Report a bug
If you spot a problem with this page, click here to create a Bugzilla issue.
Improve this page
Quickly fork, edit online, and submit a pull request for this page. Requires a signed-in GitHub account. This works well for small changes. If you'd like to make larger changes you may want to consider using a local clone.

std.experimental.ndslice.selection

This is a submodule of std.experimental.ndslice.
Selectors create new views and iteration patterns over the same data, without copying.

Subspace selectors

Subspace selectors serve to generalize and combine other selectors easily. For a slice of Slice!(N, Range) type slice.pack!K creates a slice of slices of Slice!(N-K, Slice!(K+1, Range)) type by packing the last K dimensions of the top dimension pack, and the type of element of slice.byElement is Slice!(K, Range). Another way to use pack is transposition of dimension packs using evertPack. Examples of use of subspace selectors are available for selectors, Slice.shape , and Slice.elementsCount .
Function Name Description
pack returns slice of slices
unpack merges all dimension packs
evertPack reverses dimension packs

Selectors

Function Name Description
blocks n-dimensional slice composed of n-dimensional non-overlapping blocks. If the slice has two dimensions, it is a block matrix.
byElement flat, random access range of all elements with index property
byElementInStandardSimplex an input range of all elements in standard simplex of hypercube with index property. If the slice has two dimensions, it is a range of all elements of upper left triangular matrix.
diagonal 1-dimensional slice composed of diagonal elements
indexSlice lazy slice with initial multidimensional index
iotaSlice lazy slice with initial flattened (continuous) index
mapSlice lazy multidimensional functional map
repeatSlice slice with identical values
reshape new slice with changed dimensions for the same data
windows n-dimensional slice of n-dimensional overlapping windows. If the slice has two dimensions, it is a sliding window.
Authors:
Ilya Yaroshenko
template pack(K...)
Creates a packed slice, i.e. slice of slices. The function does not carry out any calculations, it simply returns the same binary data presented differently.
Parameters:
K sizes of dimension packs
Returns:
pack!K returns Slice!(N-K, Slice!(K+1, Range)); slice.pack!(K1, K2, ..., Kn) is the same as slice.pack!K1.pack!K2. ... pack!Kn.
Examples:
import std.experimental.ndslice : sliced, Slice, pack;
import std.range : iota;

auto r = (3 * 4 * 5 * 6).iota;
auto a = r.sliced(3, 4, 5, 6);
auto b = a.pack!2;

static immutable res1 = [3, 4];
static immutable res2 = [5, 6];
assert(b.shape == res1);
assert(b[0, 0].shape == res2);
assert(a == b);
static assert(is(typeof(b) == typeof(a.pack!2)));
static assert(is(typeof(b) == Slice!(2, Slice!(3, typeof(r)))));
Slice!(N, Range).PureThis unpack(size_t N, Range)(Slice!(N, Range) slice);
Unpacks a packed slice.
The function does not carry out any calculations, it simply returns the same binary data presented differently.
Parameters:
Slice!(N, Range) slice packed slice
Returns:
unpacked slice
See Also:
Examples:
auto a = iotaSlice(3, 4, 5, 6, 7, 8, 9, 10, 11);
auto b = a.pack!(2, 3).unpack();
static assert(is(typeof(a) == typeof(b)));
assert(a == b);
SliceFromSeq!(Slice!(N, Range).PureRange, NSeqEvert!(Slice!(N, Range).NSeq)) evertPack(size_t N, Range)(Slice!(N, Range) slice);
Reverses the order of dimension packs. This function is used in a functional pipeline with other selectors.
Parameters:
Slice!(N, Range) slice packed slice
Returns:
packed slice
See Also:
Examples:
import std.experimental.ndslice.iteration : transposed;
auto slice = iotaSlice(3, 4, 5, 6, 7, 8, 9, 10, 11);
assert(slice
    .pack!2
    .evertPack
    .unpack
         == slice.transposed!(
            slice.shape.length-2,
            slice.shape.length-1));
Examples:
import std.experimental.ndslice.slice;
import std.experimental.ndslice.iteration : transposed;
import std.range.primitives : ElementType;
import std.range : iota;
import std.algorithm.comparison : equal;
auto r = (3 * 4 * 5 * 6 * 7 * 8 * 9 * 10 * 11).iota;
auto a = r.sliced(3, 4, 5, 6, 7, 8, 9, 10, 11);
auto b = a
    .pack!(2, 3)
    .evertPack;
auto c = b[8, 9];
auto d = c[5, 6, 7];
auto e = d[1, 2, 3, 4];
auto g = a[1, 2, 3, 4, 5, 6, 7, 8, 9];
assert(e == g);
assert(a == b.evertPack);
assert(c == a.transposed!(7, 8, 4, 5, 6)[8, 9]);
alias R = typeof(r);
static assert(is(typeof(b) == Slice!(2, Slice!(4, Slice!(5, R)))));
static assert(is(typeof(c) == Slice!(3, Slice!(5, R))));
static assert(is(typeof(d) == Slice!(4, R)));
static assert(is(typeof(e) == ElementType!R));
Slice!(1, Range) diagonal(size_t N, Range)(Slice!(N, Range) slice);
Returns a 1-dimensional slice over the main diagonal of an n-dimensional slice. diagonal can be generalized with other selectors such as blocks (diagonal blocks) and windows (multi-diagonal slice).
Parameters:
N dimension count
Slice!(N, Range) slice input slice
Returns:
1-dimensional slice composed of diagonal elements
Examples:
Matrix, main diagonal
//  -------
// | 0 1 2 |
// | 3 4 5 |
//  -------
//->
// | 0 4 |
static immutable d = [0, 4];
assert(iotaSlice(2, 3).diagonal == d);
Examples:
Non-square matrix
import std.algorithm.comparison : equal;
import std.range : only;

