org.hibernate.search.engine.spi.AbstractDocumentBuilder Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of hibernate-search-engine Show documentation
Show all versions of hibernate-search-engine Show documentation
Core of the Object/Lucene mapper, query engine and index management
/*
* Hibernate Search, full-text search for your domain model
*
* License: GNU Lesser General Public License (LGPL), version 2.1 or later
* See the lgpl.txt file in the root directory or .
*/
package org.hibernate.search.engine.spi;
import java.io.Serializable;
import java.util.Collection;
import java.util.Collections;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Set;
import org.hibernate.annotations.common.reflection.ReflectionManager;
import org.hibernate.annotations.common.reflection.XClass;
import org.hibernate.annotations.common.reflection.XMember;
import org.hibernate.search.backend.LuceneWork;
import org.hibernate.search.bridge.spi.ConversionContext;
import org.hibernate.search.engine.BoostStrategy;
import org.hibernate.search.engine.impl.DefaultBoostStrategy;
import org.hibernate.search.engine.impl.WorkPlan;
import org.hibernate.search.engine.metadata.impl.ContainedInMetadata;
import org.hibernate.search.engine.metadata.impl.EmbeddedTypeMetadata;
import org.hibernate.search.engine.metadata.impl.PropertyMetadata;
import org.hibernate.search.engine.metadata.impl.TypeMetadata;
import org.hibernate.search.exception.AssertionFailure;
import org.hibernate.search.spi.InstanceInitializer;
import org.hibernate.search.util.impl.ReflectionHelper;
import org.hibernate.search.util.impl.ScopedAnalyzerReference;
import org.hibernate.search.util.logging.impl.Log;
import org.hibernate.search.util.logging.impl.LoggerFactory;
/**
* Abstract base class for the document builders.
*
* @author Hardy Ferentschik
* @author Davide D'Alto
* @author Sanne Grinovero
*/
public abstract class AbstractDocumentBuilder {
private static final Log log = LoggerFactory.make();
private final XClass beanXClass;
private final Class beanClass;
private final TypeMetadata typeMetadata;
private final InstanceInitializer instanceInitializer;
private boolean isRoot;
private Set> mappedSubclasses = new HashSet>();
protected EntityState entityState;
/**
* Constructor.
*
* @param xClass The class for which to build a document builder
* @param typeMetadata metadata for the specified class
* @param reflectionManager Reflection manager to use for processing the annotations
* @param optimizationBlackList keeps track of types on which we need to disable collection events optimizations
* @param instanceInitializer a {@link org.hibernate.search.spi.InstanceInitializer} object.
*/
public AbstractDocumentBuilder(XClass xClass,
TypeMetadata typeMetadata,
ReflectionManager reflectionManager,
Set optimizationBlackList,
InstanceInitializer instanceInitializer) {
if ( xClass == null ) {
throw new AssertionFailure( "Unable to build a DocumentBuilderContainedEntity with a null class" );
}
this.instanceInitializer = instanceInitializer;
this.entityState = EntityState.CONTAINED_IN_ONLY;
this.beanXClass = xClass;
this.beanClass = reflectionManager.toClass( xClass );
this.typeMetadata = typeMetadata;
optimizationBlackList.addAll( typeMetadata.getOptimizationBlackList() );
}
public abstract void addWorkToQueue(
String tenantIdentifier,
Class entityClass,
Object entity, Serializable id,
boolean delete,
boolean add,
List queue,
ConversionContext contextualBridge);
/**
* In case of an indexed entity, return the value of it's identifier: what is marked as @Id or @DocumentId;
* in case the entity uses @ProvidedId, it's illegal to call this method.
