TYPO3-Solr /
ext-solr
Complex classes like Indexer 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. You can also have a look at the cohesion graph to spot any un-connected, or weakly-connected components.
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 Indexer, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 50 | class Indexer extends AbstractIndexer |
||
| 51 | { |
||
| 52 | |||
| 53 | # TODO change to singular $document instead of plural $documents |
||
| 54 | |||
| 55 | /** |
||
| 56 | * A Solr service instance to interact with the Solr server |
||
| 57 | * |
||
| 58 | * @var SolrService |
||
| 59 | */ |
||
| 60 | protected $solr; |
||
| 61 | |||
| 62 | /** |
||
| 63 | * @var ConnectionManager |
||
| 64 | */ |
||
| 65 | protected $connectionManager; |
||
| 66 | |||
| 67 | /** |
||
| 68 | * Holds options for a specific indexer |
||
| 69 | * |
||
| 70 | * @var array |
||
| 71 | */ |
||
| 72 | protected $options = []; |
||
| 73 | |||
| 74 | /** |
||
| 75 | * To log or not to log... #Shakespeare |
||
| 76 | * |
||
| 77 | * @var bool |
||
| 78 | */ |
||
| 79 | protected $loggingEnabled = false; |
||
| 80 | |||
| 81 | /** |
||
| 82 | * @var IdBuilder |
||
| 83 | */ |
||
| 84 | protected $variantIdBuilder; |
||
| 85 | |||
| 86 | /** |
||
| 87 | * Cache of the sys_language_overlay information |
||
| 88 | * |
||
| 89 | * @var array |
||
| 90 | */ |
||
| 91 | protected static $sysLanguageOverlay = []; |
||
| 92 | |||
| 93 | /** |
||
| 94 | * @var \ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager |
||
| 95 | */ |
||
| 96 | protected $logger = null; |
||
| 97 | |||
| 98 | /** |
||
| 99 | * Constructor |
||
| 100 | * |
||
| 101 | * @param array $options array of indexer options |
||
| 102 | * @param IdBuilder $idBuilder |
||
| 103 | */ |
||
| 104 | 16 | public function __construct(array $options = [], IdBuilder $idBuilder = null) |
|
| 111 | |||
| 112 | /** |
||
| 113 | * Indexes an item from the indexing queue. |
||
| 114 | * |
||
| 115 | * @param Item $item An index queue item |
||
| 116 | * @return bool returns true when indexed, false when not |
||
| 117 | */ |
||
| 118 | 11 | public function index(Item $item) |
|
| 145 | |||
| 146 | /** |
||
| 147 | * Creates a single Solr Document for an item in a specific language. |
||
| 148 | * |
||
| 149 | * @param Item $item An index queue item to index. |
||
| 150 | * @param int $language The language to use. |
||
| 151 | * @return bool TRUE if item was indexed successfully, FALSE on failure |
||
| 152 | */ |
||
| 153 | 11 | protected function indexItem(Item $item, $language = 0) |
|
| 154 | { |
||
| 155 | 11 | $itemIndexed = false; |
|
| 156 | 11 | $documents = []; |
|
| 157 | |||
| 158 | 11 | $itemDocument = $this->itemToDocument($item, $language); |
|
| 159 | 11 | if (is_null($itemDocument)) { |
|
| 160 | /* |
||
| 161 | * If there is no itemDocument, this means there was no translation |
||
| 162 | * for this record. This should not stop the current item to count as |
||
| 163 | * being valid because not-indexing not-translated items is perfectly |
||
| 164 | * fine. |
||
| 165 | */ |
||
| 166 | return true; |
||
| 167 | } |
||
| 168 | |||
| 169 | 11 | $documents[] = $itemDocument; |
|
