yiisoft /
db-sqlite
| Total Complexity | 49 |
| Total Lines | 449 |
| Duplicated Lines | 0 % |
| Changes | 0 | ||
Complex classes like BaseTokenizer 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 BaseTokenizer, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 34 | abstract class BaseTokenizer |
||
| 35 | { |
||
| 36 | /** |
||
| 37 | * @var string SQL code. |
||
| 38 | */ |
||
| 39 | private string $sql; |
||
| 40 | |||
| 41 | /** |
||
| 42 | * @var int SQL code string length. |
||
| 43 | */ |
||
| 44 | protected int $length; |
||
| 45 | |||
| 46 | /** |
||
| 47 | * @var int SQL code string current offset. |
||
| 48 | */ |
||
| 49 | protected int $offset; |
||
| 50 | |||
| 51 | /** |
||
| 52 | * @var SplStack stack of active tokens. |
||
| 53 | */ |
||
| 54 | private SplStack $tokenStack; |
||
| 55 | |||
| 56 | /** |
||
| 57 | * @var SqlToken|null active token. It's usually a top of the token stack. |
||
| 58 | */ |
||
| 59 | private ?SqlToken $currentToken = null; |
||
| 60 | |||
| 61 | /** |
||
| 62 | * @var string[] cached substrings. |
||
| 63 | */ |
||
| 64 | private array $substrings; |
||
| 65 | |||
| 66 | /** |
||
| 67 | * @var string string current buffer value. |
||
| 68 | */ |
||
| 69 | private string $buffer = ''; |
||
| 70 | |||
| 71 | /** |
||
| 72 | * @var SqlToken resulting token of a last {@see tokenize()} call. |
||
| 73 | */ |
||
| 74 | private ?SqlToken $token = null; |
||
| 75 | |||
| 76 | public function __construct(string $sql) |
||
| 77 | { |
||
| 78 | $this->sql = $sql; |
||
| 79 | } |
||
| 80 | |||
| 81 | /** |
||
| 82 | * Tokenizes and returns a code type token. |
||
| 83 | * |
||
| 84 | * @return SqlToken code type token. |
||
| 85 | */ |
||
| 86 | public function tokenize(): SqlToken |
||
| 87 | { |
||
| 88 | $this->length = mb_strlen($this->sql, 'UTF-8'); |
||
| 89 | $this->offset = 0; |
||
| 90 | $this->substrings = []; |
||
| 91 | $this->buffer = ''; |
||
| 92 | |||
| 93 | $this->token = (new SqlToken()) |
||
| 94 | ->type(SqlToken::TYPE_CODE) |
||
| 95 | ->content($this->sql); |
||
| 96 | |||
| 97 | $this->tokenStack = new SplStack(); |
||
| 98 | $this->tokenStack->push($this->token); |
||
| 99 | |||
| 100 | $tk = (new SqlToken()) |
||
| 101 | ->type(SqlToken::TYPE_STATEMENT); |
||
| 102 | |||
| 103 | $this->token[] = $tk; |
||
| 104 | |||
| 105 | $this->tokenStack->push($this->token[0]); |
||
| 106 | $this->currentToken = $this->tokenStack->top(); |
||
| 107 | |||
| 108 | while (!$this->isEof()) { |
||
| 109 | if ($this->isWhitespace($length) || $this->isComment($length)) { |
||
| 110 | $this->addTokenFromBuffer(); |
||
| 111 | $this->advance($length); |
||
| 112 | |||
| 113 | continue; |
||
| 114 | } |
||
| 115 | |||
| 116 | if ($this->tokenizeOperator($length) || $this->tokenizeDelimitedString($length)) { |
||
| 117 | $this->advance($length); |
||
| 118 | |||
| 119 | continue; |
||
| 120 | } |
||
| 121 | |||
| 122 | $this->buffer .= $this->substring(1); |
||
| 123 | $this->advance(1); |
||
| 124 | } |
||
| 125 | $this->addTokenFromBuffer(); |
||
| 126 | if ($this->token->getHasChildren() && !$this->token[-1]->getHasChildren()) { |
||
|
|
|||
| 127 | unset($this->token[-1]); |
||
| 128 | } |
||
| 129 | |||
| 130 | return $this->token; |
||
| 131 | } |
||
| 132 | |||
| 133 | /** |
||
| 134 | * Returns whether there's a whitespace at the current offset. |
||
| 135 | * |
||
| 136 | * If this method returns `true`, it has to set the `$length` parameter to the length of the matched string. |
