}
}
/**
* Assigns the element in the Matrix to the specified value. Performs a bounds check to make sure
* the requested element is part of the matrix.
*
* @param row The row of the element.
* @param col The column of the element.
* @param value The element's new value.
*/
public void set( int row, int col, double value ) {
ops.set(mat, row, col, value);
}
/**
* Assigns an element a value based on its index in the internal array.
*
* @param index The matrix element that is being assigned a value.
* @param value The element's new value.
*/
public void set( int index, double value ) {
if (mat.getType() == MatrixType.DDRM) {
((DMatrixRMaj)mat).set(index, value);
} else if (mat.getType() == MatrixType.FDRM) {
((FMatrixRMaj)mat).set(index, (float)value);
} else {
throw new RuntimeException("Not supported yet for this matrix type");
}
}
/**
* Used to set the complex value of a matrix element.
*
* @param row The row of the element.
* @param col The column of the element.
* @param real Real component of assigned value
* @param imaginary Imaginary component of assigned value
*/
public void set( int row, int col, double real, double imaginary ) {
if (imaginary == 0) {
set(row, col, real);
} else {
ops.set(mat, row, col, real, imaginary);
}
}
/**
*
* Assigns consecutive elements inside a row to the provided array.
*
* A(row,offset:(offset + values.length)) = values
*
*
* @param row The row that the array is to be written to.
* @param startColumn The initial column that the array is written to.
* @param values Values which are to be written to the row in a matrix.
*/
public void setRow( int row, int startColumn, double... values ) {
ops.setRow(mat, row, startColumn, values);
}
/**
*
* Assigns consecutive elements inside a column to the provided array.
*
* A(offset:(offset + values.length),column) = values
*
*
* @param column The column that the array is to be written to.
* @param startRow The initial column that the array is written to.
* @param values Values which are to be written to the row in a matrix.
*/
public void setColumn( int column, int startRow, double... values ) {
ops.setColumn(mat, column, startRow, values);
}
/**
* Returns the value of the specified matrix element. Performs a bounds check to make sure
* the requested element is part of the matrix.
*
* NOTE: Complex matrices will throw an exception
*
* @param row The row of the element.
* @param col The column of the element.
* @return The value of the element.
*/
public double get( int row, int col ) {
return ops.get(mat, row, col);
}
/**
* Returns the value of the matrix at the specified index of the 1D row major array.
*
* @param index The element's index whose value is to be returned
* @return The value of the specified element.
* @see DMatrixRMaj#get(int)
*/
public double get( int index ) {
MatrixType type = mat.getType();
if (type.isReal()) {
if (type.getBits() == 64) {
return ((DMatrixRMaj)mat).data[index];
} else {
return ((FMatrixRMaj)mat).data[index];
}
} else {
throw new IllegalArgumentException("Complex matrix. Call get(int,Complex64F) instead");
}
}
/**
* Used to get the complex value of a matrix element.
*
* @param row The row of the element.
* @param col The column of the element.
* @param output Storage for the value
*/
public void get( int row, int col, Complex_F64 output ) {
ops.get(mat, row, col, output);
}
/**
* Returns the index in the matrix's array.
*
* @param row The row number.
* @param col The column number.
* @return The index of the specified element.
* @see DMatrixRMaj#getIndex(int, int)
*/
public int getIndex( int row, int col ) {
return row*mat.getNumCols() + col;
}
/**
* Creates a new iterator for traversing through a submatrix inside this matrix. It can be traversed
* by row or by column. Range of elements is inclusive, e.g. minRow = 0 and maxRow = 1 will include rows
* 0 and 1. The iteration starts at (minRow,minCol) and ends at (maxRow,maxCol)
*
* @param rowMajor true means it will traverse through the submatrix by row first, false by columns.
* @param minRow first row it will start at.
* @param minCol first column it will start at.
* @param maxRow last row it will stop at.
* @param maxCol last column it will stop at.
* @return A new MatrixIterator
*/
public DMatrixIterator iterator( boolean rowMajor, int minRow, int minCol, int maxRow, int maxCol ) {
return new DMatrixIterator((DMatrixRMaj)mat, rowMajor, minRow, minCol, maxRow, maxCol);
}
/**
* Creates and returns a matrix which is identical to this one.