*
* @param entity the instance for which to retrieve the id
*
* @return the value, or null if it's not an indexed entity
*
* @throws IllegalStateException when used with a @ProvidedId annotated entity
*/
public abstract Serializable getId(Object entity);
public TypeMetadata getTypeMetadata() {
return typeMetadata;
}
public boolean isRoot() {
return isRoot;
}
public Class getBeanClass() {
return beanClass;
}
public XClass getBeanXClass() {
return beanXClass;
}
public TypeMetadata getMetadata() {
return typeMetadata;
}
public ScopedAnalyzerReference getAnalyzerReference() {
return typeMetadata.getDefaultAnalyzerReference();
}
public EntityState getEntityState() {
return entityState;
}
public Set> getMappedSubclasses() {
return mappedSubclasses;
}
public void postInitialize(Set> indexedClasses) {
//we initialize only once because we no longer have a reference to the reflectionManager
//in theory
Class plainClass = beanClass;
if ( entityState == EntityState.NON_INDEXABLE ) {
throw new AssertionFailure( "A non indexed entity is post processed" );
}
Set> tempMappedSubclasses = new HashSet>();
//together with the caller this creates a o(2), but I think it's still faster than create the up hierarchy for each class
for ( Class currentClass : indexedClasses ) {
if ( plainClass != currentClass && plainClass.isAssignableFrom( currentClass ) ) {
tempMappedSubclasses.add( currentClass );
}
}
this.mappedSubclasses = Collections.unmodifiableSet( tempMappedSubclasses );
Class superClass = plainClass.getSuperclass();
this.isRoot = true;
while ( superClass != null ) {
if ( indexedClasses.contains( superClass ) ) {
this.isRoot = false;
break;
}
superClass = superClass.getSuperclass();
}
}
/**
* If we have a work instance we have to check whether the instance to be indexed is contained in any other indexed entities.
*
* @see #appendContainedInWorkForInstance(Object, WorkPlan, ContainedInRecursionContext, String)
* @param instance the instance to be indexed
* @param workPlan the current work plan
* @param currentRecursionContext the current {@link org.hibernate.search.engine.spi.ContainedInRecursionContext} object used to check the graph traversal
*/
public void appendContainedInWorkForInstance(Object instance, WorkPlan workPlan, ContainedInRecursionContext currentRecursionContext) {
appendContainedInWorkForInstance( instance, workPlan, currentRecursionContext, null );
}
/**
* If we have a work instance we have to check whether the instance to be indexed is contained in any other indexed entities for a tenant.
*
* @param instance the instance to be indexed
* @param workPlan the current work plan
* @param currentRecursionContext the current {@link org.hibernate.search.engine.spi.ContainedInRecursionContext} object used to check the graph traversal
* @param tenantIdentifier the identifier of the tenant or null, if there isn't one
* @see #appendContainedInWorkForInstance(Object, WorkPlan, ContainedInRecursionContext)
*/
public void appendContainedInWorkForInstance(Object instance, WorkPlan workPlan, ContainedInRecursionContext currentRecursionContext, String tenantIdentifier) {
for ( ContainedInMetadata containedInMetadata : typeMetadata.getContainedInMetadata() ) {
XMember member = containedInMetadata.getContainedInMember();
Object unproxiedInstance = instanceInitializer.unproxy( instance );
ContainedInRecursionContext recursionContext = updateContainedInRecursionContext( unproxiedInstance, containedInMetadata, currentRecursionContext );
if ( recursionContext.isTerminal() ) {
continue;
}
Object value = ReflectionHelper.getMemberValue( unproxiedInstance, member );
if ( value == null ) {
continue;
}
if ( member.isArray() ) {
Object[] array = (Object[]) value;
for ( Object arrayValue : array ) {
processSingleContainedInInstance( workPlan, arrayValue, recursionContext, tenantIdentifier );
}
}
else if ( member.isCollection() ) {
Collection collection = null;
try {
collection = getActualCollection( member, value );
collection.size(); //load it
}
catch (Exception e) {
if ( e.getClass().getName().contains( "org.hibernate.LazyInitializationException" ) ) {
/* A deleted entity not having its collection initialized
* leads to a LIE because the collection is no longer attached to the session
*
* But that's ok as the collection update event has been processed before