| 170 | 11 | $documents = array_merge($documents, $this->getAdditionalDocuments($item, $language, $itemDocument)); |
|
| 171 | 11 | $documents = $this->processDocuments($item, $documents); |
|
| 172 | 11 | $documents = $this->preAddModifyDocuments($item, $language, $documents); |
|
| 173 | |||
| 174 | 11 | $response = $this->solr->addDocuments($documents); |
|
| 175 | 11 | if ($response->getHttpStatus() == 200) { |
|
| 176 | 11 | $itemIndexed = true; |
|
| 177 | 11 | } |
|
| 178 | |||
| 179 | 11 | $this->log($item, $documents, $response); |
|
| 180 | |||
| 181 | 11 | return $itemIndexed; |
|
| 182 | } |
||
| 183 | |||
| 184 | /** |
||
| 185 | * Gets the full item record. |
||
| 186 | * |
||
| 187 | * This general record indexer simply gets the record from the item. Other |
||
| 188 | * more specialized indexers may provide more data for their specific item |
||
| 189 | * types. |
||
| 190 | * |
||
| 191 | * @param Item $item The item to be indexed |
||
| 192 | * @param int $language Language Id (sys_language.uid) |
||
| 193 | * @return array|NULL The full record with fields of data to be used for indexing or NULL to prevent an item from being indexed |
||
| 194 | */ |
||
| 195 | 11 | protected function getFullItemRecord(Item $item, $language = 0) |
|
| 196 | { |
||
| 197 | 11 | $rootPageUid = $item->getRootPageUid(); |
|
| 198 | 11 | $overlayIdentifier = $rootPageUid . '|' . $language; |
|
| 199 | 11 | if (!isset(self::$sysLanguageOverlay[$overlayIdentifier])) { |
|
| 200 | 11 | Util::initializeTsfe($rootPageUid, $language); |
|
| 201 | 11 | self::$sysLanguageOverlay[$overlayIdentifier] = $GLOBALS['TSFE']->sys_language_contentOL; |
|
| 202 | 11 | } |
|
| 203 | |||
| 204 | 11 | $itemRecord = $item->getRecord(); |
|
| 205 | |||
| 206 | 11 | if ($language > 0) { |
|
| 207 | 1 | $page = GeneralUtility::makeInstance(PageRepository::class); |
|
| 208 | 1 | $page->init(false); |
|
| 209 | |||
| 210 | 1 | $itemRecord = $page->getRecordOverlay( |
|
| 211 | 1 | $item->getType(), |
|
| 212 | 1 | $itemRecord, |
|
| 213 | 1 | $language, |
|
| 214 | 1 | self::$sysLanguageOverlay[$overlayIdentifier] |
|
| 215 | 1 | ); |
|
| 216 | 1 | } |
|
| 217 | |||
| 218 | 11 | if (!$itemRecord) { |
|
| 219 | $itemRecord = null; |
||
| 220 | } |
||
| 221 | |||
| 222 | /* |
||
| 223 | * Skip disabled records. This happens if the default language record |
||
| 224 | * is hidden but a certain translation isn't. Then the default language |
||
| 225 | * document appears here but must not be indexed. |
||
| 226 | */ |
||
| 227 | 11 | if (!empty($GLOBALS['TCA'][$item->getType()]['ctrl']['enablecolumns']['disabled']) |
|
| 228 | 11 | && $itemRecord[$GLOBALS['TCA'][$item->getType()]['ctrl']['enablecolumns']['disabled']] |
|
| 229 | 11 | ) { |
|
| 230 | $itemRecord = null; |
||
| 231 | } |
||
| 232 | |||
| 233 | /* |
||
| 234 | * Skip translation mismatching records. Sometimes the requested language |
||
| 235 | * doesn't fit the returned language. This might happen with content fallback |
||
| 236 | * and is perfectly fine in general. |
||
| 237 | * But if the requested language doesn't match the returned language and |
||
| 238 | * the given record has no translation parent, the indexqueue_item most |
||
| 239 | * probably pointed to a non-translated language record that is dedicated |
||
