RectMDArray.H File Reference

#include "Box.H"
#include <memory>
#include <cassert>
#include <functional>
#include "RectMDArrayImplem.H"

Include dependency graph for RectMDArray.H:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
class	Interval
	A linear interval with integer bounds. More...
class	Tensor< T, C, RestD >
	Tensor object used to interface with data pointwise in a RectMDArray. More...
class	CTensor< T, C, RestD >
	Tensor object used to interface with data pointwise in a constant RectMDArray. More...
class	RectMDArray< T, C, D, E >
	Multidimensional Rectangular Array of data. More...
Functions
template<class T, unsigned int C0, unsigned int C1>
RectMDArray< T, C1 >	slice (RectMDArray< T, C0 > &a_original, const Interval &a_interval)
	Slice into a RectMDArray with vector valued (single index) data. Output is an alias to a subset of the input (shallow copy).
template<class T, unsigned int C0, unsigned int C1>
const RectMDArray< T, C1 >	slice (const RectMDArray< T, C0 > &a_original, const Interval &a_interval)
	Slice into a constant RectMDArray with vector valued (single index) data. Output is an alias to a subset of the input (shallow copy).
template<class T, unsigned int C, unsigned char D0, unsigned char D1>
RectMDArray< T, C, D1 >	slice (RectMDArray< T, C, D0 > &a_original, const Interval &a_interval)
	Slice into a RectMDArray with matrix valued (doubly indexed) data. Output is an alias to a subset of the input (shallow copy).
template<class T, unsigned int C, unsigned char D0, unsigned char D1>
const RectMDArray< T, C, D1 >	slice (const RectMDArray< T, C, D0 > &a_original, const Interval &a_interval)
	Slice into a constant RectMDArray with matrix valued (doubly indexed) data. Output is an alias to a subset of the input (shallow copy).
template<class T, unsigned int C, unsigned char D, unsigned char E0, unsigned char E1>
RectMDArray< T, C, D, E1 >	slice (RectMDArray< T, C, D, E0 > &a_original, const Interval &a_interval)
	Slice into a RectMDArray with three sets of components (three indices). Output is an alias to a subset of the input (shallow copy).
template<class T, unsigned int C, unsigned char D, unsigned char E0, unsigned char E1>
const RectMDArray< T, C, D, E1 >	slice (const RectMDArray< T, C, D, E0 > &a_original, const Interval &a_interval)
	Slice into a constant RectMDArray with three sets of components (three indices). Output is an alias to a subset of the input (shallow copy).
template<class T, unsigned int Cdest, unsigned int Csrc, typename Func>
void	forall (RectMDArray< T, Cdest > &a_dest, const RectMDArray< T, Csrc > &a_src, const Func &F, const Box &a_box)
	Pointwise operator. Applies F to the single indexed RectMDArray a_src in the domain a_box and stores the result in the singly indexed RectMDArray a_dest.
template<class T, unsigned int C, unsigned char D, unsigned char E, typename Func>
void	forall (RectMDArray< T, C > &a_dest, const RectMDArray< T, C, D, E > &a_src, const Func &F, const Box &a_box)
	Pointwise operator. Applies F to the triple indexed RectMDArray a_src in the domain a_box and stores the result in the singly indexed RectMDArray a_dest.
template<class T, unsigned int Cdest, unsigned int Csrc, unsigned char D, typename Func>
void	forall (RectMDArray< T, Cdest, D > &a_dest, const RectMDArray< T, Csrc > &a_src, const Func &F, const Box &a_box)
	Pointwise operator. Applies F to the singly indexed RectMDArray a_src in the domain a_box and stores the result in the doubly indexed RectMDArray a_dest.
template<class T, unsigned int C, unsigned char D, typename Func>
void	forall (RectMDArray< T > &a_dest, const RectMDArray< T, C, D > &a_src, const Func &F, const Box &a_box)
	Pointwise operator. Applies F to the doubly indexed RectMDArray a_src in the domain a_box and stores the result in the scalar RectMDArray a_dest.
template<class T, typename Func>
void	forall (RectMDArray< T > &a_dest, const Func &F, const Box &a_box)
	In place pointwise operator. Applies F to the (scalar) RectMDArray a_dest for every point in a_box.
template<class T>
T &	abs_max (RectMDArray< T > &a_src, const Box &a_box)
	Returns the value of a_src in a_box with the largest absolute value.
template<class T>
tuple< T &, Point >	abs_argmax (const RectMDArray< T > &a_src, const Box &a_box)
	Returns a tuple containing the value of a_src in a_box with the largest absolute value and the Point at which this value is attained.

