org.elasticsearch.search.fetch.FetchSubPhase Maven / Gradle / Ivy
The newest version!
/*
* Licensed to Elasticsearch under one or more contributor
* license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright
* ownership. Elasticsearch licenses this file to you under
* the Apache License, Version 2.0 (the "License"); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/
package org.elasticsearch.search.fetch;
import com.google.common.collect.Maps;
import org.apache.lucene.index.IndexReader;
import org.apache.lucene.index.LeafReader;
import org.apache.lucene.index.LeafReaderContext;
import org.apache.lucene.search.IndexSearcher;
import org.elasticsearch.search.SearchHit;
import org.elasticsearch.search.SearchParseElement;
import org.elasticsearch.search.internal.InternalSearchHit;
import org.elasticsearch.search.internal.SearchContext;
import java.util.Map;
/**
*
*/
public interface FetchSubPhase {
public static class HitContext {
private InternalSearchHit hit;
private IndexSearcher searcher;
private LeafReaderContext readerContext;
private int docId;
private Map cache;
public void reset(InternalSearchHit hit, LeafReaderContext context, int docId, IndexSearcher searcher) {
this.hit = hit;
this.readerContext = context;
this.docId = docId;
this.searcher = searcher;
}
public InternalSearchHit hit() {
return hit;
}
public LeafReader reader() {
return readerContext.reader();
}
public LeafReaderContext readerContext() {
return readerContext;
}
public int docId() {
return docId;
}
public IndexReader topLevelReader() {
return searcher.getIndexReader();
}
public IndexSearcher topLevelSearcher() {
return searcher;
}
public Map cache() {
if (cache == null) {
cache = Maps.newHashMap();
}
return cache;
}
public String getSourcePath(String sourcePath) {
SearchHit.NestedIdentity nested = hit().getNestedIdentity();
if (nested != null) {
// in case of nested we need to figure out what is the _source field from the perspective
// of the nested hit it self. The nested _source is isolated and the root and potentially parent objects
// are gone
StringBuilder nestedPath = new StringBuilder();
for (; nested != null; nested = nested.getChild()) {
nestedPath.append(nested.getField());
}
assert sourcePath.startsWith(nestedPath.toString());
int startIndex = nestedPath.length() + 1; // the path until the deepest nested object + '.'
return sourcePath.substring(startIndex);
} else {
return sourcePath;
}
}
}
Map parseElements();
boolean hitExecutionNeeded(SearchContext context);
/**
* Executes the hit level phase, with a reader and doc id (note, its a low level reader, and the matching doc).
*/
void hitExecute(SearchContext context, HitContext hitContext);
boolean hitsExecutionNeeded(SearchContext context);
void hitsExecute(SearchContext context, InternalSearchHit[] hits);
/**
* This interface is in the fetch phase plugin mechanism.
* Whenever a new search is executed we create a new {@link SearchContext} that holds individual contexts for each {@link org.elasticsearch.search.fetch.FetchSubPhase}.
* Fetch phases that use the plugin mechanism must provide a ContextFactory to the SearchContext that creates the fetch phase context and also associates them with a name.
* See {@link SearchContext#getFetchSubPhaseContext(FetchSubPhase.ContextFactory)}
*/
public interface ContextFactory {
/**
* The name of the context.
*/
public String getName();
/**
* Creates a new instance of a FetchSubPhaseContext that holds all information a FetchSubPhase needs to execute on hits.
*/
public SubPhaseContext newContextInstance();
}
}