*
* @return A new identical matrix.
*/
public T copy() {
T ret = createLike();
ret.getMatrix().setTo(this.getMatrix());
return ret;
}
/**
* Returns the number of rows in this matrix.
*
* @return number of rows.
*/
public int numRows() {
return mat.getNumRows();
}
/**
* Returns the number of columns in this matrix.
*
* @return number of columns.
*/
public int numCols() {
return mat.getNumCols();
}
/**
* Returns the number of elements in this matrix, which is equal to
* the number of rows times the number of columns.
*
* @return The number of elements in the matrix.
*/
public int getNumElements() {
return mat.getNumCols()*mat.getNumRows();
}
/**
* Prints the matrix to standard out.
*/
public void print() {
mat.print();
}
/**
*
* Prints the matrix to standard out given a {@link java.io.PrintStream#printf} style floating point format,
* e.g. print("%f").
*
*/
public void print( String format ) {
ops.print(System.out, mat, format);
}
/**
* Returns 2D array of doubles using the {@link SimpleBase#get(int, int)} method.
* @return 2D array of doubles.
*/
public double[][] toArray2() {
double[][] array = new double[mat.getNumRows()][mat.getNumCols()];
for (int r = 0; r < mat.getNumRows(); r++) {
for (int c = 0; c < mat.getNumCols(); c++) {
array[r][c] = get(r, c);
}
}
return array;
}
/**
*
* Converts the array into a string format for display purposes.
* The conversion is done using {@link MatrixIO#print(java.io.PrintStream, DMatrix)}.
*
*
* @return String representation of the matrix.
*/
@Override
public String toString() {
ByteArrayOutputStream stream = new ByteArrayOutputStream();
PrintStream p = new PrintStream(stream);
MatrixIO.print(p, mat);
return stream.toString();
}
/**
*
* Creates a new SimpleMatrix which is a submatrix of this matrix.
*
*
* si-y0 , j-x0 = oij for all y0 ≤ i < y1 and x0 ≤ j < x1
*
* where 'sij' is an element in the submatrix and 'oij' is an element in the
* original matrix.
*
*
*
* If any of the inputs are set to SimpleMatrix.END then it will be set to the last row
* or column in the matrix.
*
*
* @param y0 Start row.
* @param y1 Stop row + 1.
* @param x0 Start column.
* @param x1 Stop column + 1.
* @return The submatrix.
*/
public T extractMatrix( int y0, int y1, int x0, int x1 ) {
if (y0 == SimpleMatrix.END) y0 = mat.getNumRows();
if (y1 == SimpleMatrix.END) y1 = mat.getNumRows();
if (x0 == SimpleMatrix.END) x0 = mat.getNumCols();
if (x1 == SimpleMatrix.END) x1 = mat.getNumCols();
T ret = createMatrix(y1 - y0, x1 - x0, mat.getType());
ops.extract(mat, y0, y1, x0, x1, ret.mat, 0, 0);
return ret;
}
/**
*
* Extracts a row or column from this matrix. The returned vector will either be a row
* or column vector depending on the input type.
*
*
* @param extractRow If true a row will be extracted.
* @param element The row or column the vector is contained in.
* @return Extracted vector.
*/
public T extractVector( boolean extractRow, int element ) {
if (extractRow) {
return extractMatrix(element, element + 1, 0, SimpleMatrix.END);
} else {
return extractMatrix(0, SimpleMatrix.END, element, element + 1);
}
}
/**
*
* If a vector then a square matrix is returned if a matrix then a vector of diagonal ements is returned
*
*
* @return Diagonal elements inside a vector or a square matrix with the same diagonal elements.
* @see CommonOps_DDRM#extractDiag(DMatrixRMaj, DMatrixRMaj)
*/
public T diag() {
return wrapMatrix(ops.diag(mat));
}
/**
* Checks to see if matrix 'a' is the same as this matrix within the specified
* tolerance.
*
* @param a The matrix it is being compared against.
* @param tol How similar they must be to be equals.
* @return If they are equal within tolerance of each other.
*/
public boolean isIdentical( T a, double tol ) {
if (a.getType() != getType())
return false;
return ops.isIdentical(mat, a.mat, tol);
}
/**
* Checks to see if any of the elements in this matrix are either NaN or infinite.