||
| 137 | * |
||
| 138 | * @param int|null $length length of the matched string. |
||
| 139 | * |
||
| 140 | * @return bool whether there's a whitespace at the current offset. |
||
| 141 | */ |
||
| 142 | abstract protected function isWhitespace(?int &$length): bool; |
||
| 143 | |||
| 144 | /** |
||
| 145 | * Returns whether there's a commentary at the current offset. |
||
| 146 | * |
||
| 147 | * If this method returns `true`, it has to set the `$length` parameter to the length of the matched string. |
||
| 148 | * |
||
| 149 | * @param int $length length of the matched string. |
||
| 150 | * |
||
| 151 | * @return bool whether there's a commentary at the current offset. |
||
| 152 | */ |
||
| 153 | abstract protected function isComment(int &$length): bool; |
||
| 154 | |||
| 155 | /** |
||
| 156 | * Returns whether there's an operator at the current offset. |
||
| 157 | * |
||
| 158 | * If this method returns `true`, it has to set the `$length` parameter to the length of the matched string. It may |
||
| 159 | * also set `$content` to a string that will be used as a token content. |
||
| 160 | * |
||
| 161 | * @param int $length length of the matched string. |
||
| 162 | * @param string|null $content optional content instead of the matched string. |
||
| 163 | * |
||
| 164 | * @return bool whether there's an operator at the current offset. |
||
| 165 | */ |
||
| 166 | abstract protected function isOperator(int &$length, ?string &$content): bool; |
||
| 167 | |||
| 168 | /** |
||
| 169 | * Returns whether there's an identifier at the current offset. |
||
| 170 | * |
||
| 171 | * If this method returns `true`, it has to set the `$length` parameter to the length of the matched string. It may |
||
| 172 | * also set `$content` to a string that will be used as a token content. |
||
| 173 | * |
||
| 174 | * @param int $length length of the matched string. |
||
| 175 | * @param string|null $content optional content instead of the matched string. |
||
| 176 | * |
||
| 177 | * @return bool whether there's an identifier at the current offset. |
||
| 178 | */ |
||
| 179 | abstract protected function isIdentifier(int &$length, ?string &$content): bool; |
||
| 180 | |||
| 181 | /** |
||
| 182 | * Returns whether there's a string literal at the current offset. |
||
| 183 | * |
||
| 184 | * If this method returns `true`, it has to set the `$length` parameter to the length of the matched string. It may |
||
| 185 | * also set `$content` to a string that will be used as a token content. |
||
| 186 | * |
||
| 187 | * @param int $length length of the matched string. |
||
| 188 | * @param string|null $content optional content instead of the matched string. |
||
| 189 | * |
||
| 190 | * @return bool whether there's a string literal at the current offset. |
||
| 191 | */ |
||
| 192 | abstract protected function isStringLiteral(int &$length, ?string &$content): bool; |
||
| 193 | |||
| 194 | /** |
||
| 195 | * Returns whether the given string is a keyword. |
||
| 196 | * |
||
| 197 | * The method may set `$content` to a string that will be used as a token content. |
||
| 198 | * |
||
| 199 | * @param string $string string to be matched. |
||
| 200 | * @param string|null $content optional content instead of the matched string. |
||
| 201 | * |
||
| 202 | * @return bool whether the given string is a keyword. |
||
| 203 | */ |
||