---

Function Documentation

template<class T, unsigned int C0, unsigned int C1>

RectMDArray<T,C1> slice	(	RectMDArray< T, C0 > &	a_original,
		const Interval &	a_interval
	)			[inline]

Slice into a RectMDArray with vector valued (single index) data. Output is an alias to a subset of the input (shallow copy).

Note that only the lower bound of the Interval input is used. The upper bound of a_interval can be any value at all.
To obtain generate an alias to the first component of a vector-valued array, the lower bound of a_interval must be 0 and the output array must have 1 component (see Example 1).
To get the second component, the lower bound of a_interval must be 1 and the output array must have 1 component (see Example 2).
To obtain the first and second components of a 3 component array, the lower bound of a_interval must be 0 and the output data must have 2 components (see Example 3).

Example Usage: Point p1 = getZeros(); Point p2 = getOnes(); Box b = Box(p1,p2); RectMDArray<int,3> source = RectMDArray<int,3>(b);

Example 1: Create alias to first component of source: Interval I1 = Interval(0,1); RectMDArray<int,1> slice1<int,3,1>(source,I1)

Example 2: Create alias to the second component of source: Interval I2 = Interval(1,2); RectMDArray<int,1> slice1<int,3,1>(source,I2)

Example 3: Create alias to the second and third components of source: Interval I23 = Interval(1,3); RectMDArray<int,2> slice1<int,3,2>(source,I23)

Template Parameters:

	T	A Class. Type of data stored by input and output RectMDArrays
	C0	Number of components in the input RectMDArray
	C1	Number of components in the output RectMDArray

Parameters:

	a_original	The source RectMDArray with data of type T and C0 components
	a_interval	An Interval object. The lower bound of a_interval determines which components of a_original will be be sliced into the output.

References RectMDArray< T, C, D, E >::getBox(), Interval::low, RectMDArray< T, C, D, E >::m_sliceData(), RectMDArray< T, C, D, E >::m_slicePtr(), and Box::sizeOf().

template<class T, unsigned int C0, unsigned int C1>

const RectMDArray<T,C1> slice	(	const RectMDArray< T, C0 > &	a_original,
		const Interval &	a_interval
	)			[inline]

Slice into a constant RectMDArray with vector valued (single index) data. Output is an alias to a subset of the input (shallow copy).

See documentation for the non-constant version of slice for notes and example usage.

Template Parameters:

	T	A Class. Type of data stored by input and output RectMDArrays
	C0	Number of components in the input RectMDArray
	C1	Number of components in the output RectMDArray

Parameters:

	a_original	The source constant RectMDArray with data of type T and C0 components
	a_interval	An Interval object. The lower bound of a_interval determines which components of a_original will be be sliced into the output.

References RectMDArray< T, C, D, E >::getBox(), Interval::low, RectMDArray< T, C, D, E >::m_sliceData(), RectMDArray< T, C, D, E >::m_slicePtr(), and Box::sizeOf().

template<class T, unsigned int C, unsigned char D0, unsigned char D1>

RectMDArray<T,C,D1> slice	(	RectMDArray< T, C, D0 > &	a_original,
		const Interval &	a_interval
	)			[inline]

Slice into a RectMDArray with matrix valued (doubly indexed) data. Output is an alias to a subset of the input (shallow copy).

This version of slice produces alias to a C x D1 dimensional subset of the C x D0 dimensional input RectMDArray. See documentation for the vector valued version of slice() for additional example usage and notes.

Template Parameters:

	T	A Class. Type of data stored by input and output RectMDArrays
	C	Number of the first set of components of both the input and output RectMDArrays
	D0	Number of the second set of components of the input RectMDArray
	D1	Number of the second set of components of the output RectMDArray

Parameters:

	a_original	The source constant RectMDArray with data of type T and C x D0 components
	a_interval	An Interval object. The lower bound of a_interval determines which components of a_original will be be sliced into the output.

