美文网首页
LRUCache原理

LRUCache原理

作者: 小宏why | 来源:发表于2018-02-22 15:33 被阅读0次

           讲到LruCache不得不提一下LinkedHashMap,因为LruCache中Lru算法的实现就是通过LinkedHashMap来实现的。LinkedHashMap继承于HashMap,它使用了一个双向链表来存储Map中的Entry顺序关系,这种顺序有两种,一种是LRU顺序,一种是插入顺序,这可以由其构造函数public LinkedHashMap(int initialCapacity,float loadFactor, boolean accessOrder)指定。所以,对于get、put、remove等操作,LinkedHashMap除了要做HashMap做的事情,还做些调整Entry顺序链表的工作。LruCache中将LinkedHashMap的顺序设置为LRU顺序来实现LRU缓存,每次调用get(也就是从内存缓存中取图片),则将该对象移到链表的尾端。调用put插入新的对象也是存储在链表尾端,这样当内存缓存达到设定的最大值时,将链表头部的对象(近期最少用到的)移除。

    /*

     *Copyright (C) 2011 The Android Open Source Project

     *

     *Licensed 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 android.support.v4.util;

    import java.util.LinkedHashMap;

    import java.util.Map;

    /**

     *Static library version of {@link android.util.LruCache}. Used to write apps

     *that run on API levels prior to 12. When running on API level 12 or above,

     *this implementation is still used; it does not try to switch to the

     *framework's implementation. See the framework SDK documentation for a class

     *overview.

     */

    public class LruCache {

       private final LinkedHashMap map;

       /** Size of this cache in units. Not necessarily the number of elements.*/

       private int size;    //当前cache的大小

       private int maxSize; //cache最大大小

       private int putCount;       //put的次数

       private int createCount;   //create的次数

       private int evictionCount;  //回收的次数

       private int hitCount;       //命中的次数

       private int missCount;      //未命中次数

       /**

         * @param maxSize for caches that do notoverride {@link #sizeOf}, this is

        *     the maximum number ofentries in the cache. For all other caches,

        *     this is the maximum sum ofthe sizes of the entries in this cache.

        */

       public LruCache(int maxSize) {

           if (maxSize <= 0) {

               throw new IllegalArgumentException("maxSize <= 0");

           }

           this.maxSize = maxSize;

           //将LinkedHashMap的accessOrder设置为true来实现LRU

           this.map = new LinkedHashMap(0, 0.75f, true); 

        }

       /**

        * Returns the value for {@code key} if it exists in the cache or can be

        * created by {@code #create}. If a value was returned, it is moved tothe

        * head of the queue. This returns null if a value is not cached andcannot

        * be created.

        *通过key获取相应的item,或者创建返回相应的item。相应的item会移动到队列的尾部,

        *如果item的value没有被cache或者不能被创建,则返回null。

        */

       public final V get(K key) {

           if (key == null) {

               throw new NullPointerException("key == null");

           }

           V mapValue;

           synchronized (this) {

               mapValue = map.get(key);

               if (mapValue != null) {

                    //mapValue不为空表示命中,hitCount+1并返回mapValue对象

                    hitCount++;

                    return mapValue;

               }

               missCount++;  //未命中

           }

           /*

            * Attempt to create a value. This may take a long time, and the map

            * may be different when create() returns. If a conflicting value was

            * added to the map while create() was working, we leave that value in

            * the map and release the created value.

            *如果未命中,则试图创建一个对象,这里create方法返回null,并没有实现创建对象的方法

            *如果需要事项创建对象的方法可以重写create方法。因为图片缓存时内存缓存没有命中会去

            *文件缓存中去取或者从网络下载,所以并不需要创建。

            */

           V createdValue = create(key);

           if (createdValue == null) {

               return null;

           }

           //假如创建了新的对象,则继续往下执行

           synchronized (this) {

               createCount++; 

               //将createdValue加入到map中,并且将原来键为key的对象保存到mapValue

               mapValue = map.put(key, createdValue);  

               if (mapValue != null) {

                    // There was a conflict so undothat last put

                    //如果mapValue不为空,则撤销上一步的put操作。

                    map.put(key, mapValue);

               } else {

                    //加入新创建的对象之后需要重新计算size大小

                    size += safeSizeOf(key,createdValue);

               }

           }

           if (mapValue != null) {

               entryRemoved(false, key, createdValue, mapValue);

               return mapValue;

           } else {

               //每次新加入对象都需要调用trimToSize方法看是否需要回收

               trimToSize(maxSize);

               return createdValue;

           }

        }

       /**

        * Caches {@code value} for {@code key}. The value is moved to the headof

        * the queue.

        *

        * @return the previous value mapped by {@code key}.

        */

       public final V put(K key, V value) {

           if (key == null || value == null) {

               throw new NullPointerException("key == null || value ==null");

           }

           V previous;

           synchronized (this) {

               putCount++;

               size += safeSizeOf(key, value); //size加上预put对象的大小

               previous = map.put(key, value);

               if (previous != null) {

                    //如果之前存在键为key的对象,则size应该减去原来对象的大小

                   size -= safeSizeOf(key,previous);

               }

           }

           if (previous != null) {

               entryRemoved(false, key, previous, value);

           }

           //每次新加入对象都需要调用trimToSize方法看是否需要回收

           trimToSize(maxSize);

           return previous;

        }

       /**

        * @param maxSize the maximum size of the cache before returning. May be-1

        *     to evict even 0-sizedelements.