| 204 | abstract protected function isKeyword(string $string, ?string &$content): bool; |
||
| 205 | |||
| 206 | /** |
||
| 207 | * @param string $sql |
||
| 208 | */ |
||
| 209 | public function setSql(string $sql): void |
||
| 210 | { |
||
| 211 | $this->sql = $sql; |
||
| 212 | } |
||
| 213 | |||
| 214 | /** |
||
| 215 | * Returns whether the longest common prefix equals to the SQL code of the same length at the current offset. |
||
| 216 | * |
||
| 217 | * @param string[] $with strings to be tested. The method **will** modify this parameter to speed up lookups. |
||
| 218 | * @param bool $caseSensitive whether to perform a case sensitive comparison. |
||
| 219 | * @param int|null $length length of the matched string. |
||
| 220 | * @param string|null $content matched string. |
||
| 221 | * |
||
| 222 | * @return bool whether a match is found. |
||
| 223 | */ |
||
| 224 | protected function startsWithAnyLongest( |
||
| 225 | array &$with, |
||
| 226 | bool $caseSensitive, |
||
| 227 | ?int &$length = null, |
||
| 228 | ?string &$content = null |
||
| 229 | ): bool { |
||
| 230 | if (empty($with)) { |
||
| 231 | return false; |
||
| 232 | } |
||
| 233 | |||
| 234 | if (!is_array(reset($with))) { |
||
| 235 | usort($with, static function ($string1, $string2) { |
||
| 236 | return mb_strlen($string2, 'UTF-8') - mb_strlen($string1, 'UTF-8'); |
||
| 237 | }); |
||
| 238 | |||
| 239 | $map = []; |
||
| 240 | |||
| 241 | foreach ($with as $string) { |
||
| 242 | $map[mb_strlen($string, 'UTF-8')][$caseSensitive ? $string : mb_strtoupper($string, 'UTF-8')] = true; |
||
| 243 | } |
||
| 244 | |||
| 245 | $with = $map; |
||
| 246 | } |
||
| 247 | foreach ($with as $testLength => $testValues) { |
||
| 248 | $content = $this->substring($testLength, $caseSensitive); |
||
| 249 | |||
| 250 | if (isset($testValues[$content])) { |
||
| 251 | $length = $testLength; |
||
| 252 | |||
| 253 | return true; |
||
| 254 | } |
||
| 255 | } |
||
| 256 | |||
| 257 | return false; |
||
| 258 | } |
||
| 259 | |||
| 260 | /** |
||
| 261 | * Returns a string of the given length starting with the specified offset. |
||
| 262 | * |
||
| 263 | * @param int $length string length to be returned. |
||
| 264 | * @param bool $caseSensitive if it's `false`, the string will be uppercased. |
||
| 265 | * @param int|null $offset SQL code offset, defaults to current if `null` is passed. |
||
| 266 | * |
||
| 267 | * @return string result string, it may be empty if there's nothing to return. |
||
| 268 | */ |
||
| 269 | protected function substring(int $length, bool $caseSensitive = true, ?int $offset = null): string |
||
| 270 | { |
||
| 271 | if ($offset === null) { |
||
| 272 | $offset = $this->offset; |
||
| 273 | } |
||
| 274 | |||
| 275 | if ($offset + $length > $this->length) { |
||
| 276 | return ''; |
||
| 277 | } |
||
| 278 | |||
| 279 | $cacheKey = $offset . ',' . $length; |
||
| 280 | |||
| 281 | if (!isset($this->substrings[$cacheKey . ',1'])) { |
||
| 282 | $this->substrings[$cacheKey . ',1'] = mb_substr($this->sql, $offset, $length, 'UTF-8'); |
||
| 283 | } |
||
| 284 | if (!$caseSensitive && !isset($this->substrings[$cacheKey . ',0'])) { |
||
| 285 | $this->substrings[$cacheKey . ',0'] = mb_strtoupper($this->substrings[$cacheKey . ',1'], 'UTF-8'); |
||
| 286 | } |
||
| 287 | |||
| 288 | return $this->substrings[$cacheKey . ',' . (int) $caseSensitive]; |
||
| 289 | } |
||
| 290 | |||
| 291 | /** |
||
