Allows interpretting portions of larger tensor as other tensor types. More...

Classes
class	lin::internal::TensorMappingReference< D, E >
	Generic tensor reference with read and write access. More...

class	lin::internal::MatrixMappingReference< E, R, C, MR, MC >
	Generic matrix reference with read and write access. More...

class	lin::internal::VectorMappingReference< E, N, MN >
	Generic vector reference with read and write access. More...

class	lin::internal::RowVectorMappingReference< E, N, MN >
	Generic row vector reference with read and write access. More...

class	lin::internal::TensorStreamReference< D, E >
	Generic tensor reference with read only access. More...

class	lin::internal::MatrixStreamReference< E, R, C, MR, MC >
	Generic matrix reference with read-only access. More...

class	lin::internal::VectorStreamReference< E, N, MN >
	Generic vector reference with read only access. More...

class	lin::internal::RowVectorStreamReference< E, N, MN >
	Generic row vector reference with read only access. More...

Functions
template<class C , class D , typename = std::enable_if_t<internal::has_traits<C>::value>>
constexpr auto	lin::ref (internal::Mapping< D > &mapping, size_t i, size_t j)
	Creates a mapping reference with default dimensions. More...

template<class C , class D , typename = std::enable_if_t<internal::conjunction< internal::has_traits<C>, internal::is_vector<C>>::value>>
constexpr auto	lin::ref (internal::Mapping< D > &mapping, size_t i, size_t j, size_t n)
	Creates a vector mapping reference with the provided length. More...

template<class C , class D , typename = std::enable_if_t<internal::has_traits<C>::value>>
constexpr auto	lin::ref (internal::Mapping< D > &mapping, size_t i, size_t j, size_t r, size_t c)
	Creates a mapping reference with the provided dimensions. More...

template<class D >
constexpr auto	lin::col (internal::Mapping< D > &mapping, size_t j)
	Creates a mapping reference of a particular column of a given tensor. More...

template<class D >
constexpr auto	lin::row (internal::Mapping< D > &mapping, size_t i)
	Creates a mapping reference of a particular row of a given tensor. More...

template<class D , typename = std::enable_if_t<internal::conjunction< internal::is_matrix<D>, internal::is_square<D>>::value>>
constexpr auto	lin::diag (internal::Mapping< D > &mapping)
	Creates a diagonal mapping reference from the given mapping. More...

template<class C , class D , typename = std::enable_if_t<internal::has_traits<C>::value>>
constexpr auto	lin::ref (internal::Stream< D > const &stream, size_t i, size_t j)
	Creates a mapping stream with default dimensions. More...

template<class C , class D , typename = std::enable_if_t<internal::conjunction< internal::has_traits<C>, internal::is_vector<C>>::value>>
constexpr auto	lin::ref (internal::Stream< D > const &stream, size_t i, size_t j, size_t n)
	Creates a vector stream reference with the provided length. More...

template<class C , class D , typename = std::enable_if_t<internal::has_traits<C>::value>>
constexpr auto	lin::ref (internal::Stream< D > const &stream, size_t i, size_t j, size_t r, size_t c)
	Creates a stream reference with the provided dimensions. More...

template<class D >
constexpr auto	lin::col (internal::Stream< D > const &stream, size_t j)
	Creates a stream reference of a particular column of a given tensor. More...

template<class D >
constexpr auto	lin::row (internal::Stream< D > const &stream, size_t i)
	Creates a stream reference of a particular row of a given tensor. More...

template<class D , typename = std::enable_if_t<internal::conjunction< internal::is_matrix<D>, internal::is_square<D>>::value>>
constexpr auto	lin::diag (internal::Stream< D > const &stream)
	Creates a diagonal stream reference from the given stream. More...

Detailed Description

Allows interpretting portions of larger tensor as other tensor types.

This can be extremely helpful if, for example, you'd like to normalize all the columns of a matrix without copying elements in and out of the matrix itself. We'll use this an example demonstrating how to use the refernces module.

First, be sure to include the appropriate headers (that being the lin/core.hpp and lin/references.hpp headers). From there, we can implement the following function which will normalize the columns in a three by three matrix:

#include <lin/core.hpp>
#include <lin/references.hpp>
template <typename T>
void normalize(lin::Matrix<T, 3, 3> &M) {
  for (lin::size_t j = 0; j < M.cols(); j++) {
    auto m = lin::ref<lin::Vector<T, 3>>(M, 0, j);
    m = m / lin::norm(m);
  }
}

where line of interest here is clearly the call to lin::ref<lin::Vector<T, 3>> which essentially asks for a reference to be created that acts like a lin::Vector<T, 3> and whose first element matches up with M(0, j).

There are a couple things to note here:

The reference itself is invalid when the backing object (in this case M) goes out of scope.
This particular reference serves as a read and write interface. However, read only references can also be generated in the same fashion if the underlying type is const or read only.