| 240 | * to a very specific language. Now we have to avoid indexing this record |
||
| 241 | * into all language cores. |
||
| 242 | */ |
||
| 243 | 11 | $translationOriginalPointerField = 'l10n_parent'; |
|
| 244 | 11 | if (!empty($GLOBALS['TCA'][$item->getType()]['ctrl']['transOrigPointerField'])) { |
|
| 245 | 10 | $translationOriginalPointerField = $GLOBALS['TCA'][$item->getType()]['ctrl']['transOrigPointerField']; |
|
| 246 | 10 | } |
|
| 247 | |||
| 248 | 11 | $languageField = $GLOBALS['TCA'][$item->getType()]['ctrl']['languageField']; |
|
| 249 | 11 | if ($itemRecord[$translationOriginalPointerField] == 0 |
|
| 250 | 11 | && self::$sysLanguageOverlay[$overlayIdentifier] != 1 |
|
| 251 | 11 | && !empty($languageField) |
|
| 252 | 11 | && $itemRecord[$languageField] != $language |
|
| 253 | 11 | && $itemRecord[$languageField] != '-1' |
|
| 254 | 11 | ) { |
|
| 255 | $itemRecord = null; |
||
| 256 | } |
||
| 257 | |||
| 258 | 11 | if (!is_null($itemRecord)) { |
|
| 259 | 11 | $itemRecord['__solr_index_language'] = $language; |
|
| 260 | 11 | } |
|
| 261 | |||
| 262 | 11 | return $itemRecord; |
|
| 263 | } |
||
| 264 | |||
| 265 | /** |
||
| 266 | * Gets the configuration how to process an item's fields for indexing. |
||
| 267 | * |
||
| 268 | * @param Item $item An index queue item |
||
| 269 | * @param int $language Language ID |
||
| 270 | * @throws \RuntimeException |
||
| 271 | * @return array Configuration array from TypoScript |
||
| 272 | */ |
||
| 273 | 11 | protected function getItemTypeConfiguration(Item $item, $language = 0) |
|
| 286 | |||
| 287 | /** |
||
| 288 | * Converts an item array (record) to a Solr document by mapping the |
||
| 289 | * record's fields onto Solr document fields as configured in TypoScript. |
||
| 290 | * |
||
| 291 | * @param Item $item An index queue item |
||
| 292 | * @param int $language Language Id |
||
| 293 | * @return Apache_Solr_Document The Solr document converted from the record |
||
| 294 | */ |
||
| 295 | 11 | protected function itemToDocument(Item $item, $language = 0) |
|
| 308 | |||
| 309 | /** |
||
| 310 | * Creates a Solr document with the basic / core fields set already. |
||
| 311 | * |
||
| 312 | * @param Item $item The item to index |
||
| 313 | * @param array $itemRecord The record to use to build the base document |
||
| 314 | * @return Apache_Solr_Document A basic Solr document |
||
| 315 | */ |
||
| 316 | 11 | protected function getBaseDocument(Item $item, array $itemRecord) |
|
| 366 | |||
| 367 | /** |
||
| 368 | * Generates an Access Rootline for an item. |
||
| 369 | * |
||
| 370 | * @param Item $item Index Queue item to index. |
||
| 371 | * @return string The Access Rootline for the item |
||
| 372 | */ |
||
| 373 | 11 | protected function getAccessRootline(Item $item) |
|
| 391 | |||
| 392 | /** |
||
| 393 | * Sends the documents to the field processing service which takes care of |
||
| 394 | * manipulating fields as defined in the field's configuration. |
||
| 395 | * |
||
| 396 | * @param Item $item An index queue item |
||
| 397 | * @param array $documents An array of Apache_Solr_Document objects to manipulate. |
||
| 398 | * @return array Array of manipulated Apache_Solr_Document objects. |
||
| 399 | */ |
||