| 292 | * Returns an index after the given string in the SQL code starting with the specified offset. |
||
| 293 | * |
||
| 294 | * @param string $string string to be found. |
||
| 295 | * @param int|null $offset SQL code offset, defaults to current if `null` is passed. |
||
| 296 | * |
||
| 297 | * @return int index after the given string or end of string index. |
||
| 298 | */ |
||
| 299 | protected function indexAfter(string $string, ?int $offset = null): int |
||
| 300 | { |
||
| 301 | if ($offset === null) { |
||
| 302 | $offset = $this->offset; |
||
| 303 | } |
||
| 304 | |||
| 305 | if ($offset + mb_strlen($string, 'UTF-8') > $this->length) { |
||
| 306 | return $this->length; |
||
| 307 | } |
||
| 308 | |||
| 309 | $afterIndexOf = mb_strpos($this->sql, $string, $offset, 'UTF-8'); |
||
| 310 | |||
| 311 | if ($afterIndexOf === false) { |
||
| 312 | $afterIndexOf = $this->length; |
||
| 313 | } else { |
||
| 314 | $afterIndexOf += mb_strlen($string, 'UTF-8'); |
||
| 315 | } |
||
| 316 | |||
| 317 | return $afterIndexOf; |
||
| 318 | } |
||
| 319 | |||
| 320 | /** |
||
| 321 | * Determines whether there is a delimited string at the current offset and adds it to the token children. |
||
| 322 | * |
||
| 323 | * @param int $length |
||
| 324 | * |
||
| 325 | * @return bool |
||
| 326 | */ |
||
| 327 | private function tokenizeDelimitedString(int &$length): bool |
||
| 328 | { |
||
| 329 | $isIdentifier = $this->isIdentifier($length, $content); |
||
| 330 | $isStringLiteral = !$isIdentifier && $this->isStringLiteral($length, $content); |
||
| 331 | |||
| 332 | if (!$isIdentifier && !$isStringLiteral) { |
||
| 333 | return false; |
||
| 334 | } |
||
| 335 | |||
| 336 | $this->addTokenFromBuffer(); |
||
| 337 | |||
| 338 | $tk = (new SqlToken()) |
||
| 339 | ->type($isIdentifier ? SqlToken::TYPE_IDENTIFIER : SqlToken::TYPE_STRING_LITERAL) |
||
| 340 | ->content(is_string($content) ? $content : $this->substring($length)) |
||
| 341 | ->startOffset($this->offset) |
||
| 342 | ->endOffset($this->offset + $length); |
||
| 343 | |||
| 344 | $this->currentToken[] = $tk; |
||
| 345 | |||
| 346 | return true; |
||
| 347 | } |
||
| 348 | |||
| 349 | /** |
||
| 350 | * Determines whether there is an operator at the current offset and adds it to the token children. |
||
| 351 | * |
||
| 352 | * @param int $length |
||
| 353 | * |
||
| 354 | * @return bool |
||
| 355 | */ |
||
| 356 | private function tokenizeOperator(int &$length): bool |
||
| 357 | { |
||
| 358 | if (!$this->isOperator($length, $content)) { |
||
| 359 | return false; |
||
| 360 | } |
||
| 361 | |||
| 362 | $this->addTokenFromBuffer(); |
||
| 363 | |||
| 364 | switch ($this->substring($length)) { |
||
| 365 | case '(': |
||
| 366 | $tk = (new SqlToken()) |
||
| 367 | ->type(SqlToken::TYPE_OPERATOR) |
||
| 368 | ->content(is_string($content) ? $content : $this->substring($length)) |
||
| 369 | ->startOffset($this->offset) |
||
| 370 | ->endOffset($this->offset + $length); |
||
| 371 | |||
| 372 | $this->currentToken[] = $tk; |
||
| 373 | |||
| 374 | $tk1 = (new SqlToken()) |
||
| 375 | ->type(SqlToken::TYPE_PARENTHESIS); |
||
| 376 | |||
| 377 | $this->currentToken[] = $tk1; |
||
| 378 | |||
| 379 | $this->tokenStack->push($this->currentToken[-1]); |
||
| 380 | $this->currentToken = $this->tokenStack->top(); |
||
| 381 | |||
| 382 | break; |
||
| 383 | |||
| 384 | case ')': |
||
| 385 | $this->tokenStack->pop(); |
||
| 386 | $this->currentToken = $this->tokenStack->top(); |
