Package org.flag4j.core.sparse_base

Class RealSparseTensorBase<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z>

java.lang.Object

org.flag4j.core.TensorBase<T,U,W,Z,T,double[],Double>

org.flag4j.core.sparse_base.SparseTensorBase<T,U,W,Z,T,double[],Double>

org.flag4j.core.sparse_base.RealSparseTensorBase<T,U,W,Z>

Type Parameters:

T - Type of this tensor.

U - Dense Tensor type.

W - Complex Tensor type.

Z - Dense complex tensor type.

All Implemented Interfaces:

Serializable, RealTensorMixin<T,W>, SparseTensorMixin, TensorComparisonsMixin, TensorManipulationsMixin<T>, TensorOperationsMixin<T,U,W,Z,T,Double>, TensorPropertiesMixin

Direct Known Subclasses:

CooMatrix, CooTensor, CooVector, CsrMatrix

public abstract class RealSparseTensorBase<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z>
extends SparseTensorBase<T,U,W,Z,T,double[],Double>
implements RealTensorMixin<T,W>

Base class for all sparse tensor.

See Also:

• Serialized Form

Field Summary

Fields inherited from class org.flag4j.core.sparse_base.SparseTensorBase

indices, nnz

Fields inherited from class org.flag4j.core.TensorBase

DEFAULT_ROUND_TO_ZERO_THRESHOLD, entries, shape

Constructor Summary

Constructors

Modifier	Constructor	Description
protected	RealSparseTensorBase(Shape shape, int nonZeroEntries, double[] entries, int[][] indices)	Creates a sparse tensor with specified shape.
protected	RealSparseTensorBase(Shape shape, int nonZeroEntries, double[] entries, int[] initIndices, int[]... restIndices)	Creates a sparse tensor with specified shape.

Method Summary

Modifier and Type	Method	Description
T	abs()	Computes the element-wise absolute value/magnitude of a tensor.
int[]	argMax()	Finds the indices of the maximum value in this tensor.
int[]	argMin()	Finds the indices of the minimum value in this tensor.
private int[][]	copyIndices()	Creates a deep copy of the indices of this sparse tensor.
T	div(double divisor)	Computes the scalar division of a tensor.
W	div(CNumber divisor)	Computes the scalar division of a tensor.
boolean	isNeg()	Checks if this tensor contains only non-positive values.
boolean	isOnes()	Checks if this sparse tensors non-zero values are all ones.
boolean	isPos()	Checks if this tensor contains only non-negative values.
boolean	isZeros()	Checks if this tensor only contains zeros.
protected abstract W	makeComplexTensor(Shape shape, CNumber[] entries, int[][] indices)	A factory for creating a complex sparse tensor.
protected abstract U	makeDenseTensor(Shape shape, double[] entries)	A factory for creating a real dense tensor.
protected abstract T	makeTensor(Shape shape, double[] entries, int[][] indices)	A factory for creating a real sparse tensor.
double	max()	Finds the maximum value in this tensor.
double	maxAbs()	Finds the maximum value, in absolute value, in this tensor.
double	min()	Finds the minimum value in this tensor.
double	minAbs()	Finds the minimum value, in absolute value, in this tensor.
T	mult(double factor)	Computes scalar multiplication of a tensor.
W	mult(CNumber factor)	Computes scalar multiplication of a tensor.
int	nonZeroEntries()	Gets the number of non-zero entries stored in this sparse tensor.
T	recip()	Computes the reciprocals, element-wise, of a tensor.
T	reshape(int... dims)	Copies and reshapes tensor if possible.
T	round()	Rounds each entry of this tensor to the nearest whole number.
T	round(int precision)	Rounds each entry in this tensor to the nearest whole number.
T	roundToZero()	Rounds values in this tensor which are close to zero in absolute value to zero.
T	roundToZero(double threshold)	Rounds values which are close to zero in absolute value to zero.
void	sortIndices()	Sorts the indices of this tensor in lexicographical order while maintaining the associated value for each index.
T	sqrt()	Computes the element-wise square root of a tensor.
Double	sum()	Sums together all entries in the tensor.