//  -------
// | 0 1 |
// | 2 3 |
// | 4 5 |
//  -------
//->
// | 0 3 |

assert(iotaSlice(3, 2)
    .diagonal
    .equal(only(0, 3)));
Examples:
Loop through diagonal
import std.experimental.ndslice.slice;

auto slice = slice!int(3, 3);
int i;
foreach (ref e; slice.diagonal)
    e = ++i;
assert(slice == [
    [1, 0, 0],
    [0, 2, 0],
    [0, 0, 3]]);
Examples:
Matrix, subdiagonal
import std.experimental.ndslice.iteration : dropOne;
//  -------
// | 0 1 2 |
// | 3 4 5 |
//  -------
//->
// | 1 5 |
static immutable d = [1, 5];
assert(iotaSlice(2, 3).dropOne!1.diagonal == d);
Examples:
Matrix, antidiagonal
import std.experimental.ndslice.iteration : dropToHypercube, reversed;
//  -------
// | 0 1 2 |
// | 3 4 5 |
//  -------
//->
// | 1 3 |
static immutable d = [1, 3];
assert(iotaSlice(2, 3).dropToHypercube.reversed!1.diagonal == d);
Examples:
3D, main diagonal
//  -----------
// |  0   1  2 |
// |  3   4  5 |
//  - - - - - -
// |  6   7  8 |
// |  9  10 11 |
//  -----------
//->
// | 0 10 |
static immutable d = [0, 10];
assert(iotaSlice(2, 2, 3).diagonal == d);
Examples:
3D, subdiagonal
import std.experimental.ndslice.iteration : dropOne;
//  -----------
// |  0   1  2 |
// |  3   4  5 |
//  - - - - - -
// |  6   7  8 |
// |  9  10 11 |
//  -----------
//->
// | 1 11 |
static immutable d = [1, 11];
assert(iotaSlice(2, 2, 3).dropOne!2.diagonal == d);
Examples:
3D, diagonal plain
//  -----------
// |  0   1  2 |
// |  3   4  5 |
// |  6   7  8 |
//  - - - - - -
// |  9  10 11 |
// | 12  13 14 |
// | 15  16 17 |
//  - - - - - -
// | 18  20 21 |
// | 22  23 24 |
// | 24  25 26 |
//  -----------
//->
//  -----------
// |  0   4  8 |
// |  9  13 17 |
// | 18  23 26 |
//  -----------

static immutable d =
    [[ 0,  4,  8],
     [ 9, 13, 17],
     [18, 22, 26]];

auto slice = iotaSlice(3, 3, 3)
    .pack!2
    .evertPack
    .diagonal
    .evertPack;

