ghostjat /
Np
| Total Complexity | 360 |
| Total Lines | 1689 |
| Duplicated Lines | 0 % |
| Changes | 1 | ||
| Bugs | 1 | Features | 0 |
Complex classes like matrix often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
While breaking up the class, it is a good idea to analyze how other classes use matrix, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 27 | class matrix extends nd{ |
||
| 28 | |||
| 29 | /** |
||
| 30 | * create empty 2d matrix for given data type |
||
| 31 | * @param int $row num of rows |
||
| 32 | * @param int $col num of cols |
||
| 33 | * @param int $dtype matrix data type float|double |
||
| 34 | * @return \Np\matrix |
||
| 35 | */ |
||
| 36 | public static function factory(int $row, int $col, int $dtype = self::FLOAT): matrix { |
||
| 37 | return new self($row, $col, $dtype); |
||
| 38 | } |
||
| 39 | |||
| 40 | /** |
||
| 41 | * create 2d matrix using php array |
||
| 42 | * @param array $data |
||
| 43 | * @param int $dtype matrix data type float|double |
||
| 44 | * @return \Np\matrix |
||
| 45 | */ |
||
| 46 | public static function ar(array $data, int $dtype = self::FLOAT): matrix { |
||
| 47 | if (is_array($data) && is_array($data[0])) { |
||
| 48 | $ar = self::factory(count($data), count($data[0]), $dtype); |
||
| 49 | $ar->setData($data); |
||
| 50 | unset($data); |
||
| 51 | return $ar; |
||
| 52 | } else { |
||
| 53 | self::_err('given array is not rank-2 or given is not an array'); |
||
| 54 | } |
||
| 55 | } |
||
| 56 | |||
| 57 | /** |
||
| 58 | * create one like 2d matrix |
||
| 59 | * @param int $row |
||
| 60 | * @param int $col |
||
| 61 | * @return \Np\matrix |
||
| 62 | */ |
||
| 63 | public static function ones(int $row, int $col, int $dtype = self::FLOAT): matrix { |
||
| 64 | $ar = self::factory($row, $col, $dtype); |
||
| 65 | for ($i = 0; $i < $ar->ndim; ++$i) { |
||
| 66 | $ar->data[$i] = 1; |
||
| 67 | } |
||
| 68 | return $ar; |
||
| 69 | } |
||
| 70 | |||
| 71 | /** |
||
| 72 | * Create Matrix with random values |
||
| 73 | * @param int $row |
||
| 74 | * @param int $col |
||
| 75 | * @param int $dtype Float|Double |
||
| 76 | * @return \Np\matrix |
||
| 77 | */ |
||
| 78 | public static function randn(int $row, int $col, int $dtype = self::FLOAT): matrix { |
||
| 79 | $ar = self::factory($row, $col, $dtype); |
||
| 80 | $max = getrandmax(); |
||
| 81 | for ($i = 0; $i < $ar->ndim; ++$i) { |
||
| 82 | $ar->data[$i] = rand() / $max; |
||
| 83 | } |
||
| 84 | return $ar; |
||
| 85 | } |
||
| 86 | |||
| 87 | /** |
||
| 88 | * Return 2d matrix with uniform values |
||
| 89 | * @param int $row |
||
| 90 | * @param int $col |
||
| 91 | * @param int $dtype |
||
| 92 | * @return \Np\matrix |
||
| 93 | */ |
||
| 94 | public static function uniform(int $row, int $col, int $dtype = self::FLOAT): matrix { |
||
| 95 | $ar = self::factory($row, $col, $dtype); |
||
| 96 | $max = getrandmax(); |
||
| 97 | for ($i = 0; $i < $ar->ndim; ++$i) { |
||
| 98 | $ar->data[$i] = rand(-$max, $max) / $max; |
||
| 99 | } |
||
| 100 | return $ar; |
||
| 101 | } |
||
| 102 | |||
| 103 | /** |
||
| 104 | * Return a zero matrix with the given dimensions. |
||
| 105 | * @param int $row |
||
| 106 | * @param int $col |
||
| 107 | * @param int $dtype |
||
| 108 | * @return \Np\matrix |
||
| 109 | */ |
||
| 110 | public static function zeros(int $row, int $col, int $dtype = self::FLOAT): matrix { |
||
| 111 | $ar = self::factory($row, $col, $dtype); |
||
| 112 | for ($i = 0; $i < $ar->ndim; ++$i) { |
||
| 113 | $ar->data[$i] = 0.0; |
||
| 114 | } |
||
| 115 | return $ar; |
||
| 116 | } |
||
| 117 | |||
| 118 | /** |
||
| 119 | * create a null like 2d matrix |
||
| 120 | * @param int $row |
||
| 121 | * @param int $col |
||
| 122 | * @return \Np\matrix |
||
| 123 | */ |
||
| 124 | public static function null(int $row, int $col, int $dtype = self::FLOAT): matrix { |
||
| 125 | $ar = self::factory($row, $col, $dtype); |
||
| 126 | for ($i = 0; $i < $ar->ndim; ++$i) { |
||
| 127 | $ar->data[$i] = null; |
||
| 128 | } |
||
| 129 | return $ar; |
||
| 130 | } |
||
| 131 | |||
| 132 | /** |
||
| 133 | * create a 2d matrix with given scalar value |
||
| 134 | * @param int $row |
||
| 135 | * @param int $col |
||
| 136 | * @param int|float|double $val |
||
| 137 | * @return \Np\matrix |
||
| 138 | */ |
||
| 139 | public static function full(int $row, int $col, $val, int $dtype = self::FLOAT): matrix { |
||
| 140 | $ar = self::factory($row, $col, $dtype); |
||
| 141 | for ($i = 0; $i < $ar->ndim; ++$i) { |
||
| 142 | $ar->data[$i] = $val; |
||
| 143 | } |
||
| 144 | return $ar; |
||
| 145 | } |
||
| 146 | |||
| 147 | /** |
||
| 148 | * create a diagonal 2d matrix with given 1d array; |
||
| 149 | * @param array $elements |
||
| 150 | * @return \Np\matrix |
||
| 151 | */ |
||
| 152 | public static function diagonal(array $elements, int $dtype = self::FLOAT): matrix { |
||
| 153 | $n = count($elements); |
||
| 154 | $ar = self::factory($n, $n, $dtype); |
||
| 155 | for ($i = 0; $i < $n; ++$i) { |
||
| 156 | $ar->data[$i * $n + $i] = $elements[$i]; #for ($j = 0; $j < $n; ++$j) {$i === $j ? $elements[$i] : 0;#} |
||
| 157 | } |
||
| 158 | return $ar; |
||
| 159 | } |
||
| 160 | |||
| 161 | /** |
||
| 162 | * Generate a m x n matrix with elements from a Poisson distribution. |
||
| 163 | * @param int $row |
||
| 164 | * @param int $col |
||
| 165 | * @param float $lambda |
||
| 166 | * @param int $dtype |
||
| 167 | * @return \Np\matrix |
||
| 168 | */ |
||
| 169 | public static function poisson(int $row, int $col, float $lambda = 1.0, int $dtype = self::FLOAT): matrix { |
||
| 170 | $ar = self::factory($row, $col, $dtype); |
||
| 171 | $max = getrandmax(); |
||
| 172 | $l = exp(-$lambda); |
||
| 173 | for ($i = 0; $i < $row; ++$i) { |
||
| 174 | for ($j = 0; $j < $col; ++$j) { |
||
| 175 | $k = 0; |
||
| 176 | $p = 1.0; |
||
| 177 | while ($p > $l) { |
||
| 178 | ++$k; |
||
| 179 | $p = $p * rand() / $max; |
||
| 180 | } |
||
| 181 | $ar->data[$i * $col + $j] = $k - 1; |
||
| 182 | } |
||
| 183 | } |
||
| 184 | return $ar; |
||
| 185 | } |
||
| 186 | |||
| 187 | /** |
||
| 188 | * Return a standard normally distributed random matrix i.e values |
||
| 189 | * between -1 and 1. |
||
| 190 | * @param int $row |
||
| 191 | * @param int $col |
||
| 192 | * @param int $dtype Description |
||
| 193 | * @return \Np\matrix |
||
| 194 | */ |
||
| 195 | public static function gaussian(int $row, int $col, int $dtype = self::FLOAT): matrix { |
||
| 223 | } |
||
| 224 | |||
| 225 | /** |
||
| 226 | * create an identity matrix with the given dimensions. |
||
| 227 | * @param int $n |
||
| 228 | * @param int $dtype |
||
| 229 | * @return matrix |
||
| 230 | * @throws \InvalidArgumentException |
||
| 231 | */ |
||
| 232 | public static function identity(int $n, int $dtype = self::FLOAT): matrix { |
||
| 233 | if ($n < 1) { |
||
| 234 | self::_dimensionaMisMatchErr('dimensionality must be greater than 0 on all axes.'); |
||
| 235 | } |
||
| 236 | |||
| 237 | $ar = self::factory($n, $n, $dtype); |
||
| 238 | for ($i = 0; $i < $n; ++$i) { |
||
| 239 | for ($j = 0; $j < $n; ++$j) { |
||
| 240 | $ar->data[$i * $n + $j] = $i === $j ? 1 : 0; |
||
| 241 | } |
||
| 242 | } |
||
| 243 | return $ar; |
||
| 244 | } |
||
| 245 | |||
| 246 | /** |
||
| 247 | * Return the element-wise minimum of two matrices. |
||
| 248 | * |
||
| 249 | * @param \Np\matrix $m |
||
| 250 | * @return matrix |
||
| 251 | */ |
||
| 252 | public function minimum(matrix $m): matrix { |
||
| 259 | } |
||
| 260 | } |
||
| 261 | |||
| 262 | /** |
||
| 263 | * Return the element-wise maximum of two matrices. |
||
| 264 | * |
||
| 265 | * @param \Np\matrix $m |
||
| 266 | * @return matrix |
||
| 267 | */ |
||
| 268 | public function maximum(matrix $m): matrix { |
||
| 269 | if ($this->checkShape($this, $m)) { |
||
| 270 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 271 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 272 | $ar->data[$i] = max($this->data[$i], $m->data[$i]); |
||
| 273 | } |
||
| 274 | return $ar; |
||
| 275 | } |
||
| 276 | } |
||
| 277 | |||
| 278 | /** |
||
| 279 | * 2D convolution between a matrix ma and kernel kb, with a given stride. |
||
| 280 | * @param \Np\matrix $m |
||
| 281 | * @param int $stride |
||
| 282 | * @return matrix |
||
| 283 | */ |
||
| 284 | public function convolve(matrix $m, int $stride = 1): matrix { |
||
| 285 | return convolve::conv2D($this, $m, $stride); |
||
| 286 | } |
||
| 287 | |||
| 288 | /** |
||
| 289 | * Calculate the determinant of the matrix. |
||
| 290 | * @return float |
||
| 291 | */ |
||
| 292 | public function det(): float { |
||
| 293 | if (!$this->isSquare()) { |
||
| 294 | self::_err('determinant is undefined for a non square matrix'); |
||
| 295 | } |
||
| 296 | $lu = $this->lu(); |
||
| 297 | $nSwaps = $lu->p()->diagonalAsVector()->subtract($lu->p()->diagonalAsVector()->sum())->col - 1; |
||
| 298 | $detP = (-1) ** $nSwaps; |
||
| 299 | $detL = $lu->l()->diagonalAsVector()->product(); |
||
| 300 | $detU = $lu->u()->diagonalAsVector()->product(); |
||
| 301 | unset($lu); |
||
| 302 | return ($detP * $detL * $detU); |
||
| 303 | } |
||
| 304 | |||
| 305 | /** |
||
| 306 | * Return the trace of the matrix i.e the sum of all diagonal elements of a square matrix. |
||
| 307 | * @return float |
||
| 308 | */ |
||
| 309 | public function trace(): float { |
||
| 310 | if (!$this->isSquare()) { |
||
| 311 | self::_err('Error::matrix is not a squared can not Trace!'); |
||
| 312 | } |
||
| 313 | $trace = 0.0; |
||
| 314 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 315 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 316 | if ($i == $j) { |
||
| 317 | $trace += $this->data[$i * $this->col + $i]; |
||
| 318 | } |
||
| 319 | } |
||
| 320 | } |
||
| 321 | return $trace; |
||
| 322 | } |
||
| 323 | |||
| 324 | /** |
||
| 325 | * dignoalInterChange |
||
| 326 | */ |
||
| 327 | public function dignoalInterChange() { |
||
| 332 | } |
||
| 333 | } |
||
| 334 | } |
||
| 335 | |||
| 336 | //----------------Linear Algebra Opreations------------------------------- |
||
| 337 | |||
| 338 | /** |
||
| 339 | * |
||
| 340 | * get dot product of m.m or m.v |
||
| 341 | * |
||
| 342 | * @param \Np\matrix|\Np\vector $d |
||
| 343 | * @return matrix|vector |
||
| 344 | */ |
||
| 345 | public function dot(matrix|vector $d): matrix|vector { |
||
| 346 | if ($d instanceof self) { |
||
| 347 | return $this->dotMatrix($d); |
||
| 348 | } else { |
||
| 349 | return $this->dotVector($d); |
||
| 350 | } |
||
| 351 | } |
||
| 352 | |||
| 353 | /** |
||
| 354 | * get matrix & matrix dot product |
||
| 355 | * @param \Np\matrix $matrix |
||
| 356 | * @return \Np\matrix |
||
| 357 | */ |
||
| 358 | protected function dotMatrix(matrix $matrix): matrix { |
||
| 363 | } |
||
| 364 | } |
||
| 365 | |||
| 366 | /** |
||
| 367 | * get dot product of matrix & a vector |
||
| 368 | * @param \Np\vector $vector |
||
| 369 | * @return \Np\vector |
||
| 370 | */ |
||
| 371 | protected function dotVector(vector $vector): vector { |
||
| 372 | if ($this->checkDtype($this, $vector) && $this->checkDimensions($vector, $this)) { |
||
| 373 | $mvr = vector::factory($this->col, $this->dtype); |
||
| 374 | blas::gemv($this, $vector, $mvr); |
||
| 375 | return $mvr; |
||
| 376 | } |
||
| 377 | } |
||
| 378 | |||
| 379 | //---------------Arthmetic Opreations----------------------------------- |
||
| 380 | |||
| 381 | /** |
||
| 382 | * multiply this matrix with another matrix|scalar element-wise |
||
| 383 | * Matrix Scalar\Matrix multiplication |
||
| 384 | * @param int|float|matrix|vector $m |
||
| 385 | * @return matrix|vector |
||
| 386 | */ |
||
| 387 | public function multiply(int|float|matrix|vector $m): matrix|vector { |
||
| 388 | if ($m instanceof self) { |
||
| 389 | return $this->multiplyMatrix($m); |
||
| 390 | } else if ($m instanceof vector) { |
||
| 391 | return $this->multiplyVector($m); |
||
| 392 | } else { |
||
| 393 | return $this->scale($m); |
||
| 394 | } |
||
| 395 | } |
||
| 396 | |||
| 397 | /** |
||
| 398 | * |
||
| 399 | * @param \Np\vector $v |
||
| 400 | * @return matrix |
||
| 401 | */ |
||
| 402 | protected function multiplyVector(vector $v): matrix { |
||
| 403 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this, $v)) { |
||
| 404 | $ar = matrix::factory($this->row, $this->col, $this->dtype); |
||
| 405 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 406 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 407 | $ar->data[$i * $this->col + $j] = $v->data[$j] * $this->data[$i * $this->col + $j]; |
||
| 408 | } |
||
| 409 | } |
||
| 410 | return $ar; |
||
| 411 | } |
||
| 412 | } |
||
| 413 | |||
| 414 | /** |
||
| 415 | * |
||
| 416 | * @param \Np\matrix $m |
||
| 417 | * @return matrix |
||
| 418 | */ |
||
| 419 | protected function multiplyMatrix(matrix $m): matrix { |
||
| 420 | if ($this->checkDtype($this, $m) && $this->checkShape($this, $m)) { |
||
| 421 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 422 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 423 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 424 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j] * $m->data[$i * $this->col + $j]; |
||
| 425 | } |
||
| 426 | } |
||
| 427 | return $ar; |
||
| 428 | } |
||
| 429 | } |
||
| 430 | |||
| 431 | /** |
||
| 432 | * Sum of Rows of matrix |
||
| 433 | * @return vector |
||
| 434 | */ |
||
| 435 | public function sumRows(): vector { |
||
| 436 | $vr = vector::factory($this->row, $this->dtype); |
||
| 437 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 438 | $sum = 0.0; |
||
| 439 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 440 | $sum += $this->data[$i * $this->col + $j]; |
||
| 441 | } |
||
| 442 | $vr->data[$i] = $sum; |
||
| 443 | } |
||
| 444 | return $vr; |
||
| 445 | } |
||
| 446 | |||
| 447 | /** |
||
| 448 | * Sum of two matrix, vector or a scalar to current matrix |
||
| 449 | * |
||
| 450 | * @param int|float|matrix|vector $m |
||
| 451 | * @return matrix |
||
| 452 | */ |
||
| 453 | public function sum(int|float|matrix|vector $m): matrix { |
||
| 454 | if ($m instanceof self) { |
||
| 455 | return $this->sumMatrix($m); |
||
| 456 | } elseif ($m instanceof vector) { |
||
| 457 | return $this->sumVector($m); |
||
| 458 | } else { |
||
| 459 | return $this->sumScalar($m); |
