TASK: Refactor document identifier builder

The builder is now configurable via the objects.yaml. With this, the change can
got to a minor version and the new identifier and behavior can be tested.

Author: daniellienert

Classes/Domain/Model/TargetContextPath.php

@@ -13,6 +13,7 @@
  * source code.
  */
 
+use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\DocumentIdentifier\DocumentIdentifierGeneratorInterface;
 use Neos\ContentRepository\Domain\Model\NodeInterface;
 use Neos\Flow\Annotations as Flow;
 use Neos\Flow\Log\Utility\LogEnvironment;
@@ -28,6 +29,12 @@ abstract class AbstractIndexerDriver extends AbstractDriver
      */
     protected $nodeTypeMappingBuilder;
 
+    /**
+     * @Flow\Inject
+     * @var DocumentIdentifierGeneratorInterface
+     */
+    protected $documentIdentifierGenerator;
+
     /**
      * Whether the node is configured as fulltext root.
      *

Classes/Driver/AbstractIndexerDriver.php

@@ -13,12 +13,12 @@
  * source code.
  */
 
+use Neos\Flow\Annotations as Flow;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Indexer\NodeIndexer;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Driver\AbstractIndexerDriver;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Driver\IndexerDriverInterface;
 use Flowpack\ElasticSearch\Domain\Model\Document as ElasticSearchDocument;
 use Neos\ContentRepository\Domain\Model\NodeInterface;
-use Neos\Flow\Annotations as Flow;
 use Neos\Flow\Log\Utility\LogEnvironment;
 
 /**
@@ -81,7 +81,10 @@ public function document(string $indexName, NodeInterface $node, ElasticSearchDo
 
     /**
      * {@inheritdoc}
-     * @throws \Neos\Flow\Persistence\Exception\IllegalObjectTypeException
+     * @param NodeInterface $node
+     * @param array $fulltextIndexOfNode
+     * @param string|null $targetWorkspaceName
+     * @return array
      */
     public function fulltext(NodeInterface $node, array $fulltextIndexOfNode, string $targetWorkspaceName = null): array
     {
@@ -90,7 +93,7 @@ public function fulltext(NodeInterface $node, array $fulltextIndexOfNode, string
             return [];
         }
 
-        $closestFulltextNodeDocumentIdentifier = NodeIndexer::calculateDocumentIdentifier($closestFulltextNode);
+        $closestFulltextNodeDocumentIdentifier = $this->documentIdentifierGenerator->generate($closestFulltextNode);
 
         if ($closestFulltextNode->isRemoved()) {
             // fulltext root is removed, abort silently...

Classes/Driver/Version6/IndexerDriver.php

@@ -25,6 +25,7 @@
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\ErrorHandling\ErrorStorageInterface;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Exception;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\DimensionsService;
+use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\DocumentIdentifier\DocumentIdentifierGeneratorInterface;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\IndexNameService;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\NodeTypeIndexingConfiguration;
 use Flowpack\ElasticSearch\Domain\Model\Document as ElasticSearchDocument;
@@ -56,13 +57,6 @@ class NodeIndexer extends AbstractNodeIndexer implements BulkNodeIndexerInterfac
      */
     protected $nodeTypeMappingBuilder;
 
-    /**
-     * Optional postfix for the index, e.g. to have different indexes by timestamp.
-     *
-     * @var string
-     */
-    protected $indexNamePostfix = '';
-
     /**
      * @Flow\Inject
      * @var ErrorHandlingService
@@ -129,18 +123,6 @@ class NodeIndexer extends AbstractNodeIndexer implements BulkNodeIndexerInterfac
      */
     protected $indexConfiguration;
 
-    /**
-     * The current Elasticsearch bulk request, in the format required by http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/docs-bulk.html
-     *
-     * @var array
-     */
-    protected $currentBulkRequest = [];
-
-    /**
-     * @var boolean
-     */
-    protected $bulkProcessing = false;
-
     /**
      * @Flow\Inject
      * @var DimensionsService
@@ -153,12 +135,30 @@ class NodeIndexer extends AbstractNodeIndexer implements BulkNodeIndexerInterfac
      */
     protected $nodeTypeIndexingConfiguration;
 
+    /**
+     * @Flow\Inject
+     * @var DocumentIdentifierGeneratorInterface
+     */
+    protected $documentIdentifierGenerator;
+
     /**
      * @Flow\Inject
      * @var ErrorStorageInterface
      */
     protected $errorStorage;
 
+    /**
+     * The current Elasticsearch bulk request, in the format required by http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/docs-bulk.html
+     */
+    protected array $currentBulkRequest = [];
+
+    /**
+     * Optional postfix for the index, e.g. to have different indexes by timestamp.
+     */
+    protected string $indexNamePostfix = '';
+
+    protected bool $bulkProcessing = false;
+
     public function setDimensions(array $dimensionsValues): void
     {
         $this->searchClient->setDimensions($dimensionsValues);
@@ -243,7 +243,7 @@ public function indexNode(NodeInterface $node, $targetWorkspaceName = null): voi
                 }
             }
 
-            $documentIdentifier = NodeIndexer::calculateDocumentIdentifier($node, $targetWorkspaceName);
+            $documentIdentifier = $this->documentIdentifierGenerator->generate($node, $targetWorkspaceName);
             $nodeType = $node->getNodeType();
 
             $mappingType = $this->getIndex()->findType($nodeType->getName());
@@ -274,13 +274,6 @@ public function indexNode(NodeInterface $node, $targetWorkspaceName = null): voi
         $handleNode = function (NodeInterface $node, Context $context) use ($targetWorkspaceName, $indexer) {
             $nodeFromContext = $context->getNodeByIdentifier($node->getIdentifier());
             if ($nodeFromContext instanceof NodeInterface) {
-                if ($node->getPath() !== $nodeFromContext->getPath()) {
-                    // If the node from context does have a different path, purge the context cache and re-fetch
-
-                    // TODO: find a better way to handle this
-                    $context->getFirstLevelNodeCache()->flush();
-                    $nodeFromContext = $context->getNodeByIdentifier($node->getIdentifier());
-                }
                 $this->searchClient->withDimensions(static function () use ($indexer, $nodeFromContext, $targetWorkspaceName) {
                     $indexer($nodeFromContext, $targetWorkspaceName);
                 }, $nodeFromContext->getContext()->getTargetDimensions());
@@ -342,31 +335,14 @@ protected function toBulkRequest(NodeInterface $node, array $requests = null): v
         $this->flushIfNeeded();
     }
 
-    /**
-     * Returns a stable identifier for the Elasticsearch document representing the node
-     *
-     * @param NodeInterface $node
-     * @param string|null $targetWorkspaceName
-     * @return string
-     * @throws IllegalObjectTypeException
-     */
-    public static function calculateDocumentIdentifier(NodeInterface $node, $targetWorkspaceName = null): string
-    {
-        $workspaceName = $targetWorkspaceName ?: $node->getWorkspace()->getName();
-        $nodeIdentifier = $node->getIdentifier();
-
-        return sha1($nodeIdentifier . $workspaceName);
-    }
-
     /**
      * Schedule node removal into the current bulk request.
      *
      * @param NodeInterface $node
-     * @param string $targetWorkspaceName
+     * @param string|null $targetWorkspaceName
      * @return void
      * @throws Exception
      * @throws FilesException
-     * @throws IllegalObjectTypeException
      * @throws \Flowpack\ElasticSearch\Exception
      */
     public function removeNode(NodeInterface $node, string $targetWorkspaceName = null): void
@@ -384,7 +360,7 @@ public function removeNode(NodeInterface $node, string $targetWorkspaceName = nu
             }
         }
 
-        $documentIdentifier = NodeIndexer::calculateDocumentIdentifier($node, $targetWorkspaceName);
+        $documentIdentifier = $this->documentIdentifierGenerator->generate($node, $targetWorkspaceName);
 
         $this->toBulkRequest($node, $this->documentDriver->delete($node, $documentIdentifier));
         $this->toBulkRequest($node, $this->indexerDriver->fulltext($node, [], $targetWorkspaceName));
@@ -394,8 +370,8 @@ public function removeNode(NodeInterface $node, string $targetWorkspaceName = nu
 
     /**
      * @throws Exception
+     * @throws Exception\ConfigurationException
      * @throws \Flowpack\ElasticSearch\Exception
-     * @throws FilesException
      */
     protected function flushIfNeeded(): void
     {
@@ -554,6 +530,7 @@ public function updateIndexAlias(): void
      * Update the main alias to allow to query all indices at once
      * @throws Exception
      * @throws Exception\ConfigurationException
+     * @throws ApiException
      */
     public function updateMainAlias(): void
     {

