lucene/facet/src/java/org/apache/lucene/facet/taxonomy/LRUHashMap.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF 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.apache.lucene.facet.taxonomy;

import java.util.LinkedHashMap;
import java.util.Map;

/**
 * LRUHashMap is an extension of Java's HashMap, which has a bounded size();
 * When it reaches that size, each time a new element is added, the least
 * recently used (LRU) entry is removed.
 * <p>
 * Java makes it very easy to implement LRUHashMap - all its functionality is
 * already available from {@link java.util.LinkedHashMap}, and we just need to
 * configure that properly.
 * <p>
 * Note that like HashMap, LRUHashMap is unsynchronized, and the user MUST
 * synchronize the access to it if used from several threads. Moreover, while
 * with HashMap this is only a concern if one of the threads is modifies the
 * map, with LURHashMap every read is a modification (because the LRU order
 * needs to be remembered) so proper synchronization is always necessary.
 * <p>
 * With the usual synchronization mechanisms available to the user, this
 * unfortunately means that LRUHashMap will probably perform sub-optimally under
 * heavy contention: while one thread uses the hash table (reads or writes), any
 * other thread will be blocked from using it - or even just starting to use it
 * (e.g., calculating the hash function). A more efficient approach would be not
 * to use LinkedHashMap at all, but rather to use a non-locking (as much as
 * possible) thread-safe solution, something along the lines of
 * java.util.concurrent.ConcurrentHashMap (though that particular class does not
 * support the additional LRU semantics, which will need to be added separately
 * using a concurrent linked list or additional storage of timestamps (in an
 * array or inside the entry objects), or whatever).
 * 
 * @lucene.experimental
 */
public class LRUHashMap<K,V> extends LinkedHashMap<K,V> {

  private int maxSize;

  /**
   * Create a new hash map with a bounded size and with least recently
   * used entries removed.
   * @param maxSize
   *     the maximum size (in number of entries) to which the map can grow
   *     before the least recently used entries start being removed.<BR>
   *      Setting maxSize to a very large value, like
   *      {@link Integer#MAX_VALUE} is allowed, but is less efficient than
   *      using {@link java.util.HashMap} because our class needs
   *      to keep track of the use order (via an additional doubly-linked
   *      list) which is not used when the map's size is always below the
   *      maximum size. 
   */
  public LRUHashMap(int maxSize) {
    super(16, 0.75f, true);
    this.maxSize = maxSize;
  }

  /**
   * Return the max size
   */
  public int getMaxSize() {
    return maxSize;
  }

  /**
   * setMaxSize() allows changing the map's maximal number of elements
   * which was defined at construction time.
   * <P>
   * Note that if the map is already larger than maxSize, the current 
   * implementation does not shrink it (by removing the oldest elements);
   * Rather, the map remains in its current size as new elements are
   * added, and will only start shrinking (until settling again on the
   * give maxSize) if existing elements are explicitly deleted.  
   */
  public void setMaxSize(int maxSize) {
    this.maxSize = maxSize;
  }

  // We override LinkedHashMap's removeEldestEntry() method. This method
  // is called every time a new entry is added, and if we return true
  // here, the eldest element will be deleted automatically. In our case,
  // we return true if the size of the map grew beyond our limit - ignoring
  // what is that eldest element that we'll be deleting.
  @Override
  protected boolean removeEldestEntry(Map.Entry<K, V> eldest) {
    return size() > maxSize;
  }

  @SuppressWarnings("unchecked")
  @Override
  public LRUHashMap<K,V> clone() {
    return (LRUHashMap<K,V>) super.clone();
  }
  
}