||
| 460 | } |
||
| 461 | } |
||
| 462 | |||
| 463 | protected function sumScalar(int|float $s): matrix { |
||
| 464 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 465 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 466 | $ar->data[$i] = $this->data[$i] + $s; |
||
| 467 | } |
||
| 468 | return $ar; |
||
| 469 | } |
||
| 470 | |||
| 471 | protected function sumMatrix(matrix $m): matrix { |
||
| 472 | if ($this->checkShape($this, $m) && $this->checkDtype($this, $m)) { |
||
| 473 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 474 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 475 | $ar->data[$i] = $this->data[$i] + $m->data[$i]; |
||
| 476 | } |
||
| 477 | return $ar; |
||
| 478 | } |
||
| 479 | } |
||
| 480 | |||
| 481 | protected function sumVector(vector $v): matrix { |
||
| 482 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this, $v)) { |
||
| 483 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 484 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 485 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 486 | $ar->data[$i * $this->col + $j] = $v->data[$j] + $this->data[$i * $this->col + $j]; |
||
| 487 | } |
||
| 488 | } |
||
| 489 | return $ar; |
||
| 490 | } |
||
| 491 | } |
||
| 492 | |||
| 493 | /** |
||
| 494 | * subtract another matrix, vector or a scalar to this matrix |
||
| 495 | * @param int|float|matrix|vector $d matrix|$scalar to subtract this matrix |
||
| 496 | * @return \Np\matrix |
||
| 497 | */ |
||
| 498 | public function subtract(int|float|matrix|vector $d): matrix { |
||
| 499 | if ($d instanceof self) { |
||
| 500 | return $this->subtractMatrix($d); |
||
| 501 | } elseif ($d instanceof vector) { |
||
| 502 | return $this->subtractVector($d); |
||
| 503 | } else { |
||
| 504 | return $this->subtractScalar($d); |
||
| 505 | } |
||
| 506 | } |
||
| 507 | |||
| 508 | protected function subtractScalar(int|float $s): matrix { |
||
| 509 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 510 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 511 | $ar->data[$i] = $this->data[$i] - $s; |
||
| 512 | } |
||
| 513 | return $ar; |
||
| 514 | } |
||
| 515 | |||
| 516 | /** |
||
| 517 | * |
||
| 518 | * @param matrix $m |
||
| 519 | * @return matrix |
||
| 520 | */ |
||
| 521 | protected function subtractMatrix(matrix $m): matrix { |
||
| 522 | if ($this->checkShape($this, $m) && $this->checkDtype($this, $m)) { |
||
| 523 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 524 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 525 | $ar->data[$i] = $this->data[$i] - $m->data[$i]; |
||
| 526 | } |
||
| 527 | return $ar; |
||
| 528 | } |
||
| 529 | } |
||
| 530 | |||
| 531 | /** |
||
| 532 | * |
||
| 533 | * @param vector $v |
||
| 534 | * @return matrix |
||
| 535 | */ |
||
| 536 | protected function subtractVector(vector $v): matrix { |
||
| 537 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this,$v)) { |
||
| 538 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 539 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 540 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 541 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j] - $v->data[$j]; |
||
| 542 | } |
||
| 543 | } |
||
| 544 | return $ar; |
||
| 545 | } |
||
| 546 | } |
||
| 547 | |||
| 548 | /** |
||
| 549 | * |
||
| 550 | * @param vector $v |
||
| 551 | * @return matrix |
||
| 552 | */ |
||
| 553 | public function subtractColumnVector(vector $v): matrix { |
||
| 554 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this, $v)) { |
||
| 555 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 556 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 557 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 558 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j] - $v->data[$i]; |
||
| 559 | } |
||
| 560 | } |
||
| 561 | return $ar; |
||
| 562 | } |
||
| 563 | } |
||
| 564 | |||
| 565 | /** |
||
| 566 | * Return the division of two elements, element-wise. |
||
| 567 | * @param int|float|matrix $d |
||
| 568 | * @return matrix |
||
| 569 | */ |
||
| 570 | public function divide(int|float|matrix|vector $d): matrix { |
||
| 571 | if ($d instanceof self) { |
||
| 572 | return $this->divideMatrix($d); |
||
| 573 | } elseif ($d instanceof vector) { |
||
| 574 | return $this->divideVector($d); |
||
| 575 | } else { |
||
| 576 | return $this->divideScalar($d); |
||
| 577 | } |
||
| 578 | } |
||
| 579 | |||
| 580 | protected function divideMatrix(matrix $m): matrix { |
||
| 581 | if ($this->checkShape($this, $m) && $this->checkDtype($this, $m)) { |
||
| 582 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 583 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 584 | $ar->data[$i] = $this->data[$i] / $m->data[$i]; |
||
| 585 | } |
||
| 586 | return $ar; |
||
| 587 | } |
||
| 588 | } |
||
| 589 | |||
| 590 | protected function divideVector(vector $v): matrix { |
||
| 591 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this, $v)) { |
||
| 592 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 593 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 594 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 595 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j] / $v->data[$j]; |
||
| 596 | } |
||
| 597 | } |
||
| 598 | return $ar; |
||
| 599 | } |
||
| 600 | } |
||
| 601 | |||
| 602 | protected function divideScalar(int|float $s): matrix { |
||
| 603 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 604 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 605 | $ar->data[$i] = $this->data[$i] / $s; |
||
| 606 | } |
||
| 607 | return $ar; |
||
| 608 | } |
||
| 609 | |||
| 610 | /** |
||
| 611 | * |
||
| 612 | * Raise this matrix to the power of the element-wise entry in another matrix. |
||
| 613 | * |
||
| 614 | * @param int|float|matrix $m |
||
| 615 | * @return matrix |
||
| 616 | */ |
||
| 617 | public function pow(int|float|matrix|vector $d): matrix { |
||
| 618 | if ($d instanceof self) { |
||
| 619 | return $this->powMatrix($d); |
||
| 620 | } else if ($d instanceof vector) { |
||
| 621 | return $this->powVector($d); |
||
| 622 | } else { |
||
| 623 | return $this->powScalar($d); |
||
| 624 | } |
||
| 625 | } |
||
| 626 | |||
| 627 | protected function powMatrix(matrix $m): matrix { |
||
| 628 | if ($this->checkShape($this, $m) && $this->checkDtype($this, $m)) { |
||
| 629 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 630 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 631 | $ar->data[$i] = $this->data[$i] ** $m->data[$i]; |
||
| 632 | } |
||
| 633 | return $ar; |
||
| 634 | } |
||
| 635 | } |
||
| 636 | |||
| 637 | protected function powVector(vector $v): matrix { |
||
| 638 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this, $v)) { |
||
| 639 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 640 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 641 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 642 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j] ** $v->data[$j]; |
||
| 643 | } |
||
| 644 | } |
||
| 645 | return $ar; |
||
| 646 | } |
||
| 647 | } |
||
| 648 | |||
| 649 | protected function powScalar(int|float $s): matrix { |
