java.lang.Object

org.tensorflow.ndarray.impl.sparse.AbstractSparseNdArray<T,U>

org.tensorflow.ndarray.impl.sparse.SparseNdArray<T,U>

All Implemented Interfaces:: NdArray<T>, Shaped, SparseNdArray<T,U>

public class SparseNdArray<T, U extends NdArray<T>> extends AbstractSparseNdArray<T,U> implements SparseNdArray<T,U>

sparse array for the any data type

A sparse array has two separate dense arrays: indices, values, and a shape that represents the dense shape.

NOTE: all Sparse Arrays are readonly for the AbstractSparseNdArray.set(NdArray, long...) and AbstractSparseNdArray.setObject(Object, long...) methods

SparseNdArray<String> st = new SparseNdArray<>(
     StdArrays.of(new long[][] {{0, 0}, {1, 2}}),
     NdArrays.vectorOf("first", "second"),
     Shape.of(3, 4));

represents the dense array:

[[true, false, false, false]
 [false, false, true, false]
 [false, false, false, false]]

Field Summary

Fields

Modifier and Type

Field

Description

protected final DimensionalSpace

dimensions
Constructor Summary

Constructors

Modifier

Constructor

Description

protected

SparseNdArray(Class<T> type, LongNdArray indices, U values, T defaultValue, DimensionalSpace dimensions)

Creates a SparseNdArray
Method Summary

Modifier and Type

Method

Description

NdArray<T>

copyFrom(DataBuffer<T> src)

Copy the content of the source buffer into this N-dimensional array.

NdArray<T>

copyTo(DataBuffer<T> dst)

Copy the content of this N-dimensional array into the destination buffer.

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, DataBuffer<T> dataBuffer, DimensionalSpace dimensions)

Creates a new SparseNdArray from a data buffer

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, DataBuffer<T> buffer, Shape shape)

Creates a new empty SparseNdArray from a float data buffer

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, DataBuffer<T> dataBuffer, T defaultValue, DimensionalSpace dimensions)

Creates a new SparseNdArray from a data buffer

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, DataBuffer<T> buffer, T defaultValue, Shape shape)

Creates a new empty SparseNdArray from a float data buffer

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, DimensionalSpace dimensions)

Creates a new empty SparseNdArray from a data buffer

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, LongNdArray indices, U values, DimensionalSpace dimensions)

Creates a new SparseNdArray

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, LongNdArray indices, U values, T defaultValue, DimensionalSpace dimensions)

Creates a new SparseNdArray

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, T defaultValue, DimensionalSpace dimensions)

Creates a new empty SparseNdArray from a data buffer

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, U src)

Creates a new SparseNdArray from a NdArray

static <T, U extends NdArray<T>> SparseNdArray<T,U>

create(Class<T> type, U src, T defaultValue)

Creates a new SparseNdArray from a NdArray

U

createDefaultArray()

Creates the NdArray with the default value as a scalar

U

createValues(Shape shape)

Creates a NdArray of the specified shape

DimensionalSpace

dimensions()

NdArray<T>

fromDense(NdArray<T> src)

Populates this sparse array from a dense array

Class<T>

getType()

Gets the class type for this sparse array

NdArraySequence<U>

scalars()

Returns a sequence of all scalars in this array.

Shape

shape()

U

slice(long position, DimensionalSpace sliceDimensions)

protected boolean

slowEquals(NdArray<?> array)

protected int

slowHashCode()

U

toDense()

Converts the sparse array to a dense array

String

toString()

A String showing the type, default value, number of elements and the dense shape of this sparse ndarray.

Methods inherited from class AbstractSparseNdArray
copyTo, elements, equals, get, getDefaultArray, getDefaultValue, getIndices, getIndicesCoordinates, getObject, getValues, hashCode, locateIndex, positionOf, set, setDefaultValue, setIndices, setObject, setValues, slice, slowCopyTo, sortIndicesAndValues, toCoordinates, withShape

Methods inherited from class Object
clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface NdArray
copyTo, elements, equals, get, getObject, scalars, set, setObject, slice, streamOfObjects, withShape

Methods inherited from interface Shaped
rank, shape, size

Methods inherited from interface SparseNdArray
getIndices, getValues

Field Details
- dimensions
  
  protected final DimensionalSpace dimensions
Constructor Details
- SparseNdArray
  
  protected SparseNdArray(Class<T> type, LongNdArray indices, U values, T defaultValue, DimensionalSpace dimensions)
  
