Erikmitk /
kitodo-presentation
| Total Complexity | 58 |
| Total Lines | 606 |
| Duplicated Lines | 0 % |
| Changes | 1 | ||
| Bugs | 0 | Features | 0 |
Complex classes like Solr 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 Solr, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 45 | class Solr implements LoggerAwareInterface |
||
| 46 | { |
||
| 47 | use LoggerAwareTrait; |
||
| 48 | |||
| 49 | /** |
||
| 50 | * @access protected |
||
| 51 | * @var array This holds the Solr configuration |
||
| 52 | */ |
||
| 53 | protected array $config = []; |
||
| 54 | |||
| 55 | /** |
||
| 56 | * @access protected |
||
| 57 | * @var string|null This holds the core name |
||
| 58 | */ |
||
| 59 | protected ?string $core = null; |
||
| 60 | |||
| 61 | /** |
||
| 62 | * @access protected |
||
| 63 | * @var int This holds the PID for the configuration |
||
| 64 | */ |
||
| 65 | protected int $configPid = 0; |
||
| 66 | |||
| 67 | /** |
||
| 68 | * @access public |
||
| 69 | * @static |
||
| 70 | * @var string The extension key |
||
| 71 | */ |
||
| 72 | public static string $extKey = 'dlf'; |
||
| 73 | |||
| 74 | /** |
||
| 75 | * @access public |
||
| 76 | * @static |
||
| 77 | * @var array The fields for SOLR index |
||
| 78 | */ |
||
| 79 | public static array $fields = []; |
||
| 80 | |||
| 81 | /** |
||
| 82 | * @access protected |
||
| 83 | * @var int This holds the max results |
||
| 84 | */ |
||
| 85 | protected int $limit = 50000; |
||
| 86 | |||
| 87 | /** |
||
| 88 | * @access protected |
||
| 89 | * @var int This holds the number of hits for last search |
||
| 90 | */ |
||
| 91 | protected int $numberOfHits = 0; |
||
| 92 | |||
| 93 | /** |
||
| 94 | * @access protected |
||
| 95 | * @var array This holds the additional query parameters |
||
| 96 | */ |
||
| 97 | protected array $params = []; |
||
| 98 | |||
| 99 | /** |
||
| 100 | * @access protected |
||
| 101 | * @var bool Is the search instantiated successfully? |
||
| 102 | */ |
||
| 103 | protected bool $ready = false; |
||
| 104 | |||
| 105 | /** |
||
| 106 | * @access protected |
||
| 107 | * @var array(Solr) This holds the singleton search objects with their core as array key |
||
| 108 | */ |
||
| 109 | protected static array $registry = []; |
||
| 110 | |||
| 111 | /** |
||
| 112 | * @access protected |
||
| 113 | * @var Client This holds the Solr service object |
||
| 114 | */ |
||
| 115 | protected Client $service; |
||
| 116 | |||
| 117 | /** |
||
| 118 | * Add a new core to Apache Solr |
||
| 119 | * |
||
| 120 | * @access public |
||
| 121 | * |
||
| 122 | * @param string $core The name of the new core. If empty, the next available core name is used. |
||
| 123 | * |
||
| 124 | * @return string The name of the new core |
||
| 125 | */ |
||
| 126 | public static function createCore($core = ''): string |
||
| 163 | } |
||
| 164 | |||
| 165 | /** |
||
| 166 | * Escape special characters in a query string |
||
| 167 | * |
||
| 168 | * @access public |
||
| 169 | * |
||
| 170 | * @param string $query The query string |
||
| 171 | * |
||
| 172 | * @return string The escaped query string |
||
| 173 | */ |
||
| 174 | public static function escapeQuery(string $query): string |
||
| 175 | { |
||
| 176 | // Escape query by disallowing range and field operators |
||
| 177 | // Permit operators: wildcard, boolean, fuzzy, proximity, boost, grouping |
||
| 178 | // https://solr.apache.org/guide/solr/latest/query-guide/standard-query-parser.html |
||