||
| 387 | |||
| 388 | $tk = (new SqlToken()) |
||
| 389 | ->type(SqlToken::TYPE_OPERATOR) |
||
| 390 | ->content(')') |
||
| 391 | ->startOffset($this->offset) |
||
| 392 | ->endOffset($this->offset + $length); |
||
| 393 | |||
| 394 | $this->currentToken[] = $tk; |
||
| 395 | |||
| 396 | break; |
||
| 397 | case ';': |
||
| 398 | if (!$this->currentToken->getHasChildren()) { |
||
| 399 | break; |
||
| 400 | } |
||
| 401 | |||
| 402 | $tk = (new SqlToken()) |
||
| 403 | ->type(SqlToken::TYPE_OPERATOR) |
||
| 404 | ->content(is_string($content) ? $content : $this->substring($length)) |
||
| 405 | ->startOffset($this->offset) |
||
| 406 | ->endOffset($this->offset + $length); |
||
| 407 | |||
| 408 | $this->currentToken[] = $tk; |
||
| 409 | |||
| 410 | $this->tokenStack->pop(); |
||
| 411 | $this->currentToken = $this->tokenStack->top(); |
||
| 412 | |||
| 413 | $tk1 = (new SqlToken()) |
||
| 414 | ->type(SqlToken::TYPE_STATEMENT); |
||
| 415 | |||
| 416 | $this->currentToken[] = $tk1; |
||
| 417 | $this->tokenStack->push($this->currentToken[-1]); |
||
| 418 | $this->currentToken = $this->tokenStack->top(); |
||
| 419 | |||
| 420 | break; |
||
| 421 | default: |
||
| 422 | $tk = (new SqlToken()) |
||
| 423 | ->type(SqlToken::TYPE_OPERATOR) |
||
| 424 | ->content(is_string($content) ? $content : $this->substring($length)) |
||
| 425 | ->startOffset($this->offset) |
||
| 426 | ->endOffset($this->offset + $length); |
||
| 427 | |||
| 428 | $this->currentToken[] = $tk; |
||
| 429 | |||
| 430 | break; |
||
| 431 | } |
||
| 432 | |||
| 433 | return true; |
||
| 434 | } |
||
| 435 | |||
| 436 | /** |
||
| 437 | * Determines a type of text in the buffer, tokenizes it and adds it to the token children. |
||
| 438 | */ |
||
| 439 | private function addTokenFromBuffer(): void |
||
| 440 | { |
||
| 441 | if ($this->buffer === '') { |
||
| 442 | return; |
||
| 443 | } |
||
| 444 | |||
| 445 | $isKeyword = $this->isKeyword($this->buffer, $content); |
||
| 446 | |||
| 447 | $tk = (new SqlToken()) |
||
| 448 | ->type($isKeyword ? SqlToken::TYPE_KEYWORD : SqlToken::TYPE_TOKEN) |
||
| 449 | ->content(is_string($content) ? $content : $this->buffer) |
||
| 450 | ->startOffset($this->offset - mb_strlen($this->buffer, 'UTF-8')) |
||
| 451 | ->endOffset($this->offset); |
||
| 452 | |||
| 453 | $this->currentToken[] = $tk; |
||
| 454 | |||
| 455 | $this->buffer = ''; |
||
| 456 | } |
||
| 457 | |||
| 458 | /** |
||
| 459 | * Adds the specified length to the current offset. |
||
| 460 | * |
||
| 461 | * @param int $length |
||
| 462 | * |
||
| 463 | * @throws InvalidArgumentException |
||
| 464 | */ |
||
| 465 | private function advance(int $length): void |
||
| 466 | { |
||
| 467 | if ($length <= 0) { |
||
| 468 | throw new InvalidArgumentException('Length must be greater than 0.'); |
||
| 469 | } |
||
| 470 | |||
| 471 | $this->offset += $length; |
||
| 472 | $this->substrings = []; |
||
| 473 | } |
||
| 474 | |||
| 475 | /** |
||
| 476 | * Returns whether the SQL code is completely traversed. |
||
| 477 | * |
||
| 478 | * @return bool |
||
| 479 | */ |
||
| 480 | private function isEof(): bool |
||
| 483 | } |
||
| 484 | } |
||
| 485 |
Loading history...
This check looks for calls to methods that do not seem to exist on a given type. It looks for the method on the type itself as well as in inherited classes or implemented interfaces.
This is most likely a typographical error or the method has been renamed.