| 400 | 11 | protected function processDocuments(Item $item, array $documents) |
|
| 417 | |||
| 418 | /** |
||
| 419 | * Allows third party extensions to provide additional documents which |
||
| 420 | * should be indexed for the current item. |
||
| 421 | * |
||
| 422 | * @param Item $item The item currently being indexed. |
||
| 423 | * @param int $language The language uid currently being indexed. |
||
| 424 | * @param Apache_Solr_Document $itemDocument The document representing the item for the given language. |
||
| 425 | * @return array An array of additional Apache_Solr_Document objects to index. |
||
| 426 | */ |
||
| 427 | 15 | protected function getAdditionalDocuments( |
|
| 428 | Item $item, |
||
| 429 | $language, |
||
| 430 | Apache_Solr_Document $itemDocument |
||
| 431 | ) { |
||
| 432 | 15 | $documents = []; |
|
| 433 | |||
| 434 | 15 | if (is_array($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['solr']['IndexQueueIndexer']['indexItemAddDocuments'])) { |
|
| 435 | 4 | foreach ($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['solr']['IndexQueueIndexer']['indexItemAddDocuments'] as $classReference) { |
|
| 436 | 4 | if (!class_exists($classReference)) { |
|
| 437 | 2 | throw new \InvalidArgumentException('Class does not exits' . $classReference, 1490363487); |
|
| 438 | } |
||
| 439 | 2 | $additionalIndexer = GeneralUtility::makeInstance($classReference); |
|
| 440 | 2 | if ($additionalIndexer instanceof AdditionalIndexQueueItemIndexer) { |
|
| 441 | 1 | $additionalDocuments = $additionalIndexer->getAdditionalItemDocuments($item, |
|
| 442 | 1 | $language, $itemDocument); |
|
| 443 | |||
| 444 | 1 | if (is_array($additionalDocuments)) { |
|
| 445 | 1 | $documents = array_merge($documents, |
|
| 446 | 1 | $additionalDocuments); |
|
| 447 | 1 | } |
|
| 448 | 1 | } else { |
|
| 449 | 1 | throw new \UnexpectedValueException( |
|
| 450 | 1 | get_class($additionalIndexer) . ' must implement interface ' . AdditionalIndexQueueItemIndexer::class, |
|
| 451 | 1326284551 |
||
| 452 | 1 | ); |
|
| 453 | } |
||
| 454 | 1 | } |
|
| 455 | 1 | } |
|
| 456 | 12 | return $documents; |
|
| 457 | } |
||
| 458 | |||
| 459 | /** |
||
| 460 | * Provides a hook to manipulate documents right before they get added to |
||
| 461 | * the Solr index. |
||
| 462 | * |
||
| 463 | * @param Item $item The item currently being indexed. |
||
| 464 | * @param int $language The language uid of the documents |
||
| 465 | * @param array $documents An array of documents to be indexed |
||
| 466 | * @return array An array of modified documents |
||
| 467 | */ |
||
| 468 | 11 | protected function preAddModifyDocuments( |
|
| 494 | |||
| 495 | // Initialization |
||
| 496 | |||
| 497 | /** |
||
| 498 | * Gets the Solr connections applicable for an item. |
||
| 499 | * |
||
| 500 | * The connections include the default connection and connections to be used |
||
| 501 | * for translations of an item. |
||
| 502 | * |
||
| 503 | * @param Item $item An index queue item |
||
| 504 | * @return array An array of ApacheSolrForTypo3\Solr\SolrService connections, the array's keys are the sys_language_uid of the language of the connection |
||
| 505 | */ |
||
| 506 | 12 | protected function getSolrConnectionsByItem(Item $item) |
|
| 507 | { |