| 179 | return preg_replace('/(\{|}|\[|]|:|\/|\\\)/', '\\\$1', $query); |
||
| 180 | } |
||
| 181 | |||
| 182 | /** |
||
| 183 | * Escape all special characters in a query string while retaining valid field queries |
||
| 184 | * |
||
| 185 | * @access public |
||
| 186 | * |
||
| 187 | * @param string $query The query string |
||
| 188 | * @param int $pid The PID for the field configuration |
||
| 189 | * |
||
| 190 | * @return string The escaped query string |
||
| 191 | */ |
||
| 192 | public static function escapeQueryKeepField(string $query, int $pid): string |
||
| 193 | { |
||
| 194 | // Is there a field query? |
||
| 195 | if (preg_match('/^[[:alnum:]]+_[tu][su]i:\(?.*\)?$/', $query)) { |
||
| 196 | $queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class) |
||
| 197 | ->getQueryBuilderForTable('tx_dlf_metadata'); |
||
| 198 | |||
| 199 | // Get all indexed fields. |
||
| 200 | $fields = []; |
||
| 201 | $result = $queryBuilder |
||
| 202 | ->select( |
||
| 203 | 'tx_dlf_metadata.index_name AS index_name', |
||
| 204 | 'tx_dlf_metadata.index_tokenized AS index_tokenized', |
||
| 205 | 'tx_dlf_metadata.index_stored AS index_stored' |
||
| 206 | ) |
||
| 207 | ->from('tx_dlf_metadata') |
||
| 208 | ->where( |
||
| 209 | $queryBuilder->expr()->eq('tx_dlf_metadata.index_indexed', 1), |
||
| 210 | $queryBuilder->expr()->eq('tx_dlf_metadata.pid', (int) $pid), |
||
| 211 | $queryBuilder->expr()->orX( |
||
| 212 | $queryBuilder->expr()->in('tx_dlf_metadata.sys_language_uid', [-1, 0]), |
||
| 213 | $queryBuilder->expr()->eq('tx_dlf_metadata.l18n_parent', 0) |
||
| 214 | ), |
||
| 215 | Helper::whereExpression('tx_dlf_metadata') |
||
| 216 | ) |
||
| 217 | ->execute(); |
||
| 218 | |||
| 219 | while ($resArray = $result->fetchAssociative()) { |
||
| 220 | $fields[] = $resArray['index_name'] . '_' . ($resArray['index_tokenized'] ? 't' : 'u') . ($resArray['index_stored'] ? 's' : 'u') . 'i'; |
||
| 221 | } |
||
| 222 | |||
| 223 | // Check if queried field is valid. |
||
| 224 | $splitQuery = explode(':', $query, 2); |
||
| 225 | if (in_array($splitQuery[0], $fields)) { |
||
| 226 | $query = $splitQuery[0] . ':(' . self::escapeQuery(trim($splitQuery[1], '()')) . ')'; |
||
| 227 | } else { |
||
| 228 | $query = self::escapeQuery($query); |
||
| 229 | } |
||
| 230 | } else { |
||
| 231 | $query = self::escapeQuery($query); |
||
| 232 | } |
||
| 233 | return $query; |
||
| 234 | } |
||
| 235 | |||
| 236 | /** |
||
| 237 | * Get fields for index. |
||
| 238 | * |
||
| 239 | * @access public |
||
| 240 | * |
||
| 241 | * @return array fields |
||
| 242 | */ |
||
| 243 | public static function getFields(): array |
||
| 244 | { |
||
| 245 | if (empty(self::$fields)) { |
||
| 246 | $conf = GeneralUtility::makeInstance(ExtensionConfiguration::class)->get(self::$extKey, 'solr'); |
||
| 247 | $solrFields = $conf['fields']; |
||
| 248 | self::$fields['id'] = $solrFields['id']; |
||
| 249 | self::$fields['uid'] = $solrFields['uid']; |
||
| 250 | self::$fields['pid'] = $solrFields['pid']; |
||
| 251 | self::$fields['page'] = $solrFields['page']; |
||
| 252 | self::$fields['partof'] = $solrFields['partof']; |
||
| 253 | self::$fields['root'] = $solrFields['root']; |
||
| 254 | self::$fields['sid'] = $solrFields['sid']; |
||
| 255 | self::$fields['toplevel'] = $solrFields['toplevel']; |
||
| 256 | self::$fields['type'] = $solrFields['type']; |
||
| 257 | self::$fields['title'] = $solrFields['title']; |
||
| 258 | self::$fields['volume'] = $solrFields['volume']; |
||