Methods inherited from class org.flag4j.core.sparse_base.SparseTensorBase

density, sparsity, toDense

Methods inherited from class org.flag4j.core.TensorBase

allClose, allClose, getEntries, getRank, getSelf, getShape, sameLength, sameShape, tensorEquals, totalEntries

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.flag4j.core.RealTensorMixin

toComplex

Methods inherited from interface org.flag4j.core.TensorComparisonsMixin

equals

Methods inherited from interface org.flag4j.core.TensorManipulationsMixin

flatten, flatten, reshape, set

Methods inherited from interface org.flag4j.core.TensorOperationsMixin

add, add, add, copy, elemDiv, elemMult, get, sub, sub, sub, T, transpose

Constructor Details

RealSparseTensorBase

protected RealSparseTensorBase(Shape shape,
 int nonZeroEntries,
 double[] entries,
 int[][] indices)

Creates a sparse tensor with specified shape. Note, this constructor stores indices for each element in the same array. That is, for a shape with rank m and n non-zero entries, the indices array will have shape n-by-m. This is the opposite of RealSparseTensorBase(Shape, int, double[], int[], int[][]).

Parameters:

shape - Shape of this tensor.

nonZeroEntries - Number of non-zero entries in the sparse tensor.

entries - Non-zero entries of this sparse tensor.

indices - Indices of the non-zero entries.

Throws:

IllegalArgumentException - If the number of rows in the indices array is not equal to the number of elements in the entries array.

IllegalArgumentException - If the number of columns in the entries array is not equal to the rank of this tensor.

RealSparseTensorBase

protected RealSparseTensorBase(Shape shape,
 int nonZeroEntries,
 double[] entries,
 int[] initIndices,
 int[]... restIndices)

Creates a sparse tensor with specified shape. Note, this constructor stores indices for each element in different arrays. That is, for a shape with rank m and n non-zero entries, the indices array will have shape m-by-n. This is the opposite of RealSparseTensorBase(Shape, int, double[], int[][]).

Parameters:

shape - Shape

nonZeroEntries - The number of non-zero entries of the tensor.

entries - Non-zero entries of the sparse tensor.

initIndices - Non-zero indices of the first axis of the tensor.

restIndices - Non-zero indices of the rest of this tensor's axes.

Method Details

copyIndices

private int[][] copyIndices()

Creates a deep copy of the indices of this sparse tensor.

Returns:

A deep copy of the indices of this sparse tensor.

makeTensor

protected abstract T makeTensor(Shape shape,
 double[] entries,
 int[][] indices)

A factory for creating a real sparse tensor.

Parameters:

shape - Shape of the sparse tensor to make.

entries - Non-zero entries of the sparse tensor to make.

indices - Non-zero indices of the sparse tensor to make.

Returns:

A tensor created from the specified parameters.

makeDenseTensor

protected abstract U makeDenseTensor(Shape shape,
 double[] entries)

A factory for creating a real dense tensor.

Parameters:

shape - Shape of the tensor to make.

entries - Entries of the dense tensor to make.

Returns:

A tensor created from the specified parameters.

makeComplexTensor

protected abstract W makeComplexTensor(Shape shape,
 CNumber[] entries,
 int[][] indices)

A factory for creating a complex sparse tensor.

Parameters:

shape - Shape of the tensor to make.

entries - Non-zero entries of the sparse tensor to make.

indices - Non-zero indices of the sparse tensor to make.

Returns:

A tensor created from the specified parameters.

sortIndices

public void sortIndices()

Sorts the indices of this tensor in lexicographical order while maintaining the associated value for each index.

Specified by:

sortIndices in interface SparseTensorMixin

nonZeroEntries

public int nonZeroEntries()

Gets the number of non-zero entries stored in this sparse tensor.

Overrides:

nonZeroEntries in class SparseTensorBase<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,double[],Double>

Returns:

The number of non-zero entries stored in this tensor.

min

public double min()

Description copied from interface: TensorPropertiesMixin

Finds the minimum value in this tensor. If this tensor is complex, then this method finds the smallest value in magnitude.

