Package com.yahoo.tensor

Class TensorType.Builder

java.lang.Object

com.yahoo.tensor.TensorType.Builder

Enclosing class:

TensorType

public static class TensorType.Builder
extends Object

Constructor Summary

Constructors

Constructor	Description
Builder()	Creates an empty builder with cells of type double
Builder(boolean allowDifferentSizes, TensorType... types)
Builder(TensorType... types)	Creates a builder containing a combination of the dimensions of the given types If the same dimension is indexed with different size restrictions the smallest size will be used.
Builder(TensorType.Value valueType)
Builder(TensorType.Value valueType, Iterable<TensorType.Dimension> dimensions)	Creates a builder from the given value type and dimensions
Builder(Iterable<TensorType.Dimension> dimensions)	Creates a builder from the given dimensions, having double as the value type

Method Summary

Modifier and Type	Method	Description
TensorType	build()
TensorType.Builder	dimension(TensorType.Dimension dimension)	Adds the given dimension
TensorType.Builder	dimension(String name, TensorType.Dimension.Type type)
Optional<TensorType.Dimension>	getDimension(String dimension)	Returns the given dimension, or empty if none is present
TensorType.Builder	indexed(String name)	Adds an unbound indexed dimension to this
TensorType.Builder	indexed(String name, long size)	Adds a bound indexed dimension to this
TensorType.Builder	mapped(String name)	Adds a mapped dimension to this
int	rank()	Returns the current number of dimensions in this
TensorType.Builder	set(TensorType.Dimension dimension)	Adds or replaces a dimension in this

Methods inherited from class java.lang.Object

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

Constructor Details

Builder

public Builder()

Creates an empty builder with cells of type double

Builder

public Builder(TensorType.Value valueType)

Builder

public Builder(TensorType... types)

Creates a builder containing a combination of the dimensions of the given types If the same dimension is indexed with different size restrictions the smallest size will be used. If it is size restricted in one argument but not the other it will not be size restricted. If it is indexed in one and mapped in the other it will become mapped. The value type will be the largest of the value types of the input types

Builder

public Builder(boolean allowDifferentSizes,
 TensorType... types)

Builder

public Builder(Iterable<TensorType.Dimension> dimensions)

Creates a builder from the given dimensions, having double as the value type

Builder

public Builder(TensorType.Value valueType,
 Iterable<TensorType.Dimension> dimensions)

Creates a builder from the given value type and dimensions

Method Details

rank

public int rank()

Returns the current number of dimensions in this

set

public TensorType.Builder set(TensorType.Dimension dimension)

Adds or replaces a dimension in this

indexed

public TensorType.Builder indexed(String name,
 long size)

Adds a bound indexed dimension to this

Throws:

IllegalArgumentException - if the dimension is already present

indexed

public TensorType.Builder indexed(String name)

Adds an unbound indexed dimension to this

Throws:

IllegalArgumentException - if the dimension is already present

mapped

public TensorType.Builder mapped(String name)

Adds a mapped dimension to this

Throws:

IllegalArgumentException - if the dimension is already present

dimension

public TensorType.Builder dimension(TensorType.Dimension dimension)

Adds the given dimension

getDimension

public Optional<TensorType.Dimension> getDimension(String dimension)

Returns the given dimension, or empty if none is present

dimension

public TensorType.Builder dimension(String name,
 TensorType.Dimension.Type type)

build

public TensorType build()