View Javadoc
1   /**
2    * 
3    */
4   package org.apache.any23.plugin.htmlscraper;
5   
6   import java.util.Arrays;
7   
8   import org.apache.any23.extractor.ExtractorDescription;
9   import org.apache.any23.extractor.ExtractorFactory;
10  import org.apache.any23.extractor.SimpleExtractorFactory;
11  import org.apache.any23.rdf.Prefixes;
12  
13  /**
14   * @author Peter Ansell p_ansell@yahoo.com
15   *
16   */
17  public class HTMLScraperExtractorFactory extends SimpleExtractorFactory<HTMLScraperExtractor> implements
18          ExtractorFactory<HTMLScraperExtractor> {
19  
20      public static final String NAME = "html-scraper";
21      
22      public static final Prefixes PREFIXES = null;
23  
24      private static final ExtractorDescription descriptionInstance = new HTMLScraperExtractorFactory();
25      
26      public HTMLScraperExtractorFactory() {
27          super(
28                  HTMLScraperExtractorFactory.NAME, 
29                  HTMLScraperExtractorFactory.PREFIXES,
30                  Arrays.asList("text/html;q=0.02", "application/xhtml+xml;q=0.02"),
31                  null);
32      }
33      
34      @Override
35      public HTMLScraperExtractor createExtractor() {
36          return new HTMLScraperExtractor();
37      }
38  
39      public static ExtractorDescription getDescriptionInstance() {
40          return descriptionInstance;
41      }
42  }