コード例 #1
0
ファイル: Spider.java プロジェクト: huzorro/webmagic
 /**
  * Process specific urls without url discovering.
  *
  * @param urls urls to process
  */
 public void test(String... urls) {
   checkComponent();
   if (urls.length > 0) {
     for (String url : urls) {
       processRequest(new Request(url));
     }
   }
 }
コード例 #2
0
ファイル: Spider.java プロジェクト: huzorro/webmagic
 @Override
 public void run() {
   if (!stat.compareAndSet(STAT_INIT, STAT_RUNNING)) {
     throw new IllegalStateException("Spider is already running!");
   }
   checkComponent();
   if (startUrls != null) {
     for (String startUrl : startUrls) {
       scheduler.push(new Request(startUrl), this);
     }
   }
   Request request = scheduler.poll(this);
   // singel thread
   if (executorService == null) {
     while (request != null) {
       processRequest(request);
       request = scheduler.poll(this);
     }
   } else {
     // multi thread
     final AtomicInteger threadAlive = new AtomicInteger(0);
     while (true) {
       if (request == null) {
         // when no request found but some thread is alive, sleep a while.
         try {
           Thread.sleep(100);
         } catch (InterruptedException e) {
         }
       } else {
         final Request requestFinal = request;
         threadAlive.incrementAndGet();
         executorService.execute(
             new Runnable() {
               @Override
               public void run() {
                 processRequest(requestFinal);
                 threadAlive.decrementAndGet();
               }
             });
       }
       request = scheduler.poll(this);
       if (threadAlive.get() == 0) {
         request = scheduler.poll(this);
         if (request == null) {
           break;
         }
       }
     }
     executorService.shutdown();
   }
   stat.compareAndSet(STAT_RUNNING, STAT_STOPPED);
   // release some resources
   destroy();
 }