Preparations for configurable bloom index analyzer

neo4j · Sep 12, 2017 · 8fca632 · 8fca632
1 parent fbc12d9
commit 8fca632
Show file tree

Hide file tree

Showing 6 changed files with 27 additions and 22 deletions.
diff --git a/.../bloom/BloomInsightDocumentStructure.java → ...pi/impl/bloom/BloomDocumentStructure.java b/.../bloom/BloomInsightDocumentStructure.java → ...pi/impl/bloom/BloomDocumentStructure.java
@@ -34,13 +34,13 @@
 
 import static org.apache.lucene.document.Field.Store.YES;
 
-class BloomInsightDocumentStructure
+class BloomDocumentStructure
 {
     static final String ID_KEY = "id";
 
     private static final ThreadLocal<DocWithId> perThreadDocument = ThreadLocal.withInitial( DocWithId::new );
 
-    private BloomInsightDocumentStructure()
+    private BloomDocumentStructure()
     {
     }
 

diff --git a/enterprise/bloom-index/src/main/java/org/neo4j/kernel/api/impl/bloom/BloomIndex.java b/enterprise/bloom-index/src/main/java/org/neo4j/kernel/api/impl/bloom/BloomIndex.java
@@ -45,21 +45,24 @@ public class BloomIndex implements AutoCloseable
     public BloomIndex( FileSystemAbstraction fileSystem, File file, Config config ) throws IOException
     {
         this.properties = config.get( GraphDatabaseSettings.bloom_indexed_properties ).toArray( new String[0] );
-        Factory<IndexWriterConfig> population = () -> IndexWriterConfigs.population( new EnglishAnalyzer() );
+        EnglishAnalyzer analyzer = new EnglishAnalyzer();
+        Factory<IndexWriterConfig> population = () -> {
+            return IndexWriterConfigs.population( analyzer );
+        };
         WritableIndexPartitionFactory partitionFactory = new WritableIndexPartitionFactory( population );
 
         LuceneIndexStorageBuilder storageBuilder = LuceneIndexStorageBuilder.create();
         storageBuilder.withFileSystem( fileSystem ).withIndexIdentifier( "insightNodes" )
                 .withDirectoryFactory( directoryFactory( false, fileSystem ) )
                 .withIndexRootFolder( Paths.get( file.getAbsolutePath(),"insightindex" ).toFile() );
-        nodeIndex = new BloomLuceneIndex( storageBuilder.build(), partitionFactory, this.properties );
+        nodeIndex = new BloomLuceneIndex( storageBuilder.build(), partitionFactory, this.properties, analyzer);
         nodeIndex.open();
 
         storageBuilder = LuceneIndexStorageBuilder.create();
         storageBuilder.withFileSystem( fileSystem ).withIndexIdentifier( "insightRelationships" )
                 .withDirectoryFactory( directoryFactory( false, fileSystem ) )
                 .withIndexRootFolder( Paths.get( file.getAbsolutePath(),"insightindex" ).toFile() );
-        relationshipIndex = new BloomLuceneIndex( storageBuilder.build(), partitionFactory, properties );
+        relationshipIndex = new BloomLuceneIndex( storageBuilder.build(), partitionFactory, properties,  analyzer);
         relationshipIndex.open();
     }
 

diff --git a/...ndex/src/main/java/org/neo4j/kernel/api/impl/bloom/BloomIndexTransactionEventUpdater.java b/...ndex/src/main/java/org/neo4j/kernel/api/impl/bloom/BloomIndexTransactionEventUpdater.java
@@ -103,10 +103,10 @@ private void updatePropertyData( Iterable<PropertyEntry<Node>> propertyEntries,
                 return;
             }
 
-            Document document = BloomInsightDocumentStructure.documentRepresentingProperties( nodeId, allProperties );
+            Document document = BloomDocumentStructure.documentRepresentingProperties( nodeId, allProperties );
             try
             {
-                nodeIndex.getIndexWriter().updateDocument( BloomInsightDocumentStructure.newTermForChangeOrRemove( nodeId ), document );
+                nodeIndex.getIndexWriter().updateDocument( BloomDocumentStructure.newTermForChangeOrRemove( nodeId ), document );
             }
             catch ( IOException e )
             {
@@ -129,7 +129,7 @@ private void deleteIndexData( Iterable<Node> nodes, WritableDatabaseBloomIndex n
         {
             try
             {
-                nodeIndex.getIndexWriter().deleteDocuments( BloomInsightDocumentStructure.newTermForChangeOrRemove( node.getId() ) );
+                nodeIndex.getIndexWriter().deleteDocuments( BloomDocumentStructure.newTermForChangeOrRemove( node.getId() ) );
             }
             catch ( IOException e )
             {

diff --git a/enterprise/bloom-index/src/main/java/org/neo4j/kernel/api/impl/bloom/BloomLuceneIndex.java b/enterprise/bloom-index/src/main/java/org/neo4j/kernel/api/impl/bloom/BloomLuceneIndex.java
@@ -19,6 +19,7 @@
  */
 package org.neo4j.kernel.api.impl.bloom;
 