  Creates a SparseNdArray
  
  Parameters:
  
  indices - A 2-D LongNdArray of shape [N, ndims], that specifies the indices of the elements in the sparse array that contain non-default values (elements are zero-indexed). For example, indices=[[1,3], [2,4]] specifies that the elements with indexes of [1,3] and [2,4] have non-default values.
  
  values - A 1-D NdArray of Boolean type and shape [N], which supplies the values for each element in indices. For example, given indices=[[1,3], [2,4]], the parameter values=[18, 3.6] specifies that element [1,3] of the sparse NdArray has a value of 18, and element [2,4] of the NdArray has a value of 3.6.
  
  defaultValue - Scalar value to set for indices not specified in AbstractSparseNdArray.getIndices()
  
  dimensions - the dimensional space for the dense object represented by this sparse array,
Method Details
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, LongNdArray indices, U values, DimensionalSpace dimensions)
  
  Creates a new SparseNdArray
  
  Parameters:
  
  indices - A 2-D LongNdArray of shape [N, ndims], that specifies the indices of the elements in the sparse array that contain non-default values (elements are zero-indexed). For example, indices=[[1,3], [2,4]] specifies that the elements with indexes of [1,3] and [2,4] have non-default values.
  
  values - A 1-D NdArray of any type and shape [N], which supplies the values for each element in indices. For example, given indices=[[1,3], [2,4]], the parameter values=[18, 3.6] specifies that element [1,3] of the sparse NdArray has a value of 18, and element [2,4] of the NdArray has a value of 3.6.
  
  dimensions - the dimensional space for the dense object represented by this sparse array.
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, LongNdArray indices, U values, T defaultValue, DimensionalSpace dimensions)
  
  Creates a new SparseNdArray
  
  Parameters:
  
  indices - A 2-D LongNdArray of shape [N, ndims], that specifies the indices of the elements in the sparse array that contain non-default values (elements are zero-indexed). For example, indices=[[1,3], [2,4]] specifies that the elements with indexes of [1,3] and [2,4] have non-default values.
  
  values - A 1-D NdArray of any type and shape [N], which supplies the values for each element in indices. For example, given indices=[[1,3], [2,4]], the parameter values=[18, 3.6] specifies that element [1,3] of the sparse NdArray has a value of 18, and element [2,4] of the NdArray has a value of 3.6.
  
  defaultValue - Scalar value to set for indices not specified in AbstractSparseNdArray.getIndices()
  
  dimensions - the dimensional space for the dense object represented by this sparse array.
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, DataBuffer<T> dataBuffer, DimensionalSpace dimensions)
  
  Creates a new SparseNdArray from a data buffer
  
  Parameters:
  
  dataBuffer - the databuffer containing the dense array
  
  dimensions - the dimensional space for the sparse array
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, DataBuffer<T> dataBuffer, T defaultValue, DimensionalSpace dimensions)
  
  Creates a new SparseNdArray from a data buffer
  
  Parameters:
  
  dataBuffer - the databuffer containing the dense array
  
  defaultValue - Scalar value to set for indices not specified in AbstractSparseNdArray.getIndices()
  
  dimensions - the dimensional space for the sparse array
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, DimensionalSpace dimensions)
  
  Creates a new empty SparseNdArray from a data buffer
  
  Parameters:
  
  dimensions - the dimensions array
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, T defaultValue, DimensionalSpace dimensions)
  
  Creates a new empty SparseNdArray from a data buffer
  
  Parameters:
  
  defaultValue - Scalar value to set for indices not specified in AbstractSparseNdArray.getIndices()
  
  dimensions - the dimensions array
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, DataBuffer<T> buffer, Shape shape)
  
  Creates a new empty SparseNdArray from a float data buffer
  
  Parameters:
  
  buffer - the data buffer
  
  shape - the shape of the sparse array.
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, DataBuffer<T> buffer, T defaultValue, Shape shape)
  
  Creates a new empty SparseNdArray from a float data buffer
  
  Parameters:
  
  buffer - the data buffer
  
  defaultValue - Scalar value to set for indices not specified in AbstractSparseNdArray.getIndices()
  
  shape - the shape of the sparse array.
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, U src)
  
  Creates a new SparseNdArray from a NdArray
  
  Parameters:
  
  src - the NdArray
  
  Returns:
  
  the new Sparse Array
- create
  
  public static <T, U extends NdArray<T>> SparseNdArray<T,U> create(Class<T> type, U src, T defaultValue)
  
