platine-php /
expression
| Total Complexity | 90 |
| Total Lines | 469 |
| Duplicated Lines | 0 % |
| Changes | 2 | ||
| Bugs | 0 | Features | 0 |
Complex classes like Tokenizer 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 Tokenizer, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 59 | class Tokenizer |
||
| 60 | { |
||
| 61 | /** |
||
| 62 | * List of token |
||
| 63 | * @var Token[] |
||
| 64 | */ |
||
| 65 | protected array $tokens = []; |
||
| 66 | |||
| 67 | /** |
||
| 68 | * The expression to evaluate |
||
| 69 | * @var string |
||
| 70 | */ |
||
| 71 | protected string $input = ''; |
||
| 72 | |||
| 73 | /** |
||
| 74 | * The number buffering |
||
| 75 | * @var string |
||
| 76 | */ |
||
| 77 | protected string $numberBuffer = ''; |
||
| 78 | |||
| 79 | /** |
||
| 80 | * The string buffering |
||
| 81 | * @var string |
||
| 82 | */ |
||
| 83 | protected string $stringBuffer = ''; |
||
| 84 | |||
| 85 | /** |
||
| 86 | * Whether to allow negative or not |
||
| 87 | * @var bool |
||
| 88 | */ |
||
| 89 | protected bool $allowNegative = true; |
||
| 90 | |||
| 91 | /** |
||
| 92 | * The list of operator |
||
| 93 | * @var array<string, Operator> |
||
| 94 | */ |
||
| 95 | protected array $operators = []; |
||
| 96 | |||
| 97 | /** |
||
| 98 | * Whether the current pointer is inside single quoted string |
||
| 99 | * @var bool |
||
| 100 | */ |
||
| 101 | protected bool $inSingleQuotedString = false; |
||
| 102 | |||
| 103 | /** |
||
| 104 | * Whether the current pointer is inside double quoted string |
||
| 105 | * @var bool |
||
| 106 | */ |
||
| 107 | protected bool $inDoubleQuotedString = false; |
||
| 108 | |||
| 109 | /** |
||
| 110 | * Create new instance |
||
| 111 | * @param string $input |
||
| 112 | * @param array<string, Operator> $operators |
||
| 113 | */ |
||
| 114 | public function __construct(string $input, array $operators) |
||
| 118 | } |
||
| 119 | |||
| 120 | /** |
||
| 121 | * Produce the tokens |
||
| 122 | * @return $this |
||
| 123 | */ |
||
| 124 | public function tokenize(): self |
||
| 320 | } |
||
| 321 | |||
| 322 | /** |
||
| 323 | * Build the reverse polish notation |
||
| 324 | * @return Token[] |
||
| 325 | */ |
||
| 326 | public function buildReversePolishNotation(): array |
||
| 327 | { |
||
| 328 | $tokens = []; |
||
| 329 | /** @var SplStack<Token> $stack */ |
||
| 330 | $stack = new SplStack(); |
||
| 331 | |||
| 332 | /** @var SplStack<int> $paramCounter */ |
||
| 333 | $paramCounter = new SplStack(); |
||
| 334 | foreach ($this->tokens as $token) { |
||
| 335 | switch ($token->getType()) { |
||
| 336 | case Token::LITERAL: |
||
| 337 | case Token::VARIABLE: |
||
| 338 | case Token::STRING: |
||
| 339 | $tokens[] = $token; |
||
| 340 | |||
| 341 | if ($paramCounter->count() > 0 && $paramCounter->top() === 0) { |
||
| 342 | $paramCounter->push($paramCounter->pop() + 1); |
||
| 343 | } |
||
| 344 | |||
| 345 | break; |
||
| 346 | |||
| 347 | case Token::FUNCTION: |
||
| 348 | if ($paramCounter->count() > 0 && $paramCounter->top() === 0) { |
||
| 349 | $paramCounter->push($paramCounter->pop() + 1); |
||
| 350 | } |
||
| 351 | $stack->push($token); |
||
| 352 | $paramCounter->push(0); |
||
| 353 | |||
| 354 | break; |
||
| 355 | |||
| 356 | case Token::LEFT_PARENTHESIS: |
||
| 357 | $stack->push($token); |
||
| 358 | |||
| 359 | break; |
||
| 360 | |||
| 361 | case Token::PARAM_SEPARATOR: |
||
| 362 | while (Token::LEFT_PARENTHESIS !== $stack->top()->getType()) { |
||
| 363 | if ($stack->count() === 0) { |
||
| 364 | throw new IncorrectBracketsException('Incorrect brackets'); |
||
| 365 | } |
||
| 366 | $tokens[] = $stack->pop(); |
||
| 367 | } |
||
| 368 | |||
| 369 | $paramCounter->push($paramCounter->pop() + 1); |
||
| 370 | break; |
||
| 371 | |||
| 372 | case Token::OPERATOR: |
||
| 373 | if (!array_key_exists($token->getValue(), $this->operators)) { |
||
| 374 | throw new UnknownOperatorException(sprintf( |
||
| 375 | 'Unknown operator [%s]', |
||
| 376 | $token->getValue() |
||
| 377 | )); |
||
| 378 | } |
||
| 379 | $op1 = $this->operators[$token->getValue()]; |
||
| 380 | while ($stack->count() > 0 && Token::OPERATOR === $stack->top()->getType()) { |
||
| 381 | if (!array_key_exists($stack->top()->getValue(), $this->operators)) { |
||
| 382 | throw new UnknownOperatorException(sprintf( |
||
| 383 | 'Unknown operator [%s]', |
||
| 384 | $stack->top()->getValue() |
||
| 385 | )); |
||
| 386 | } |
||
| 387 | $op2 = $this->operators[$stack->top()->getValue()]; |
||
| 388 | if ($op2->getPriority() >= $op1->getPriority()) { |
||
| 389 | $tokens[] = $stack->pop(); |
||
| 390 | |||
| 391 | continue; |
||
| 392 | } |
||
| 393 | |||
| 394 | break; |
||
| 395 | } |
||
| 396 | $stack->push($token); |
||
| 397 | break; |
||
| 398 | |||
| 399 | case Token::RIGHT_PARENTHESIS: |
||
| 400 | while (true) { |
||
| 401 | try { |
||
| 402 | $ctoken = $stack->pop(); |
||
| 403 | if (Token::LEFT_PARENTHESIS === $ctoken->getType()) { |
||
| 404 | break; |
||
| 405 | } |
||
| 406 | $tokens[] = $ctoken; |
||
| 407 | } catch (RuntimeException $ex) { |
||
| 408 | throw new IncorrectBracketsException('Incorrect brackets'); |
||
| 409 | } |
||
| 410 | } |
||
| 411 | if ($stack->count() > 0 && Token::FUNCTION === $stack->top()->getType()) { |
||
| 412 | /** @var Token $funcToken */ |
||
| 413 | $funcToken = $stack->pop(); |
||
| 414 | $funcToken->setParamCount($paramCounter->pop()); |
||
| 415 | $tokens[] = $funcToken; |
||
| 416 | } |
||
| 417 | break; |
||
| 418 | case Token::SPACE: |
||
| 419 | // do nothing |
||
| 420 | } |
||
| 421 | } |
||
| 422 | |||
| 423 | while ($stack->count() !== 0) { |
||
| 424 | if ( |
||
| 425 | in_array( |
||
| 426 | $stack->top()->getType(), |
||
| 427 | [Token::LEFT_PARENTHESIS, Token::RIGHT_PARENTHESIS] |
||
| 428 | ) |
||
| 429 | ) { |
||
| 430 | throw new IncorrectBracketsException('Incorrect brackets'); |
||
| 431 | } |
||
| 432 | |||
| 433 | if (Token::SPACE === $stack->top()->getType()) { |
||
| 434 | $stack->pop(); |
||
| 435 | |||
| 436 | continue; |
||
| 437 | } |
||
| 438 | $tokens[] = $stack->pop(); |
||
| 439 | } |
||
| 440 | |||
| 441 | return $tokens; |
||
| 442 | } |
||
| 443 | |||
| 444 | /** |
||
| 445 | * Put the current number buffer content to token |
||
| 446 | * as literal |
||
| 447 | * @return void |
||
| 448 | */ |
||
| 449 | protected function emptyNumberBufferAsLiteral(): void |
||
| 450 | { |
||
| 451 | if (strlen($this->numberBuffer) > 0) { |
||
| 452 | $this->tokens[] = new Token(Token::LITERAL, $this->numberBuffer); |
||
| 453 | $this->numberBuffer = ''; |
||
| 454 | } |
||
| 455 | } |
||
| 456 | |||
| 457 | /** |
||
| 458 | * Put the current string buffer content to token |
||
| 459 | * as variable |
||
| 460 | * @return void |
||
| 461 | */ |
||
| 462 | protected function emptyStringBufferAsVariable(): void |
||
| 463 | { |
||
| 464 | if (strlen($this->stringBuffer) > 0) { |
||
| 465 | $this->tokens[] = new Token(Token::VARIABLE, $this->stringBuffer); |
||
| 466 | $this->stringBuffer = ''; |
||
| 467 | } |
||
| 468 | } |
||
| 469 | |||
| 470 | /** |
||
| 471 | * Whether the give argument is number |
||
| 472 | * @param string $chr |
||
| 473 | * @return bool |
||
| 474 | */ |
||
| 475 | protected function isNumber(string $chr): bool |
||
| 476 | { |
||
| 477 | return $chr >= '0' && $chr <= '9'; |
||
| 478 | } |
||
| 479 | |||
| 480 | /** |
||
| 481 | * Whether the give argument is alphabetic |
||
| 482 | * @param string $chr |
||
| 483 | * @return bool |
||
| 484 | */ |
||
| 485 | protected function isAlpha(string $chr): bool |
||
| 486 | { |
||
| 487 | return $chr >= 'a' && $chr <= 'z' || $chr >= 'A' && $chr <= 'Z' || $chr === '_'; |
||
| 488 | } |
||
| 489 | |||
| 490 | /** |
||
| 491 | * Whether the give argument is dot |
||
| 492 | * @param string $chr |
||
| 493 | * @return bool |
||
| 494 | */ |
||
| 495 | protected function isDot(string $chr): bool |
||
| 496 | { |
||
| 497 | return $chr === '.'; |
||
| 498 | } |
||
| 499 | |||
| 500 | /** |
||
| 501 | * Whether the give argument is left parenthesis |
||
| 502 | * @param string $chr |
||
| 503 | * @return bool |
||
| 504 | */ |
||
| 505 | protected function isLeftParenthesis(string $chr): bool |
||
| 506 | { |
||
| 507 | return $chr === '('; |
||
| 508 | } |
||
| 509 | |||
| 510 | /** |
||
| 511 | * Whether the give argument is right parenthesis |
||
| 512 | * @param string $chr |
||
| 513 | * @return bool |
||
| 514 | */ |
||
| 515 | protected function isRightParenthesis(string $chr): bool |
||
| 518 | } |
||
| 519 | |||
| 520 | /** |
||
| 521 | * Whether the give argument is comma |
||
| 522 | * @param string $chr |
||
| 523 | * @return bool |
||
| 524 | */ |
||
| 525 | protected function isComma(string $chr): bool |
||
| 528 | } |
||
| 529 | } |
||
| 530 |