assert(slice == d);
Slice!(N, Slice!(N + 1, Range)) blocks(size_t N, Range)(Slice!(N, Range) slice, size_t[N] lengths...);
Returns an n-dimensional slice of n-dimensional non-overlapping blocks. blocks can be generalized with other selectors. For example, blocks in combination with diagonal can be used to get a slice of diagonal blocks.
Parameters:
N dimension count
Slice!(N, Range) slice slice to be split into blocks
size_t[N] lengths dimensions of block, residual blocks are ignored
Returns:
packed N-dimensional slice composed of N-dimensional slices
Examples:
import std.experimental.ndslice.slice;
auto slice = slice!int(5, 8);
auto blocks = slice.blocks(2, 3);
int i;
foreach (block; blocks.byElement)
    block[] = ++i;

assert(blocks ==
    [[[[1, 1, 1], [1, 1, 1]],
      [[2, 2, 2], [2, 2, 2]]],
     [[[3, 3, 3], [3, 3, 3]],
      [[4, 4, 4], [4, 4, 4]]]]);

assert(    slice ==
    [[1, 1, 1,  2, 2, 2,  0, 0],
     [1, 1, 1,  2, 2, 2,  0, 0],

     [3, 3, 3,  4, 4, 4,  0, 0],
     [3, 3, 3,  4, 4, 4,  0, 0],

     [0, 0, 0,  0, 0, 0,  0, 0]]);
Examples:
Diagonal blocks
import std.experimental.ndslice.slice;
auto slice = slice!int(5, 8);
auto blocks = slice.blocks(2, 3);
auto diagonalBlocks = blocks.diagonal.unpack;

diagonalBlocks[0][] = 1;
diagonalBlocks[1][] = 2;

assert(diagonalBlocks ==
    [[[1, 1, 1], [1, 1, 1]],
     [[2, 2, 2], [2, 2, 2]]]);

assert(blocks ==
    [[[[1, 1, 1], [1, 1, 1]],
      [[0, 0, 0], [0, 0, 0]]],
     [[[0, 0, 0], [0, 0, 0]],
      [[2, 2, 2], [2, 2, 2]]]]);

assert(slice ==
    [[1, 1, 1,  0, 0, 0,  0, 0],
     [1, 1, 1,  0, 0, 0,  0, 0],

     [0, 0, 0,  2, 2, 2,  0, 0],
     [0, 0, 0,  2, 2, 2,  0, 0],

     [0, 0, 0, 0, 0, 0, 0, 0]]);
Examples:
Matrix divided into vertical blocks
import std.experimental.ndslice.slice;
auto slice = slice!int(5, 13);
auto blocks = slice
    .pack!1
    .evertPack
    .blocks(3)
    .unpack
    .pack!2;

int i;
foreach (block; blocks.byElement)
    block[] = ++i;

assert(slice ==
    [[1, 1, 1,  2, 2, 2,  3, 3, 3,  4, 4, 4,  0],
     [1, 1, 1,  2, 2, 2,  3, 3, 3,  4, 4, 4,  0],
     [1, 1, 1,  2, 2, 2,  3, 3, 3,  4, 4, 4,  0],
     [1, 1, 1,  2, 2, 2,  3, 3, 3,  4, 4, 4,  0],
     [1, 1, 1,  2, 2, 2,  3, 3, 3,  4, 4, 4,  0]]);
Slice!(N, Slice!(N + 1, Range)) windows(size_t N, Range)(Slice!(N, Range) slice, size_t[N] lengths...);
Returns an n-dimensional slice of n-dimensional overlapping windows. windows can be generalized with other selectors. For example, windows in combination with diagonal can be used to get a multi-diagonal slice.
Parameters:
N dimension count
Slice!(N, Range) slice slice to be iterated
size_t[N] lengths dimensions of windows
Returns:
packed N-dimensional slice composed of N-dimensional slices
Examples:
import std.experimental.ndslice.slice;
auto slice = slice!int(5, 8);
auto windows = slice.windows(2, 3);
foreach (window; windows.byElement)
    window[] += 1;

assert(slice ==
    [[1,  2,  3, 3, 3, 3,  2,  1],

     [2,  4,  6, 6, 6, 6,  4,  2],
     [2,  4,  6, 6, 6, 6,  4,  2],
     [2,  4,  6, 6, 6, 6,  4,  2],