||
| 508 | 12 | $solrConnections = []; |
|
| 509 | |||
| 510 | 12 | $pageId = $item->getRootPageUid(); |
|
| 511 | 12 | if ($item->getType() == 'pages') { |
|
| 512 | 2 | $pageId = $item->getRecordUid(); |
|
| 513 | 2 | } |
|
| 514 | |||
| 515 | // Solr configurations possible for this item |
||
| 516 | 12 | $site = $item->getSite(); |
|
| 517 | 12 | $solrConfigurationsBySite = $this->connectionManager->getConfigurationsBySite($site); |
|
| 518 | |||
| 519 | 12 | $siteLanguages = []; |
|
| 520 | 12 | foreach ($solrConfigurationsBySite as $solrConfiguration) { |
|
| 521 | 12 | $siteLanguages[] = $solrConfiguration['language']; |
|
| 522 | 12 | } |
|
| 523 | |||
| 524 | 12 | $translationOverlays = $this->getTranslationOverlaysForPage($pageId, $site->getSysLanguageMode()); |
|
| 525 | 12 | foreach ($translationOverlays as $key => $translationOverlay) { |
|
| 526 | 1 | if (!in_array($translationOverlay['sys_language_uid'], |
|
| 527 | 1 | $siteLanguages) |
|
| 528 | 1 | ) { |
|
| 529 | unset($translationOverlays[$key]); |
||
| 530 | } |
||
| 531 | 12 | } |
|
| 532 | |||
| 533 | 12 | $defaultConnection = $this->connectionManager->getConnectionByPageId($pageId); |
|
| 534 | 12 | $translationConnections = $this->getConnectionsForIndexableLanguages($translationOverlays); |
|
| 535 | |||
| 536 | 12 | $solrConnections[0] = $defaultConnection; |
|
| 537 | 12 | foreach ($translationConnections as $systemLanguageUid => $solrConnection) { |
|
| 538 | 1 | $solrConnections[$systemLanguageUid] = $solrConnection; |
|
| 539 | 12 | } |
|
| 540 | |||
| 541 | 12 | return $solrConnections; |
|
| 542 | } |
||
| 543 | |||
| 544 | /** |
||
| 545 | * Finds the alternative page language overlay records for a page based on |
||
| 546 | * the sys_language_mode. |
||
| 547 | * |
||
| 548 | * Possible Language Modes: |
||
| 549 | * 1) content_fallback --> all languages |
||
| 550 | * 2) strict --> available languages with page overlay |
||
| 551 | * 3) ignore --> available languages with page overlay |
||
| 552 | * 4) unknown mode or blank --> all languages |
||
| 553 | * |
||
| 554 | * @param int $pageId Page ID. |
||
| 555 | * @param string $languageMode |
||
| 556 | * @return array An array of translation overlays (or fake overlays) found for the given page. |
||
| 557 | */ |
||
| 558 | 12 | protected function getTranslationOverlaysForPage($pageId, $languageMode) |
|
| 559 | { |
||
| 560 | 12 | $translationOverlays = []; |
|
| 561 | 12 | $pageId = intval($pageId); |
|
| 562 | |||
| 563 | 12 | $languageModes = ['content_fallback', 'strict', 'ignore']; |
|
| 564 | 12 | $hasOverlayMode = in_array($languageMode, $languageModes, |
|
| 565 | 12 | true); |
|
| 566 | 12 | $isContentFallbackMode = ($languageMode === 'content_fallback'); |
|
| 567 | |||
| 568 | 12 | if ($hasOverlayMode && !$isContentFallbackMode) { |
|
| 569 | 1 | $translationOverlays = $GLOBALS['TYPO3_DB']->exec_SELECTgetRows( |
|
| 570 | 1 | 'pid, sys_language_uid', |
|
| 571 | 1 | 'pages_language_overlay', |
|
| 572 | 'pid = ' . $pageId |
||
| 573 | 1 | . BackendUtility::deleteClause('pages_language_overlay') |
|
| 574 | 1 | . BackendUtility::BEenableFields('pages_language_overlay') |
|
| 575 | 1 | ); |
|
| 576 | 1 | } else { |
|