||
| 650 | $ar = $this->copyMatrix(); |
||
| 651 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 652 | $ar->data[$i] **= $s; |
||
| 653 | } |
||
| 654 | return $ar; |
||
| 655 | } |
||
| 656 | |||
| 657 | /** |
||
| 658 | * Calculate the modulus i.e remainder of division between this matrix and another matrix. |
||
| 659 | * @param int|float|matrix|vector $d |
||
| 660 | * @return matrix |
||
| 661 | */ |
||
| 662 | public function mod(int|float|matrix|vector $d): matrix { |
||
| 663 | if ($d instanceof self) { |
||
| 664 | $this->modMatrix($d); |
||
| 665 | } else if ($d instanceof vector) { |
||
| 666 | $this->modVector($d); |
||
| 667 | } else { |
||
| 668 | $this->modScalar($d); |
||
| 669 | } |
||
| 670 | } |
||
| 671 | |||
| 672 | protected function modMatrix(matrix $m): matrix { |
||
| 673 | if ($this->checkShape($this, $m) && $this->checkDtype($this, $m)) { |
||
| 674 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 675 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 676 | $ar->data[$i] = $this->data[$i] % $m->data[$i]; |
||
| 677 | } |
||
| 678 | return $ar; |
||
| 679 | } |
||
| 680 | } |
||
| 681 | |||
| 682 | protected function modVector(vector $v): matrix { |
||
| 683 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this, $v)) { |
||
| 684 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 685 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 686 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 687 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j] % $v->data[$j]; |
||
| 688 | } |
||
| 689 | } |
||
| 690 | return $ar; |
||
| 691 | } |
||
| 692 | } |
||
| 693 | |||
| 694 | protected function modScalar(int|float $s): matrix { |
||
| 695 | $ar = $this->copyMatrix(); |
||
| 696 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 697 | $ar->data[$i] %= $s; |
||
| 698 | } |
||
| 699 | return $ar; |
||
| 700 | } |
||
| 701 | |||
| 702 | /** |
||
| 703 | * Return the element-wise reciprocal of the matrix. |
||
| 704 | * |
||
| 705 | * @return matrix |
||
| 706 | */ |
||
| 707 | public function reciprocal(): matrix { |
||
| 708 | return self::ones($this->row, $this->col, $this->dtype)->divideMatrix($this); |
||
| 709 | } |
||
| 710 | |||
| 711 | /** |
||
| 712 | * |
||
| 713 | * @param int|float $d |
||
| 714 | * @return bool |
||
| 715 | */ |
||
| 716 | public static function is_zero($d): bool { |
||
| 717 | if (abs($d) < self::EPSILON) { |
||
| 718 | return true; |
||
| 719 | } |
||
| 720 | return false; |
||
| 721 | } |
||
| 722 | |||
| 723 | /** |
||
| 724 | * is row zero |
||
| 725 | * @param int $row |
||
| 726 | * @return bool |
||
| 727 | */ |
||
| 728 | public function is_rowZero(int $row): bool { |
||
| 729 | for ($i = 0; $i < $this->col; ++$i) { |
||
| 730 | if ($this->data[$row * $this->col + $i] != 0) { |
||
| 731 | return false; |
||
| 732 | } |
||
| 733 | } |
||
| 734 | return true; |
||
| 735 | } |
||
| 736 | |||
| 737 | /** |
||
| 738 | * |
||
| 739 | * @return bool |
||
| 740 | */ |
||
| 741 | public function has_ZeroRow(): bool { |
||
| 742 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 743 | if ($this->is_rowZero($i)) { |
||
| 744 | return true; |
||
| 745 | } |
||
| 746 | } |
||
| 747 | return false; |
||
| 748 | } |
||
| 749 | |||
| 750 | /** |
||
| 751 | * Transpose the matrix i.e row become cols and cols become rows. |
||
| 752 | * @return \Np\matrix |
||
| 753 | */ |
||
| 754 | public function transpose(): matrix { |
||
| 755 | $ar = self::factory($this->col, $this->row, $this->dtype); |
||
| 756 | for ($i = 0; $i < $ar->row; ++$i) { |
||
| 757 | for ($j = 0; $j < $ar->col; ++$j) { |
||
| 758 | $ar->data[$i * $ar->col + $j] = $this->data[$j * $ar->col + $i]; |
||
| 759 | } |
||
| 760 | } |
||
| 761 | return $ar; |
||
| 762 | } |
||
| 763 | |||
| 764 | /** |
||
| 765 | * swap specific values in matrix |
||
| 766 | * @param int $i1 |
||
| 767 | * @param int $i2 |
||
| 768 | */ |
||
| 769 | public function swapValue(int $i1, int $i2) { |
||
| 770 | $tmp = $this->data[$i1]; |
||
| 771 | $this->data[$i1] = $this->data[$i2]; |
||
| 772 | $this->data[$i2] = $tmp; |
||
| 773 | } |
||
| 774 | |||
| 775 | /** |
||
| 776 | * swap specific rows in matrix |
||
| 777 | * @param int $r1 |
||
| 778 | * @param int $r2 |
||
| 779 | */ |
||
| 780 | public function swapRows(int $r1, int $r2) { |
||
| 781 | for ($i = 0; $i < $this->col; ++$i) { |
||
| 782 | $tmp = $this->data[$r1 * $this->col + $i]; |
||
| 783 | $this->data[$r1 * $this->col + $i] = $this->data[$r2 * $this->col + $i]; |
||
| 784 | $this->data[$r2 * $this->col + $i] = $tmp; |
||
| 785 | } |
||
| 786 | } |
||
| 787 | |||
| 788 | /** |
||
| 789 | * swap specific cols in matrix |
||
| 790 | * @param int $c1 |
||
| 791 | * @param int $c2 |
||
| 792 | */ |
||
| 793 | public function swapCols(int $c1, int $c2) { |
||
| 794 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 795 | $tmp = $this->data[$i * $this->row + $c1]; |
||
| 796 | $this->data[$i * $this->row + $c1] = $this->data[$i * $this->row + $c2]; |
||
| 797 | $this->data[$i * $this->row + $c2] = $tmp; |
||
| 798 | } |
||
| 799 | } |
||
| 800 | |||
| 801 | /** |
||
| 802 | * |
||
| 803 | * @param int|float $scalar |
||
| 804 | * @return matrix |
||
| 805 | */ |
||
| 806 | public function scale(int|float $scalar): matrix { |
||
| 807 | if ($scalar == 0) { |
||
| 808 | return self::zeros($this->row, $this->col, $this->dtype); |
||
| 809 | } |
||
| 810 | |||
| 811 | $ar = $this->copyMatrix(); |
||
| 812 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 813 | $ar->data[$i] *= $scalar; |
||
| 814 | } |
||
| 815 | |||
| 816 | return $ar; |
||
| 817 | } |
||
| 818 | |||
| 819 | /** |
||
| 820 | * scale all the elements of a row |
||
| 821 | * @param int $row |
||
| 822 | * @param int|float $c |
||
| 823 | */ |
||
| 824 | public function scaleRow(int $row, int|float $c) { |
||
| 825 | for ($i = 0; $i < $this->col; ++$i) { |
||
| 826 | $this->data[$row * $this->col + $i] *= $c; |
||
| 827 | } |
||
| 828 | } |
||
| 829 | |||
| 830 | /** |
||
| 831 | * scale all the elements of |
||
| 832 | * @param int $col |
||
| 833 | * @param int|float $c |
||
| 834 | */ |
||
| 835 | public function scaleCol(int $col, int|float $c) { |
||
| 836 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 837 | $this->data[$i * $this->col + $col] *= $c; |
||
| 838 | } |
||
| 839 | } |
||
| 840 | |||
| 841 | /** |
||
| 842 | * Scale digonally |
||
| 843 | * @param int|float $c |
||
| 844 | * @param bool $lDig |
||
| 845 | */ |
||
| 846 | public function scaleDigonalCol(int|float $c, bool $lDig = true) { |
||
| 847 | if($lDig){ |
||
| 848 | for ($i = 0; $i < $this->row ; ++$i) { |
||
| 849 | $this->data[$i * $this->col + $i] *= $c; |
||
| 850 | } |
||
| 851 | } |
||
| 852 | else{ |
||
| 853 | for ($i = $this->row; $i > 0 ; --$i) { |
||
| 854 | $this->data[$i * $this->col - $i] *= $c; |
||
| 855 | } |
||
| 856 | } |
||
| 857 | } |
||
| 858 | |||
| 859 | /** |
||
| 860 | * |
||
| 861 | * @param int $r1 |
||
| 862 | * @param int $r2 |
||
| 863 | * @param float $c |
||
| 864 | */ |
||
| 865 | public function addScaleRow(int $r1, int $r2, float $c) { |
||
| 866 | for ($i = 0; $i < $this->col; ++$i) { |
||
| 867 | $this->data[$r2 * $this->col + $i] += $this->data[$r1 * $this->col + $i] * $c; |
||
| 868 | } |
||
| 869 | } |
||
| 870 | |||
| 871 | /** |
||
| 872 | * Attach given matrix to the left of this matrix. |
||
| 873 | * |
||
| 874 | * @param \Np\matrix $m |
||
| 875 | * @return \Np\matrix |
||
| 876 | */ |
||
| 877 | public function joinLeft(matrix $m): matrix { |
||
| 878 | if ($this->row != $m->row && !$this->checkDtype($this, $m)) { |
||
| 879 | self::_err('Error::Invalid size! or DataType!'); |
||
| 880 | } |
||
| 881 | $col = $this->col + $m->col; |
||
| 882 | $ar = self::factory($this->row, $col, $this->dtype); |
||
| 883 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 884 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 885 | $ar->data[$i * $col + $j] = $this->data[$i * $this->col + $j]; |
||
| 886 | } |
||
| 887 | for ($j = 0; $j < $m->col; ++$j) { |
||
| 888 | $ar->data[$i * $col + ($this->col + $j)] = $m->data[$i * $m->col + $j]; |
||
| 889 | } |
||
| 890 | } |
||
| 891 | return $ar; |
||
| 892 | } |
||
| 893 | |||
| 894 | /** |
||
| 895 | * Join matrix m to the Right of this matrix. |
||
| 896 | * @param \Np\matrix $m |
||
| 897 | * @return matrix |
||
| 898 | */ |
||
| 899 | public function joinRight(matrix $m): matrix { |
||
| 900 | if ($this->row != $m->row && !$this->checkDtype($this,$m)) { |
||
| 901 | self::_err('Error::Invalid size! or DataType!'); |
||
| 902 | } |
||
| 903 | $col = $this->col + $m->col; |
||
| 904 | $ar = self::factory($this->row, $col, $this->dtype); |
||
| 905 | for ($i = 0; $i < $m->row; ++$i) { |
||
| 906 | for ($j = 0; $j < $m->col; ++$j) { |
||
| 907 | $ar->data[$i * $col + $j] = $m->data[$i * $m->col + $j]; |
||
| 908 | } |
||
| 909 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 910 | $ar->data[$i * $col + ($this->col + $j)] = $this->data[$i * $this->col + $j]; |
||
| 911 | } |
||
| 912 | } |
||
| 913 | return $ar; |
||
| 914 | } |
||
| 915 | |||
| 916 | /** |
||
| 917 | * Join matrix m Above this matrix. |
||
| 918 | * @param \Np\matrix $m |
||
| 919 | * @return matrix |
||
| 920 | */ |
||
| 921 | public function joinAbove(matrix $m): matrix { |
||
| 922 | if ($this->col !== $m->col && !$this->checkDtype($this, $m)) { |
||
| 923 | self::_err('Error::Invalid size! or DataType!'); |
||
| 924 | } |
||
| 925 | $row = $this->row + $m->row; |
||
| 926 | $ar = self::factory($row, $this->col, $this->dtype); |
||
| 927 | for ($i = 0; $i < $m->row; ++$i) { |
||
| 928 | for ($j = 0; $j < $m->col; ++$j) { |
||
| 929 | $ar->data[$i * $m->col + $j] = $m->data[$i * $m->col + $j]; |
||
| 930 | } |
||
| 931 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 932 | $ar->data[($i + $this->row) * $this->col + $j] = $this->data[$i * $this->col + $j]; |
||
| 933 | } |
||
| 934 | } |
||
| 935 | return $ar; |
||
| 936 | } |
||
| 937 | |||
| 938 | /** |
||
| 939 | * Join matrix m below this matrix. |
||
| 940 | * @param \Np\matrix $m |
||
| 941 | * @return matrix |
||
| 942 | */ |
||
| 943 | public function joinBelow(matrix $m): matrix { |
||
| 944 | if ($this->col !== $m->col && !$this->checkDtype($this, $m)) { |
||
| 945 | self::_err('Error::Invalid size! or DataType!'); |
||
| 946 | } |
||
| 947 | $row = $this->row + $m->row; |
||
| 948 | $ar = self::factory($row, $this->col, $this->dtype); |
||
| 949 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 950 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 951 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j]; |
||
| 952 | } |
||
| 953 | for ($j = 0; $j < $m->col; ++$j) { |
||
| 954 | $ar->data[($i + $m->row) * $m->col + $j] = $m->data[$i * $m->col + $j]; |
||
| 955 | } |
||
| 956 | } |
||
| 957 | return $ar; |
||
| 958 | } |
||
| 959 | |||
| 960 | /** |
||
| 961 | * Calculate the row echelon form of the matrix. |
||
| 962 | * Return the reduced matrix. |
||
| 963 | * |
||
| 964 | * @return matrix|null |
||
| 965 | */ |
||
| 966 | public function ref(): matrix|null { |
||
| 967 | return ref::factory($this); |
||
| 968 | } |
||
| 969 | |||
| 970 | /** |
||
| 971 | * Return the lower triangular matrix of the Cholesky decomposition. |
||
| 972 | * |
||
| 973 | * @return matrix|null |
||
| 974 | */ |
||
| 975 | public function cholesky(): matrix|null { |
||
| 976 | return cholesky::factory($this); |
||
| 977 | } |
||
| 978 | |||
| 979 | /** |
||
| 980 | * FIXME-------------- |
||
| 981 | * RREF |
||
| 982 | * The reduced row echelon form (RREF) of a matrix. |
||
| 983 | * @return \Np\matrix |
||
| 984 | */ |
||
| 985 | public function rref(): matrix { |
||
| 986 | return rref::factory($this); |
||
| 987 | } |
||
| 988 | |||
| 989 | /** |
||
| 990 | * make copy of the matrix |
||
| 991 | * @return \Np\matrix |
||
| 992 | */ |
||
| 993 | public function copyMatrix(): matrix { |
||
| 994 | return clone $this; |
||
| 995 | } |
||
| 996 | |||
| 997 | /** |
||
| 998 | * |
||
| 999 | * @param int $cols |
||
| 1000 | * @return \Np\matrix |
||
| 1001 | */ |
||
| 1002 | public function diminish_left(int $cols): matrix { |
||
| 1003 | $ar = self::factory($this->row, $cols, $this->dtype); |
||
| 1004 | for ($i = 0; $i < $ar->row; ++$i) { |
||
| 1005 | for ($j = 0; $j < $ar->col; ++$j) { |
||
| 1006 | $ar->data[$i * $ar->col + $j] = $this->data[$i * $this->col + $j]; |
||
| 1007 | } |
||
| 1008 | } |
||
| 1009 | return $ar; |
||
| 1010 | } |
||
| 1011 | |||
| 1012 | /** |
||
| 1013 | * |
||
| 1014 | * @param int $cols |
||
| 1015 | * @return \Np\matrix |
||
| 1016 | */ |
||
| 1017 | public function diminish_right(int $cols): matrix { |
||
| 1018 | $ar = self::factory($this->row, $cols, $this->dtype); |
||
| 1019 | for ($i = 0; $i < $ar->row; ++$i) { |
||
| 1020 | for ($j = 0; $j < $ar->col; ++$j) { |
||
| 1021 | $ar->data[$i * $ar->col + $j] = $this->data[$i * $this->col - $cols + $j]; |
||
| 1022 | } |
||
| 1023 | } |
||
| 1024 | return $ar; |
||
| 1025 | } |
||
| 1026 | |||
| 1027 | /** |
||
| 1028 | * Return the index of the maximum element in every row of the matrix. |
||
| 1029 | * @return \Np\vector int |
||
| 1030 | */ |
||
| 1031 | public function argMax(): vector { |
||
| 1032 | $v = vector::factory($this->row, vector::INT); |
||
| 1033 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1034 | $v->data[$i] = blas::max($this->rowAsVector($i)); |
||
| 1035 | } |
||
| 1036 | return $v; |
||
| 1037 | } |
||
| 1038 | |||
| 1039 | /** |
||
| 1040 | * Return the index of the minimum element in every row of the matrix. |
||
| 1041 | * @return \Np\vector int |
||
| 1042 | */ |
||
| 1043 | public function argMin(): vector { |
||
| 1044 | $v = vector::factory($this->row, vector::INT); |
||
| 1045 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1046 | $v->data[$i] = blas::min($this->rowAsVector($i)); |
||
| 1047 | } |
||
| 1048 | |||
| 1049 | return $v; |
||
| 1050 | } |
||
| 1051 | |||
| 1052 | /** |
||
| 1053 | * Set given data in matrix |
||
| 1054 | * @param int|float|array $data |
||
| 1055 | * @param bool $dignoal |
||
| 1056 | * @return void |
||
| 1057 | */ |
||
| 1058 | public function setData(int|float|array $data, bool $dignoal = false): void { |
||
| 1059 | if ($dignoal == false) { |