References RectMDArray< T, C, D, E >::getBox(), Interval::low, RectMDArray< T, C, D, E >::m_sliceData(), RectMDArray< T, C, D, E >::m_slicePtr(), and Box::sizeOf().

template<class T, unsigned int C, unsigned char D0, unsigned char D1>

const RectMDArray<T,C,D1> slice	(	const RectMDArray< T, C, D0 > &	a_original,
		const Interval &	a_interval
	)			[inline]

Slice into a constant RectMDArray with matrix valued (doubly indexed) data. Output is an alias to a subset of the input (shallow copy).

See the non-constant input version for notes on usage. See documentation for the vector valued version of slice() for additional example usage and notes.

Template Parameters:

	T	A Class. Type of data stored by input and output RectMDArrays
	C	Number of the first set of components of both the input and output RectMDArrays
	D0	Number of the second set of components of the input RectMDArray
	D1	Number of the second set of components of the output RectMDArray

Parameters:

	a_original	The source constant RectMDArray with data of type T and C x D0 components
	a_interval	An Interval object. The lower bound of a_interval determines which components of a_original will be be sliced into the output.

References RectMDArray< T, C, D, E >::getBox(), Interval::low, RectMDArray< T, C, D, E >::m_sliceData(), RectMDArray< T, C, D, E >::m_slicePtr(), and Box::sizeOf().

template<class T, unsigned int C, unsigned char D, unsigned char E0, unsigned char E1>

RectMDArray<T,C,D,E1> slice	(	RectMDArray< T, C, D, E0 > &	a_original,
		const Interval &	a_interval
	)			[inline]

Slice into a RectMDArray with three sets of components (three indices). Output is an alias to a subset of the input (shallow copy).

This version of slice produces alias to a C x D x E1 dimensional subset of the C x D x E0 dimensional input RectMDArray. See documentation for the vector valued version of slice() for additional example usage and notes.

Template Parameters:

	T	A Class. Type of data stored by input and output RectMDArrays
	C	Number of the first set of components of both the input and output RectMDArrays
	D	Number of the second set of components of both the input and output RectMDArrays
	E0	Number of the third set of components of the input RectMDArray
	E1	Number of the third set of components of the output RectMDArray

Parameters:

	a_original	The source constant RectMDArray with data of type T and C x D x E0 components
	a_interval	An Interval object. The lower bound of a_interval determines which components of a_original will be be sliced into the output.

References RectMDArray< T, C, D, E >::getBox(), Interval::low, RectMDArray< T, C, D, E >::m_sliceData(), RectMDArray< T, C, D, E >::m_slicePtr(), and Box::sizeOf().

template<class T, unsigned int C, unsigned char D, unsigned char E0, unsigned char E1>

const RectMDArray<T,C,D,E1> slice	(	const RectMDArray< T, C, D, E0 > &	a_original,
		const Interval &	a_interval
	)			[inline]

Slice into a constant RectMDArray with three sets of components (three indices). Output is an alias to a subset of the input (shallow copy).

See the non-constant input version for notes on usage. See documentation for the vector valued version of slice() for additional example usage and notes.

Template Parameters:

	T	A Class. Type of data stored by input and output RectMDArrays
	C	Number of the first set of components of both the input and output RectMDArrays
	D	Number of the second set of components of both the input and output RectMDArrays
	E0	Number of the third set of components of the input RectMDArray
	E1	Number of the third set of components of the output RectMDArray

Parameters:

	a_original	The source constant RectMDArray with data of type T and C x D x E0 components
	a_interval	An Interval object. The lower bound of a_interval determines which components of a_original will be be sliced into the output.

References RectMDArray< T, C, D, E >::getBox(), Interval::low, RectMDArray< T, C, D, E >::m_sliceData(), RectMDArray< T, C, D, E >::m_slicePtr(), and Box::sizeOf().

template<class T, unsigned int Cdest, unsigned int Csrc, typename Func>

void forall	(	RectMDArray< T, Cdest > &	a_dest,
		const RectMDArray< T, Csrc > &	a_src,
		const Func &	F,
		const Box &	a_box
	)			[inline]

Pointwise operator. Applies F to the single indexed RectMDArray a_src in the domain a_box and stores the result in the singly indexed RectMDArray a_dest.

Template Parameters:

	T	Data type of the RectMDArray arguments
	Cdest	Number of components in the destination RectMDArray
	Csrc	Number of components in the source RectMDArray
	Func	Function datatype. Usually this parameter will be filled with decltype(F).

