mikera.arrayz

Class Arrayz

java.lang.Object

mikera.arrayz.Arrayz

public class Arrayz
extends java.lang.Object

Static function class for array operations

Method Summary

Modifier and Type	Method and Description
static void	checkShape(INDArray array, int dimension, int index) Checks that a specified index exists along a specified dimension.
static INDArray	create(INDArray a) Create a new mutable array instance with the given source data.
static INDArray	create(java.util.List<?> slices) Creates an array from the given List of slices.
static INDArray	create(java.lang.Object object) Creates an array from the given data.
static <T> INDArray	create(T[] data) Creates an array using the given data as slices.
static INDArray	createFromVector(AVector a, int... shape) Creates a new array using the elements in the specified vector.
static INDArray	createSparse(INDArray a) Creates a sparse copy of the given array.
static <T> INDArray	createSparse(java.util.List<T> o) Creates a sparse array given the provided List of slice objects
static INDArray	createSparse(java.lang.Object o) Creates a sparse array given the provided input object Supports - Numbers (wrapped as scalars) - Existing INDArrays (copied into sparse format) - Iterable objects (interpreted as ordered lists of major slices) - Java Arrays (interpreted as arrays of slices Objects)
static <T> INDArray	createSparse(T... slices) Creates a sparse copy of the given data, given an array of slices to make sparse.
static INDArray	createSparseArray(int... shape) Creates a mutable sparse array of the specified shape.
static INDArray	createSparseMutable(INDArray a) Creates a fully mutable sparse clone of the given array
static INDArray	createZeroArray(int... shape) Creates an immutable zero-filled array of the given shape
static void	fillNormal(INDArray a, long seed)
static void	fillNormal(INDArray a, java.util.Random random)
static void	fillRandom(INDArray a, long seed)
static void	fillRandom(INDArray a, java.util.Random random)
static boolean	isPackedLayout(double[] data, int offset, int[] shape, int[] strides) Returns true if the offset, shape and strides describe a fully packed, row-major dense layout for the given data array.
static boolean	isPackedStrides(int[] shape, int[] strides) Checks if the given set of strides represents a fully packed, row major layout for the given shape
static INDArray	load(java.io.Reader reader)
static INDArray	newArray(int... shape) Create a new mutable array instance with the given shape.
static INDArray	parse(java.lang.String ednString) Parse an array from a String.
static INDArray	wrap(double[] data, int[] shape) Creates an INDArray instance wrapping the given double data, with the provided shape.
static INDArray	wrapStrided(double[] data, int offset, int[] shape, int[] strides) Wraps a Java double[] as a general, mutable strided array.

Methods inherited from class java.lang.Object

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

Method Detail

create

public static INDArray create(java.lang.Object object)

Creates an array from the given data. Makes a copy of underlying data as necessary. Handles double arrays, Java arrays, INDArray instances, and lists

Parameters:

object -

Returns:

create

public static INDArray create(java.util.List<?> slices)

Creates an array from the given List of slices. Calls create recursively on underlying slices if needed, so that nested structures can be used

Parameters:

object -

Returns:

newArray

public static INDArray newArray(int... shape)

Create a new mutable array instance with the given shape. New array will be filled with zeros. Uses the most efficient densely packed format where possible.

Parameters:

shape -

Returns:

create

public static INDArray create(INDArray a)

Create a new mutable array instance with the given source data. Uses the most efficient densely packed format where possible.

Parameters:

shape -

Returns:

create

public static <T> INDArray create(T[] data)

Creates an array using the given data as slices.

Parameters:

data -

Returns:

wrap

public static INDArray wrap(double[] data,
                            int[] shape)

Creates an INDArray instance wrapping the given double data, with the provided shape.

Parameters:

data -

shape -

Returns:

createFromVector

public static INDArray createFromVector(AVector a,
                                        int... shape)

Creates a new array using the elements in the specified vector. Truncates or zero-pads the data as required to fill the new array

Parameters:

data -

rows -

columns -

Returns:

load

public static INDArray load(java.io.Reader reader)

parse

public static INDArray parse(java.lang.String ednString)

Parse an array from a String. String should be in edn format

Parameters:

ednString -

Returns:

wrapStrided

public static INDArray wrapStrided(double[] data,
                                   int offset,
                                   int[] shape,
                                   int[] strides)

Wraps a Java double[] as a general, mutable strided array. Selects the most appropriate Vectorz type (Vector, Matrix, Array etc.)

Parameters:

data -

offset -

shape -

strides -

Returns:

isPackedLayout

public static boolean isPackedLayout(double[] data,
                                     int offset,
                                     int[] shape,
                                     int[] strides)

Returns true if the offset, shape and strides describe a fully packed, row-major dense layout for the given data array.

Parameters:

data -

offset -

shape -

strides -

Returns:

isPackedStrides

public static boolean isPackedStrides(int[] shape,
                                      int[] strides)

Checks if the given set of strides represents a fully packed, row major layout for the given shape

Parameters:

shape -

strides -

Returns:

createSparse

public static INDArray createSparse(INDArray a)

Creates a sparse copy of the given array. May or may not be mutable.

Parameters:

a -

Returns:

createSparse

public static <T> INDArray createSparse(T... slices)

Creates a sparse copy of the given data, given an array of slices to make sparse.

Parameters:

a -

Returns:

createSparse

public static INDArray createSparse(java.lang.Object o)

Creates a sparse array given the provided input object Supports - Numbers (wrapped as scalars) - Existing INDArrays (copied into sparse format) - Iterable objects (interpreted as ordered lists of major slices) - Java Arrays (interpreted as arrays of slices Objects)

createSparse

public static <T> INDArray createSparse(java.util.List<T> o)

Creates a sparse array given the provided List of slice objects

createSparseMutable

public static INDArray createSparseMutable(INDArray a)

Creates a fully mutable sparse clone of the given array

Parameters:

a -

Returns:

createZeroArray

public static INDArray createZeroArray(int... shape)

Creates an immutable zero-filled array of the given shape

Parameters:

shape -

Returns:

createSparseArray

public static INDArray createSparseArray(int... shape)

Creates a mutable sparse array of the specified shape.

Parameters:

shape -

Returns:

fillRandom

public static void fillRandom(INDArray a,
                              long seed)

fillRandom

public static void fillRandom(INDArray a,
                              java.util.Random random)

fillNormal

public static void fillNormal(INDArray a,
                              long seed)

fillNormal

public static void fillNormal(INDArray a,
                              java.util.Random random)

checkShape

public static void checkShape(INDArray array,
                              int dimension,
                              int index)

Checks that a specified index exists along a specified dimension. Throws an exception if the index does not exist.

Parameters:

array -

dimension -

index -

Throws:

java.lang.IndexOutOfBoundsException - if the specified index or dimension does not exist.