||
| 1060 | if (is_array($data) && is_array($data[0])) { |
||
| 1061 | $f = $this->flattenArray($data); |
||
| 1062 | foreach ($f as $k => $v) { |
||
| 1063 | $this->data[$k] = $v; |
||
| 1064 | } |
||
| 1065 | } elseif (is_numeric($data)) { |
||
| 1066 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1067 | $this->data[$i] = $data; |
||
| 1068 | } |
||
| 1069 | } |
||
| 1070 | } elseif (is_numeric($data) || is_array($data) && !is_array($data[0])) { |
||
| 1071 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1072 | $this->data[$i * $this->col * $i] = $data; |
||
| 1073 | } |
||
| 1074 | } |
||
| 1075 | } |
||
| 1076 | |||
| 1077 | /** |
||
| 1078 | * get the matrix data type |
||
| 1079 | * @return type |
||
| 1080 | */ |
||
| 1081 | public function getDtype() { |
||
| 1083 | } |
||
| 1084 | |||
| 1085 | /** |
||
| 1086 | * get the shape of matrix |
||
| 1087 | * @return object |
||
| 1088 | */ |
||
| 1089 | public function getShape(): object { |
||
| 1090 | return (object) ['m' => $this->row, 'n' => $this->col]; |
||
| 1091 | } |
||
| 1092 | |||
| 1093 | /** |
||
| 1094 | * get the number of elements in the matrix. |
||
| 1095 | * @return int |
||
| 1096 | */ |
||
| 1097 | public function getSize(): int { |
||
| 1098 | return $this->ndim; |
||
| 1099 | } |
||
| 1100 | |||
| 1101 | /** |
||
| 1102 | * is matrix squred |
||
| 1103 | * @return bool |
||
| 1104 | */ |
||
| 1105 | public function isSquare(): bool { |
||
| 1106 | if ($this->row === $this->col) { |
||
| 1107 | return true; |
||
| 1108 | } |
||
| 1109 | return false; |
||
| 1110 | } |
||
| 1111 | |||
| 1112 | /** |
||
| 1113 | * Return a row as vector from the matrix. |
||
| 1114 | * @param int $index |
||
| 1115 | * @return \Np\vector |
||
| 1116 | */ |
||
| 1117 | public function rowAsVector(int $index): vector { |
||
| 1118 | $vr = vector::factory($this->col, $this->dtype); |
||
| 1119 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 1120 | $vr->data[$j] = $this->data[$index * $this->col + $j]; |
||
| 1121 | } |
||
| 1122 | return $vr; |
||
| 1123 | } |
||
| 1124 | |||
| 1125 | /** |
||
| 1126 | * Return a col as vector from the matrix. |
||
| 1127 | * @param int $index |
||
| 1128 | * @return \Np\vector |
||
| 1129 | */ |
||
| 1130 | public function colAsVector(int $index): vector { |
||
| 1131 | $vr = vector::factory($this->row, $this->dtype); |
||
| 1132 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1133 | $vr->data[$i] = $this->data[$i * $this->row + $index]; |
||
| 1134 | } |
||
| 1135 | return $vr; |
||
| 1136 | } |
||
| 1137 | |||
| 1138 | /** |
||
| 1139 | * Return the diagonal elements of a square matrix as a vector. |
||
| 1140 | * @return \Np\vector |
||
| 1141 | */ |
||
| 1142 | public function diagonalAsVector(): vector { |
||
| 1143 | if (!$this->isSquare()) { |
||
| 1144 | self::_err('Can not trace of a none square matrix'); |
||
| 1145 | } |
||
| 1146 | $vr = vector::factory($this->row, $this->dtype); |
||
| 1147 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1148 | $vr->data[$i] = $this->getDiagonalVal($i); |
||
| 1149 | } |
||
| 1150 | return $vr; |
||
| 1151 | } |
||
| 1152 | |||
| 1153 | /** |
||
| 1154 | * Flatten i.e unravel the matrix into a vector. |
||
| 1155 | * |
||
| 1156 | * @return \Np\vector |
||
| 1157 | */ |
||
| 1158 | public function asVector(): vector { |
||
| 1159 | $vr = vector::factory($this->ndim, $this->dtype); |
||
| 1160 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1161 | $vr->data[$i] = $this->data[$i]; |
||
| 1162 | } |
||
| 1163 | return $vr; |
||
| 1164 | } |
||
| 1165 | |||
| 1166 | /** |
||
| 1167 | * Return the elements of the matrix in a 2-d array. |
||
| 1168 | * @return array |
||
| 1169 | */ |
||
| 1170 | public function asArray(): array { |
||
| 1171 | $ar = array_fill(0, $this->row, array_fill(0, $this->col, null)); |
||
| 1172 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1173 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 1174 | $ar[$i][$j] = $this->data[$i * $this->col + $j]; |
||
| 1175 | } |
||
| 1176 | } |
||
| 1177 | return $ar; |
||
| 1178 | } |
||
| 1179 | |||
| 1180 | /** |
||
| 1181 | * get a diagonal value from matrix |
||
| 1182 | * @param int $i |
||
| 1183 | * @return float |
||
| 1184 | */ |
||
| 1185 | public function getDiagonalVal(int $i) { |
||
| 1186 | if ($this->isSquare()) { |
||
| 1187 | return $this->data[$i * $this->row + $i]; |
||
| 1188 | } |
||
| 1189 | } |
||
| 1190 | |||
| 1191 | /** |
||
| 1192 | * |
||
| 1193 | * Compute the multiplicative inverse of the matrix. |
||
| 1194 | * @return matrix |
||
| 1195 | */ |
||
| 1196 | public function inverse(): matrix { |
||
| 1197 | if (!$this->isSquare()) { |
||
| 1198 | self::_err('Error::invalid Size of matrix!'); |
||
| 1199 | } |
||
| 1200 | $imat = $this->copyMatrix(); |
||
| 1201 | $ipiv = vector::factory($this->row, vector::INT); |
||
| 1202 | $lp = lapack::getrf($imat, $ipiv); |
||
| 1203 | if ($lp != 0) { |
||
| 1204 | return null; |
||
| 1205 | } |
||
| 1206 | $lp = lapack::getri($imat, $ipiv); |
||
| 1207 | if ($lp != 0) { |
||
| 1208 | return null; |
||
| 1209 | } |
||
| 1210 | unset($ipiv); |
||
| 1211 | unset($lp); |
||
| 1212 | return $imat; |
||
| 1213 | } |
||
| 1214 | |||
| 1215 | /** |
||
| 1216 | * Compute the (Moore-Penrose) pseudo inverse of the general matrix. |
||
| 1217 | * @return matrix|null |
||
| 1218 | */ |
||
| 1219 | public function pseudoInverse(): matrix|null { |
||
| 1220 | $k = min($this->row, $this->col); |
||
| 1221 | $s = vector::factory($k, $this->dtype); |
||
| 1222 | $u = self::factory($this->row, $this->row, $this->dtype); |
||
| 1223 | $vt = self::factory($this->col, $this->col, $this->dtype); |
||
| 1224 | $imat = $this->copyMatrix(); |
||
| 1225 | $lp = lapack::gesdd($imat, $s, $u, $vt); |
||
| 1226 | if ($lp != 0) { |
||
| 1227 | return null; |
||
| 1228 | } |
||
| 1229 | for ($i = 0; $i < $k; ++$i) { |
||
| 1230 | blas::scale(1.0 / $s->data[$i], $vt->rowAsVector($i)); |
||
| 1231 | } |
||
| 1232 | unset($imat); |
||
| 1233 | unset($k); |
||
| 1234 | unset($lp); |
||
| 1235 | unset($s); |
||
| 1236 | $mr = self::factory($this->col, $this->row, $this->dtype); |
||
| 1237 | blas::gemm($vt, $u, $mr); |
||
| 1238 | unset($u); |
||
| 1239 | unset($vt); |
||
| 1240 | return $mr; |
||
| 1241 | } |
||
| 1242 | |||
| 1243 | /** |
||
| 1244 | * Compute the singular value decomposition of a matrix and |
||
| 1245 | * return an object of the singular values and unitary matrices |
||
| 1246 | * |
||
| 1247 | * @return object (u,s,v) |
||
| 1248 | */ |
||
| 1249 | public function svd(): svd { |
||
| 1250 | return svd::factory($this); |
||
| 1251 | } |
||
| 1252 | |||
| 1253 | /** |
||
| 1254 | * Compute the eigen decomposition of a general matrix. |
||
| 1255 | * return the eigenvalues and eigenvectors as object |
||
| 1256 | * |
||
| 1257 | * @param bool $symmetric |
||
| 1258 | * @return eigen |
||
| 1259 | */ |
||
| 1260 | public function eign(bool $symmetric = false): eigen { |
||
| 1261 | return eigen::factory($this, $symmetric); |
||
| 1262 | } |
||
| 1263 | |||
| 1264 | /** |
||
| 1265 | * |
||
| 1266 | * Compute the LU factorization of matrix. |