*
* @return True of an element is NaN or infinite. False otherwise.
*/
public boolean hasUncountable() {
return ops.hasUncountable(mat);
}
/**
* Computes a full Singular Value Decomposition (SVD) of this matrix with the
* eigenvalues ordered from largest to smallest.
*
* @return SVD
*/
public SimpleSVD svd() {
return new SimpleSVD(mat, false);
}
/**
* Computes the SVD in either compact format or full format.
*
* @return SVD of this matrix.
*/
public SimpleSVD svd( boolean compact ) {
return new SimpleSVD(mat, compact);
}
/**
* Returns the Eigen Value Decomposition (EVD) of this matrix.
*/
public SimpleEVD eig() {
return new SimpleEVD(mat);
}
/**
* Copy matrix B into this matrix at location (insertRow, insertCol).
*
* @param insertRow First row the matrix is to be inserted into.
* @param insertCol First column the matrix is to be inserted into.
* @param B The matrix that is being inserted.
*/
public void insertIntoThis( int insertRow, int insertCol, T B ) {
convertType.specify(this, B);
B = convertType.convert(B);
// See if this type's need to be changed or not
if (convertType.commonType == getType()) {
insert(B.mat, mat, insertRow, insertCol);
} else {
T A = convertType.convert(this);
A.insert(B.mat, A.mat, insertRow, insertCol);
setMatrix(A.mat);
}
}
void insert( Matrix src, Matrix dst, int destY0, int destX0 ) {
ops.extract(src, 0, src.getNumRows(), 0, src.getNumCols(), dst, destY0, destX0);
}
/**
*
* Creates a new matrix that is a combination of this matrix and matrix B. B is
* written into A at the specified location if needed the size of A is increased by
* growing it. A is grown by padding the new area with zeros.
*
*
*
* While useful when adding data to a matrix which will be solved for it is also much
* less efficient than predeclaring a matrix and inserting data into it.
*
*
*
* If insertRow or insertCol is set to SimpleMatrix.END then it will be combined
* at the last row or column respectively.
*
*
* @param insertRow Row where matrix B is written in to.
* @param insertCol Column where matrix B is written in to.
* @param B The matrix that is written into A.
* @return A new combined matrix.
*/
public T combine( int insertRow, int insertCol, T B ) {
convertType.specify(this, B);
T A = convertType.convert(this);
B = convertType.convert(B);
if (insertRow == SimpleMatrix.END) {
insertRow = mat.getNumRows();
}
if (insertCol == SimpleMatrix.END) {
insertCol = mat.getNumCols();
}
int maxRow = insertRow + B.numRows();
int maxCol = insertCol + B.numCols();
T ret;
if (maxRow > mat.getNumRows() || maxCol > mat.getNumCols()) {
int M = Math.max(maxRow, mat.getNumRows());
int N = Math.max(maxCol, mat.getNumCols());
ret = A.createMatrix(M, N, A.getType());
ret.insertIntoThis(0, 0, A);
} else {
ret = A.copy();
}
ret.insertIntoThis(insertRow, insertCol, B);
return ret;
}
/**
* Returns the maximum real value of all the elements in this matrix.
*
* @return Largest real value of any element.
*/
public double elementMax() {
return ops.elementMax(mat);
}
/**
* Returns the minimum real value of all the elements in this matrix.
*
* @return Smallest real value of any element.
*/
public double elementMin() {
return ops.elementMin(mat);
}
/**
* Returns the maximum absolute value of all the elements in this matrix. This is
* equivalent to the infinite p-norm of the matrix.
*
* @return Largest absolute value of any element.
*/
public double elementMaxAbs() {
return ops.elementMaxAbs(mat);
}
/**
* Returns the minimum absolute value of all the elements in this matrix.
*
* @return Smallest absolute value of any element.
*/
public double elementMinAbs() {
return ops.elementMinAbs(mat);
}
/**
* Computes the sum of all the elements in the matrix.
*
* @return Sum of all the elements.
*/
public double elementSum() {
return ops.elementSum(mat);
}
/**
*
* Returns a matrix which is the result of an element by element multiplication of 'this' and 'b':
* ci,j = ai,j*bi,j
*
*
* @param b A simple matrix.
* @return The element by element multiplication of 'this' and 'b'.
