picasso-强大的Android图片下载缓存库

picasso是Square公司开源的一个Android图形缓存库，地址http://square.github.io/picasso/，可以实现图片下载和缓存功能。仅仅只需要一行代码就能完全实现图片的异步加载：

1	Picasso. with (context).load( "http://i.imgur.com/DvpvklR.png" ).into(imageView);

  Api看起来非常独特，是吧。

    Picasso不仅实现了图片异步加载的功能，还解决了android中加载图片时需要解决的一些常见问题：

   1.在adapter中需要取消已经不在视野范围的ImageView图片资源的加载，否则会导致图片错位，Picasso已经解决了这个问题。

   2.使用复杂的图片压缩转换来尽可能的减少内存消耗

   3.自带内存和硬盘二级缓存功能

 特性以及示例代码：

        ADAPTER 中的下载：Adapter的重用会被自动检测到，Picasso会取消上次的加载

1 2 3 4 5 6 7 8

@Override public void getView(int position, View convertView, ViewGroup parent) {    SquaredImageView view = (SquaredImageView) convertView;    if (view == null ) {      view = new SquaredImageView(context);    }    String url = getItem(position);    Picasso. with (context).load(url).into(view); }

   图片转换：转换图片以适应布局大小并减少内存占用

1 2 3 4 5

Picasso. with (context)    .load(url)    .resize(50, 50)    .centerCrop()    .into(imageView);

   你还可以自定义转换：

1 2 3 4 5 6 7 8 9 10 11 12 13

public class CropSquareTransformation implements Transformation {    @Override public Bitmap transform(Bitmap source) {      int size = Math.min(source.getWidth(), source.getHeight());      int x = (source.getWidth() - size) / 2;      int y = (source.getHeight() - size) / 2;      Bitmap result = Bitmap.createBitmap(source, x, y, size, size);      if (result != source) {        source.recycle();      }      return result;    }    @Override public String key() { return "square()" ; } }

   将CropSquareTransformation 的对象传递给transform 方法即可。

   Place holders-空白或者错误占位图片：picasso提供了两种占位图片，未加载完成或者加载发生错误的时需要一张图片作为提示。

1 2 3 4 5

Picasso. with (context)      .load(url)      .placeholder(R.drawable.user_placeholder)      .error(R.drawable.user_placeholder_error) .into(imageView);

   如果加载发生错误会重复三次请求，三次都失败才会显示erro Place holder

   资源文件的加载：除了加载网络图片picasso还支持加载Resources, assets, files, content providers中的资源文件。

1 2	Picasso. with (context).load(R.drawable.landing_screen).into(imageView1); Picasso. with (context).load( new File(...)).into(imageView2);

下面是picasso源码的解析（不看不影响使用）

Cache，缓存类

Lrucacha，主要是get和set方法，存储的结构采用了LinkedHashMap，这种map内部实现了lru算法（Least Recently Used 近期最少使用算法）。

1	this .map = new LinkedHashMap<String, Bitmap>(0, 0.75f, true );

最后一个参数的解释：

true if the ordering should be done based on the last access (from least-recently accessed to most-recently accessed), and false if the ordering should be the order in which the entries were inserted.

因为可能会涉及多线程，所以在存取的时候都会加锁。而且每次set操作后都会判断当前缓存区是否已满，如果满了就清掉最少使用的图形。代码如下

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

private void trimToSize(int maxSize) {          while ( true ) {              String key;              Bitmap value;              synchronized ( this ) {                  if (size < 0 || (map.isEmpty() && size != 0)) {                      throw new IllegalStateException(getClass().getName()                              + ".sizeOf() is reporting inconsistent results!" );                  }                                                                                                                                                                                                                                                                    if (size <= maxSize || map.isEmpty()) {                      break ;                  }                                                                                                                                                                                                                                                                    Map.Entry<String, Bitmap> toEvict = map.entrySet().iterator()                          .next();                  key = toEvict.getKey();                  value = toEvict.getValue();                  map.remove(key);                  size -= Utils.getBitmapBytes(value);                  evictionCount++;              }          } }

Request，操作封装类

所有对图形的操作都会记录在这里，供之后图形的创建使用，如重新计算大小，旋转角度，也可以自定义变换，只需要实现Transformation，一个bitmap转换的接口。

1 2 3 4 5 6 7 8 9 10 11 12 13 14

public interface Transformation {    /**     * Transform the source bitmap into a new bitmap. If you create a new bitmap instance, you must     * call {@link android.graphics.Bitmap#recycle()} on {@code source}. You may return the original     * if no transformation is required.     */    Bitmap transform(Bitmap source);                                                                                                                                                                                                               /**     * Returns a unique key for the transformation, used for caching purposes. If the transformation     * has parameters (e.g. size, scale factor, etc) then these should be part of the key.     */    String key(); }

当操作封装好以后，会将Request传到另一个结构中Action。

Action

Action代表了一个具体的加载任务，主要用于图片加载后的结果回调，有两个抽象方法，complete和error，也就是当图片解析为bitmap后用户希望做什么。最简单的就是将bitmap设置给imageview，失败了就将错误通过回调通知到上层。

ImageViewAction实现了Action，在complete中将bitmap和imageview组成了一个PicassoDrawable，里面会实现淡出的动画效果。

