okhttp原理分析(持续更新)，包含okio了解,拦截器以及断点下载的使用

1. 原理

1. okio

最基本的接口只有两个：Sink(writer)、Source（Read），大概相当于OutputStream和InputStream在原生接口中的地位。这两个接口中只定义了一些最基础的IO操作方法

1. 图解okio

okio的骨架。下面是几个核心的类：
Okio：提供生成Sink和Source的方法
Sink : 接口类，功能上对应OutputStream
Source ：接口类，功能上对应InputStream
BufferedSink：接口类继承自Sink，内部有一个Buffer的Sink
BufferedSource：接口类继承自Source，内部有一个Buffer的Source
Buffer：BufferedSink和BufferedSource的最终实现实现类, 实现缓存功能，内部有一个Segment链表
Segment：里面有个byte数组,通过pos,limit控制读写的位置(从byte[]哪里开始读，哪里开始写入)，next, prev实现导航到前面或后面的Segment（实现Segment链表结构）

2. 拦截器流程

|---RealCall
|   |--- Response getResponseWithInterceptorChain()

getResponseWithInterceptorChain()

Response getResponseWithInterceptorChain() throws IOException {// Build a full stack of interceptors.List<Interceptor> interceptors = new ArrayList<>();// 客户端的所有拦截器// 使用了责任链的设计模式 每一个拦截器只处理与他相关的拦截器interceptors.addAll(client.interceptors());interceptors.add(new RetryAndFollowUpInterceptor(client));interceptors.add(new BridgeInterceptor(client.cookieJar()));interceptors.add(new CacheInterceptor(client.internalCache()));interceptors.add(new ConnectInterceptor(client));if (!forWebSocket) {interceptors.addAll(client.networkInterceptors());}interceptors.add(new CallServerInterceptor(forWebSocket));Interceptor.Chain chain = new RealInterceptorChain(interceptors, transmitter, null, 0,originalRequest, this, client.connectTimeoutMillis(),client.readTimeoutMillis(), client.writeTimeoutMillis());boolean calledNoMoreExchanges = false;try {Response response = chain.proceed(originalRequest);if (transmitter.isCanceled()) {closeQuietly(response);throw new IOException("Canceled");}return response;} catch (IOException e) {calledNoMoreExchanges = true;throw transmitter.noMoreExchanges(e);} finally {if (!calledNoMoreExchanges) {transmitter.noMoreExchanges(null);}}}

请求完成之后，通过getResponseWithInterceptorChain()方法把请求变成了一个响应。
分析几个拦截器的作用：
- RetryAndFollowUpInterceptor：处理重试的拦截器，首先会处理一些异常，只要不是一些致命的异常，它就会continue,重新发起一次请求（把request给下级），如果是致命的异常，抛给上一层。会处理一些重定向
- BridgeInterceptor 设置一些通用的请求头：Content-Type,Content-Length,Cookie,做一些返回的处理：如果返回得数据被压缩了，那木采用ZipSource，保存Cookie
- CacheInterceptor (重点) 在缓存可用的情况下，读取本地的缓存的数据，如果有首先判断有没有缓存陈列，然后判断有没有过期，如果没有过期直接拿缓存，如果过期了需要添加一些之前头部信息，如conditionName = "If-Modified-Since"
- ConnectInterceptor
  findhealthyConnection()找一个连接，首先判断有没有健康的，没有就创建（建立Socket，握手连接，）连接缓存。OkHttp基于原生的Socket.RealConnection
- CallServerInterceptor :从服务器读数据以及写数据。写头部信息，写body信息
重定向：
- 返回码是307
- 然后从response的Header中获取location
- 重新请求
Http缓存
1. Cache-Control（缓存策略）public private no-cache maxage no-sotre
2. Expires (缓存的过期策略)指明了缓存数据有效的绝对时间，告诉客户端到了这个时间点（比照客户端时间）后本地就作废了
连接三个核心类：连接复用
- RealConnection
- ConnectionPool
no-cache， no-store以及max-age=0辨析
- no-cache：相当于是重载
- no-store相当于是不缓存
- max-age=0; 相当于是刷新