Parameters:

	a_dest	Destination RectMDArray. This is where the result of F(a_src) is stored.
	a_src	Source RectMDArray. This is the array that F acts on.
	F	A pointwise function with a Tensor and a CTensor as arguments which acts on a_src. F can be any externally defined function that bears the following signature: void F(Tensor<T,Cdest>& a_OUT, CTensor<T, Csrc>& a_IN);
	a_box	A Box which defines the domain in point space of the operation.

When writing F, the Tensor object may be considered to represent all of the data in a RectMDArray at a given point. This version of forall() assumes that both RectMDArrays are singly indexed (or scalar) arrays. Below is a simple example of a valid function input to forall(). This function permutes the components of a 3 component vector datatype:

  void PermuteF(Tensor<double,3>& a_OUT, CTensor<double, 3>& a_IN) {
    //a_IN(i) is the ith component of the vector data at some point, and similarly with a_OUT(i).
    a_OUT(0) = a_IN(1);
    a_OUT(1) = a_IN(2);
    a_OUT(2) = a_IN(0);
  }

References Box::getLowCorner(), Box::increment(), and Box::notDone().

template<class T, unsigned int C, unsigned char D, unsigned char E, typename Func>

void forall	(	RectMDArray< T, C > &	a_dest,
		const RectMDArray< T, C, D, E > &	a_src,
		const Func &	F,
		const Box &	a_box
	)			[inline]

Pointwise operator. Applies F to the triple indexed RectMDArray a_src in the domain a_box and stores the result in the singly indexed RectMDArray a_dest.

In this case, F has a return value. The maximum of this return value over a_box will be returned by this function.

Template Parameters:

	T	Data type of the RectMDArray arguments
	Cdest	Number of components in the destination RectMDArray
	Csrc	Number of components in the source RectMDArray
	Func	Function datatype. Usually this parameter will be filled with decltype(F).

Parameters:

	a_dest	Destination RectMDArray. This is where the result of F(a_src) is stored.
	a_src	Source RectMDArray. This is the array that F acts on.
	F	A pointwise function with two T-type arrays of size Cdest and Csrc respectively. F can be any externally defined function that bears the following signature: // void F(T destv[Cdest], T srcv[Csrc]); //
	a_box	A Box which defines the domain in point space of the operation. Identical to the forall() function of the same signature except that this version is not implemented with Tensors.

Template Parameters:

	T	Data type of the RectMDArray arguments
	Cdest	Number of components in the destination RectMDArray
	Csrc	Number of components in the source RectMDArray
	Func	Function datatype. Usually this parameter will be filled with decltype(F).

Parameters:

	a_dest	Destination RectMDArray. This is where the result of F(a_src) is stored.
	a_src	Source RectMDArray. This is the array that F acts on.
	F	A pointwise function with two T-type arrays of size Cdest and Csrc respectively. F can be any externally defined function that bears the following signature: // void F(T destv[Cdest], T srcv[Csrc]); //
	a_box	A Box which defines the domain in point space of the operation. Alteration of forall() that operates on a triply indexed RectMDArray. Implemented using Tensors.

Template Parameters:

	T	Data type of the RectMDArray arguments
	C	Number of components in the first index of BOTH the input and output RectMDArrays.
	D	Number of components in the second index of the input RectMDArray.
	E	Number of components in the third index of the input RectMDArray.
	Func	Function datatype. Usually this parameter will be filled with decltype(F).

Parameters:

	a_dest	Destination RectMDArray. This is where the result of F(a_src) is stored.
	a_src	Source RectMDArray. This is the array that F acts on.
	F	A pointwise function with a Tensor and a CTensor as arguments which acts on a_src. F can be any externally defined function that bears the following signature: void F(Tensor<T,C>& a_OUT, CTensor<T,C,D,E>& a_IN);
	a_box	A Box which defines the domain in point space of the operation.

References Box::getLowCorner(), Box::increment(), and Box::notDone().

template<class T, unsigned int Cdest, unsigned int Csrc, unsigned char D, typename Func>

void forall	(	RectMDArray< T, Cdest, D > &	a_dest,
		const RectMDArray< T, Csrc > &	a_src,
		const Func &	F,
		const Box &	a_box
	)			[inline]

Pointwise operator. Applies F to the singly indexed RectMDArray a_src in the domain a_box and stores the result in the doubly indexed RectMDArray a_dest.

