MatrixOperator

  • namespace: Rindow\Math\Matrix
  • classname: MatrixOperator

General operations on N-dimensional arrays. If you have loaded the “OpenBLAS”, enable it.

Methods

constructor

public function __construct($blas=null,$math=null)

Arguments

  • blas: blas library instance.
    • Default is Rindow\Math\Matrix\PhpBlas
  • math: math library instance.
    • Default is Rindow\Math\Matrix\PhpMath

Example

$mo = new Rindow\Math\Matrix\MatrixOperator();

setDefaultIntType

public function setDefaultIntType($dtype)

Set default of Integer type of NDArray

If not set, the default is int32.

Example

use Interop\Polite\Math\Matrix\NDArray;
$mo->setDefaultIntType(NDArray::int32);

setDefaultFloatType

public function setDefaultFloatType($dtype)

Set default of Float type of NDArray

If not set, the default is float32.

Example

use Interop\Polite\Math\Matrix\NDArray;
$mo->setDefaultFloatType(NDArray::float32);

array

public function array($array,$dtype=null) : NDArray

NDArray with value specified by PHP array type.

Arguments

  • array: PHP array.
  • dtype: data type.
    • If omitted, default float type is used.

Example

$a = $mo->array([[1,2],[3,4]]);
## a => [[1.0, 2.0], [3.0, 4.0]]  ... float32

$a = $mo->array([[1,2],[3,4]], NDArray::int32);
## a => [[1, 2], [3, 4]]  ... int32

$a = $mo->array([true,false], NDArray::bool);
## a => [true,false]  ... bool

zeros

public function zeros(array $shape, $dtype=null) : NDArray

Create an NDArray of the shape specified by “shape” and fill it with zeros.

Arguments

  • shape: NDArray shape.
  • dtype: data type.
    • If omitted, default float type is used.

Example

$a = $mo->zeros([2,3]);
## a => [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0]]  ... float32

$a = $mo->zeros([2,3],NDArray::int32);
## a => [[0, 0, 0], [0, 0, 0]]  ... int32

ones

public function ones(array $shape, $dtype=null) : NDArray

Create an NDArray of the shape specified by “shape” and fill it with ones.

Arguments

  • shape: NDArray shape.
  • dtype: data type.
    • If omitted, default float type is used.

Example

$a = $mo->zeros([2,3]);
## a => [[1.0, 1.0, 1.0], [1.0, 1.0, 1.0]]  ... float32

$a = $mo->zeros([2,3],NDArray::int32);
## a => [[1, 1, 1], [1, 1, 1]]  ... int32

full

public function full(array $shape, $value, $dtype=null) : NDArray

Create an NDArray of the shape specified by “shape” and fill it with “value”.

Arguments

  • shape: NDArray shape.
  • value: Fill value
  • dtype: data type.
    • If omitted, According to the specified “value” type.

Example

$a = $mo->full([2,3],2.0);
## a => [[2.0, 2.0, 2.0], [2.0, 2.0, 2.0]]  ... float32
$a = $mo->full([2,3],3);
## a => [[3, 3, 3], [3, 3, 3]]  ... int32
$a = $mo->full([2,3],false);
## a => [[false, false, false], [false, false, false]]  ... bool

$a = $mo->full([2,3],2.0,NDArray::float64);
## a => [[2.0, 2.0, 2.0], [2.0, 2.0, 2.0]]  ... float64
$a = $mo->full([2,3],3,NDArray::uint8);
## a => [[3, 3, 3], [3, 3, 3]]  ... uint8

zerosLike

public function zerosLike(NDArray $array)

Create an NDArray with the same shape as the specified NDArray and fill it with zeros

Arguments

  • array: NDArray.

Example

$org = $mo->array([[1,2],[3,4]]);
$a = $mo->zerosLike($org);
## a => [[0.0, 0.0], [0.0, 0.0]]  ... float32

$org = $mo->array([[1,2],[3,4]], NDArray::int32);
$a = $mo->zerosLike($org);
## a => [[0, 0], [0, 0]]  ... int32

fullLike

public function fullLike(NDArray $array,$value)

Create an NDArray with the same shape as the specified NDArray and fill it with “value”

Arguments

  • array: NDArray.