*/
public T elementMult( T b ) {
convertType.specify(this, b);
T A = convertType.convert(this);
b = convertType.convert(b);
T c = A.createLike();
A.ops.elementMult(A.mat, b.mat, c.mat);
return c;
}
/**
*
* Returns a matrix which is the result of an element by element division of 'this' and 'b':
* ci,j = ai,j/bi,j
*
*
* @param b A simple matrix.
* @return The element by element division of 'this' and 'b'.
*/
public T elementDiv( T b ) {
convertType.specify(this, b);
T A = convertType.convert(this);
b = convertType.convert(b);
T c = A.createLike();
A.ops.elementDiv(A.mat, b.mat, c.mat);
return c;
}
/**
*
* Returns a matrix which is the result of an element by element power of 'this' and 'b':
* ci,j = ai,j ^ bi,j
*
*
* @param b A simple matrix.
* @return The element by element power of 'this' and 'b'.
*/
public T elementPower( T b ) {
convertType.specify(this, b);
T A = convertType.convert(this);
b = convertType.convert(b);
T c = A.createLike();
A.ops.elementPower(A.mat, b.mat, c.mat);
return c;
}
/**
*
* Returns a matrix which is the result of an element by element power of 'this' and 'b':
* ci,j = ai,j ^ b
*
*
* @param b Scalar
* @return The element by element power of 'this' and 'b'.
*/
public T elementPower( double b ) {
T c = createLike();
ops.elementPower(mat, b, c.mat);
return c;
}
/**
*
* Returns a matrix which is the result of an element by element exp of 'this'
* ci,j = Math.exp(ai,j)
*
*
* @return The element by element power of 'this' and 'b'.
*/
public T elementExp() {
T c = createLike();
ops.elementExp(mat, c.mat);
return c;
}
/**
*
* Returns a matrix which is the result of an element by element exp of 'this'
* ci,j = Math.log(ai,j)
*
*
* @return The element by element power of 'this' and 'b'.
*/
public T elementLog() {
T c = createLike();
ops.elementLog(mat, c.mat);
return c;
}
/**
*
* Returns a new matrix whose elements are the negative of 'this' matrix's elements.
*
* bij = -aij
*
*
* @return A matrix that is the negative of the original.
*/
public T negative() {
T A = copy();
ops.changeSign(A.mat);
return A;
}
/**
* Allows you to perform an equation in-place on this matrix by specifying the right hand side. For information on how to define an equation
* see {@link org.ejml.equation.Equation}. The variable sequence alternates between variable and it's label String.
* This matrix is by default labeled as 'A', but is a string is the first object in 'variables' then it will take
* on that value. The variable passed in can be any data type supported by Equation can be passed in.
* This includes matrices and scalars.
*
* Examples:
*
* perform("A = A + B",matrix,"B"); // Matrix addition
* perform("A + B",matrix,"B"); // Matrix addition with implicit 'A = '
* perform("A(5,:) = B",matrix,"B"); // Insert a row defined by B into A
* perform("[A;A]"); // stack A twice with implicit 'A = '
* perform("Q = B + 2","Q",matrix,"B"); // Specify the name of 'this' as Q
*
*
*
* @param equation String representing the symbol equation
* @param variables List of variable names and variables
*/
public void equation( String equation, Object... variables ) {
if (variables.length >= 25)
throw new IllegalArgumentException("Too many variables! At most 25");
if (!(mat instanceof DMatrixRMaj))
return;
Equation eq = new Equation();
String nameThis = "A";
int offset = 0;
if (variables.length > 0 && variables[0] instanceof String) {
nameThis = (String)variables[0];
offset = 1;
if (variables.length%2 != 1)
throw new IllegalArgumentException("Expected and odd length for variables");
} else {
if (variables.length%2 != 0)
throw new IllegalArgumentException("Expected and even length for variables");
}
eq.alias((DMatrixRMaj)mat, nameThis);
for (int i = offset; i < variables.length; i += 2) {
if (!(variables[i + 1] instanceof String))
throw new IllegalArgumentException("String expected at variables index " + i);
Object o = variables[i];
String name = (String)variables[i + 1];
if (SimpleBase.class.isAssignableFrom(o.getClass())) {
eq.alias(((SimpleBase)o).getDDRM(), name);
} else if (o instanceof DMatrixRMaj) {
eq.alias((DMatrixRMaj)o, name);
} else if (o instanceof Double) {
eq.alias((Double)o, name);
} else if (o instanceof Integer) {
eq.alias((Integer)o, name);
} else {
String type = o == null ? "null" : o.getClass().getSimpleName();
throw new IllegalArgumentException("Variable type not supported by Equation! " + type);
}
}
// see if the assignment is implicit
if (!equation.contains("=")) {
equation = nameThis + " = " + equation;
}
eq.process(equation);
}
/**
*
* Saves this matrix to a file as a serialized binary object.
