cern.colt.matrix.tint
Class IntFactory2D

java.lang.Object
  cern.colt.PersistentObject
      cern.colt.matrix.tint.IntFactory2D

All Implemented Interfaces:

Serializable, Cloneable

public class IntFactory2D
extends PersistentObject

Factory for convenient construction of 2-d matrices holding int cells. Also provides convenient methods to compose (concatenate) and decompose (split) matrices from/to constituent blocks.

Construction	Use idioms like IntFactory2D.dense.make(4,4) to construct dense matrices, IntFactory2D.sparse.make(4,4) to construct sparse matrices.
Construction with initial values	Use other make methods to construct matrices with given initial values.
Appending rows and columns	Use methods appendColumns, appendRows and repeat to append rows and columns.
General block matrices	Use methods compose and decompose to work with general block matrices.
Diagonal matrices	Use methods diagonal(vector), diagonal(matrix) and identity to work with diagonal matrices.
Diagonal block matrices	Use method composeDiagonal to work with diagonal block matrices.
Random	Use methods random and sample to construct random matrices.

If the factory is used frequently it might be useful to streamline the notation. For example by aliasing:

  IntFactory2D F = IntFactory2D.dense;
  F.make(4,4);
  F.descending(10,20);
  F.random(4,4);
  ...
 

Version:

1.0, 09/24/99

Author:

wolfgang.hoschek@cern.ch

See Also:

Serialized Form

Field Summary

static IntFactory2D	dense A factory producing dense matrices.
static IntFactory2D	rowCompressed A factory producing sparse row compressed matrices.
static IntFactory2D	sparse A factory producing sparse hash matrices.

Fields inherited from class cern.colt.PersistentObject

serialVersionUID

Method Summary

IntMatrix2D	appendColumn(IntMatrix2D A, IntMatrix1D b)
IntMatrix2D	appendColumns(IntMatrix2D A, IntMatrix2D B) C = A||B; Constructs a new matrix which is the column-wise concatenation of two other matrices.
IntMatrix2D	appendRow(IntMatrix2D A, IntMatrix1D b)
IntMatrix2D	appendRows(IntMatrix2D A, IntMatrix2D B) C = A||B; Constructs a new matrix which is the row-wise concatenation of two other matrices.
IntMatrix2D	ascending(int rows, int columns) Constructs a matrix with cells having ascending values.
IntMatrix2D	compose(IntMatrix2D[][] parts) Constructs a block matrix made from the given parts.
IntMatrix2D	composeBidiagonal(IntMatrix2D A, IntMatrix2D B)
IntMatrix2D	composeDiagonal(IntMatrix2D A, IntMatrix2D B) Constructs a diagonal block matrix from the given parts (the direct sum of two matrices).
IntMatrix2D	composeDiagonal(IntMatrix2D A, IntMatrix2D B, IntMatrix2D C) Constructs a diagonal block matrix from the given parts.
void	decompose(IntMatrix2D[][] parts, IntMatrix2D matrix) Splits a block matrix into its constituent blocks; Copies blocks of a matrix into the given parts.
void	demo1() Demonstrates usage of this class.
void	demo2() Demonstrates usage of this class.
IntMatrix2D	descending(int rows, int columns) Constructs a matrix with cells having descending values.
IntMatrix2D	diagonal(int[] vector) Constructs a new diagonal matrix whose diagonal elements are the elements of vector.
IntMatrix2D	diagonal(IntMatrix1D vector) Constructs a new diagonal matrix whose diagonal elements are the elements of vector.
IntMatrix1D	diagonal(IntMatrix2D A) Constructs a new vector consisting of the diagonal elements of A .
IntMatrix2D	identity(int rowsAndColumns) Constructs an identity matrix (having ones on the diagonal and zeros elsewhere).
IntMatrix2D	make(int[][] values) Constructs a matrix with the given cell values.
IntMatrix2D	make(int[] values, int rows) Construct a matrix from a one-dimensional column-major packed array, ala Fortran.
IntMatrix2D	make(int rows, int columns) Constructs a matrix with the given shape, each cell initialized with zero.
IntMatrix2D	make(int rows, int columns, int initialValue) Constructs a matrix with the given shape, each cell initialized with the given value.
IntMatrix2D	random(int rows, int columns) Constructs a matrix with uniformly distributed values in (0,1) (exclusive).
IntMatrix2D	repeat(IntMatrix2D A, int rowRepeat, int columnRepeat) C = A||A||..||A; Constructs a new matrix which is duplicated both along the row and column dimension.
IntMatrix2D	reshape(IntMatrix1D a, int rows, int columns)
IntMatrix2D	sample(int rows, int columns, int value, int nonZeroFraction) Constructs a randomly sampled matrix with the given shape.
IntMatrix2D	sample(IntMatrix2D matrix, int value, int nonZeroFraction) Modifies the given matrix to be a randomly sampled matrix.

Methods inherited from class cern.colt.PersistentObject

clone

Methods inherited from class java.lang.Object

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

Field Detail

dense

public static final IntFactory2D dense

A factory producing dense matrices.

sparse

public static final IntFactory2D sparse

A factory producing sparse hash matrices.

rowCompressed

public static final IntFactory2D rowCompressed

A factory producing sparse row compressed matrices.