Alteration of forall() that operates on a singly indexed RectMDArray and produces a doubly indexed RectMDArray. Implemented using Tensors.
An example of an F that would do this is the calculation of the velocity gradient (a second order tensor) from the velocity (a first order tensor).

Template Parameters:

	T	Data type of the RectMDArray arguments
	Cdest	Number of components in the first index of the output RectMDArray
	Csrc	Number of components in the first index of the input RectMDArray
	D	Number of components in the second index of the output RectMDArray.
	Func	Function datatype. Usually this parameter will be filled with decltype(F).

Parameters:

	a_dest	Destination RectMDArray. This is where the result of F(a_src) is stored.
	a_src	Source RectMDArray. This is the array that F acts on.
	F	A pointwise function with a Tensor and a CTensor as arguments which acts on a_src. F can be any externally defined function that bears the following signature: void F(Tensor<T,C>& a_OUT, CTensor<T,C,D,E>& a_IN);
	a_box	A Box which defines the domain in point space of the operation.

References Box::getLowCorner(), Box::increment(), and Box::notDone().

template<class T, unsigned int C, unsigned char D, typename Func>

void forall	(	RectMDArray< T > &	a_dest,
		const RectMDArray< T, C, D > &	a_src,
		const Func &	F,
		const Box &	a_box
	)			[inline]

Pointwise operator. Applies F to the doubly indexed RectMDArray a_src in the domain a_box and stores the result in the scalar RectMDArray a_dest.

Alteration of forall() that operates on a doubly indexed RectMDArray and produces a scalar RectMDArray. Implemented using Tensors.
An example of an F that would do this is the calculation of the pointwise trace of a 2D matrix valued array.

Template Parameters:

	T	Data type of the RectMDArray arguments
	C	Number of components in the first index of the input RectMDArray
	D	Number of components in the second index of the input RectMDArray.
	Func	Function datatype. Usually this parameter will be filled with decltype(F).

Parameters:

	a_dest	Destination RectMDArray. This is where the result of F(a_src) is stored.
	a_src	Source RectMDArray. This is the array that F acts on.
	F	A pointwise function with a scalar T and a CTensor as arguments which acts on a_src. F can be any externally defined function that bears the following signature: void F(T& a_OUT, CTensor<T,C,D>& a_IN);
	a_box	A Box which defines the domain in point space of the operation.

References RectMDArray< T, C, D, E >::get(), Box::getLowCorner(), Box::increment(), and Box::notDone().

template<class T, typename Func>

void forall	(	RectMDArray< T > &	a_dest,
		const Func &	F,
		const Box &	a_box
	)			[inline]

In place pointwise operator. Applies F to the (scalar) RectMDArray a_dest for every point in a_box.

Alteration of forall() that applies F in place on a scalar RectMDArray. No Tensors are necessary for this version.

Template Parameters:

	T	Data type of the RectMDArray arguments
	Func	Function datatype. Usually this parameter will be filled with decltype(F).

Parameters:

	a_dest	Destination RectMDArray. This is where the result of F(a_src) is stored.
	F	A pointwise function with a Tensor and a CTensor as arguments which acts on a_src. F can be any externally defined function that bears the following signature: void F(T& a_IN);
	a_box	A Box which defines the domain in point space of the operation.

References RectMDArray< T, C, D, E >::get(), Box::getLowCorner(), Box::increment(), and Box::notDone().

template<class T>

T& abs_max	(	RectMDArray< T > &	a_src,
		const Box &	a_box
	)			[inline]

Returns the value of a_src in a_box with the largest absolute value.

Template Parameters:

T

Data type of a_src

Parameters:

	a_src	Input data
	a_box	Domain

References Box::getLowCorner(), Box::increment(), and Box::notDone().

template<class T>

tuple<T&, Point> abs_argmax	(	const RectMDArray< T > &	a_src,
		const Box &	a_box
	)			[inline]

Returns a tuple containing the value of a_src in a_box with the largest absolute value and the Point at which this value is attained.

Template Parameters:

T

Data type of a_src

Parameters:

	a_src	Input data
	a_box	Domain

References Box::getLowCorner(), Box::increment(), and Box::notDone().