*
*
* @see MatrixIO#saveBin(DMatrix, String)
*/
public void saveToFileBinary( String fileName )
throws IOException {
MatrixIO.saveBin((DMatrixRMaj)mat, fileName);
}
/**
*
* Loads a new matrix from a serialized binary file.
*
*
* @param fileName File which is to be loaded.
* @return The matrix.
* @see MatrixIO#loadBin(String)
*/
public static SimpleMatrix loadBinary( String fileName )
throws IOException {
DMatrix mat = MatrixIO.loadBin(fileName);
// see if its a DMatrixRMaj
if (mat instanceof DMatrixRMaj) {
return SimpleMatrix.wrap((DMatrixRMaj)mat);
} else {
// if not convert it into one and wrap it
return SimpleMatrix.wrap(new DMatrixRMaj(mat));
}
}
/**
*
* Saves this matrix to a file in a CSV format. For the file format see {@link MatrixIO}.
*
*
* @see MatrixIO#saveBin(DMatrix, String)
*/
public void saveToFileCSV( String fileName )
throws IOException {
MatrixIO.saveDenseCSV((DMatrixRMaj)mat, fileName);
}
/**
*
* Loads a new matrix from a CSV file. For the file format see {@link MatrixIO}.
*
*
* @param fileName File which is to be loaded.
* @return The matrix.
* @see MatrixIO#loadCSV(String, boolean)
*/
public T loadCSV( String fileName )
throws IOException {
DMatrix mat = MatrixIO.loadCSV(fileName, true);
T ret = createMatrix(1, 1, mat.getType());
ret.setMatrix(mat);
return ret;
}
/**
* Returns true of the specified matrix element is valid element inside this matrix.
*
* @param row Row index.
* @param col Column index.
* @return true if it is a valid element in the matrix.
*/
public boolean isInBounds( int row, int col ) {
return row >= 0 && col >= 0 && row < mat.getNumRows() && col < mat.getNumCols();
}
/**
* Prints the number of rows and column in this matrix.
*/
public void printDimensions() {
System.out.println("[rows = " + numRows() + " , cols = " + numCols() + " ]");
}
/**
* Size of internal array elements. 32 or 64 bits
*/
public int bits() {
return mat.getType().getBits();
}
/**
* Concatenates all the matrices together along their columns. If the rows do not match the upper elements
* are set to zero.
*
* A = [ this, m[0] , ... , m[n-1] ]
*
* @param matrices Set of matrices
* @return Resulting matrix
*/
public T concatColumns( SimpleBase... matrices ) {
convertType.specify0(this, matrices);
T A = convertType.convert(this);
int numCols = A.numCols();
int numRows = A.numRows();
for (int i = 0; i < matrices.length; i++) {
numRows = Math.max(numRows, matrices[i].numRows());
numCols += matrices[i].numCols();
}
SimpleMatrix combined = SimpleMatrix.wrap(convertType.commonType.create(numRows, numCols));
A.ops.extract(A.mat, 0, A.numRows(), 0, A.numCols(), combined.mat, 0, 0);
int col = A.numCols();
for (int i = 0; i < matrices.length; i++) {
Matrix m = convertType.convert(matrices[i]).mat;
int cols = m.getNumCols();
int rows = m.getNumRows();
A.ops.extract(m, 0, rows, 0, cols, combined.mat, 0, col);
col += cols;
}
return (T)combined;
}
/**
* Concatenates all the matrices together along their columns. If the rows do not match the upper elements
* are set to zero.