Method Detail

appendColumns

public IntMatrix2D appendColumns(IntMatrix2D A,
                                 IntMatrix2D B)

C = A||B; Constructs a new matrix which is the column-wise concatenation of two other matrices.

   0 1 2
   3 4 5
   appendColumns
   6 7
   8 9
   -->
   0 1 2 6 7 
   3 4 5 8 9
 
 

appendColumn

public IntMatrix2D appendColumn(IntMatrix2D A,
                                IntMatrix1D b)

appendRows

public IntMatrix2D appendRows(IntMatrix2D A,
                              IntMatrix2D B)

C = A||B; Constructs a new matrix which is the row-wise concatenation of two other matrices.

   0 1 
   2 3 
   4 5
   appendRows
   6 7
   8 9
   -->
   0 1 
   2 3 
   4 5
   6 7
   8 9
 
 

appendRow

public IntMatrix2D appendRow(IntMatrix2D A,
                             IntMatrix1D b)

ascending

public IntMatrix2D ascending(int rows,
                             int columns)

Constructs a matrix with cells having ascending values. For debugging purposes. Example:

   0 1 2 
   3 4 5
 
 

reshape

public IntMatrix2D reshape(IntMatrix1D a,
                           int rows,
                           int columns)

compose

public IntMatrix2D compose(IntMatrix2D[][] parts)

Constructs a block matrix made from the given parts. The inverse to method decompose(IntMatrix2D[][], IntMatrix2D).

All matrices of a given column within parts must have the same number of columns. All matrices of a given row within parts must have the same number of rows. Otherwise an IllegalArgumentException is thrown. Note that nulls within parts[row,col] are an exception to this rule: they are ignored. Cells are copied. Example:

Code	Result
IntMatrix2D[][] parts1 = { { null, make(2, 2, 1), null }, { make(4, 4, 2), null, make(4, 3, 3) }, { null, make(2, 2, 4), null } }; System.out.println(compose(parts1));	8 x 9 matrix 0 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 2 2 2 2 0 0 3 3 3 2 2 2 2 0 0 3 3 3 2 2 2 2 0 0 3 3 3 2 2 2 2 0 0 3 3 3 0 0 0 0 4 4 0 0 0 0 0 0 0 4 4 0 0 0
IntMatrix2D[][] parts3 = { { identity(3), null, }, { null, identity(3).viewColumnFlip() }, { identity(3).viewRowFlip(), null } }; System.out.println("\n" + make(parts3));	9 x 6 matrix 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0
IntMatrix2D A = ascending(2, 2); IntMatrix2D B = descending(2, 2); IntMatrix2D _ = null; IntMatrix2D[][] parts4 = { { A, _, A, _ }, { _, A, _, B } }; System.out.println("\n" + make(parts4));	4 x 8 matrix 1 2 0 0 1 2 0 0 3 4 0 0 3 4 0 0 0 0 1 2 0 0 3 2 0 0 3 4 0 0 1 0
IntMatrix2D[][] parts2 = { { null, make(2, 2, 1), null }, { make(4, 4, 2), null, make(4, 3, 3) }, { null, make(2, 3, 4), null } }; System.out.println("\n" + Factory2D.make(parts2));	IllegalArgumentException A[0,1].cols != A[2,1].cols (2 != 3)

Throws:

IllegalArgumentException - subject to the conditions outlined above.

composeDiagonal

public IntMatrix2D composeDiagonal(IntMatrix2D A,
                                   IntMatrix2D B)

Constructs a diagonal block matrix from the given parts (the direct sum of two matrices). That is the concatenation

   A 0
   0 B
 
 

(The direct sum has A.rows()+B.rows() rows and A.columns()+B.columns() columns). Cells are copied.

Returns:

a new matrix which is the direct sum.

composeDiagonal

public IntMatrix2D composeDiagonal(IntMatrix2D A,
                                   IntMatrix2D B,
                                   IntMatrix2D C)

Constructs a diagonal block matrix from the given parts. The concatenation has the form

   A 0 0
   0 B 0
   0 0 C
 
 

from the given parts. Cells are copied.

composeBidiagonal

public IntMatrix2D composeBidiagonal(IntMatrix2D A,
                                     IntMatrix2D B)

decompose

public void decompose(IntMatrix2D[][] parts,
                      IntMatrix2D matrix)

Splits a block matrix into its constituent blocks; Copies blocks of a matrix into the given parts. The inverse to method compose(IntMatrix2D[][]).