Classes/Indexer/NodeIndexer.php

@@ -0,0 +1,27 @@
+<?php
+declare(strict_types=1);
+
+namespace Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\DocumentIdentifier;
+
+/*
+ * This file is part of the Flowpack.ElasticSearch.ContentRepositoryAdaptor package.
+ *
+ * (c) Contributors of the Neos Project - www.neos.io
+ *
+ * This package is Open Source Software. For the full copyright and license
+ * information, please view the LICENSE file which was distributed with this
+ * source code.
+ */
+
+use Neos\ContentRepository\Domain\Model\NodeInterface;
+
+interface DocumentIdentifierGeneratorInterface
+{
+    /**
+     * Generates a stable identifier out of the given node
+     *
+     * @param NodeInterface $node
+     * @return string
+     */
+    public function generate(NodeInterface $node, ?string $targetWorkspaceName = null): string;
+}

Classes/Service/DocumentIdentifier/DocumentIdentifierGeneratorInterface.php

@@ -0,0 +1,28 @@
+<?php
+declare(strict_types=1);
+
+namespace Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\DocumentIdentifier;
+
+/*
+ * This file is part of the Flowpack.ElasticSearch.ContentRepositoryAdaptor package.
+ *
+ * (c) Contributors of the Neos Project - www.neos.io
+ *
+ * This package is Open Source Software. For the full copyright and license
+ * information, please view the LICENSE file which was distributed with this
+ * source code.
+ */
+
+use Neos\ContentRepository\Domain\Model\NodeInterface;
+
+class NodeIdentifierBasedDocumentIdentifierGenerator implements DocumentIdentifierGeneratorInterface
+{
+
+    public function generate(NodeInterface $node, ?string $targetWorkspaceName = null): string
+    {
+        $workspaceName = $targetWorkspaceName ?: $node->getWorkspace()->getName();
+        $nodeIdentifier = $node->getIdentifier();
+
+        return sha1($nodeIdentifier . $workspaceName);
+    }
+}

Classes/Service/DocumentIdentifier/NodeIdentifierBasedDocumentIdentifierGenerator.php

@@ -0,0 +1,31 @@
+<?php
+declare(strict_types=1);
+
+namespace Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\DocumentIdentifier;
+
+/*
+ * This file is part of the Flowpack.ElasticSearch.ContentRepositoryAdaptor package.
+ *
+ * (c) Contributors of the Neos Project - www.neos.io
+ *
+ * This package is Open Source Software. For the full copyright and license
+ * information, please view the LICENSE file which was distributed with this
+ * source code.
+ */
+
+use Neos\ContentRepository\Domain\Model\NodeInterface;
+
+class NodePathBasedDocumentIdentifierGenerator implements DocumentIdentifierGeneratorInterface
+{
+
+    public function generate(NodeInterface $node, ?string $targetWorkspaceName = null): string
+    {
+        $contextPath = $node->getContextPath();
+
+        if ($targetWorkspaceName !== null) {
+            $contextPath = str_replace($node->getContext()->getWorkspace()->getName(), $targetWorkspaceName, $contextPath);
+        }
+
+        return sha1($contextPath);
+    }
+}

Classes/Service/DocumentIdentifier/NodePathBasedDocumentIdentifierGenerator.php

@@ -7,6 +7,9 @@ Neos\ContentRepository\Search\Indexer\NodeIndexerInterface:
 Neos\ContentRepository\Search\AssetExtraction\AssetExtractorInterface:
   className: 'Flowpack\ElasticSearch\ContentRepositoryAdaptor\AssetExtraction\IngestAttachmentAssetExtractor'
 
+'Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\DocumentIdentifier\DocumentIdentifierGeneratorInterface':
+  className: 'Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\DocumentIdentifier\NodePathBasedDocumentIdentifierGenerator'
+
 Flowpack\ElasticSearch\ContentRepositoryAdaptor\Driver\QueryInterface:
   scope: prototype
   factoryObjectName: 'Flowpack\ElasticSearch\ContentRepositoryAdaptor\Factory\QueryFactory'