        *此方法根据maxSize来调整内存cache的大小,如果maxSize传入-1,则清空缓存中的所有对象

        */

       private void trimToSize(int maxSize) {

           while (true) {

               K key;

               V value;

               synchronized (this) {

                    if (size < 0 ||(map.isEmpty() && size != 0)) {

                        throw newIllegalStateException(getClass().getName()

                                + ".sizeOf() is reportinginconsistent results!");

                    }

                    //如果当前size小于maxSize或者map没有任何对象,则结束循环

                    if (size <= maxSize ||map.isEmpty()) {

                        break;

                    }

                    //移除链表头部的元素,并进入下一次循环

                    Map.Entry toEvict =map.entrySet().iterator().next();

                    key = toEvict.getKey();

                    value = toEvict.getValue();

                    map.remove(key);

                    size -= safeSizeOf(key, value);

                    evictionCount++;  //回收次数+1

               }

               entryRemoved(true, key, value, null);

           }

        }

       /**

        * Removes the entry for {@code key} if it exists.

        *

        * @return the previous value mapped by {@code key}.

        *从内存缓存中根据key值移除某个对象并返回该对象

        */

       public final V remove(K key) {

           if (key == null) {

               throw new NullPointerException("key == null");

           }

           V previous;

           synchronized (this) {

               previous = map.remove(key);

               if (previous != null) {

                    size -= safeSizeOf(key,previous);

               }

           }

           if (previous != null) {

               entryRemoved(false, key, previous, null);

           }

           return previous;

        }

       /**

        * Called for entries that have been evicted or removed. This method is

        * invoked when a value is evicted to make space, removed by a call to

        * {@link #remove}, or replaced by a call to {@link #put}. The default

        * implementation does nothing.

        *

        *

    The method is called without synchronization: other threadsmay

        * access the cache while this method is executing.

        *

        * @param evicted true if the entry is being removed to make space, false

        *     if the removal was caused bya {@link #put} or {@link #remove}.

        * @param newValue the new value for {@code key}, if it exists. Ifnon-null,

        *     this removal was caused by a{@link #put}. Otherwise it was caused by

        *     an eviction or a {@link#remove}.

         */

       protected void entryRemoved(boolean evicted, K key, V oldValue, VnewValue) {}

       /**

        * Called after a cache miss to compute a value for the correspondingkey.

        * Returns the computed value or null if no value can be computed. The

         * default implementation returns null.

        *

        *

    The method is called without synchronization: other threadsmay

        * access the cache while this method is executing.

        *

        *

    If a value for {@code key} exists in the cache when thismethod

        * returns, the created value will be released with {@link #entryRemoved}

        * and discarded. This can occur when multiple threads request the samekey

        * at the same time (causing multiple values to be created), or when one

        * thread calls {@link #put} while another is creating a value for thesame

        * key.

        */

       protected V create(K key) {

           return null;

        }

       private int safeSizeOf(K key, V value) {

           int result = sizeOf(key, value);

           if (result < 0) {

               throw new IllegalStateException("Negative size: " + key +"=" + value);

           }

           return result;

        }

       /**

        * Returns the size of the entry for {@code key} and {@code value} in

        * user-defined units.  The defaultimplementation returns 1 so that size

        * is the number of entries and max size is the maximum number ofentries.

        *

        *

    An entry's size must not change while it is in the cache.

        *用来计算单个对象的大小,这里默认返回1,一般需要重写该方法来计算对象的大小

        * xUtils中创建LruMemoryCache时就重写了sizeOf方法来计算bitmap的大小

        * mMemoryCache = new LruMemoryCache(globalConfig.getMemoryCacheSize()) {

        *       @Override

        *       protected intsizeOf(MemoryCacheKey key, Bitmap bitmap) {

        *           if (bitmap == null)return 0;

        *           returnbitmap.getRowBytes() * bitmap.getHeight();

        *       }

        *   };

        *

        */

       protected int sizeOf(K key, V value) {

           return 1;

        }

       /**

        * Clear the cache, calling {@link #entryRemoved} on each removed entry.

        *清空内存缓存

        */

       public final void evictAll() {

           trimToSize(-1); // -1 will evict 0-sized elements

        }

       /**

        * For caches that do not override {@link #sizeOf}, this returns thenumber

        * of entries in the cache. For all other caches, this returns the sum of

        * the sizes of the entries in this cache.

        */

       public synchronized final int size() {

           return size;

        }

       /**

        * For caches that do not override {@link #sizeOf}, this returns themaximum

        * number of entries in the cache. For all other caches, this returns the

        * maximum sum of the sizes of the entries in this cache.

        */

       public synchronized final int maxSize() {

           return maxSize;

        }

       /**

        * Returns the number of times {@link #get} returned a value.

        */

       public synchronized final int hitCount() {

           return hitCount;

        }

       /**

        * Returns the number of times {@link #get} returned null or required anew

        * value to be created.

        */

       public synchronized final int missCount() {

           return missCount;

        }

       /**

        * Returns the number of times {@link #create(Object)} returned a value.

        */

       public synchronized final int createCount() {

           return createCount;

        }

       /**

        * Returns the number of times {@link #put} was called.

        */

       public synchronized final int putCount() {

           return putCount;

        }

       /**

        * Returns the number of values that have been evicted.

        */

       public synchronized final int evictionCount() {

           return evictionCount;

        }

       /**

        * Returns a copy of the current contents of the cache, ordered fromleast

        * recently accessed to most recently accessed.

        */

       public synchronized final Map snapshot() {

           return new LinkedHashMap(map);

        }

       @Override public synchronized final String toString() {

            int accesses = hitCount + missCount;

           int hitPercent = accesses != 0 ? (100 * hitCount / accesses) : 0;

           returnString.format("LruCache[maxSize=%d,hits=%d,misses=%d,hitRate=%d%%]",

                    maxSize, hitCount, missCount,hitPercent);

        }

    }

    相关文章

      网友评论

          本文标题:LRUCache原理

          本文链接:https://www.haomeiwen.com/subject/ohkztftx.html