| 259 | self::$fields['date'] = $solrFields['date']; |
||
| 260 | self::$fields['thumbnail'] = $solrFields['thumbnail']; |
||
| 261 | self::$fields['default'] = $solrFields['default']; |
||
| 262 | self::$fields['timestamp'] = $solrFields['timestamp']; |
||
| 263 | self::$fields['autocomplete'] = $solrFields['autocomplete']; |
||
| 264 | self::$fields['fulltext'] = $solrFields['fulltext']; |
||
| 265 | self::$fields['record_id'] = $solrFields['recordId']; |
||
| 266 | self::$fields['purl'] = $solrFields['purl']; |
||
| 267 | self::$fields['urn'] = $solrFields['urn']; |
||
| 268 | self::$fields['location'] = $solrFields['location']; |
||
| 269 | self::$fields['collection'] = $solrFields['collection']; |
||
| 270 | self::$fields['license'] = $solrFields['license']; |
||
| 271 | self::$fields['terms'] = $solrFields['terms']; |
||
| 272 | self::$fields['restrictions'] = $solrFields['restrictions']; |
||
| 273 | self::$fields['geom'] = $solrFields['geom']; |
||
| 274 | } |
||
| 275 | |||
| 276 | return self::$fields; |
||
| 277 | } |
||
| 278 | |||
| 279 | /** |
||
| 280 | * This is a singleton class, thus instances must be created by this method |
||
| 281 | * |
||
| 282 | * @access public |
||
| 283 | * |
||
| 284 | * @param mixed $core Name or UID of the core to load or null to get core admin endpoint |
||
| 285 | * |
||
| 286 | * @return Solr Instance of this class |
||
| 287 | */ |
||
| 288 | public static function getInstance($core = null): Solr |
||
| 289 | { |
||
| 290 | // Get core name if UID is given. |
||
| 291 | if (MathUtility::canBeInterpretedAsInteger($core)) { |
||
| 292 | $core = Helper::getIndexNameFromUid($core, 'tx_dlf_solrcores'); |
||
|
Bug
introduced
by
|
|||
| 293 | } |
||
| 294 | // Check if core is set or null. |
||
| 295 | if ( |
||
| 296 | empty($core) |
||
| 297 | && $core !== null |
||
| 298 | ) { |
||
| 299 | Helper::log('Invalid core UID or name given for Apache Solr', LOG_SEVERITY_ERROR); |
||
| 300 | } |
||
| 301 | if (!empty($core)) { |
||
| 302 | // Check if there is an instance in the registry already. |
||
| 303 | if ( |
||
| 304 | array_key_exists($core, self::$registry) |
||
| 305 | && self::$registry[$core] instanceof self |
||
| 306 | ) { |
||
| 307 | // Return singleton instance if available. |
||
| 308 | return self::$registry[$core]; |
||
| 309 | } |
||
| 310 | } |
||
| 311 | // Create new instance... |
||
| 312 | $instance = new self($core); |
||
| 313 | // ...and save it to registry. |
||
| 314 | if (!empty($instance->core)) { |
||
| 315 | self::$registry[$instance->core] = $instance; |
||
| 316 | } |
||
| 317 | return $instance; |
||
| 318 | } |
||
| 319 | |||
| 320 | /** |
||
| 321 | * Get next unused Solr core number |
||
| 322 | * |
||
| 323 | * @access public |
||
| 324 | * |
||
| 325 | * @param int $number Number to start with |
||
| 326 | * |
||
| 327 | * @return int First unused core number found |
||
| 328 | */ |
||
| 329 | public static function getNextCoreNumber(int $number = 0): int |
||
| 330 | { |
||
| 331 | $number = max($number, 0); |
||
| 332 | // Check if core already exists. |
||
| 333 | $solr = self::getInstance('dlfCore' . $number); |
||
| 334 | if (!$solr->ready) { |
||
| 335 | return $number; |
||
| 336 | } else { |
||
| 337 | return self::getNextCoreNumber($number + 1); |
||
| 338 | } |
||
| 339 | } |
||
| 340 | |||
| 341 | /** |
||
| 342 | * Sets the connection information for Solr |
||
| 343 | * |
||
| 344 | * @access protected |
||
| 345 | * |
||
| 346 | * @return void |