Example

$org = $mo->array([[1,2],[3,4]]);
$a = $mo->fullLike($org, 1);
## a => [[1.0, 1.0], [1.0, 1.0]]  ... float32

$org = $mo->array([[1,2],[3,4]], NDArray::int32);
$a = $mo->fullLike($org, 1);
## a => [[1, 1], [1, 1]]  ... int32

astype

public function astype(NDArray $array, $dtype) : NDArray

Create NDArray with changed data type, convert array elements and copy

Arguments

  • array: NDArray of copy source.
  • dtype: Data type to be converted

Example

$org = $mo->array([[1,2],[3,4]],NDArray:int32);
$a = $mo->astype($org, NDArray::float32);
## a => [[1.0, 2.0], [3.0, 4.0]]  ... float32

arange

public function arange(int $count ,$start=null, $step=null, $dtype=null) : NDArray

Create an NDArray containing consecutive values.

Note that it is different from numpy.

Arguments

  • count: Element count.
  • start: Starting value
    • Default is 0
  • step: incremental
    • Default is 1
  • dtype: Data type of NDArray
    • Default depends on start value

Example

$a = $mo->arange(3);
## a => [0, 1, 2]  ... int32
$a = $mo->arange(3, 0.0);
## a => [0.0, 1.0, 2.0]  ... float32

$a = $mo->arange(3, 1, 2);
## a => [1, 3, 5]  ... int32
$a = $mo->arange(3, 0, 10, NDArray::float32);
## a => [0.0, 10.0, 20.0]  ... float32

copy

public function copy(NDArray $array) : NDArray

Copies the NDArray and the internal buffer instance.

Arguments

  • array: NDArray of copy source.

Example

$org = $mo->array([[1,2],[3,4]]);
$a = $mo->copy($org);
## a => [[1.0, 2.0], [3.0, 4.0]]  ... float32

cross

public function cross(NDArray $x, NDArray $y) : NDArray

Calculates the cross product.

Operations on arrays of three or more dimensions conform to numpy. See the section “Dimension Operations” for details.

Arguments

  • x: Left matrix
  • y: Right matrix

Example

$x = $mo->array([[1,2],[3,4]]);
$y = $mo->array([[1,0],[0,1]]);
$z = $mo->cross($x,$y);
## z => [[1.0, 2.0], [3.0, 4.0]]

dot

public function dot(NDArray $x, NDArray $y)

Calculates the dot product. See the section “Dimension Operations” for details.

Example

$a = $mo->array([1,2,3,4,5,6]);
$b = $mo->array([3,4,5,6,7,8]);
$c = $mo->dot($a,$b);
## c ==> 133.0

transpose

public function transpose(NDArray $x) : NDArray

Get the transpose matrix.

Example

$x = $mo->array([[1,2],[3,4],[5,6]]);
$y = $mo->transpose($x);
## y => [[1.0, 3.0, 5.0],[2.0, 4.0, 6.0]]

add

public function add(NDArray $x, NDArray $y) : NDArray

Sum of Matrix

Example

$x = $mo->array([[1,2],[3,4]]);
$y = $mo->array([[10,20],[30,40]]);
$z = $mo->add($x,$y);
## z => [[11.0, 22.0],[33.0, 44.0]]

scale

public function scale($a, NDArray $X) : NDArray

Scale of Matrix

Example

$x = $mo->array([[1,2],[3,4]]);
$y = $mo->add(10,$x);
## y => [[10.0, 20.0],[30.0, 40.0]]

sum

public function sum(NDArray $x,int $axis=null)

Sum of matrix elements.