     [1,  2,  3, 3, 3, 3,  2,  1]]);
Examples:
import std.experimental.ndslice.slice;
auto slice = slice!int(5, 8);
auto windows = slice.windows(2, 3);
windows[1, 2][] = 1;
windows[1, 2][0, 1] += 1;
windows.unpack[1, 2, 0, 1] += 1;

assert(slice ==
    [[0, 0,  0, 0, 0,  0, 0, 0],

     [0, 0,  1, 3, 1,  0, 0, 0],
     [0, 0,  1, 1, 1,  0, 0, 0],

     [0, 0,  0, 0, 0,  0, 0, 0],
     [0, 0,  0, 0, 0,  0, 0, 0]]);
Examples:
Multi-diagonal matrix
import std.experimental.ndslice.slice;
auto slice = slice!int(8, 8);
auto windows = slice.windows(3, 3);

auto multidiagonal = windows
    .diagonal
    .unpack;
foreach (window; multidiagonal)
    window[] += 1;

assert(slice ==
    [[ 1, 1, 1,  0, 0, 0, 0, 0],
     [ 1, 2, 2, 1,  0, 0, 0, 0],
     [ 1, 2, 3, 2, 1,  0, 0, 0],
     [0,  1, 2, 3, 2, 1,  0, 0],
     [0, 0,  1, 2, 3, 2, 1,  0],
     [0, 0, 0,  1, 2, 3, 2, 1],
     [0, 0, 0, 0,  1, 2, 2, 1],
     [0, 0, 0, 0, 0,  1, 1, 1]]);
Examples:
Sliding window over matrix columns
import std.experimental.ndslice.slice;
auto slice = slice!int(5, 8);
auto windows = slice
    .pack!1
    .evertPack
    .windows(3)
    .unpack
    .pack!2;


foreach (window; windows.byElement)
    window[] += 1;

assert(slice ==
    [[1,  2,  3, 3, 3, 3,  2,  1],
     [1,  2,  3, 3, 3, 3,  2,  1],
     [1,  2,  3, 3, 3, 3,  2,  1],
     [1,  2,  3, 3, 3, 3,  2,  1],
     [1,  2,  3, 3, 3, 3,  2,  1]]);
Slice!(M, Range) reshape(size_t N, Range, size_t M)(Slice!(N, Range) slice, size_t[M] lengths...);
Returns a new slice for the same data with different dimensions.
Parameters:
Slice!(N, Range) slice slice to be reshaped
size_t[M] lengths list of new dimensions. One of the lengths can be set to -1. In this case, the corresponding dimension is inferable.
Returns:
reshaped slice
Throws:
ReshapeException if the slice cannot be reshaped with the input lengths.
Examples:
import std.experimental.ndslice.iteration : allReversed;
auto slice = iotaSlice(3, 4)
    .allReversed
    .reshape(-1, 3);
assert(slice ==
    [[11, 10, 9],
     [ 8,  7, 6],
     [ 5,  4, 3],
     [ 2,  1, 0]]);
Examples:
Reshaping with memory allocation
import std.experimental.ndslice.slice;
import std.experimental.ndslice.iteration : reversed;
import std.array : array;

auto reshape2(S, size_t M)(S slice, size_t[M] lengths...)
{
    // Tries to reshape without allocation
    try return slice.reshape(lengths);
    catch (ReshapeException e)
        // Allocates
        return slice.slice.reshape(lengths);
}

auto slice =
    [0, 1,  2,  3,
     4, 5,  6,  7,
     8, 9, 10, 11]
    .sliced(3, 4)
    .reversed!0;

assert(reshape2(slice, 4, 3) ==
    [[ 8, 9, 10],
     [11, 4,  5],
     [ 6, 7,  0],
     [ 1, 2,  3]]);
class ReshapeException: std.experimental.ndslice.slice.SliceException;
See Also:
size_t[] lengths;
Old lengths
sizediff_t[] strides;
Old strides
size_t[] newLengths;
New lengths
pure nothrow @nogc @safe this(size_t[] lengths, sizediff_t[] strides, size_t[] newLengths, string msg, string file = __FILE__, uint line = cast(uint)904, Throwable next = null);
auto byElement(size_t N, Range)(Slice!(N, Range) slice);