* or the fk would have been cleared and thus triggering the cleaning
*/
collection = null;
}
}
if ( collection != null ) {
for ( Object collectionValue : collection ) {
processSingleContainedInInstance( workPlan, collectionValue, recursionContext, tenantIdentifier );
}
}
}
else {
processSingleContainedInInstance( workPlan, value, recursionContext, tenantIdentifier );
}
}
}
protected InstanceInitializer getInstanceInitializer() {
return instanceInitializer;
}
private ContainedInRecursionContext updateContainedInRecursionContext(Object containedInstance, ContainedInMetadata containedInMetadata,
ContainedInRecursionContext containedContext) {
int maxDepth;
int depth;
// Handle @IndexedEmbedded.depth-induced limits
Integer metadataMaxDepth = containedInMetadata.getMaxDepth();
if ( containedInstance != null && metadataMaxDepth != null ) {
maxDepth = metadataMaxDepth;
}
else {
maxDepth = containedContext != null ? containedContext.getMaxDepth() : Integer.MAX_VALUE;
}
depth = containedContext != null ? containedContext.getDepth() : 0;
if ( depth < Integer.MAX_VALUE ) { // Avoid integer overflow
++depth;
}
/*
* Handle @IndexedEmbedded.includePaths-induced limits If the context for the contained element has a
* comprehensive set of included paths, and if the @IndexedEmbedded matching the @ContainedIn we're currently
* processing also has a comprehensive set of embedded paths, *then* we can compute the resulting set of
* embedded fields (which is the intersection of those two sets). If this resulting set is empty, we can safely
* stop the @ContainedIn processing: any changed field wouldn't be included in the Lucene document for
* "containerInstance" anyway.
*/
Set comprehensivePaths;
Set metadataIncludePaths = containedInMetadata.getIncludePaths();
/*
* See @IndexedEmbedded.depth: it should be considered as zero if it has its default value and if includePaths
* contains elements
*/
if ( metadataIncludePaths != null && !metadataIncludePaths.isEmpty()
&& metadataMaxDepth != null && metadataMaxDepth.equals( Integer.MAX_VALUE ) ) {
String metadataPrefix = containedInMetadata.getPrefix();
/*
* If the contained context Filter by contained context's included paths if they are comprehensive This
* allows to detect when a @ContainedIn is irrelevant because the matching @IndexedEmbedded would not
* capture any property.
*/
Set containedComprehensivePaths =
containedContext != null ? containedContext.getComprehensivePaths() : null;
comprehensivePaths = new HashSet<>();
for ( String includedPath : metadataIncludePaths ) {
/*
* If the contained context has a comprehensive list of included paths, use it to filter out our own
* list
*/
if ( containedComprehensivePaths == null || containedComprehensivePaths.contains( includedPath ) ) {
comprehensivePaths.add( metadataPrefix + includedPath );
}
}
}
else {
comprehensivePaths = null;
}
return new ContainedInRecursionContext( maxDepth, depth, comprehensivePaths );
}
@Override
public String toString() {
return "DocumentBuilder for {" + beanClass.getName() + "}";
}
/**
* A {@code XMember } instance treats a map as a collection as well in which case the map values are returned as
* collection.
*
* @param member The member instance
* @param value The value
*
* @return The {@code value} cast to collection or in case of {@code value} being a map the map values as collection.
*/
@SuppressWarnings("unchecked")
private Collection getActualCollection(XMember member, Object value) {
Collection collection;
if ( Map.class.equals( member.getCollectionClass() ) ) {
collection = ( (Map) value ).values();
}
else {
collection = (Collection) value;
}
return collection;
}
private void processSingleContainedInInstance(WorkPlan workplan, T value, ContainedInRecursionContext depth, String tenantId) {
workplan.recurseContainedIn( value, depth, tenantId );
}
/**
* Hibernate entities might be dirty (their state has changed), but none of these changes would affect
* the index state. This method will return {@code true} if any of changed entity properties identified
* by their names ({@code dirtyPropertyNames}) will effect the index state.
*
* @param dirtyPropertyNames array of property names for the changed entity properties, {@code null} in case the
* changed properties cannot be specified.