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,2],[3,4]]);
$y = $mo->sum($x);
## y => 10.0
$y = $mo->sum($x,0);
## y => [4.0, 6.0]
$y = $mo->sum($x,1);
## y => [3.0, 7.0]

sum

public function asum(NDArray $x,int $axis=null)

Sum of absolute values of array elements

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->asum($x);
## y => 10.0
$y = $mo->asum($x,0);
## y => [4.0, 6.0]
$y = $mo->asum($x,1);
## y => [3.0, 7.0]

max

public function max(NDArray $x,int $axis=null)

Maximum value of array element

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->max($x);
## y => 3.0
$y = $mo->max($x,0);
## y => [3.0, -2.0]
$y = $mo->max($x,1);
## y => [1.0, 3.0]

argMax

public function argMax(NDArray $x,int $axis=null)

Index of maximum value array element. If the array has two or more dimensions, returns the position represented in one dimension.

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->argMax($x);
## y => 2
$y = $mo->argMax($x,0);
## y => [1, 0]
$y = $mo->argMax($x,1);
## y => [0, 0]

amax

public function amax(NDArray $x,int $axis=null)

Maximum absolute value of array element

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->amax($x);
## y => -4.0
$y = $mo->amax($x,0);
## y => [3.0, -4.0]
$y = $mo->amax($x,1);
## y => [-2.0, -4.0]

argAmax

public function argAmax(NDArray $x,int $axis=null)

Index of maximum absolute value array element. If the array has two or more dimensions, returns the position represented in one dimension.

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->argAmax($x);
## y => 3
$y = $mo->argAmax($x,0);
## y => [1, 1]
$y = $mo->argAmax($x,1);
## y => [1, 1]

min

public function min(NDArray $x,int $axis=null)

Minimum value of array element

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->min($x);
## y => -4.0
$y = $mo->min($x,0);
## y => [1.0, -4.0]
$y = $mo->min($x,1);
## y => [-2.0, -4.0]

argMin

public function argMin(NDArray $x,int $axis=null)

Index of Minimum value array element. If the array has two or more dimensions, returns the position represented in one dimension.

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->argMin($x);
## y => 3
$y = $mo->argMin($x,0);
## y => [0, 1]
$y = $mo->argMin($x,1);
## y => [1, 1]

amin

public function amin(NDArray $x,int $axis=null)

Minimum absolute value of array element

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->amin($x);
## y => 1.0
$y = $mo->amin($x,0);
## y => [1.0, -2.0]
$y = $mo->amin($x,1);
## y => [1.0, 3.0]

argAmin

public function argAmin(NDArray $x,int $axis=null)

Index of Minimum absolute value array element. If the array has two or more dimensions, returns the position represented in one dimension.

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->argAmin($x);
## y => 0
$y = $mo->argAmin($x,0);
## y => [0, 0]
$y = $mo->argAmin($x,1);
## y => [0, 0]

mean

public function mean(NDArray $x,int $axis=null)

Minimum absolute value of array element

Arguments

  • x: Input matrix.
  • axis: which dimension to summarize.

Example

$x = $mo->array([[1,-2],[3,-4]]);
$y = $mo->mean($x);
## y => -0.5
$y = $mo->mean($x,0);
## y => [2.0, -3.0]
$y = $mo->mean($x,1);
## y => [-0.5, -0.5]

f

public function f(callable $func,NDArray $x, ...$args) : NDArray

Apply functions to array elements.

Arguments

  • func: All type of functions available for call_user_func()
  • x: Input matrix.
  • args: Argument specified by call_user_func()

Example

$x = $mo->array([[0,1],[4,9]]);
$y = $mo->f('sqrt',$x);
## y => [[0.0, 1.0], [2.0, 3.0]]
$y = $mo->f(fn($a,$b)=>$a*$a+$b, $x, 100);
## y => [[100.0, 101.0],[116.0, 181.0]]

u

public function u(NDArray $x, callable $func, ...$args) : NDArray

Update by applying function to array elements

Arguments

  • x: Input matrix.
  • func: All type of functions available for call_user_func()
  • args: Argument specified by call_user_func()

Return

  • Returns the updated x.