||
| 1267 | * return lower, upper, and permutation matrices as object. |
||
| 1268 | * |
||
| 1269 | * @return lu |
||
| 1270 | */ |
||
| 1271 | public function lu(): lu { |
||
| 1272 | return lu::factory($this); |
||
| 1273 | } |
||
| 1274 | |||
| 1275 | /** |
||
| 1276 | * Return the L1 norm of the matrix. |
||
| 1277 | * @return float |
||
| 1278 | */ |
||
| 1279 | public function normL1(): float { |
||
| 1280 | return lapack::lange('l', $this); |
||
| 1281 | } |
||
| 1282 | |||
| 1283 | /** |
||
| 1284 | * Return the L2 norm of the matrix. |
||
| 1285 | * @return float |
||
| 1286 | */ |
||
| 1287 | public function normL2(): float { |
||
| 1288 | return lapack::lange('f', $this); |
||
| 1289 | } |
||
| 1290 | |||
| 1291 | /** |
||
| 1292 | * Return the L1 norm of the matrix. |
||
| 1293 | * @return float |
||
| 1294 | */ |
||
| 1295 | public function normINF(): float { |
||
| 1296 | return lapack::lange('i', $this); |
||
| 1297 | } |
||
| 1298 | |||
| 1299 | /** |
||
| 1300 | * Return the Frobenius norm of the matrix. |
||
| 1301 | * @return float |
||
| 1302 | */ |
||
| 1303 | public function normFrob(): float { |
||
| 1304 | return $this->normL2(); |
||
| 1305 | } |
||
| 1306 | |||
| 1307 | /** |
||
| 1308 | * Run a function over all of the elements in the matrix. |
||
| 1309 | * @param callable $func |
||
| 1310 | * @return \Np\matrix |
||
| 1311 | */ |
||
| 1312 | public function map(callable $func): matrix { |
||
| 1313 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1314 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1315 | $ar->data[$i] = $func($this->data[$i]); |
||
| 1316 | } |
||
| 1317 | return $ar; |
||
| 1318 | } |
||
| 1319 | |||
| 1320 | public function abs(): matrix { |
||
| 1321 | return $this->map('abs'); |
||
| 1322 | } |
||
| 1323 | |||
| 1324 | public function sqrt(): matrix { |
||
| 1325 | return $this->map('sqrt'); |
||
| 1326 | } |
||
| 1327 | |||
| 1328 | public function exp(): matrix { |
||
| 1329 | return $this->map('exp'); |
||
| 1330 | } |
||
| 1331 | |||
| 1332 | public function exp1(): matrix { |
||
| 1333 | return $this->map('exp1'); |
||
| 1334 | } |
||
| 1335 | |||
| 1336 | public function log(float $b = M_E): matrix { |
||
| 1337 | $ar = $this->copyMatrix(); |
||
| 1338 | for ($i = 0; $i < $ar->ndim; ++$i) { |
||
| 1339 | log($ar->data[$i], $b); |
||
| 1340 | } |
||
| 1341 | return $ar; |
||
| 1342 | } |
||
| 1343 | |||
| 1344 | public function log1p(): matrix { |
||
| 1345 | return $this->map('log1p'); |
||
| 1346 | } |
||
| 1347 | |||
| 1348 | public function sin(): matrix { |
||
| 1349 | return $this->map('sin'); |
||
| 1350 | } |
||
| 1351 | |||
| 1352 | public function asin(): matrix { |
||
| 1353 | return $this->map('asin'); |
||
| 1354 | } |
||
| 1355 | |||
| 1356 | public function cos(): matrix { |
||
| 1357 | return $this->map('cos'); |
||
| 1358 | } |
||
| 1359 | |||
| 1360 | public function acos(): matrix { |
||
| 1361 | return $this->map('acos'); |
||
| 1362 | } |
||
| 1363 | |||
| 1364 | public function tan(): matrix { |
||
| 1365 | return $this->map('tan'); |
||
| 1366 | } |
||
| 1367 | |||
| 1368 | public function atan(): matrix { |
||
| 1369 | return $this->map('atan'); |
||
| 1370 | } |
||
| 1371 | |||
| 1372 | public function radToDeg(): matrix { |
||
| 1373 | return $this->map('rad2deg'); |
||
| 1374 | } |
||
| 1375 | |||
| 1376 | public function degToRad(): matrix { |
||
| 1377 | return $this->map('deg2rad'); |
||
| 1378 | } |
||
| 1379 | |||
| 1380 | public function floor(): matrix { |
||
| 1381 | return $this->map('floor'); |
||
| 1382 | } |
||
| 1383 | |||
| 1384 | public function ceil(): matrix { |
||
| 1385 | return $this->map('ceil'); |
||
| 1386 | } |
||
| 1387 | |||
| 1388 | /** |
||
| 1389 | * Compute the means of each row and return them in a vector. |
||
| 1390 | * |
||
| 1391 | * @return vector |
||
| 1392 | */ |
||
| 1393 | public function mean(): vector { |
||
| 1394 | return $this->sumRows()->divide($this->col); |
||
| 1395 | } |
||
| 1396 | |||
| 1397 | /** |
||
| 1398 | * Compute the row variance of the matrix. |
||
| 1399 | * |
||
| 1400 | * @param vector|null $mean |
||
| 1401 | * @return vector |
||
| 1402 | */ |
||
| 1403 | public function variance(vector|null $mean = null): vector { |
||
| 1404 | if (isset($mean)) { |
||
| 1405 | if (!$mean instanceof vector) { |
||
| 1406 | self::_invalidArgument('mean must be a vector!'); |
||
| 1407 | } |
||
| 1408 | if ($this->row !== $mean->col) { |
||
| 1409 | self::_err('Err:: given mean vector dimensionality mismatched!'); |
||
| 1410 | } |
||
| 1411 | } else { |
||
| 1412 | $mean = $this->mean(); |
||
| 1413 | } |
||
| 1414 | return $this->subtractColumnVector($mean)->square() |
||
| 1415 | ->sumRows()->divide($this->row); |
||
| 1416 | } |
||
| 1417 | |||
| 1418 | /** |
||
| 1419 | * Return the median vector of this matrix. |
||
| 1420 | * @return vector |
||
| 1421 | */ |
||
| 1422 | public function median(): vector { |
||
| 1423 | $mid = intdiv($this->col, 2); |
||
| 1424 | $odd = $this->col % 2 === 1; |
||
| 1425 | $vr = vector::factory($this->row, $this->dtype); |
||
| 1426 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1427 | $a = $this->rowAsVector($i)->sort(); |
||
| 1428 | if ($odd) { |
||
| 1429 | $median = $a->data[$mid]; |
||
| 1430 | } else { |
||
| 1431 | $median = ($a->data[$mid - 1] + $a->data[$mid]) / 2.0; |
||
| 1432 | } |
||
| 1433 | $vr->data[$i] = $median; |
||
| 1434 | } |
||
| 1435 | unset($a); |
||
| 1436 | return $vr; |
||
| 1437 | } |
||
| 1438 | |||
| 1439 | /** |
||
| 1440 | * Compute the covariance matrix. |
||
| 1441 | * |
||
| 1442 | * @param vector|null $mean |
||
| 1443 | * @return matrix |
||
| 1444 | */ |
||
| 1445 | public function covariance(vector|null $mean = null): matrix { |
||
| 1446 | if (isset($mean)) { |
||
| 1447 | if ($mean->col !== $this->row) { |
||
| 1448 | self::_err('Err:: given mean vector dimensionality mismatched!'); |
||
| 1449 | } |
||
| 1450 | } else { |
||
| 1451 | $mean = $this->mean(); |
||
| 1452 | } |
||
| 1453 | |||
| 1454 | $b = $this->subtractColumnVector($mean); |
||
| 1455 | |||
| 1456 | return $b->dot($b->transpose()) |
||
| 1457 | ->divideScalar($this->row); |
||
| 1458 | } |
||
| 1459 | |||
| 1460 | /** |
||
| 1461 | * Clip the elements in the matrix to be between given minimum and maximum |
||
| 1462 | * and return a new matrix. |
||
| 1463 | * |
||
| 1464 | * @param float $min |
||
| 1465 | * @param float $max |
||
| 1466 | * @return matrix |
||
| 1467 | */ |
||
| 1468 | public function clip(float $min, float $max): matrix { |
||
| 1469 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1470 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1471 | if ($this->data[$i] > $max) { |
||
| 1472 | $ar->data[$i] = $max; |
||
| 1473 | continue; |
||
| 1474 | } |
||
| 1475 | if ($this->data[$i] < $min) { |
||
| 1476 | $ar->data[$i] = $min; |
||
| 1477 | continue; |
||
| 1478 | } |
||
| 1479 | $ar->data[$i] = $this->data[$i]; |
||
| 1480 | } |
||
| 1481 | return $ar; |
||
| 1482 | } |
||
| 1483 | |||
| 1484 | /** |
||
| 1485 | * Clip the matrix to be lower bounded by a given minimum. |
||
| 1486 | * @param float $min |
||
| 1487 | * @return matrix |
||
| 1488 | */ |
||
| 1489 | public function clipLower(float $min): matrix { |
||
| 1490 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1491 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1492 | if ($this->data[$i] < $min) { |
||
| 1493 | $ar->data[$i] = $min; |
||
| 1494 | continue; |
||
| 1495 | } |
||
| 1496 | $ar->data[$i] = $this->data[$i]; |
||
| 1497 | } |
||
| 1498 | return $ar; |
||
| 1499 | } |
||
| 1500 | |||
| 1501 | /** |
||
| 1502 | * Clip the matrix to be upper bounded by a given maximum. |
||
| 1503 | * |
||
| 1504 | * @param float $max |
||
| 1505 | * @return matrix |
||
| 1506 | */ |
||
| 1507 | public function clipUpper(float $max): matrix { |
||
| 1508 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1509 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1510 | if ($this->data[$i] > $max) { |
||
| 1511 | $ar->data[$i] = $max; |
||
| 1512 | continue; |
||
| 1513 | } |
||
| 1514 | $ar->data[$i] = $this->data[$i]; |
||
| 1515 | } |
||
| 1516 | return $ar; |
||
| 1517 | } |
||
| 1518 | |||
| 1519 | /** |
||
| 1520 | * Square of matrix |
||
| 1521 | * @return matrix |
||
| 1522 | */ |
||
| 1523 | public function square(): matrix { |
||
| 1524 | return $this->multiplyMatrix($this); |
||
| 1525 | } |
||
| 1526 | |||
| 1527 | /** |
||
| 1528 | * |
||
| 1529 | * @param int|float|matrix|vector $d |
||
| 1530 | * @return matrix |
||
| 1531 | */ |
||
| 1532 | public function equal(int|float|matrix|vector $d): matrix { |
||
| 1533 | if ($d instanceof self) { |
||
| 1534 | return $this->equalMatrix($d); |
||
| 1535 | } elseif ($d instanceof vector) { |
||
| 1536 | return $this->equalVector($d); |
||
| 1537 | } else { |
||
| 1538 | return $this->equalScalar($d); |
||
| 1539 | } |
||
| 1540 | } |
||
| 1541 | |||
| 1542 | protected function equalMatrix(matrix $m): matrix { |
||
| 1543 | if ($this->checkShape($this, $m) && $this->checkDtype($this, $m)) { |
||
| 1544 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1545 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1546 | $ar->data[$i] = $this->data[$i] == $m->data[$i] ? 1 : 0; |
||
| 1547 | } |
||
| 1548 | return $ar; |
||
| 1549 | } |
||
| 1550 | } |
||
| 1551 | |||
| 1552 | protected function equalVector(vector $v): matrix { |
||
| 1553 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this, $v)) { |
||
| 1554 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1555 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1556 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 1557 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j] == $v->data[$j] ? 1 : 0; |
||
| 1558 | } |
||
| 1559 | } |
||
| 1560 | return $ar; |
||
| 1561 | } |
||
| 1562 | } |
||
| 1563 | |||
| 1564 | protected function equalScalar(int|float $s): matrix { |
||
| 1565 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1566 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1567 | $ar->data[$i] = $this->data[$i] == $s ? 1 : 0; |
||
| 1568 | } |
||
| 1569 | return $ar; |
||
| 1570 | } |
||
| 1571 | |||
| 1572 | /** |
||
| 1573 | * |
||
| 1574 | * @param int|float|matrix|vector $d |
||
| 1575 | * @return matrix |
||
| 1576 | */ |
||
| 1577 | public function greater(int|float|matrix|vector $d): matrix { |
||
| 1584 | } |
||
| 1585 | } |
||
| 1586 | |||
| 1587 | protected function greaterMatrix(matrix $m): matrix { |
||
| 1588 | if ($this->checkShape($this, $m) && $this->checkDtype($this,$m)) { |
||
| 1589 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1590 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1591 | $ar->data[$i] = $this->data[$i] > $m->data[$i] ? 1 : 0; |
||
| 1592 | } |
||
| 1593 | return $ar; |
||
| 1594 | } |
||
| 1595 | } |
||
| 1596 | |||
| 1597 | protected function greaterVector(vector $v): matrix { |
||
| 1598 | if ($this->checkDimensions($v, $this) && $this->checkDtype($this,$v)) { |
||
| 1599 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1600 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1601 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 1602 | $ar->data[$i * $this->col + $j] = $this->data[$i * $this->col + $j] > $v->data[$j] ? 1 : 0; |
||
| 1603 | } |
||
| 1604 | } |
||
| 1605 | return $ar; |
||
| 1606 | } |
||
| 1607 | } |
||
| 1608 | |||
| 1609 | protected function greaterScalar(int|float $s): matrix { |
||
| 1610 | $ar = self::factory($this->row, $this->col, $this->dtype); |
||
| 1611 | for ($i = 0; $i < $this->ndim; ++$i) { |
||
| 1612 | $ar->data[$i] = $this->data[$i] > $s ? 1 : 0; |
||
| 1613 | } |
||
| 1614 | return $ar; |
||
| 1615 | } |
||
| 1616 | |||
| 1617 | /** |
||
| 1618 | * |
||
| 1619 | * @param int|float|matrix $m |
||
| 1620 | * @return matrix |
||
| 1621 | */ |
||
| 1622 | public function less(int|float|matrix $m): matrix { |
||
| 1636 | } |
||
| 1637 | } |
||
| 1638 | |||
| 1639 | /** |
||
| 1640 | * Is the matrix symmetric i.e. is it equal to its own transpose? |
||
| 1641 | * |
||
| 1642 | * @return bool |
||
| 1643 | */ |
||
| 1644 | public function isSymmetric(): bool { |
||
| 1645 | if (!$this->isSquare()) { |
||
| 1646 | return false; |
||
| 1647 | } |
||
| 1648 | $ar = $this->transpose(); |
||
| 1649 | for ($i = 0; $i < $ar->ndim; ++$i) { |
||
| 1650 | if ($ar->data[$i] != $this->data[$i]) { |
||
| 1651 | unset($ar); |
||
| 1652 | return false; |
||
| 1653 | } |
||
| 1654 | } |
||
| 1655 | unset($ar); |
||
| 1656 | return true; |
||
| 1657 | } |
||
| 1658 | |||
| 1659 | /** |
||
| 1660 | * Reshape current matrix. |
||
| 1661 | * @param int $row |
||
| 1662 | * @param int $col |
||
| 1663 | * @return matrix |
||
| 1664 | */ |
||
| 1665 | public function reshape(int $row, int $col):matrix { |
||
| 1666 | if($this->ndim != $row * $col) { |
||
| 1667 | self::_dimensionaMisMatchErr('given dimenssion is not valid for current bufferData'); |
||
| 1668 | } |
||
| 1669 | $this->row = $row; |
||
| 1670 | $this->col = $col; |
||
| 1671 | return $this; |
||
| 1672 | } |
||
| 1673 | |||
| 1674 | /** |
||
| 1675 | * print the matrix in consol |
||
| 1676 | */ |
||
| 1677 | public function printMatrix() { |
||
| 1678 | echo __CLASS__ . PHP_EOL; |
||
| 1679 | for ($i = 0; $i < $this->row; ++$i) { |
||
| 1680 | for ($j = 0; $j < $this->col; ++$j) { |
||
| 1681 | printf('%lf ', $this->data[$i * $this->col + $j]); |
||
| 1682 | } |
||
| 1683 | echo PHP_EOL; |
||
| 1684 | } |
||
| 1685 | } |
||
| 1686 | |||
| 1687 | public function __toString() { |
||
| 1689 | } |
||
| 1690 | |||
| 1691 | private function flattenArray(array $ar) { |
||
| 1692 | if (is_array($ar) && is_array($ar[0])) { |
||
| 1693 | $a = []; |
||
| 1694 | foreach ($ar as $y => $value) { |
||
| 1695 | foreach ($value as $k => $v) { |
||
| 1696 | $a[] = $v; |
||
| 1697 | } |
||
| 1698 | } |
||
| 1699 | return $a; |
||
| 1700 | } |
||
| 1701 | } |
||
| 1702 | |||
| 1703 | /** |
||
| 1704 | * |
||
| 1705 | * @param int $row |
||
| 1706 | * @param int $col |
||
| 1707 | * @param int $dtype |
||
| 1708 | * @return $this |
||
| 1709 | */ |
||
| 1710 | protected function __construct(public int $row, public int $col, int $dtype = self::Float) { |
||
| 1716 | } |
||
| 1717 | } |
||
| 1718 |
Loading history...