*
* @return {@code true} if the entity changes will effect the index state, {@code false} otherwise
*
* @since 3.4
*/
public boolean isDirty(String[] dirtyPropertyNames) {
if ( dirtyPropertyNames == null || dirtyPropertyNames.length == 0 ) {
return true; // it appears some collection work has no oldState -> reindex
}
if ( !stateInspectionOptimizationsEnabled() ) {
return true;
}
for ( String dirtyPropertyName : dirtyPropertyNames ) {
PropertyMetadata propertyMetadata = typeMetadata.getPropertyMetadataForProperty( dirtyPropertyName );
if ( propertyMetadata != null ) {
// if there is a property metadata it means that there is at least one @Field.
// Fields are either indexed or stored, so we need to re-index
return true;
}
// consider IndexedEmbedded:
for ( EmbeddedTypeMetadata embeddedTypeMetadata : typeMetadata.getEmbeddedTypeMetadata() ) {
String name = embeddedTypeMetadata.getEmbeddedFieldName();
if ( name.equals( dirtyPropertyName ) ) {
return true;
}
}
}
return false;
}
/**
* To be removed, see org.hibernate.search.engine.DocumentBuilderIndexedEntity.isIdMatchingJpaId()
*
* @return true if a providedId needs to be provided for indexing
*/
public boolean requiresProvidedId() {
return false;
}
/**
* To be removed, see org.hibernate.search.engine.DocumentBuilderIndexedEntity.isIdMatchingJpaId()
*
* @return true if @DocumentId and @Id are found on the same property
*/
public boolean isIdMatchingJpaId() {
return true;
}
/**
* Returns {@code true} if the collection event is going to affect the index state. In this case
* the indexing event can be ignored. {@code false} otherwise.
*
* @param collectionRoleName a {@link java.lang.String} object.
*
* @return {@code true} if an update to the collection identified by the given role name effects the index
* state, {@code false} otherwise.
*/
public boolean collectionChangeRequiresIndexUpdate(String collectionRoleName) {
if ( collectionRoleName == null ) {
// collection name will only be non null for PostCollectionUpdateEvents
return true;
}
// don't check stateInspectionOptimizationsEnabled() as it might ignore depth limit:
// it will disable optimization even if we have class bridges, but we're too deep
// to be reachable. The evaluation of stateInspectionOptimizationsEnabled() was
// actually stored in stateInspectionOptimizationsEnabled, but limiting to depth recursion.
if ( !typeMetadata.areStateInspectionOptimizationsEnabled() ) {
// if optimizations are not enabled we need to re-index
return true;
}
return this.typeMetadata.containsCollectionRole( collectionRoleName );
}
/**
* Verifies entity level preconditions to know if it's safe to skip index updates based
* on specific field or collection updates.
*
* @return true if it seems safe to apply such optimizations
*/
boolean stateInspectionOptimizationsEnabled() {
if ( !typeMetadata.areStateInspectionOptimizationsEnabled() ) {
return false;
}
if ( typeMetadata.areClassBridgesUsed() ) {
log.tracef(
"State inspection optimization disabled as entity %s uses class bridges",
this.beanClass.getName()
);
return false; // can't know what a class bridge is going to look at -> reindex // TODO nice new feature to have?
}
BoostStrategy boostStrategy = typeMetadata.getDynamicBoost();
if ( boostStrategy != null && !( boostStrategy instanceof DefaultBoostStrategy ) ) {
log.tracef(
"State inspection optimization disabled as DynamicBoost is enabled on entity %s",
this.beanClass.getName()
);
return false; // as with class bridge: might be affected by any field // TODO nice new feature to have?
}
return true;
}
/**
* Makes sure isCollectionRoleExcluded will always return false, so that
* collection update events are always processed.
*
* @see #collectionChangeRequiresIndexUpdate(String)
*/
public void forceStateInspectionOptimizationsDisabled() {
typeMetadata.disableStateInspectionOptimizations();
}
/**
* Closes any resource
*/
public void close() {
typeMetadata.getDefaultAnalyzerReference().close();
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy