Package org.la4j.vector.sparse

Class CompressedVector

java.lang.Object

org.la4j.Vector

org.la4j.vector.SparseVector

org.la4j.vector.sparse.CompressedVector

All Implemented Interfaces:

java.lang.Iterable<java.lang.Double>

public class CompressedVector
extends SparseVector

A basic sparse vector implementation using underlying value and index arrays. A sparse data structure does not store blank elements, and instead just stores elements with values. A sparse data structure can be initialized with a large length but take up no storage until the space is filled with non-zero elements. However, there is a performance cost. Fetch/store operations take O(log n) time instead of the O(1) time of a dense data structure. CompressedVector stores the underlying data in a two arrays: A values array and a indices array. The values array matches the indices array, and they're both sorted by the indices array. To get a value at an index, the index is found in the indices array with a binary search and the respective value is obtained from the values array.

Field Summary

Fields

Modifier and Type	Field	Description
private int[]	indices
private static int	MINIMUM_SIZE
private double[]	values
private static byte	VECTOR_TAG

Fields inherited from class org.la4j.vector.SparseVector

cardinality

Fields inherited from class org.la4j.Vector

length

Constructor Summary

Constructors

Constructor	Description
CompressedVector()
CompressedVector(int length)
CompressedVector(int length, int capacity)
CompressedVector(int length, int cardinality, double[] values, int[] indices)

Method Summary

Modifier and Type	Method	Description
private int	align(int length, int capacity)
Vector	blankOfLength(int length)	Creates a blank (an empty vector) copy of this vector with the given length.
Vector	copyOfLength(int length)	Copies this vector into the new vector with specified length.
void	each(VectorProcedure procedure)	Applies given procedure to each element of this vector.
void	eachNonZero(VectorProcedure procedure)	Applies given procedure to each non-zero element of this vector.
static CompressedVector	fromArray(double[] array)	Creates a new CompressedVector from the given array with compressing (copying) the underlying array.
static CompressedVector	fromBinary(byte[] array)	Decodes CompressedVector from the given byte array.
static CompressedVector	fromCollection(java.util.Collection<? extends java.lang.Number> list)	Creates new CompressedVector from collection
static CompressedVector	fromCSV(java.lang.String csv)	Parses CompressedVector from the given CSV string.
static CompressedVector	fromMap(java.util.Map<java.lang.Integer,? extends java.lang.Number> map, int length)	Creates new CompressedVector from index-value map
static CompressedVector	fromMatrixMarket(java.lang.String mm)	Parses CompressedVector from the given Matrix Market string.
double	getOrElse(int i, double defaultValue)	Gets the specified element, or a defaultValue if there is no actual element at index i in this sparse vector.
private void	growUp()
private void	insert(int k, int i, double value)
VectorIterator	iterator()	Returns a vector iterator.
boolean	nonZeroAt(int i)	* Whether or not the specified element is not zero.
VectorIterator	nonZeroIterator()	Returns a non-zero vector iterator.
static CompressedVector	random(int length, double density, java.util.Random random)	Creates a random CompressedVector of the given length with the given density and Random.
private void	remove(int k)
private int	searchForIndex(int i)	Does the binary searching to find the position in the value array given it's index.
void	set(int i, double value)	Sets the specified element of this matrix to given value.
void	setAll(double value)	Sets all elements of this vector to given value.
void	swapElements(int i, int j)	Swaps the specified elements of this vector.
<T extends Vector> T	to(VectorFactory<T> factory)	Converts this vector using the given factory.
byte[]	toBinary()	Encodes this vector into a byte array.
void	updateAt(int i, VectorFunction function)	Updates the specified element of this vector by applying given function.
static CompressedVector	zero(int length)	Creates a zero CompressedVector of the given length.
static CompressedVector	zero(int length, int capacity)	Creates a zero CompressedVector of the given length with the given capacity.

Methods inherited from class org.la4j.vector.SparseVector

add, apply, apply, apply, cardinality, density, ensureIndexIsInBounds, euclideanNorm, foldNonZero, get, hashCode, infinityNorm, isZeroAt, manhattanNorm, max, min, multiply, toColumnMatrix, toDiagonalMatrix, toMatrixMarket, toRowMatrix

Methods inherited from class org.la4j.Vector

add, blank, constant, copy, divide, ensureLengthIsCorrect, equals, equals, fail, fold, hadamardProduct, innerProduct, is, length, mkString, mkString, multiply, non, norm, outerProduct, product, random, select, shuffle, slice, sliceLeft, sliceRight, subtract, subtract, sum, toCSV, toCSV, toDenseVector, toMatrixMarket, toSparseVector, toString, transform, unit, update

Methods inherited from class java.lang.Object

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

Methods inherited from interface java.lang.Iterable

forEach, spliterator

Field Detail

VECTOR_TAG

private static final byte VECTOR_TAG

See Also:

Constant Field Values

MINIMUM_SIZE

private static final int MINIMUM_SIZE

See Also:

Constant Field Values

values

private double[] values

indices

private int[] indices

Constructor Detail

CompressedVector

public CompressedVector()

CompressedVector

public CompressedVector(int length)

CompressedVector

public CompressedVector(int length,
                        int capacity)

CompressedVector

public CompressedVector(int length,
                        int cardinality,
                        double[] values,
                        int[] indices)

Method Detail

zero

public static CompressedVector zero(int length)

Creates a zero CompressedVector of the given length.

zero

public static CompressedVector zero(int length,
                                    int capacity)

Creates a zero CompressedVector of the given length with the given capacity.

random

public static CompressedVector random(int length,
                                      double density,
                                      java.util.Random random)

Creates a random CompressedVector of the given length with the given density and Random.

fromArray

public static CompressedVector fromArray(double[] array)

Creates a new CompressedVector from the given array with compressing (copying) the underlying array.

fromBinary

public static CompressedVector fromBinary(byte[] array)

Decodes CompressedVector from the given byte array.