||
| 347 | */ |
||
| 348 | protected function loadSolrConnectionInfo(): void |
||
| 349 | { |
||
| 350 | if (empty($this->config)) { |
||
| 351 | $config = []; |
||
| 352 | // Extract extension configuration. |
||
| 353 | $conf = GeneralUtility::makeInstance(ExtensionConfiguration::class)->get(self::$extKey, 'solr'); |
||
| 354 | // Derive Solr scheme |
||
| 355 | $config['scheme'] = empty($conf['https']) ? 'http' : 'https'; |
||
| 356 | // Derive Solr host name. |
||
| 357 | $config['host'] = ($conf['host'] ? $conf['host'] : '127.0.0.1'); |
||
| 358 | // Set username and password. |
||
| 359 | $config['username'] = $conf['user']; |
||
| 360 | $config['password'] = $conf['pass']; |
||
| 361 | // Set port if not set. |
||
| 362 | $config['port'] = MathUtility::forceIntegerInRange($conf['port'], 1, 65535, 8983); |
||
| 363 | // Trim path of slashes and (re-)add trailing slash if path not empty. |
||
| 364 | $config['path'] = trim($conf['path'], '/'); |
||
| 365 | if (!empty($config['path'])) { |
||
| 366 | $config['path'] .= '/'; |
||
| 367 | } |
||
| 368 | |||
| 369 | // Set connection timeout lower than PHP's max_execution_time. |
||
| 370 | $maxExecutionTime = (int) ini_get('max_execution_time') ? : 30; |
||
| 371 | $config['timeout'] = MathUtility::forceIntegerInRange($conf['timeout'], 1, $maxExecutionTime, 10); |
||
| 372 | $this->config = $config; |
||
| 373 | } |
||
| 374 | } |
||
| 375 | |||
| 376 | /** |
||
| 377 | * Processes a search request and returns the raw Apache Solr Documents. |
||
| 378 | * |
||
| 379 | * @access public |
||
| 380 | * |
||
| 381 | * @param array $parameters Additional search parameters |
||
| 382 | * |
||
| 383 | * @return array The Apache Solr Documents that were fetched |
||
| 384 | */ |
||
| 385 | public function searchRaw(array $parameters = []): array |
||
| 408 | } |
||
| 409 | |||
| 410 | /** |
||
| 411 | * This returns $this->core via __get() |
||
| 412 | * |
||
| 413 | * @access protected |
||
| 414 | * |
||
| 415 | * @return string|null The core name of the current query endpoint or null if core admin endpoint |
||
| 416 | */ |
||
| 417 | protected function magicGetCore(): ?string |
||
| 418 | { |
||
| 419 | return $this->core; |
||
| 420 | } |
||
| 421 | |||
| 422 | /** |
||
| 423 | * This returns $this->limit via __get() |
||
| 424 | * |
||
| 425 | * @access protected |
||
| 426 | * |
||
| 427 | * @return int The max number of results |
||
| 428 | */ |
||
| 429 | protected function magicGetLimit(): int |
||
| 430 | { |
||
| 431 | return $this->limit; |
||
| 432 | } |
||
| 433 | |||
| 434 | /** |
||
| 435 | * This returns $this->numberOfHits via __get() |
||
| 436 | * |
||
| 437 | * @access protected |
||
| 438 | * |
||
| 439 | * @return int Total number of hits for last search |
||
| 440 | */ |
||
| 441 | protected function magicGetNumberOfHits(): int |
||
| 442 | { |
||
| 443 | return $this->numberOfHits; |
||
| 444 | } |
||
| 445 | |||
| 446 | /** |
||
| 447 | * This returns $this->ready via __get() |
||
| 448 | * |
||
| 449 | * @access protected |
||
| 450 | * |
||
| 451 | * @return bool Is the search instantiated successfully? |
||
| 452 | */ |
||
| 453 | protected function magicGetReady(): bool |
||
| 454 | { |
||
| 455 | return $this->ready; |
||
| 456 | } |
||
| 457 | |||
| 458 | /** |
||
| 459 | * This returns $this->service via __get() |
||
| 460 | * |
||
| 461 | * @access protected |
||
| 462 | * |
||
| 463 | * @return Client Apache Solr service object |
||
| 464 | */ |
||
| 465 | protected function magicGetService(): Client |
||
| 468 | } |
||
| 469 | |||
| 470 | /** |
||
| 471 | * This sets $this->configPid via __set() |
||
| 472 | * |
||
| 473 | * @access protected |
||
| 474 | * |
||
| 475 | * @param int $value The new PID for the metadata definitions |
||
| 476 | * |
||
| 477 | * @return void |
||
| 478 | */ |
||
| 479 | protected function magicSetConfigPid(int $value): void |
||
| 480 | { |
||
| 481 | $this->configPid = max($value, 0); |
||
| 482 | } |
||
| 483 | |||
| 484 | /** |
||
| 485 | * This sets $this->limit via __set() |
||
| 486 | * |
||
| 487 | * @access protected |
||
| 488 | * |
||
| 489 | * @param int $value The max number of results |
||
| 490 | * |
||
| 491 | * @return void |
||
| 492 | */ |
||
| 493 | protected function magicSetLimit(int $value): void |
||
| 494 | { |
||
| 495 | $this->limit = max($value, 0); |
||
| 496 | } |
||
| 497 | |||
| 498 | /** |
||
| 499 | * This sets $this->params via __set() |
||
| 500 | * |
||
| 501 | * @access protected |
||
| 502 | * |
||
| 503 | * @param array $value The query parameters |
||
| 504 | * |
||
| 505 | * @return void |
||
| 506 | */ |
||
| 507 | protected function magicSetParams(array $value): void |
||
| 508 | { |
||
| 509 | $this->params = $value; |
||
| 510 | } |
||
| 511 | |||
| 512 | /** |
||
| 513 | * This magic method is called each time an invisible property is referenced from the object |
||
| 514 | * |
||
| 515 | * @access public |
||
| 516 | * |
||
| 517 | * @param string $var Name of variable to get |
||
| 518 | * |
||
| 519 | * @return mixed Value of $this->$var |
||
| 520 | */ |
||
| 521 | public function __get(string $var) |
||
| 522 | { |
||
| 523 | $method = 'magicGet' . ucfirst($var); |
||
| 524 | if ( |
||
| 525 | !property_exists($this, $var) |
||
| 526 | || !method_exists($this, $method) |
||
| 527 | ) { |
||
| 528 | $this->logger->warning('There is no getter function for property "' . $var . '"'); |
||
| 529 | return null; |
||
| 530 | } else { |
||
| 531 | return $this->$method(); |
||
| 532 | } |
||
| 533 | } |
||
| 534 | |||
| 535 | /** |
||
| 536 | * This magic method is called each time an invisible property is checked for isset() or empty() |
||
| 537 | * |
||
| 538 | * @access public |
||
| 539 | * |
||
| 540 | * @param string $var Name of variable to check |
||
| 541 | * |
||
| 542 | * @return bool true if variable is set and not empty, false otherwise |
||
| 543 | */ |
||
| 544 | public function __isset(string $var): bool |
||
| 547 | } |
||
| 548 | |||
| 549 | /** |
||
| 550 | * This magic method is called each time an invisible property is referenced from the object |
||
| 551 | * |
||
| 552 | * @access public |
||
| 553 | * |
||
| 554 | * @param string $var Name of variable to set |
||
| 555 | * @param mixed $value New value of variable |
||
| 556 | * |
||
| 557 | * @return void |
||
| 558 | */ |
||
| 559 | public function __set(string $var, $value): void |
||
| 569 | } |
||
| 570 | } |
||
| 571 | |||
| 572 | /** |
||
| 573 | * This is a singleton class, thus the constructor should be private/protected |
||
| 574 | * |
||
| 575 | * @access protected |
||
| 576 | * |
||
| 577 | * @param string|null $core The name of the core to use or null for core admin endpoint |
||
| 578 | * |
||
| 579 | * @return void |
||
| 580 | */ |
||
| 581 | protected function __construct(?string $core) |
||
| 651 | // Nothing to do here. |
||
| 652 | } |
||
| 653 | } |
||
| 654 | } |
||
| 655 |
Loading history...