OkHttpClient okHttpClient = new OkHttpClient();
OkHttpClient newClient = okHttpClient.newBuilder().cache(new Cache(mContext.getCacheDir(), 10240*1024)).connectTimeout(20, TimeUnit.SECONDS).readTimeout(20, TimeUnit.SECONDS).build();

缓存相关使用实例

/*** 一、无论有无网路都添加缓存。* 目前的情况是我们这个要addNetworkInterceptor* 这样才有效。经过本人测试（chan）测试有效.* 60S后如果没有网络将获取不到数据，显示连接失败*/static Interceptor netInterceptor = new Interceptor() {@Overridepublic Response intercept(Chain chain) throws IOException {Request request = chain.request();Response response = chain.proceed(request);/*String cacheControl = request.header("Cache-Control");if (TextUtils.isEmpty(cacheControl)) {cacheControl = "public, max-age=60";}*/int maxAge = 60;return response.newBuilder().removeHeader("Pragma")// 清除头信息，因为服务器如果不支持，会返回一些干扰信息，不清除下面无法生效.removeHeader("Cache-Control").header("Cache-Control", "public, max-age=" + maxAge).build();}};File cacheFile = new File(BaseApp.getInstance().getCacheDir(), "caheData");//设置缓存大小Cache cache = new Cache(cacheFile, DEFAULT_DIR_CACHE);//google建议放到这里OkHttpClient client = new OkHttpClient.Builder().retryOnConnectionFailure(true)//连接失败后是否重新连接.connectTimeout(15, TimeUnit.SECONDS)//超时时间15S.addNetworkInterceptor(cacheInterceptor)//这里大家一定要注意了是addNetworkOnterceptor别搞错了啊。.cache(cache).build();

拦截器的调用顺序

自己的
retry
Bridge
Cache
ConnectInterceptor
CallServer

题目

自己写一个基于okhttp的缓存，有网络10s内可以读取缓存，没网每次请求读取缓存。

参考文献

简书上的分享

2. socket

文件上传进度监听

首先需要找到okhttp中是怎样进行上传的。

首先来看一段基础的代码

 private void uploadFile() {// 这个是 Okhttp 上传文件的用法String url = "https://api.baidu.com/api/upload";File file = new File(Environment.getExternalStorageDirectory(), "test.apk");OkHttpClient httpClient = new OkHttpClient();// 构建请求 Body , 这个我们之前自己动手写过MultipartBody.Builder builder = new MultipartBody.Builder().setType(MultipartBody.FORM);builder.addFormDataPart("platform", "android");builder.addFormDataPart("file", file.getName(),RequestBody.create(MediaType.parse(guessMimeType(file.getAbsolutePath())), file));ExMultipartBody exMultipartBody = new ExMultipartBody(builder.build(),new UploadProgressListener(){@Overridepublic void onProgress(long total, long current) {showToast(total,current);}});// 怎么监听上传文件的进度？// 构建一个请求final Request request = new Request.Builder().url(url).post(exMultipartBody).build();// new RealCall 发起请求Call call = httpClient.newCall(request);call.enqueue(new Callback() {@Overridepublic void onFailure(Call call, IOException e) {e.printStackTrace();}@Overridepublic void onResponse(Call call, Response response) throws IOException {Log.e("TAG", response.body().string());}});}

分析流程

|---RealCall
|   |---CallServerInterceptor
|   |--- MultipartBody
|   |   |---writeTo(BufferedSink sink)

由于MultipartBody 是final类型的方法，我们无法直接拿到这个方法，那木我们就可以考虑静态代理的形式，编写一个扩展类，传入代理。监听到contentlength以及currentLength这两个参数之后，通过编写一个借口回调，就可以拿到值，然后在主线程中进行一些操作就可以完成。
ForwardingSink 这是Sink中的一个代理类。

断点下载

流程讲解：

大概的流程图如上所示：

|---MainActivity()
|   |---DownloadFacade(Context context) //用于 管理
|   |   |---startDownload(String url,DownloadCallback callback)
|   |   |   |---DownloadDispatcher // okhttp执行请求的类
|   |   |   |   |---startDownload(final String url, final DownloadCallback callback)  //真正执行
|   |   |   |   |   |---DownloadTask // download 线程池的分发类
|   |   |   |   |   |   |---init() //实例化，开始计算哪个线程下载哪部分
|   |   |   |   |   |   |   |---DownloadRunnable(String url, int threadId, long start, long end, long progress, DownloadEntity downloadEntity,DownloadCallback callback) //执行下载的线程

