total++;
    }
    assertEquals(0, in.available());
    assertEquals(20, total);
  }

  @SuppressWarnings("CheckReturnValue") // these calls to skip always return 0
  public void testSkip() throws Exception {
    MultiInputStream multi =
        new MultiInputStream(
            Collections.singleton(
                    new ByteSource() {
                      @Override

      total++;
    }
    assertEquals(0, in.available());
    assertEquals(20, total);
  }

  @SuppressWarnings("CheckReturnValue") // these calls to skip always return 0
  public void testSkip() throws Exception {
    MultiInputStream multi =
        new MultiInputStream(
            Collections.singleton(
                    new ByteSource() {
                      @Override

    }

    public void test_getTika_xml_sjis() {
        final InputStream in = ResourceUtil.getResourceAsStream("extractor/test_sjis.xml");
        final Map<String, String> params = new HashMap<String, String>();
        params.put("resourceName", "test_sjis.xml");
        final ExtractData extractData = tikaExtractor.getText(in, params);
        final String content = extractData.getContent();

        // Add some test data
        crawler.addUrl("http://example.com/");

        // Perform cleanup - should not throw exception
        crawler.cleanup(sessionId);
    }

    public void test_stop() throws Exception {
        final CrawlerWebServer server = new CrawlerWebServer(7070);
        server.start();

        try {
            final String url = "http://localhost:7070/";

        `(batch_size, 1)`. This enables `Model` subclasses to process scalar
        data in their `train_step()`/`test_step()`/`predict_step()` methods. \
        Note that this change may break certain subclassed models. You can
        revert back to the previous behavior by adding upranking yourself in the
        `train_step()`/`test_step()`/`predict_step()` methods, e.g. `if
        x.shape.rank == 1: x = tf.expand_dims(x, axis=-1)`. Functional models as