Slice!(1, Range) byElement(size_t N : 1, Range)(Slice!(N, Range) slice);
Returns a random access range of all elements of a slice. The order of elements is preserved. byElement can be generalized with other selectors.
Parameters:
N dimension count
Slice!(N, Range) slice slice to be iterated
Returns:
random access range composed of elements of the slice
Examples:
Regular slice
import std.algorithm.comparison : equal;
import std.range : iota;
assert(iotaSlice(4, 5)
    .byElement
    .equal(20.iota));
Examples:
Packed slice
import std.experimental.ndslice.slice;
import std.experimental.ndslice.iteration;
import std.range : drop;
assert(iotaSlice(3, 4, 5, 6, 7)
    .pack!2
    .byElement()
    .drop(1)
    .front
     == iotaSlice([6, 7], 6 * 7));
Examples:
Properties
auto elems = iotaSlice(3, 4).byElement;

elems.popFrontExactly(2);
assert(elems.front == 2);
assert(elems.index == [0, 2]);

elems.popBackExactly(2);
assert(elems.back == 9);
assert(elems.length == 8);
Examples:
Index property
import std.experimental.ndslice.slice;
auto slice = new long[20].sliced(5, 4);

for (auto elems = slice.byElement; !elems.empty; elems.popFront)
{
    size_t[2] index = elems.index;
    elems.front = index[0] * 10 + index[1] * 3;
}
assert(slice ==
    [[ 0,  3,  6,  9],
     [10, 13, 16, 19],
     [20, 23, 26, 29],
     [30, 33, 36, 39],
     [40, 43, 46, 49]]);
Examples:
Random access and slicing
import std.experimental.ndslice.slice;
import std.algorithm.comparison : equal;
import std.array : array;
import std.range : iota, repeat;
static data = 20.iota.array;
auto elems = data.sliced(4, 5).byElement;

elems = elems[11 .. $ - 2];

assert(elems.length == 7);
assert(elems.front == 11);
assert(elems.back == 17);

foreach (i; 0 .. 7)
    assert(elems[i] == i + 11);

// assign an element
elems[2 .. 6] = -1;
assert(elems[2 .. 6].equal(repeat(-1, 4)));

// assign an array
static ar = [-1, -2, -3, -4];
elems[2 .. 6] = ar;
assert(elems[2 .. 6].equal(ar));

// assign a slice
ar[] *= 2;
auto sl = ar.sliced(ar.length);
elems[2 .. 6] = sl;
assert(elems[2 .. 6].equal(sl));
Examples:
Forward access works faster than random access or backward access. Use allReversed  in pipeline before byElement to achieve fast backward access.
import std.range : retro;
import std.experimental.ndslice.iteration : allReversed;

auto slice = iotaSlice(3, 4, 5);

/// Slow backward iteration #1
foreach (ref e; slice.byElement.retro)
{
    //...
}

/// Slow backward iteration #2
foreach_reverse (ref e; slice.byElement)
{
    //...
}

/// Fast backward iteration
foreach (ref e; slice.allReversed.byElement)
{
    //...
}
Examples:
byElement just return the slice for 1-dimensional case.
import std.experimental.ndslice.slice : slice;
static assert(is(typeof(slice!int(100)) == Slice!(1, int*)));
auto byElementInStandardSimplex(size_t N, Range)(Slice!(N, Range) slice, size_t maxHypercubeLength = size_t.max);

Slice!(1, Range) byElementInStandardSimplex(size_t N : 1, Range)(Slice!(N, Range) slice, size_t maxHypercubeLength = size_t.max);
Returns an forward range of all elements of standard simplex of a slice. In case the slice has two dimensions, it is composed of elements of upper left triangular matrix. The order of elements is preserved. byElementInStandardSimplex can be generalized with other selectors.
Parameters:
N dimension count
Slice!(N, Range) slice slice to be iterated
size_t maxHypercubeLength maximal length of simplex hypercube.
Returns:
forward range composed of all elements of standard simplex of the slice
Examples:
import std.experimental.ndslice.slice;
auto slice = slice!int(4, 5);
auto elems = slice
    .byElementInStandardSimplex;
int i;
foreach (ref e; elems)
    e = ++i;
assert(slice ==
    [[ 1, 2, 3, 4, 0],
     [ 5, 6, 7, 0, 0],
     [ 8, 9, 0, 0, 0],
     [10, 0, 0, 0, 0]]);
Examples:
import std.experimental.ndslice.slice;
import std.experimental.ndslice.iteration;
auto slice = slice!int(4, 5);
auto elems = slice
    .transposed
    .allReversed
    .byElementInStandardSimplex;
int i;
foreach (ref e; elems)
    e = ++i;
assert(slice ==
    [[0,  0, 0, 0, 4],
     [0,  0, 0, 7, 3],
     [0,  0, 9, 6, 2],
     [0, 10, 8, 5, 1]]);
Examples:
Properties
import std.range.primitives : popFrontN;