| 577 | // ! If no sys_language_mode is configured, all languages will be indexed ! |
||
| 578 | 11 | $languages = $this->getSystemLanguages(); |
|
| 579 | |||
| 580 | 11 | foreach ($languages as $language) { |
|
| 581 | 11 | if ($language['uid'] <= 0) { |
|
| 582 | 11 | continue; |
|
| 583 | } |
||
| 584 | $translationOverlays[] = [ |
||
| 585 | 'pid' => $pageId, |
||
| 586 | 'sys_language_uid' => $language['uid'], |
||
| 587 | ]; |
||
| 588 | 11 | } |
|
| 589 | } |
||
| 590 | |||
| 591 | 12 | return $translationOverlays; |
|
| 592 | } |
||
| 593 | |||
| 594 | /** |
||
| 595 | * Returns an array of system languages. |
||
| 596 | * |
||
| 597 | * @return array |
||
| 598 | */ |
||
| 599 | 11 | protected function getSystemLanguages() |
|
| 603 | |||
| 604 | /** |
||
| 605 | * Checks for which languages connections have been configured and returns |
||
| 606 | * these connections. |
||
| 607 | * |
||
| 608 | * @param array $translationOverlays An array of translation overlays to check for configured connections. |
||
| 609 | * @return array An array of ApacheSolrForTypo3\Solr\SolrService connections. |
||
| 610 | */ |
||
| 611 | 12 | protected function getConnectionsForIndexableLanguages( |
|
| 612 | array $translationOverlays |
||
| 613 | ) { |
||
| 614 | 12 | $connections = []; |
|
| 615 | |||
| 616 | 12 | foreach ($translationOverlays as $translationOverlay) { |
|
| 617 | 1 | $pageId = $translationOverlay['pid']; |
|
| 618 | 1 | $languageId = $translationOverlay['sys_language_uid']; |
|
| 619 | |||
| 620 | try { |
||
| 621 | 1 | $connection = $this->connectionManager->getConnectionByPageId($pageId, |
|
| 622 | 1 | $languageId); |
|
| 623 | 1 | $connections[$languageId] = $connection; |
|
| 624 | 1 | } catch (NoSolrConnectionFoundException $e) { |
|
| 625 | // ignore the exception as we seek only those connections |
||
| 626 | // actually available |
||
| 627 | } |
||
| 628 | 12 | } |
|
| 629 | |||
| 630 | 12 | return $connections; |
|
| 631 | } |
||
| 632 | |||
| 633 | // Utility methods |
||
| 634 | |||
| 635 | // FIXME extract log() and setLogging() to ApacheSolrForTypo3\Solr\IndexQueue\AbstractIndexer |
||
| 636 | // FIXME extract an interface Tx_Solr_IndexQueue_ItemInterface |
||
| 637 | |||
| 638 | /** |
||
| 639 | * Enables logging dependent on the configuration of the item's site |
||
| 640 | * |
||
| 641 | * @param Item $item An item being indexed |
||
| 642 | * @return void |
||
| 643 | */ |
||
| 644 | 12 | protected function setLogging(Item $item) |
|
| 645 | { |
||
| 646 | 12 | $solrConfiguration = Util::getSolrConfigurationFromPageId($item->getRootPageUid()); |
|
| 647 | 12 | $this->loggingEnabled = $solrConfiguration->getLoggingIndexingQueueOperationsByConfigurationNameWithFallBack( |
|
| 648 | 12 | $item->getIndexingConfigurationName() |
|
| 649 | 12 | ); |
|
| 650 | 12 | } |
|
| 651 | |||
| 652 | /** |
||
| 653 | * Logs the item and what document was created from it |
||
| 654 | * |
||
| 655 | * @param Item $item The item that is being indexed. |
||
| 656 | * @param array $itemDocuments An array of Solr documents created from the item's data |
||
| 657 | * @param Apache_Solr_Response $response The Solr response for the particular index document |
||
| 658 | */ |
||
| 659 | 11 | protected function log( |
|
| 700 | } |
||
| 701 |