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

@Override      public void complete(Bitmap result, Picasso.LoadedFrom from) {          if (result == null ) {              throw new AssertionError(String.format(                      "Attempted to complete action with no result!\n%s" , this ));          }                                                                                                                                                                                          ImageView target = this .target.get();          if (target == null ) {              return ;          }                                                                                                                                                                                          Context context = picasso.context;          boolean debugging = picasso.debugging;          PicassoDrawable.setBitmap(target, context, result, from, noFade,                  debugging);                                                                                                                                                                                          if (callback != null ) {              callback.onSuccess();          }      }

有了加载任务，具体的图片下载与解析是在哪里呢？这些都是耗时的操作，应该放在异步线程中进行，就是下面的BitmapHunter。

BitmapHunter

BitmapHunter是一个Runnable，其中有一个decode的抽象方法，用于子类实现不同类型资源的解析。

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

@Override      public void run() {          try {              Thread.currentThread()                      .setName(Utils.THREAD_PREFIX + data.getName());                                                                                                                                                                       result = hunt();                                                                                                                                                                       if (result == null ) {                  dispatcher.dispatchFailed( this );              } else {                  dispatcher.dispatchComplete( this );              }          } catch (IOException e) {              exception = e;              dispatcher.dispatchRetry( this );          } catch (Exception e) {              exception = e;              dispatcher.dispatchFailed( this );          } finally {              Thread.currentThread().setName(Utils.THREAD_IDLE_NAME);          }      }                                                                                                                                                               abstract Bitmap decode(Request data) throws IOException;                                                                                                                                                               Bitmap hunt() throws IOException {          Bitmap bitmap;                                                                                                                                                                   if (!skipMemoryCache) {              bitmap = cache.get(key);              if (bitmap != null ) {                  stats.dispatchCacheHit();                  loadedFrom = MEMORY;                  return bitmap;              }          }                                                                                                                                                                   bitmap = decode(data);                                                                                                                                                                   if (bitmap != null ) {              stats.dispatchBitmapDecoded(bitmap);              if (data.needsTransformation() || exifRotation != 0) {                  synchronized (DECODE_LOCK) {                      if (data.needsMatrixTransform() || exifRotation != 0) {                          bitmap = transformResult(data, bitmap, exifRotation);                      }                      if (data.hasCustomTransformations()) {                          bitmap = applyCustomTransformations(                                  data.transformations, bitmap);                      }                  }                  stats.dispatchBitmapTransformed(bitmap);              }          }                                                                                                                                                                   return bitmap;      }

可以看到，在decode生成原始bitmap，之后会做需要的转换transformResult和applyCustomTransformations。最后在将最终的结果传递到上层dispatcher.dispatchComplete(this)。

基本的组成元素有了，那这一切是怎么连接起来运行呢，答案是Dispatcher。

Dispatcher任务调度器

在bitmaphunter成功得到bitmap后，就是通过dispatcher将结果传递出去的，当然让bitmaphunter执行也要通过Dispatcher。

Dispatcher内有一个HandlerThread，所有的请求都会通过这个thread转换，也就是请求也是异步的，这样应该是为了Ui线程更加流畅，同时保证请求的顺序，因为handler的消息队列。
外部调用的是dispatchXXX方法，然后通过handler将请求转换到对应的performXXX方法。
例如生成Action以后就会调用dispather的dispatchSubmit()来请求执行，

1 2 3

void dispatchSubmit(Action action) {          handler.sendMessage(handler.obtainMessage(REQUEST_SUBMIT, action));      }

handler接到消息后转换到performSubmit方法

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

void performSubmit(Action action) {          BitmapHunter hunter = hunterMap.get(action.getKey());          if (hunter != null ) {              hunter.attach(action);              return ;          }                                                                                                                                  if (service.isShutdown()) {              return ;          }                                                                                                                                  hunter = forRequest(context, action.getPicasso(), this , cache, stats,                  action, downloader);          hunter.future = service.submit(hunter);          hunterMap.put(action.getKey(), hunter);      }

这里将通过action得到具体的BitmapHunder，然后交给ExecutorService执行。

下面是Picasso.with(context).load("http://i.imgur.com/DvpvklR.png").into(imageView)的过程，

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

public static Picasso with (Context context) {          if (singleton == null ) {              singleton = new Builder(context).build();          }          return singleton;      }                                                                                                                          public Picasso build() {              Context context = this .context;                                                                                                                              if (downloader == null ) {                  downloader = Utils.createDefaultDownloader(context);              }              if (cache == null ) {                  cache = new LruCache(context);              }              if (service == null ) {                  service = new PicassoExecutorService();              }              if (transformer == null ) {                  transformer = RequestTransformer.IDENTITY;              }                                                                                                                              Stats stats = new Stats(cache);                                                                                                                              Dispatcher dispatcher = new Dispatcher(context, service, HANDLER,                      downloader, cache, stats);                                                                                                                              return new Picasso(context, dispatcher, cache, listener,                      transformer, stats, debugging);          }

在Picasso.with()的时候会将执行所需的所有必备元素创建出来，如缓存cache、执行executorService、调度dispatch等，在load()时创建Request，在into()中创建action、bitmapHunter，并最终交给dispatcher执行。

参考：http://blog.csdn.net/xu_fu/article/details/17043231