*
* A = [ this; m[0] ; ... ; m[n-1] ]
*
* @param matrices Set of matrices
* @return Resulting matrix
*/
public T concatRows( SimpleBase... matrices ) {
convertType.specify0(this, matrices);
T A = convertType.convert(this);
int numCols = A.numCols();
int numRows = A.numRows();
for (int i = 0; i < matrices.length; i++) {
numRows += matrices[i].numRows();
numCols = Math.max(numCols, matrices[i].numCols());
}
SimpleMatrix combined = SimpleMatrix.wrap(convertType.commonType.create(numRows, numCols));
A.ops.extract(A.mat, 0, A.numRows(), 0, A.numCols(), combined.mat, 0, 0);
int row = A.numRows();
for (int i = 0; i < matrices.length; i++) {
Matrix m = convertType.convert(matrices[i]).mat;
int cols = m.getNumCols();
int rows = m.getNumRows();
A.ops.extract(m, 0, rows, 0, cols, combined.mat, row, 0);
row += rows;
}
return (T)combined;
}
/**
* Extracts the specified rows from the matrix.
*
* @param begin First row (inclusive).
* @param end Last row (exclusive).
* @return Submatrix that contains the specified rows.
*/
public T rows( int begin, int end ) {
return extractMatrix(begin, end, 0, SimpleMatrix.END);
}
/**
* Extracts the specified columns from the matrix.
*
* @param begin First column (inclusive).
* @param end Last column (exclusive).
* @return Submatrix that contains the specified columns.
*/
public T cols( int begin, int end ) {
return extractMatrix(0, SimpleMatrix.END, begin, end);
}
/**
* Returns the type of matrix it is wrapping.
*/
public MatrixType getType() {
return mat.getType();
}
/**
* Creates a matrix that is the same type and shape
*
* @return New matrix
*/
public T createLike() {
return createMatrix(numRows(), numCols(), getType());
}
protected void setMatrix( Matrix mat ) {
this.mat = mat;
this.ops = lookupOps(mat.getType());
}
@Nullable Method findAlternative( String method, Object... arguments ) {
Method[] methods = ops.getClass().getMethods();
for (int methodIdx = 0; methodIdx < methods.length; methodIdx++) {
if (!methods[methodIdx].getName().equals(method))
continue;
Class[] paramTypes = methods[methodIdx].getParameterTypes();
if (paramTypes.length != arguments.length)
continue;
// look for an exact match only
boolean match = true;
for (int j = 0; j < paramTypes.length; j++) {
if (arguments[j] instanceof Class) {
if (paramTypes[j] != arguments[j]) {
match = false;
break;
}
} else if (paramTypes[j] != arguments[j].getClass()) {
match = false;
break;
}
}
if (match) {
return methods[methodIdx];
}
}
return null;
}
public void invoke( Method m, Object... inputs ) {
try {
m.invoke(ops, inputs);
} catch (IllegalAccessException | InvocationTargetException e) {
throw new RuntimeException(e);
}
}
/**
* Switches from a dense to sparse matrix
*/
public void convertToSparse() {
switch (mat.getType()) {
case DDRM: {
DMatrixSparseCSC m = new DMatrixSparseCSC(mat.getNumRows(), mat.getNumCols());
DConvertMatrixStruct.convert((DMatrixRMaj)mat, m, 0);
setMatrix(m);
}
break;
case FDRM: {
FMatrixSparseCSC m = new FMatrixSparseCSC(mat.getNumRows(), mat.getNumCols());
FConvertMatrixStruct.convert((FMatrixRMaj)mat, m, 0);
setMatrix(m);
}
break;
case DSCC:
case FSCC:
break;
default:
throw new RuntimeException("Conversion not supported!");
}
}
/**
* Switches from a sparse to dense matrix
*/
public void convertToDense() {
switch (mat.getType()) {
case DSCC: {
DMatrix m = new DMatrixRMaj(mat.getNumRows(), mat.getNumCols());
DConvertMatrixStruct.convert((DMatrix)mat, m);
setMatrix(m);
}
break;
case FSCC: {
FMatrix m = new FMatrixRMaj(mat.getNumRows(), mat.getNumCols());
FConvertMatrixStruct.convert((FMatrix)mat, m);
setMatrix(m);
}
break;
case DDRM:
case FDRM:
case ZDRM:
case CDRM:
break;
default:
throw new RuntimeException("Not a sparse matrix!");
}
}
}