auto elems = iotaSlice(3, 4).byElementInStandardSimplex;

elems.popFront;
assert(elems.front == 1);
assert(elems.index == cast(size_t[2])[0, 1]);

elems.popFrontN(3);
assert(elems.front == 5);
Examples:
Save
auto elems = iotaSlice(3, 4).byElementInStandardSimplex;
import std.range : dropOne, popFrontN;
elems.popFrontN(4);

assert(elems.save.dropOne.front == 8);
assert(elems.front == 5);
assert(elems.index == cast(size_t[2])[1, 1]);
assert(elems.length == 2);
Examples:
byElementInStandardSimplex returns a corresponding part of the slice for 1-dimensional case.
auto sl = iotaSlice(7);
auto simplex = sl.byElementInStandardSimplex;

assert(simplex == [0, 1, 2, 3, 4, 5, 6]);

static assert(is(typeof(simplex) == typeof(sl)));

assert(sl.byElementInStandardSimplex(3) == [0, 1, 2]);
IndexSlice!N indexSlice(size_t N)(size_t[N] lengths...);
Returns a slice, the elements of which are equal to the initial multidimensional index value. This is multidimensional analog of std.range.iota. For a flattened (continuous) index, see iotaSlice.
Parameters:
N dimension count
size_t[N] lengths list of dimension lengths
Returns:
N-dimensional slice composed of indexes
Examples:
auto slice = indexSlice(2, 3);
static immutable array =
    [[[0, 0], [0, 1], [0, 2]],
     [[1, 0], [1, 1], [1, 2]]];

assert(slice == array);


static assert(is(IndexSlice!2 : Slice!(2, Range), Range));
static assert(is(DeepElementType!(IndexSlice!2) == size_t[2]));
Examples:
auto im = indexSlice(7, 9);

assert(im[2, 1] == [2, 1]);

//slicing works correctly
auto cm = im[1 .. $, 4 .. $];
assert(cm[2, 1] == [3, 5]);
template IndexSlice(size_t N) if (N)
Slice composed of indexes.
See Also:
IotaSlice!N iotaSlice(size_t N)(size_t[N] lengths...);

IotaSlice!N iotaSlice(size_t N)(size_t[N] lengths, size_t shift);

IotaSlice!N iotaSlice(size_t N)(size_t[N] lengths, size_t shift, size_t step);
Returns a slice, the elements of which are equal to the initial flattened index value. For a multidimensional index, see indexSlice.
Parameters:
N dimension count
size_t[N] lengths list of dimension lengths
size_t shift value of the first element in a slice (optional)
size_t step value of the step between elements (optional)
Returns:
N-dimensional slice composed of indexes
Examples:
auto slice = iotaSlice(2, 3);
static immutable array =
    [[0, 1, 2],
     [3, 4, 5]];

assert(slice == array);

import std.range.primitives : isRandomAccessRange;
static assert(isRandomAccessRange!(IotaSlice!2));
static assert(is(IotaSlice!2 : Slice!(2, Range), Range));
static assert(is(DeepElementType!(IotaSlice!2) == size_t));
Examples:
auto im = iotaSlice([10, 5], 100);

assert(im[2, 1] == 111); // 100 + 2 * 5 + 1

//slicing works correctly
auto cm = im[1 .. $, 3 .. $];
assert(cm[2, 1] == 119); // 119 = 100 + (1 + 2) * 5 + (3 + 1)
Examples:
iotaSlice with step
auto sl = iotaSlice([2, 3], 10, 10);

assert(sl == [[10, 20, 30],
              [40, 50, 60]]);
template IotaSlice(size_t N) if (N)
Slice composed of flattened indexes.
See Also:
RepeatSlice!(M, T) repeatSlice(T, size_t M)(T value, size_t[M] lengths...)
if (!is(T : Slice!(N, Range), size_t N, Range));