Parameters:

array - the byte array representing a vector

Returns:

a decoded vector

fromCSV

public static CompressedVector fromCSV(java.lang.String csv)

Parses CompressedVector from the given CSV string.

Parameters:

csv - the CSV string representing a vector

Returns:

a parsed vector

fromMatrixMarket

public static CompressedVector fromMatrixMarket(java.lang.String mm)

Parses CompressedVector from the given Matrix Market string.

Parameters:

mm - the string in Matrix Market format

Returns:

a parsed vector

fromCollection

public static CompressedVector fromCollection(java.util.Collection<? extends java.lang.Number> list)

Creates new CompressedVector from collection

Parameters:

list - value list

Returns:

created vector

fromMap

public static CompressedVector fromMap(java.util.Map<java.lang.Integer,? extends java.lang.Number> map,
                                       int length)

Creates new CompressedVector from index-value map

Parameters:

map - index-value map

length - vector length

Returns:

created vector

getOrElse

public double getOrElse(int i,
                        double defaultValue)

Description copied from class: SparseVector

Gets the specified element, or a defaultValue if there is no actual element at index i in this sparse vector.

Specified by:

getOrElse in class SparseVector

Parameters:

i - the element's index

defaultValue - the default value

Returns:

the element of this vector or a default value

set

public void set(int i,
                double value)

Description copied from class: Vector

Sets the specified element of this matrix to given value.

Specified by:

set in class Vector

Parameters:

i - element's index

value - element's new value

setAll

public void setAll(double value)

Description copied from class: Vector

Sets all elements of this vector to given value.

Overrides:

setAll in class Vector

Parameters:

value - the element's new value

swapElements

public void swapElements(int i,
                         int j)

Description copied from class: Vector

Swaps the specified elements of this vector.

Overrides:

swapElements in class Vector

Parameters:

i - element's index

j - element's index

copyOfLength

public Vector copyOfLength(int length)

Description copied from class: Vector

Copies this vector into the new vector with specified length.

Specified by:

copyOfLength in class Vector

Parameters:

length - the length of new vector

Returns:

the copy of this vector with new length

each

public void each(VectorProcedure procedure)

Description copied from class: Vector

Applies given procedure to each element of this vector.

Overrides:

each in class Vector

Parameters:

procedure - the vector procedure

eachNonZero

public void eachNonZero(VectorProcedure procedure)

Description copied from class: SparseVector

Applies given procedure to each non-zero element of this vector.

Overrides:

eachNonZero in class SparseVector

Parameters:

procedure - the vector procedure

updateAt

public void updateAt(int i,
                     VectorFunction function)

Description copied from class: Vector

Updates the specified element of this vector by applying given function.

Overrides:

updateAt in class Vector

Parameters:

i - element's index

function - the vector function

nonZeroAt

public boolean nonZeroAt(int i)

Description copied from class: SparseVector

* Whether or not the specified element is not zero.

Specified by:

nonZeroAt in class SparseVector

Parameters:

i - element's index

Returns:

true if specified element is zero, false otherwise

to

public <T extends Vector> T to(VectorFactory<T> factory)

Description copied from class: Vector

Converts this vector using the given factory.

Overrides:

to in class SparseVector

Type Parameters:

T - type of the result vector

Parameters:

factory - the factory that creates an output vector

Returns:

a converted vector

blankOfLength

public Vector blankOfLength(int length)

Description copied from class: Vector

Creates a blank (an empty vector) copy of this vector with the given length.

Specified by:

blankOfLength in class Vector

Parameters:

length - the length of the blank vector

Returns:

blank vector

toBinary

public byte[] toBinary()

Description copied from class: Vector

Encodes this vector into a byte array.

Specified by:

toBinary in class Vector

Returns:

a byte array representing this vector

searchForIndex

private int searchForIndex(int i)

Does the binary searching to find the position in the value array given it's index.

Parameters:

i - the index to search for

Returns:

the position in the value array

insert

private void insert(int k,
                    int i,
                    double value)

remove

private void remove(int k)

growUp

private void growUp()

align

private int align(int length,
                  int capacity)

nonZeroIterator

public VectorIterator nonZeroIterator()

Description copied from class: SparseVector

Returns a non-zero vector iterator.

Specified by:

nonZeroIterator in class SparseVector

Returns:

a non-zero vector iterator

iterator

public VectorIterator iterator()

Description copied from class: Vector

Returns a vector iterator.

Specified by:

iterator in interface java.lang.Iterable<java.lang.Double>

Overrides:

iterator in class Vector

Returns:

a vector iterator.