Lastly, there are also convenience functions lin::col, lin::row, and lin::diag to more easily reference the entire column, row, or diagonal of a tensor. In fact, if you don't mind using the lin::internal namespace a little, a simple function normalizing the columns of any matrix can be implemented as shown below:

#include <lin/core.hpp>
#include <lin/references.hpp>
#include <type_traits>
template <class D, typename = std::enable_if_t<lin::internal::disjunction<
    lin::internal::is_matrix<D>, lin::internal::is_col_vector<D>>::value>>
void normalize(lin::internal::Mapping<D> &M) {
  for (lin::size_t j = 0; j < M.cols(); j++) {
    auto m = lin::col(M, j);
    m = m / lin::norm(m);
  }
}

Function Documentation

template<class C , class D , typename = std::enable_if_t<internal::has_traits<C>::value>>

constexpr auto lin::ref	(	internal::Mapping< D > &	mapping,
		size_t	i,
		size_t	j
	)

Creates a mapping reference with default dimensions.

Template Parameters

C	Tensor type whose traits are replicated.
D	Underlying referenced type.

Parameters

mapping	Underlying mapping.
i	Anchor point row index.
j	Anchor point column index.

Returns: Instace of an internal::MatrixMappingReference, internal::RowVectorMappingReference, or internal::VectorMappingReference

This serves essentially as a wrapper around the reference constructors using default dimensions.

The anchor points specifies where the top left corner of the reference maps to in the underlying mapping.

template<class C , class D , typename = std::enable_if_t<internal::conjunction< internal::has_traits<C>, internal::is_vector<C>>::value>>

constexpr auto lin::ref	(	internal::Mapping< D > &	mapping,
		size_t	i,
		size_t	j,
		size_t	n
	)

Creates a vector mapping reference with the provided length.

Template Parameters

C	Vector type whose traits are replicated.
D	Underlying referenced type.

Parameters

mapping	Underlying mapping.
i	Anchor point row index.
j	Anchor point column index.
n	Provided length.

Returns: Instance of an internal::RowVectorMappingReference or internal::VectorMappingReference

This serves essentially as a wrapper around the vector reference constructors requesting a length.

The anchor points specifies where the top left corner of the reference maps to in the underlying mapping.

Lin assertions errors will be triggered if the requested dimensions aren't possible given the reference's traits.

See also: internal::traits

template<class C , class D , typename = std::enable_if_t<internal::has_traits<C>::value>>

constexpr auto lin::ref	(	internal::Mapping< D > &	mapping,
		size_t	i,
		size_t	j,
		size_t	r,
		size_t	c
	)

Creates a mapping reference with the provided dimensions.

Template Parameters

C	Tensor type whose traits are replicated.
D	Underlying referenced type.

Parameters

mapping	Underlying mapping.
i	Anchor point row index.
j	Anchor point column index.
r	Provided row dimension.
c	Provided column dimension.

Returns: Instance of an internal::MatrixMappingReference, internal::RowVectorMappingReference, or internal::VectorMappingReference

This serves essentially as a wrapper around the reference constructors using default dimensions.

The anchor points specifies where the top left corner of the reference maps to in the underlying mapping.

Lin assertions errors will be triggered if the requested dimensions aren't possible given the reference's traits.

See also: internal::traits

template<class D >

constexpr auto lin::col	(	internal::Mapping< D > &	mapping,
		size_t	j
	)

Creates a mapping reference of a particular column of a given tensor.

Template Parameters

D	Underlying referened type.

Parameters

mapping	Underlying mapping.
j	Index of the referenced column.

Returns: Instance of an internal::VectorMappingReference or internal::MatrixMappingReference

This is a convenience function and the same result can be obtained with a call to another reference function.

The dimensions of a variably sized column are set depending on the provided mapping's row dimension at run time.

If calling this function on a mapping representing a row vector, a one by one matrix reference is returned.

template<class D >

constexpr auto lin::row	(	internal::Mapping< D > &	mapping,
		size_t	i
	)

Creates a mapping reference of a particular row of a given tensor.

Template Parameters

D	Underlying referened type.

Parameters

mapping	Underlying mapping.
j	Index of the referenced row.

Returns: Instance of an internal::RowVectorMappingReference or internal::MatrixMappingReference.

This is a convenience function and the same result can be obtained with a call to another reference function.

The dimensions of a variably sized row are set depending on the provided mapping's column dimension at run time.

If calling this function on a mapping representing a column vector, a one by one matrix reference is returned.

template<class D , typename = std::enable_if_t<internal::conjunction< internal::is_matrix<D>, internal::is_square<D>>::value>>

constexpr auto lin::diag ( internal::Mapping< D > & mapping )

Creates a diagonal mapping reference from the given mapping.

Template Parameters

D	Underlying referenced type.