+import org.apache.lucene.analysis.Analyzer;
 import org.apache.lucene.index.DirectoryReader;
 import org.apache.lucene.index.IndexWriter;
 import org.apache.lucene.store.Directory;
@@ -41,12 +42,13 @@ class BloomLuceneIndex extends AbstractLuceneIndex
 {
 
     private String[] properties;
+    private final Analyzer analyzer;
 
-    BloomLuceneIndex( PartitionedIndexStorage indexStorage, IndexPartitionFactory partitionFactory,
+    BloomLuceneIndex( PartitionedIndexStorage indexStorage, IndexPartitionFactory partitionFactory, String[] properties, Analyzer analyzer )
-            String[] properties )
     {
         super( indexStorage, partitionFactory );
         this.properties = properties;
+        this.analyzer = analyzer;
     }
 
     private static final String KEY_STATUS = "status";
@@ -129,14 +131,14 @@ public void markAsFailed( String failure ) throws IOException
     private SimpleBloomIndexReader createSimpleReader( List<AbstractIndexPartition> partitions ) throws IOException
     {
         AbstractIndexPartition singlePartition = getFirstPartition( partitions );
-        return new SimpleBloomIndexReader( singlePartition.acquireSearcher(), properties );
+        return new SimpleBloomIndexReader( singlePartition.acquireSearcher(), properties, analyzer );
     }
 
     private PartitionedBloomIndexReader createPartitionedReader( List<AbstractIndexPartition> partitions )
             throws IOException
     {
         List<PartitionSearcher> searchers = acquireSearchers( partitions );
-        return new PartitionedBloomIndexReader( searchers, properties );
+        return new PartitionedBloomIndexReader( searchers, properties, analyzer );
     }
 
 }
diff --git a/...loom-index/src/main/java/org/neo4j/kernel/api/impl/bloom/PartitionedBloomIndexReader.java b/...loom-index/src/main/java/org/neo4j/kernel/api/impl/bloom/PartitionedBloomIndexReader.java
@@ -19,6 +19,8 @@
  */
 package org.neo4j.kernel.api.impl.bloom;
 
+import org.apache.lucene.analysis.Analyzer;
+
 import java.io.IOException;
 import java.util.List;
 import java.util.function.Function;
@@ -41,10 +43,9 @@ class PartitionedBloomIndexReader implements BloomIndexReader
 
     private final List<BloomIndexReader> indexReaders;
 
-    PartitionedBloomIndexReader( List<PartitionSearcher> partitionSearchers, String[] properties )
+    PartitionedBloomIndexReader( List<PartitionSearcher> partitionSearchers, String[] properties, Analyzer analyzer )
     {
-        this( partitionSearchers.stream().map( partitionSearcher -> new SimpleBloomIndexReader( partitionSearcher,
+        this( partitionSearchers.stream().map( partitionSearcher -> new SimpleBloomIndexReader( partitionSearcher, properties, analyzer ) )
-                properties ) )
                 .collect( Collectors.toList() ) );
     }
 

diff --git a/...ise/bloom-index/src/main/java/org/neo4j/kernel/api/impl/bloom/SimpleBloomIndexReader.java b/...ise/bloom-index/src/main/java/org/neo4j/kernel/api/impl/bloom/SimpleBloomIndexReader.java
@@ -19,7 +19,7 @@
  */
 package org.neo4j.kernel.api.impl.bloom;
 
-import org.apache.lucene.analysis.en.EnglishAnalyzer;
+import org.apache.lucene.analysis.Analyzer;
 import org.apache.lucene.queryparser.classic.MultiFieldQueryParser;
 import org.apache.lucene.queryparser.classic.QueryParser;
 import org.apache.lucene.search.BooleanClause;
@@ -45,22 +45,21 @@
 class SimpleBloomIndexReader implements BloomIndexReader
 {
     private final PartitionSearcher partitionSearcher;
-    private final EnglishAnalyzer analyzer;
+    private final Analyzer analyzer;
     private String[] properties;
-    private final QueryParser multiFieldQueryParser;
 
-    SimpleBloomIndexReader( PartitionSearcher partitionSearcher, String[] properties )
+    SimpleBloomIndexReader( PartitionSearcher partitionSearcher, String[] properties, Analyzer analyzer )
     {
         this.partitionSearcher = partitionSearcher;
         this.properties = properties;
-        analyzer = new EnglishAnalyzer();
+        this.analyzer = analyzer;
-        multiFieldQueryParser = new MultiFieldQueryParser( properties, analyzer );
-        multiFieldQueryParser.setDefaultOperator( QueryParser.Operator.OR );
     }
 
     public PrimitiveLongIterator query( String... query )
     {
         BooleanQuery.Builder builder = new BooleanQuery.Builder();
+        QueryParser multiFieldQueryParser = new MultiFieldQueryParser( properties, analyzer );
+        multiFieldQueryParser.setDefaultOperator( QueryParser.Operator.OR );
         for ( String s : query )
         {
             for ( String property : properties )