if (mimeTypeHelper == null) {
            throw new CrawlerSystemException("MimeTypeHelper is unavailable.");
        }
        return mimeTypeHelper;
    }

    protected ExtractorFactory getExtractorFactory() {
        final ExtractorFactory extractorFactory = crawlerContainer.getComponent("extractorFactory");
        if (extractorFactory == null) {
            throw new CrawlerSystemException("ExtractorFactory is unavailable.");

        if (in == null) {
            throw new CrawlerSystemException("The inputstream is null.");
        }
        try {
            @SuppressWarnings("resource")
            final PublisherTextExtractor publisherTextExtractor = new PublisherTextExtractor(in);
            return new ExtractData(publisherTextExtractor.getText());
        } catch (final IOException e) {
            throw new ExtractException(e);
        }
    }

    public void addMaxLength(final String mimeType, final long maxLength) {
        if (StringUtil.isBlank(mimeType)) {
            throw new CrawlerSystemException("MIME type is a blank.");
        }
        if (maxLength < 0) {
            throw new CrawlerSystemException("The value of maxLength is invalid.");
        }
        maxLengthMap.put(mimeType, maxLength);
    }

            throws ArtifactMetadataRetrievalException {
        RepositoryMetadata metadata = new ArtifactRepositoryMetadata(request.getArtifact());

        try {
            repositoryMetadataManager.resolve(metadata, request);
        } catch (RepositoryMetadataResolutionException e) {
            throw new ArtifactMetadataRetrievalException(e.getMessage(), e, request.getArtifact());

    public MimeTypeHelperImpl() {
        try {
            mimeTypes = MimeTypesFactory.create(MIME_TYPES_RESOURCE_NAME);
        } catch (final Exception e) {
            throw new CrawlerSystemException("Could not initialize MimeTypeHelper.", e);
        }
    }

    @Override
    public String getContentType(final InputStream is, final String filename) {

    public interface StreamSupplier {
        InputStream open() throws IOException;
    }

    public interface ReaderSupplier {
        Reader open() throws IOException;
    }

    /**
     * @deprecated use {@link #build(ReaderSupplier)}
     */
    @Deprecated
    public PluginDescriptor build(Reader reader) throws PlexusConfigurationException {
        return build(reader, null);
    }

    /**

   * ContiguousSet}.
   *
   * @throws UnsupportedOperationException always
   * @deprecated Use {@link #create}.
   */
  @Deprecated
  @DoNotCall("Always throws UnsupportedOperationException")
  public static <E> ImmutableSortedSet.Builder<E> builder() {
    throw new UnsupportedOperationException();
  }

  // redeclare to help optimizers with b/310253115

    return new Callable<T>() {
      @Override
      public T call() throws Exception {
        throw exception;
      }
    };
  }

  private static Runnable runnableThrowing(final RuntimeException e) {
    return new Runnable() {
      @Override
      public void run() {
        throw e;
      }
    };
  }

  @SuppressWarnings("serial")

    @Override
    public boolean hasNext() {
      if (modCount != expectedModCount) {
        throw new ConcurrentModificationException();
      }
      return next != null && remaining > 0;
    }

    @Override
    public T next() {
      if (!hasNext()) {
        throw new NoSuchElementException();
      }

      // requireNonNull is safe because of the hasNext check.

        if (StringUtil.isBlank(url)) {
            throw new PluginException("url is blank: " + artifact.getName());
        }
        if (url.startsWith("http:") || url.startsWith("https:")) {
            try (final CurlResponse response = createCurlRequest(url).execute()) {
                if (response.getHttpStatusCode() != 200) {