代码流程分析

1. MainActivity 在这里仅仅做实例化以及传入需要下载的对象的链接。

2. DownloadFacade 做一些实例化的准备，以及下载入口

public class DownloadFacade {private static final DownloadFacade sFacade = new DownloadFacade();private DownloadFacade(){}public static DownloadFacade getFacade() {return sFacade;}public void init(Context context){FileManager.manager().init(context);DaoManagerHelper.getManager().init(context);}public void startDownload(String url,DownloadCallback callback){DownloadDispatcher.getDispatcher().startDownload(url,callback);}public void startDownload(String url){// DownloadDispatcher.getDispatcher().startDownload(url);}
}

DownloadDispatcher

okhttp的执行类，同时也是下载任务的管理类，在这里开始执行下载

package com.darren.architect_day28.download;import com.darren.architect_day28.OkHttpManager;import java.io.IOException;
import java.util.ArrayDeque;
import java.util.Deque;import okhttp3.Call;
import okhttp3.Callback;
import okhttp3.Response;final class DownloadDispatcher {private static final DownloadDispatcher sDispatcher = new DownloadDispatcher();private DownloadDispatcher(){}public static DownloadDispatcher getDispatcher() {return sDispatcher;}/** Ready async calls in the order they'll be run. */private final Deque<DownloadTask> readyTasks = new ArrayDeque<>();/** Running asynchronous calls. Includes canceled calls that haven't finished yet. */private final Deque<DownloadTask> runningTasks = new ArrayDeque<>();/** Running synchronous calls. Includes canceled calls that haven't finished yet. */private final Deque<DownloadTask> stopTasks = new ArrayDeque<>();// 最大只能下载多少个 3 5public void startDownload(final String url, final DownloadCallback callback){// 获取文件的大小Call call = OkHttpManager.getManager().asyncCall(url);call.enqueue(new Callback() {@Overridepublic void onFailure(Call call, IOException e) {callback.onFailure(e);}@Overridepublic void onResponse(Call call, Response response) throws IOException {// 获取文件的大小long contentLength = response.body().contentLength();if(contentLength <= -1){// 没有获取到文件的大小，// 1. 跟后台商量// 2. 只能采用单线程去下载return;}// 计算每个线程负责哪一块？DownloadTask downloadTask = new DownloadTask(url,contentLength,callback);downloadTask.init();runningTasks.add(downloadTask);}});}public void recyclerTask(DownloadTask downloadTask) {runningTasks.remove(downloadTask);// 参考 OkHttp 的 Dispatcher 的源码,如果还有需要下载的开始下一个的下载}public void stopDownload(String url){// 这个停止的是不是正在下载的}// 开个单独的线程去执行 所有下载的回调}