Parameters

mapping Underlying mapping.

Returns: Instance of an internal::DiagonalMappingReference.

The underlying mapping must have the traits of a square matrix and lin assertion errors will be thrown if the underlying mapping isn't square at runtime.

See also: internal::is_matrix; internal::is_square; internal::traits

template<class C , class D , typename = std::enable_if_t<internal::has_traits<C>::value>>

constexpr auto lin::ref	(	internal::Stream< D > const &	stream,
		size_t	i,
		size_t	j
	)

Creates a mapping stream with default dimensions.

Template Parameters

C	Tensor type whose traits are replicated.
D	Underlying referenced type.

Parameters

stream	Underlying stream.
i	Anchor point row index.
j	Anchor point column index.

Returns: Instace of an internal::MatrixStreamReference, internal::RowVectorStreamReference, or internal::VectorStreamReference

This serves essentially as a wrapper around the reference constructors using default dimensions.

The anchor points specifies where the top left corner of the reference maps to in the underlying stream.

template<class C , class D , typename = std::enable_if_t<internal::conjunction< internal::has_traits<C>, internal::is_vector<C>>::value>>

constexpr auto lin::ref	(	internal::Stream< D > const &	stream,
		size_t	i,
		size_t	j,
		size_t	n
	)

Creates a vector stream reference with the provided length.

Template Parameters

C	Vector type whose traits are replicated.
D	Underlying referenced type.

Parameters

stream	Underlying stream.
i	Anchor point row index.
j	Anchor point column index.
n	Provided length.

Returns: Instance of an internal::RowVectorStreamReference or internal::VectorStreamReference

This serves essentially as a wrapper around the vector reference constructors requesting a length.

The anchor points specifies where the top left corner of the reference maps to in the underlying stream.

Lin assertions errors will be triggered if the requested dimensions aren't possible given the reference's traits.

See also: internal::traits

template<class C , class D , typename = std::enable_if_t<internal::has_traits<C>::value>>

constexpr auto lin::ref	(	internal::Stream< D > const &	stream,
		size_t	i,
		size_t	j,
		size_t	r,
		size_t	c
	)

Creates a stream reference with the provided dimensions.

Template Parameters

C	Tensor type whose traits are replicated.
D	Underlying referenced type.

Parameters

stream	Underlying stream.
i	Anchor point row index.
j	Anchor point column index.
r	Provided row dimension.
c	Provided column dimension.

Returns: Instance of an internal::MatrixStreamReference, internal::RowVectorStreamReference, or internal::VectorStreamReference

This serves essentially as a wrapper around the reference constructors using default dimensions.

The anchor points specifies where the top left corner of the reference maps to in the underlying mapping.

Lin assertions errors will be triggered if the requested dimensions aren't possible given the reference's traits.

See also: internal::traits

template<class D >

constexpr auto lin::col	(	internal::Stream< D > const &	stream,
		size_t	j
	)

Creates a stream reference of a particular column of a given tensor.

Template Parameters

D	Underlying referened type.

Parameters

stream	Underlying stream.
j	Index of the referenced column.

Returns: Instance of an internal::VectorStreamReference or internal::MatrixStreamReference.

This is a convenience function and the same result can be obtained with a call to another reference function.

The dimensions of a variably sized column are set depending on the provided mapping's row dimension at run time.

If calling this function on a stream representing a row vector, a one by one matrix reference is returned.

template<class D >

constexpr auto lin::row	(	internal::Stream< D > const &	stream,
		size_t	i
	)

Creates a stream reference of a particular row of a given tensor.

Template Parameters

D	Underlying referened type.

Parameters

stream	Underlying stream.
j	Index of the referenced row.

Returns: Instance of an internal::RowVectorStreamReference or internal::MatrixStreamReference.

This is a convenience function and the same result can be obtained with a call to another reference function.

The dimensions of a variably sized row are set depending on the provided mapping's column dimension at run time.

If calling this function on a stream representing a column vector, a one by one matrix reference is returned.

template<class D , typename = std::enable_if_t<internal::conjunction< internal::is_matrix<D>, internal::is_square<D>>::value>>

constexpr auto lin::diag ( internal::Stream< D > const & stream )

Creates a diagonal stream reference from the given stream.

Creates a diagonal stream from a vector stream.

Template Parameters

D	Underlying referenced type.

Parameters

stream Underlying stream.

Returns: Instance of an internal::DiagonalMappingReference.

The underlying mapping must have the traits of a square matrix and lin assertion errors will be thrown if the underlying mapping isn't square at runtime.

See also: internal::is_matrix; internal::is_square; internal::traits

Template Parameters

D	Underlying type.

Parameters

stream Underlying vector stream.

Returns: Instance of an internal::StreamDiagonal

See also: internal::is_vector

Classes

Functions

Detailed Description

Function Documentation