Merge pull request #234 from dfeyer/feature-disable-indexin-per-nodetype-or-namespace

FEATURE: Skip specific node type or namespace from indexing

Author: daniellienert

Classes/Driver/Version5/Mapping/NodeTypeMappingBuilder.php

@@ -14,6 +14,7 @@
  * source code.
  */
 
+use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\NodeTypeIndexingConfiguration;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Driver\AbstractNodeTypeMappingBuilder;
 use Flowpack\ElasticSearch\Domain\Model\Index;
 use Flowpack\ElasticSearch\Domain\Model\Mapping;
@@ -32,6 +33,12 @@
  */
 class NodeTypeMappingBuilder extends AbstractNodeTypeMappingBuilder
 {
+    /**
+     * @var NodeTypeIndexingConfiguration
+     * @Flow\Inject
+     */
+    protected $nodeTypeIndexingConfiguration;
+
     /**
      * Called by the Flow object framework after creating the object and resolving all dependencies.
      *
@@ -51,6 +58,7 @@ public function initializeObject($cause): void
      *
      * @param Index $index
      * @return MappingCollection<\Flowpack\ElasticSearch\Domain\Model\Mapping>
+     * @throws \Flowpack\ElasticSearch\ContentRepositoryAdaptor\Exception
      */
     public function buildMappingInformation(Index $index): MappingCollection
     {
@@ -64,6 +72,10 @@ public function buildMappingInformation(Index $index): MappingCollection
                 continue;
             }
 
+            if ($this->nodeTypeIndexingConfiguration->isIndexable($nodeType) === false) {
+                continue;
+            }
+
             $type = $index->findType($this->convertNodeTypeNameToMappingName($nodeTypeName));
             $mapping = new Mapping($type);
             $fullConfiguration = $nodeType->getFullConfiguration();

Classes/Indexer/NodeIndexer.php

@@ -24,6 +24,7 @@
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Indexer\Error\BulkIndexingError;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Indexer\Error\MalformedBulkRequestError;
 use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\ErrorHandlingService;
+use Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service\NodeTypeIndexingConfiguration;
 use Flowpack\ElasticSearch\Domain\Model\Document as ElasticSearchDocument;
 use Flowpack\ElasticSearch\Domain\Model\Index;
 use Flowpack\ElasticSearch\Transfer\Exception\ApiException;
@@ -132,6 +133,12 @@ class NodeIndexer extends AbstractNodeIndexer implements BulkNodeIndexerInterfac
      */
     protected $bulkProcessing = false;
 
+    /**
+     * @var NodeTypeIndexingConfiguration
+     * @Flow\Inject
+     */
+    protected $nodeTypeIndexingConfiguration;
+
     /**
      * Returns the index name to be used for indexing, with optional indexNamePostfix appended.
      *
@@ -183,6 +190,11 @@ public function getIndex(): Index
      */
     public function indexNode(NodeInterface $node, $targetWorkspaceName = null)
     {
+        if ($this->nodeTypeIndexingConfiguration->isIndexable($node->getNodeType()) === false) {
+            $this->logger->log(sprintf('NodeIndexer - Node "%s" (%s) skipped, Node Type is not allowed in the index.', $node->getContextPath(), $node->getNodeType()), LOG_DEBUG, null, 'ElasticSearch (CR)');
+            return;
+        }
+
         $indexer = function (NodeInterface $node, $targetWorkspaceName = null) {
             $contextPath = $node->getContextPath();
 

Classes/Service/NodeTypeIndexingConfiguration.php

@@ -0,0 +1,55 @@
+<?php
+namespace Flowpack\ElasticSearch\ContentRepositoryAdaptor\Service;
+
+/*
+ * 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 Flowpack\ElasticSearch\ContentRepositoryAdaptor\Exception;
+use Neos\ContentRepository\Domain\Model\NodeType;
+use Neos\Flow\Annotations as Flow;
+
+/**
+ * @Flow\Scope("singleton")
+ */
+final class NodeTypeIndexingConfiguration
+{
+    /**
+     * @var array
+     * @Flow\InjectConfiguration(path="defaultConfigurationPerNodeType", package="Neos.ContentRepository.Search")
+     */
+    protected $settings;
+
+    /**
+     * @param NodeType $nodeType
+     * @return bool
+     * @throws Exception
+     */
+    public function isIndexable(NodeType $nodeType)
+    {
+        if ($this->settings === null || !is_array($this->settings)) {
+            return true;
+        }
+
+        if (isset($this->settings[$nodeType->getName()]['indexed'])) {
+            return (bool)$this->settings[$nodeType->getName()]['indexed'];
+        }
+
+        $nodeTypeParts = explode(':', $nodeType->getName());
+        $namespace = reset($nodeTypeParts) . ':*';
+        if (isset($this->settings[$namespace]['indexed'])) {
+            return (bool)$this->settings[$namespace]['indexed'];
+        }
+        if (isset($this->settings['*']['indexed'])) {
+            return (bool)$this->settings['*']['indexed'];
+        }
+
+        return false;
+    }
+}

Configuration/Settings.yaml

@@ -1,6 +1,10 @@
 Flowpack:
   ElasticSearch:
     ContentRepositoryAdaptor:
+      configuration:
+        nodeTypes:
+          '*':
+            indexed: true
       driver:
         version: '5.x'
         mapping:

README.md

@@ -72,7 +72,35 @@ changes the index alias.
 ./flow nodeindex:cleanup
 ```
 
+### Advanced Configuration
+
+By default the indexing processes all NodeTypes, but you can change this in your *Settings.yaml*:
+
+```yaml
+Neos:
+  ContentRepository:
+    Search:
+      defaultConfigurationPerNodeType:
+        '*':
+          indexed: true
+        'Neos.Neos:FallbackNode':
+          indexed: false
+        'Neos.Neos:Shortcut':
+          indexed: false
+        'Neos.Neos:ContentCollection':
+          indexed: false
+```
+
+You need to explicitly configure the individual NodeTypes (this feature does not check the Super Type configuration).
+But you  can use a special notation to configure a full namespace, `Acme.AcmeCom:*` will be applied for all node
+types in the `Acme.AcmeCom` namespace. The most specific configuration is used in this order: 
+
+- NodeType name (`Neos.Neos:Shortcut`)
+- Full namespace notation (`Neos.Neos:*`)
+- Catch all (`*`)
+
 ### Advanced Index Settings
+
 If you need advanced settings you can define them in your *Settings.yaml*:
 
 Example is from the Documentation of the used *Flowpack.ElasticSearch* Package