Specified by:

min in interface TensorPropertiesMixin

Returns:

The minimum value (smallest in magnitude for a complex valued tensor) in this tensor.

max

public double max()

Description copied from interface: TensorPropertiesMixin

Finds the maximum value in this tensor. If this tensor is complex, then this method finds the largest value in magnitude.

Specified by:

max in interface TensorPropertiesMixin

Returns:

The maximum value (largest in magnitude for a complex valued tensor) in this tensor.

minAbs

public double minAbs()

Description copied from interface: TensorPropertiesMixin

Finds the minimum value, in absolute value, in this tensor. If this tensor is complex, then this method is equivalent to TensorPropertiesMixin.min().

Specified by:

minAbs in interface TensorPropertiesMixin

Returns:

The minimum value, in absolute value, in this tensor.

maxAbs

public double maxAbs()

Description copied from interface: TensorPropertiesMixin

Finds the maximum value, in absolute value, in this tensor. If this tensor is complex, then this method is equivalent to TensorPropertiesMixin.max().

Specified by:

maxAbs in interface TensorPropertiesMixin

Returns:

The maximum value, in absolute value, in this tensor.

isPos

public boolean isPos()

Description copied from interface: RealTensorMixin

Checks if this tensor contains only non-negative values.

Specified by:

isPos in interface RealTensorMixin<T extends TensorBase<T,?,?,?,?,?,?>,U>

Returns:

True if this tensor only contains non-negative values. Otherwise, returns false.

isNeg

public boolean isNeg()

Description copied from interface: RealTensorMixin

Checks if this tensor contains only non-positive values.

Specified by:

isNeg in interface RealTensorMixin<T extends TensorBase<T,?,?,?,?,?,?>,U>

Returns:

trie if this tensor only contains non-positive values. Otherwise, returns false.

isZeros

public boolean isZeros()

Description copied from interface: TensorComparisonsMixin

Checks if this tensor only contains zeros.

Specified by:

isZeros in interface TensorComparisonsMixin

Returns:

True if this tensor only contains zeros. Otherwise, returns false.

isOnes

public boolean isOnes()

Checks if this sparse tensors non-zero values are all ones.

Specified by:

isOnes in interface TensorComparisonsMixin

Returns:

True if this sparse tensors non-zero values are all ones. False otherwise.

sum

public Double sum()

Description copied from interface: TensorOperationsMixin

Sums together all entries in the tensor.

Specified by:

sum in interface TensorOperationsMixin<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,Double>

Returns:

The sum of all entries in this tensor.

mult

public T mult(double factor)

Description copied from interface: TensorOperationsMixin

Computes scalar multiplication of a tensor.

Specified by:

mult in interface TensorOperationsMixin<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,Double>

Parameters:

factor - Scalar value to multiply with tensor.

Returns:

The result of multiplying this tensor by the specified scalar.

mult

public W mult(CNumber factor)

Description copied from interface: TensorOperationsMixin

Computes scalar multiplication of a tensor.

Specified by:

mult in interface TensorOperationsMixin<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,Double>

Parameters:

factor - Scalar value to multiply with tensor.

Returns:

The result of multiplying this tensor by the specified scalar.

div

public T div(double divisor)

Description copied from interface: TensorOperationsMixin

Computes the scalar division of a tensor.

Specified by:

div in interface TensorOperationsMixin<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,Double>

Parameters:

divisor - The scalar value to divide tensor by.

Returns:

The result of dividing this tensor by the specified scalar.

div

public W div(CNumber divisor)

Description copied from interface: TensorOperationsMixin

Computes the scalar division of a tensor.

Specified by:

div in interface TensorOperationsMixin<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,Double>

Parameters:

divisor - The scalar value to divide tensor by.

Returns:

The result of dividing this tensor by the specified scalar.

sqrt

public T sqrt()

Description copied from interface: TensorOperationsMixin

Computes the element-wise square root of a tensor.

Specified by:

sqrt in interface TensorOperationsMixin<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,Double>

Returns:

The result of applying an element-wise square root to this tensor. Note, this method will compute the principle square root i.e. the square root with positive real part.

abs

public T abs()

Description copied from interface: TensorOperationsMixin

Computes the element-wise absolute value/magnitude of a tensor. If the tensor contains complex values, the magnitude will be computed.