DownloadTask

真正执行断点下载的管理类，为执行断点下载的线程分发任务。这个task执行完毕后从downloaddispatcher中的list中移除任务。


public class DownloadTask {private String mUrl;private long mContentLength;private List<DownloadRunnable> mRunnables;// OkHttp 为什么搞一个能被回收的线程池？OkHttpClient client = new OkHttpClient();/*** Executes calls. Created lazily.*/private@NullableExecutorService executorService;private volatile int mSucceedNumber;private DownloadCallback mCallback;private static final ThreadFactory sThreadFactory = new ThreadFactory() {private final AtomicInteger mCount = new AtomicInteger(1);public Thread newThread(Runnable r) {return new Thread(r, "DownThread #" + mCount.getAndIncrement());}};public synchronized ExecutorService executorService() {if (executorService == null) {executorService = new ThreadPoolExecutor(0, THREAD_SIZE, 30, TimeUnit.SECONDS,new SynchronousQueue<Runnable>(), new ThreadFactory() {@Overridepublic Thread newThread(@NonNull Runnable r) {Thread thread = new Thread(r, "DownloadTask");thread.setDaemon(false);return thread;}});}return executorService;}private static final int CPU_COUNT = Runtime.getRuntime().availableProcessors();private static final int THREAD_SIZE = Math.max(2, Math.min(CPU_COUNT - 1, 4));public DownloadTask(String url, long contentLength, DownloadCallback callback) {this.mUrl = url;this.mContentLength = contentLength;mRunnables = new ArrayList<>();this.mCallback = callback;}/*** 初始化*/public void init() {for (int i = 0; i < THREAD_SIZE; i++) {// 计算出每个线程要下载的内容long threadSize = mContentLength / THREAD_SIZE;// 初始化的时候 这里要去读取数据库long start = i * threadSize;long end = (i + threadSize) - 1;if (i == THREAD_SIZE - 1) {end = mContentLength - 1;}List<DownloadEntity> entities = DaoManagerHelper.getManager().queryAll(mUrl);DownloadEntity downloadEntity = getEntity(i, entities);if (downloadEntity == null) {downloadEntity = new DownloadEntity(start, end, mUrl, i, 0, mContentLength);}DownloadRunnable downloadRunnable = new DownloadRunnable(mUrl, i, start, end,downloadEntity.getProgress(), downloadEntity, new DownloadCallback() {@Overridepublic void onFailure(IOException e) {// 一个apk 下载里面有一个线程异常了，处理异常,把其他线程停止掉mCallback.onFailure(e);}@Overridepublic void onSucceed(File file) {// 线程同步一下，synchronized (DownloadTask.this) {mSucceedNumber += 1;if (mSucceedNumber == THREAD_SIZE) {mCallback.onSucceed(file);DownloadDispatcher.getDispatcher().recyclerTask(DownloadTask.this);// 清楚数据库的这个文件下载存储}}}});// 通过线程池去执行executorService().execute(downloadRunnable);}}private DownloadEntity getEntity(int threadId, List<DownloadEntity> entities) {for (DownloadEntity entity : entities) {if (threadId == entity.getThreadId()) {return entity;}}return null;}public void stop() {for (DownloadRunnable runnable : mRunnables) {runnable.stop();}}

DownloadRunnable 执行断点下载的线程

public class DownloadRunnable implements Runnable{private static final int STATUS_DOWNLOADING = 1;private static final int STATUS_STOP = 2;private final long start;private final long end;private final int threadId;private final String url;private final DownloadCallback mCallback;private int mStatus = STATUS_DOWNLOADING;private long mProgress = 0;private DownloadEntity mDownloadEntity;public DownloadRunnable(String url, int threadId, long start, long end, long progress, DownloadEntity downloadEntity,DownloadCallback callback) {this.threadId = threadId;this.url = url;this.start = start + progress;// 1M-2M 0.5M  1.5M - 2Mthis.end = end;mCallback = callback;this.mProgress = progress;this.mDownloadEntity = downloadEntity;}@Overridepublic void run() {// 只读写我自己的内容，RangeRandomAccessFile accessFile = null;InputStream inputStream = null;try {Response response = OkHttpManager.getManager().syncResponse(url,start,end);Log.e("TAG",this.toString());inputStream = response.body().byteStream();// 写数据File file = FileManager.manager().getFile(url);// 从这里开始accessFile.seek(start);int len = 0;byte[] buffer = new byte[1024*10];while ((len = inputStream.read(buffer))!=-1){if(mStatus == STATUS_STOP)break;// 保存进度，做断点 , 100kbmProgress += len;accessFile.write(buffer,0,len);}mCallback.onSucceed(file);} catch (IOException e) {mCallback.onFailure(e);}finally {Utils.close(inputStream);Utils.close(accessFile);// 存到数据库，数据库怎么存？mDownloadEntity.setProgress(mProgress);DaoManagerHelper.getManager().addEntity(mDownloadEntity);}}@Overridepublic String toString() {return "DownloadRunnable{" +"start=" + start +", end=" + end +", threadId=" + threadId +", url='" + url + '\'' +'}';}public void stop() {mStatus = STATUS_STOP;}

参考文献

hongyang
okhttp的使用教程