  Creates a new SparseNdArray from a NdArray
  
  Parameters:
  
  src - the NdArray
  
  defaultValue - Scalar value to set for indices not specified in AbstractSparseNdArray.getIndices()
  
  Returns:
  
  the new Sparse Array
- createDefaultArray
  
  public U createDefaultArray()
  
  Creates the NdArray with the default value as a scalar
  
  Specified by:
  
  createDefaultArray in class AbstractSparseNdArray<T, U extends NdArray<T>>
  
  Returns:
  
  the default NdArray of the default value as a scalar
- createValues
  
  public U createValues(Shape shape)
  
  Creates a NdArray of the specified shape
  
  Specified by:
  
  createValues in class AbstractSparseNdArray<T, U extends NdArray<T>>
  
  Parameters:
  
  shape - the shape of the dense array.
  
  Returns:
  
  a NdArray of the specified shape
- slice
  
  public U slice(long position, DimensionalSpace sliceDimensions)
- copyTo
  public NdArray<T> copyTo(DataBuffer<T> dst)
  
  Copy the content of this N-dimensional array into the destination buffer.
  The size of the buffer must be equal or greater to the Shaped.size() of this array, or an exception is thrown. After the copy, content of the buffer and of the array can be altered independently, without affecting each other.
  Note: in version 0.4.0 and earlier, this method was named read(DataBuffer<T>). It has been renamed to explicitly indicate the direction of the data flow to avoid confusion.
  
  Specified by:
  
  copyTo in interface NdArray<T>
  
  Parameters:
  
  dst - the destination buffer
  
  Returns:
  
  this array
  
  See Also:
  
  DataBuffer.size()
- copyFrom
  public NdArray<T> copyFrom(DataBuffer<T> src)
  
  Copy the content of the source buffer into this N-dimensional array.
  The size of the buffer must be equal or greater to the Shaped.size() of this array, or an exception is thrown. After the copy, content of the buffer and of the array can be altered independently, without affecting each other.
  Note: in version 0.4.0 and earlier, this method was named write(DataBuffer<T>). It has been renamed to explicitly indicate the direction of the data flow to avoid confusion.
  
  Specified by:
  
  copyFrom in interface NdArray<T>
  
  Parameters:
  
  src - the source buffer
  
  Returns:
  
  this array
  
  See Also:
  
  DataBuffer.size()
- toDense
  
  public U toDense()
  
  Converts the sparse array to a dense array
  
  Specified by:
  
  toDense in class AbstractSparseNdArray<T, U extends NdArray<T>>
  
  Returns:
  
  the dense array
- fromDense
  
  public NdArray<T> fromDense(NdArray<T> src)
  
  Populates this sparse array from a dense array
  
  Parameters:
  
  src - the dense array
  
  Returns:
  
  this sparse array
- getType
  
  public Class<T> getType()
  
  Gets the class type for this sparse array
  
  Returns:
  
  the class type for this sparse array.
- toString
  
  public String toString()
  
  A String showing the type, default value, number of elements and the dense shape of this sparse ndarray.
  
  Overrides:
  
  toString in class AbstractSparseNdArray<T, U extends NdArray<T>>
  
  Returns:
  
  A string containing the type, default value, number of elements and shape.
- dimensions
  
  public DimensionalSpace dimensions()
- shape
  
  public Shape shape()
  
  Specified by:
  
  shape in interface Shaped
  
  Returns:
  
  the shape of this container
- scalars
  
  public NdArraySequence<U> scalars()
  
  Description copied from interface: NdArray
  
  Returns a sequence of all scalars in this array.
  This is equivalent to call elements(shape().numDimensions() - 1)
  
  Specified by:
  
  scalars in interface NdArray<T>
  
  Returns:
  
  an NdArray sequence
- slowHashCode
  
  protected int slowHashCode()
- slowEquals
  
  protected boolean slowEquals(NdArray<?> array)

Class SparseNdArray<T, U extends NdArray<T>>

Field Summary

Constructor Summary

Method Summary

Methods inherited from class AbstractSparseNdArray

Methods inherited from class Object

Methods inherited from interface NdArray

Methods inherited from interface Shaped

Methods inherited from interface SparseNdArray

Field Details

dimensions

Constructor Details

SparseNdArray

Method Details

create

create

create

create

create

create

create

create

create

create

createDefaultArray

createValues

slice

copyTo

copyFrom

toDense

fromDense

getType

toString

dimensions

shape

scalars

slowHashCode

slowEquals