Slice!(M, Slice!(N + 1, Range)) repeatSlice(size_t N, Range, size_t M)(Slice!(N, Range) slice, size_t[M] lengths...);
Returns a slice with identical elements. RepeatSlice stores only single value.
Parameters:
size_t[M] lengths list of dimension lengths
Returns:
n-dimensional slice composed of identical values, where n is dimension count.
Examples:
auto sl = iotaSlice(3)
    .repeatSlice(4);
assert(sl == [[0, 1, 2],
              [0, 1, 2],
              [0, 1, 2],
              [0, 1, 2]]);
Examples:
import std.experimental.ndslice.iteration : transposed;

auto sl = iotaSlice(3)
    .repeatSlice(4)
    .unpack
    .transposed;

assert(sl == [[0, 0, 0, 0],
              [1, 1, 1, 1],
              [2, 2, 2, 2]]);
Examples:
import std.experimental.ndslice.slice : slice;

auto sl = iotaSlice([3], 6).slice;
auto slC = sl.repeatSlice(2, 3);
sl[1] = 4;
assert(slC == [[[6, 4, 8],
                [6, 4, 8],
                [6, 4, 8]],
               [[6, 4, 8],
                [6, 4, 8],
                [6, 4, 8]]]);
Examples:
auto sl = repeatSlice(4.0, 2, 3);
assert(sl == [[4.0, 4.0, 4.0],
              [4.0, 4.0, 4.0]]);

static assert(is(DeepElementType!(typeof(sl)) == double));

sl[1, 1] = 3;
assert(sl == [[3.0, 3.0, 3.0],
              [3.0, 3.0, 3.0]]);
template RepeatSlice(size_t N, T) if (N)
Slice composed of identical values.
template mapSlice(fun...) if (fun.length)
Implements the homonym function (also known as transform) present in many languages of functional flavor. The call mapSlice!(fun)(tensor) returns a tensor of which elements are obtained by applying fun for all elements in tensor. The original tensors are not changed. Evaluation is done lazily.

Note: transposed  and pack  can be used to specify dimensions.

Parameters:
fun One or more functions.
Slice!(N, Range) tensor An input tensor.
Returns:
a tensor with each fun applied to all the elements. If there is more than one fun, the element type will be Tuple containing one element for each fun.
Examples:
import std.experimental.ndslice.selection : iotaSlice;

auto s = iotaSlice(2, 3).mapSlice!(a => a * 3);
assert(s == [[ 0,  3,  6],
             [ 9, 12, 15]]);
Examples:
Packed tensors.
import std.experimental.ndslice.selection : iotaSlice, windows;

//  iotaSlice        windows     mapSlice  sums ( ndFold!"a + b" )
//                --------------
//  -------      |  ---    ---  |      ------
// | 0 1 2 |  => || 0 1 || 1 2 ||  => | 8 12 |
// | 3 4 5 |     || 3 4 || 4 5 ||      ------
//  -------      |  ---    ---  |
//                --------------
auto s = iotaSlice(2, 3)
    .windows(2, 2)
    .mapSlice!((a) {
        size_t s;
        foreach (r; a)
            foreach (e; r)
                s += e;
        return s;
        });

assert(s == [[8, 12]]);
Examples:
Zipped tensors
import std.experimental.ndslice.slice : assumeSameStructure;
import std.experimental.ndslice.selection : iotaSlice;

// 0 1 2
// 3 4 5
auto sl1 = iotaSlice(2, 3);
// 1 2 3
// 4 5 6
auto sl2 = iotaSlice([2, 3], 1);

// tensors must have the same strides
assert(sl1.structure == sl2.structure);

auto zip = assumeSameStructure!("a", "b")(sl1, sl2);

auto lazySum = zip.mapSlice!(z => z.a + z.b);

assert(lazySum == [[ 1,  3,  5],
                   [ 7,  9, 11]]);
Examples:
Multiple functions can be passed to mapSlice. In that case, the element type of mapSlice is a tuple containing one element for each function.
import std.experimental.ndslice.selection : iotaSlice;

auto s = iotaSlice(2, 3).mapSlice!("a + a", "a * a");

auto sums     = [[0, 2, 4], [6,  8, 10]];
auto products = [[0, 1, 4], [9, 16, 25]];

foreach (i; 0..s.length!0)
foreach (j; 0..s.length!1)
{
    auto values = s[i, j];
    assert(values[0] == sums[i][j]);
    assert(values[1] == products[i][j]);
}
Examples:
You may alias mapSlice with some function(s) to a symbol and use it separately:
import std.conv : to;
import std.experimental.ndslice.selection : iotaSlice;

alias stringize = mapSlice!(to!string);
assert(stringize(iotaSlice(2, 3)) == [["0", "1", "2"], ["3", "4", "5"]]);
auto mapSlice(size_t N, Range)(Slice!(N, Range) tensor);