Example

$x = $mo->array([[0,1],[4,9]]);
$mo->u($x,'sqrt');
## x => [[0.0, 1.0], [2.0, 3.0]]
$x = $mo->array([[0,1],[4,9]]);
$mo->u($x, fn($a,$b)=>$a*$a+$b, 100);
## x => [[100.0, 101.0],[116.0, 181.0]]
$x = $mo->zeros([2,2]);
$mo->u($mo->u($x,fn($a)=>$a+1),fn($a)=>$a+1);
## x => [[2.0, 2.0], [2.0, 2.0]]

op

public function op($X, string $operator, $Y, NDArray $R=null) : NDArray

Broadcast operators for array elements.

See the section “Broadcasting” for details.

Arguments

  • X,Y: Input matrix.
  • operator: Operator characters such as arithmetic operations
  • R: Array space for storing results
    • Specify this when you want to store the operation result in a prepared array.
    • Create an array to automatically store if not specified

Valid operator characters

  • Numerical calculations: + - * / % **
  • Conditional operator: == != > >= < <=

Return

  • Calculation result

Example

$x = $mo->array([[1,2],[3,4]]);
$y = $mo->array([[10,20],[30,40]]);
$z = $mo->op($x,'+',$y);
## z => [[11.0, 22.0], [33.0, 44.0]]
$x = $mo->array([[1,2],[3,4]]);
$y = $mo->array([10,100]);
$z = $mo->op($x,'*',$y);
## z => [[10.0,200.0],[30.0,400.0]]

select

public function select(NDArray $X, NDArray ...$MASKs) : NDArray

Selective operation.

See the section “Selective operation” for details.

Arguments

  • X: Input matrix.
  • MASKs: Array of indices to select or mask array.
    • If you specify more than one, repeat the selection operation multiple times

Example

## Extract from set x by index y
$x = $mo->array([[1,2,3],[4,5,6],[7,8,9]]);
$y = $mo->array([0,1,2],NDArray::int32);
$z = $mo->select($x, $mo->arange($x->shape()[0]), $y);
## z => [1.0, 5.0, 9.0]

## Extract only positive numbers
$x = $mo->array([1,-2,3,-4]);
$z = $mo->select($x, $mo->op($x,'>',0));
## z => [1.0, 3.0]

update

public function update(NDArray $X, string $operator, $value, NDArray ...$MASKs) : NDArray

Selective update.

See the section “Selective operation” for details.

Arguments

  • X: Input matrix.
  • operator: update operator character.
  • value: update value.
  • MASKs: Array of indices to select or mask array.
    • If you specify more than one, repeat the selection operation multiple times

Valid operator characters - Numerical calculations: = += -= *= /= %= **=

Return

  • Returns the updated x.

Example

## fill zero by index y
$x = $mo->array([[1,2,3],[4,5,6],[7,8,9]]);
$y = $mo->array([0,1,2]);
$mo->update($x, '=', 0, $mo->arange($x->shape()[0]), $y);
## x => [[0,2,3],[4,0,6],[7,8,0]]

## limit x by 5
$x = $mo->array([[1,4,7],[2,5,8],[3,6,9]]);
$mo->update($x, '=', 5, $mo->op($x,'>',5));
## x => [[1,4,5],[2,5,5],[3,5,5]]

dtypeToString

public function dtypeToString($dtype) : string

Convert dtype of NDarray to string

Example

$x = $mo->array([[1,2,3]]);
echo $mo->dtypeToString($x->dtype());
## float32
$x = $mo->arange(3);
echo $mo->dtypeToString($x->dtype());
## int32

toString

public function toString(NDArray $array) : string

Convert NDarray to string

Example

$x = $mo->array([[1,2,3]]);
echo $mo->toString($x);
## [[1,2,3]]
$x = $mo->arange(3);
echo $mo->toString($x);
## [0,1,2]

random

public function random()

Get random number library

See the section “Random library” for details.

Example

$x = $mo->random()->rand(10);

la

public function la()

Get linear algebra library

See the section “Linear algebra library” for details.

Example

$x = $mo->array([1,2,3]);
$mo->la()->scal(10,$x);