All matrices of a given column within parts must have the same number of columns. All matrices of a given row within parts must have the same number of rows. Otherwise an IllegalArgumentException is thrown. Note that nulls within parts[row,col] are an exception to this rule: they are ignored. Cells are copied. Example:

Code	matrix	--> parts
IntMatrix2D matrix = ... ; IntMatrix2D _ = null; IntMatrix2D A,B,C,D; A = make(2,2); B = make (4,4); C = make(4,3); D = make (2,2); IntMatrix2D[][] parts = { { _, A, _ }, { B, _, C }, { _, D, _ } }; decompose(parts,matrix); System.out.println("\nA = "+A); System.out.println("\nB = "+B); System.out.println("\nC = "+C); System.out.println("\nD = "+D);	8 x 9 matrix 9 9 9 9 1 1 9 9 9 9 9 9 9 1 1 9 9 9 2 2 2 2 9 9 3 3 3 2 2 2 2 9 9 3 3 3 2 2 2 2 9 9 3 3 3 2 2 2 2 9 9 3 3 3 9 9 9 9 4 4 9 9 9 9 9 9 9 4 4 9 9 9	A = 2 x 2 matrix 1 1 1 1 B = 4 x 4 matrix 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 C = 4 x 3 matrix 3 3 3 3 3 3 3 3 3 3 3 3 D = 2 x 2 matrix 4 4 4 4

Throws:

IllegalArgumentException - subject to the conditions outlined above.

demo1

public void demo1()

Demonstrates usage of this class.

demo2

public void demo2()

Demonstrates usage of this class.

descending

public IntMatrix2D descending(int rows,
                              int columns)

Constructs a matrix with cells having descending values. For debugging purposes. Example:

   5 4 3 
   2 1 0
 
 

diagonal

public IntMatrix2D diagonal(IntMatrix1D vector)

Constructs a new diagonal matrix whose diagonal elements are the elements of vector. Cells values are copied. The new matrix is not a view. Example:

   5 4 3 -->
   5 0 0
   0 4 0
   0 0 3
 
 

Returns:

a new matrix.

diagonal

public IntMatrix2D diagonal(int[] vector)

Constructs a new diagonal matrix whose diagonal elements are the elements of vector. Cells values are copied. The new matrix is not a view. Example:

   5 4 3 -->
   5 0 0
   0 4 0
   0 0 3
 
 

Returns:

a new matrix.

diagonal

public IntMatrix1D diagonal(IntMatrix2D A)

Constructs a new vector consisting of the diagonal elements of A . Cells values are copied. The new vector is not a view. Example:

   5 0 0 9
   0 4 0 9
   0 0 3 9
   --> 5 4 3
 
 

Parameters:

A - the matrix, need not be square.

Returns:

a new vector.

identity

public IntMatrix2D identity(int rowsAndColumns)

Constructs an identity matrix (having ones on the diagonal and zeros elsewhere).

make

public IntMatrix2D make(int[][] values)

Constructs a matrix with the given cell values. values is required to have the form values[row][column] and have exactly the same number of columns in every row.

The values are copied. So subsequent changes in values are not reflected in the matrix, and vice-versa.

Parameters:

values - The values to be filled into the new matrix.

Throws:

IllegalArgumentException - if for any 1 <= row < values.length: values[row].length != values[row-1].length .

make

public IntMatrix2D make(int[] values,
                        int rows)

Construct a matrix from a one-dimensional column-major packed array, ala Fortran. Has the form matrix.get(row,column) == values[row + column*rows]. The values are copied.

Parameters:

values - One-dimensional array of doubles, packed by columns (ala Fortran).

rows - the number of rows.

Throws:

IllegalArgumentException - values.length must be a multiple of rows .

make

public IntMatrix2D make(int rows,
                        int columns)

Constructs a matrix with the given shape, each cell initialized with zero.

make

public IntMatrix2D make(int rows,
                        int columns,
                        int initialValue)

Constructs a matrix with the given shape, each cell initialized with the given value.

random

public IntMatrix2D random(int rows,
                          int columns)

Constructs a matrix with uniformly distributed values in (0,1) (exclusive).

repeat

public IntMatrix2D repeat(IntMatrix2D A,
                          int rowRepeat,
                          int columnRepeat)

C = A||A||..||A; Constructs a new matrix which is duplicated both along the row and column dimension. Example:

   0 1
   2 3
   repeat(2,3) -->
   0 1 0 1 0 1
   2 3 2 3 2 3
   0 1 0 1 0 1
   2 3 2 3 2 3
 
 

sample

public IntMatrix2D sample(int rows,
                          int columns,
                          int value,
                          int nonZeroFraction)

Constructs a randomly sampled matrix with the given shape. Randomly picks exactly Math.round(rows*columns*nonZeroFraction) cells and initializes them to value, all the rest will be initialized to zero. Note that this is not the same as setting each cell with probability nonZeroFraction to value. Note: The random seed is a constant.

Throws:

IllegalArgumentException - if nonZeroFraction < 0 || nonZeroFraction > 1.

See Also:

DoubleRandomSamplingAssistant

sample

public IntMatrix2D sample(IntMatrix2D matrix,
                          int value,
                          int nonZeroFraction)

Modifies the given matrix to be a randomly sampled matrix. Randomly picks exactly Math.round(rows*columns*nonZeroFraction) cells and initializes them to value, all the rest will be initialized to zero. Note that this is not the same as setting each cell with probability nonZeroFraction to value. Note: The random seed is a constant.

Throws:

IllegalArgumentException - if nonZeroFraction < 0 || nonZeroFraction > 1.

See Also:

DoubleRandomSamplingAssistant