Specified by:

abs in interface TensorOperationsMixin<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,Double>

Returns:

The result of applying an element-wise absolute value/magnitude to this tensor.

reshape

public T reshape(int... dims)

Description copied from interface: TensorManipulationsMixin

Copies and reshapes tensor if possible. The total number of entries in this tensor must match the total number of entries in the reshaped tensor.

Specified by:

reshape in interface TensorManipulationsMixin<T extends TensorBase<T,?,?,?,?,?,?>>

Parameters:

dims - Shape of the new tensor.

Returns:

A tensor which is equivalent to this tensor but with the specified shape.

argMax

public int[] argMax()

Description copied from interface: TensorPropertiesMixin

Finds the indices of the maximum value in this tensor.

Specified by:

argMax in interface TensorPropertiesMixin

Returns:

The indices of the maximum value in this tensor. If this value occurs multiple times, the indices of the first entry (in row-major ordering) are returned.

argMin

public int[] argMin()

Description copied from interface: TensorPropertiesMixin

Finds the indices of the minimum value in this tensor.

Specified by:

argMin in interface TensorPropertiesMixin

Returns:

The indices of the minimum value in this tensor. If this value occurs multiple times, the indices of the first entry (in row-major ordering) are returned.

recip

public T recip()

Description copied from interface: TensorOperationsMixin

Computes the reciprocals, element-wise, of a tensor.

Specified by:

recip in interface TensorOperationsMixin<T extends TensorBase<T,?,?,?,?,?,?>,U,W,Z,T extends TensorBase<T,?,?,?,?,?,?>,Double>

Returns:

A tensor containing the reciprocal elements of this tensor.

round

public T round()

Description copied from interface: TensorManipulationsMixin

Rounds each entry of this tensor to the nearest whole number.

Specified by:

round in interface TensorManipulationsMixin<T extends TensorBase<T,?,?,?,?,?,?>>

Returns:

A copy of this tensor with each entry rounded to the nearest whole number.

See Also:

• TensorManipulationsMixin.round(int)
• TensorManipulationsMixin.roundToZero()
• TensorManipulationsMixin.roundToZero(double)

round

public T round(int precision)

Description copied from interface: TensorManipulationsMixin

Rounds each entry in this tensor to the nearest whole number.

Specified by:

round in interface TensorManipulationsMixin<T extends TensorBase<T,?,?,?,?,?,?>>

Parameters:

precision - The number of decimal places to round to. This value must be non-negative.

Returns:

A copy of this matrix with rounded values.

See Also:

• TensorManipulationsMixin.round()
• TensorManipulationsMixin.roundToZero()
• TensorManipulationsMixin.roundToZero(double)

roundToZero

public T roundToZero()

Description copied from interface: TensorManipulationsMixin

Rounds values in this tensor which are close to zero in absolute value to zero. If the matrix is complex, both the real and imaginary components will be rounded independently. By default, the values must be within 1.0*E^-12 of zero. To specify a threshold value see TensorManipulationsMixin.roundToZero(double).

Specified by:

roundToZero in interface TensorManipulationsMixin<T extends TensorBase<T,?,?,?,?,?,?>>

Returns:

A copy of this matrix with rounded values.

See Also:

• TensorManipulationsMixin.roundToZero(double)
• TensorManipulationsMixin.round()
• TensorManipulationsMixin.round(int)

roundToZero

public T roundToZero(double threshold)

Description copied from interface: TensorManipulationsMixin

Rounds values which are close to zero in absolute value to zero. If the matrix is complex, both the real and imaginary components will be rounded independently.

Specified by:

roundToZero in interface TensorManipulationsMixin<T extends TensorBase<T,?,?,?,?,?,?>>

Parameters:

threshold - Threshold for rounding values to zero. That is, if a value in this matrix is less than the threshold in absolute value then it will be rounded to zero. This value must be non-negative.

Returns:

A copy of this matrix with rounded values.

See Also:

• TensorManipulationsMixin.roundToZero()
• TensorManipulationsMixin.round()
• TensorManipulationsMixin.round(int)