// BSD License (http://www.galagosearch.org/license)

 package org.galagosearch.core.parse;

 import java.io.IOException;

 import java.util.ArrayList;

 import java.util.HashMap;

 import org.galagosearch.tupleflow.InputClass;

 import org.galagosearch.tupleflow.OutputClass;

 import org.galagosearch.tupleflow.StandardStep;

 import org.galagosearch.tupleflow.execution.Verified;

 import org.galagosearch.core.types.IdentifiedLink;

 /**

  * From an IdentifiedLink object, this class constructs a document containing

  * only anchor text.

  * 

  * @author trevor

  */

 @Verified

 @InputClass(className = "org.galagosearch.core.types.IdentifiedLink")

 @OutputClass(className = "org.galagosearch.core.parse.Document")

 public class AnchorTextDocumentCreator extends StandardStep<IdentifiedLink, Document> {

     TagTokenizer tokenizer = new TagTokenizer();

     ArrayList<IdentifiedLink> links = new ArrayList<IdentifiedLink>();

     String lastDocument = null;

     /**

      * This method takes the text from a link object, tokenizes it,

      * then adds it to a document object.

      */

     public void process(IdentifiedLink link) throws IOException {

         if (lastDocument != null && !lastDocument.equals(link.identifier)) {

             flush();

         }

         links.add(link);

         lastDocument = link.identifier;

     }

     public void flush() throws IOException {

         Document document = new Document();

         if (links.size() == 0) {

             return;

         } else if (links.size() == 1) {

             IdentifiedLink link = links.get(0);

             document.text = link.anchorText;

             document.identifier = link.identifier;

         } else {

             StringBuilder builder = new StringBuilder();

             for (IdentifiedLink link : links) {

                 builder.append(link.anchorText);

                 builder.append(' ');

             }

             document.text = builder.toString();

             document.identifier = links.get(0).identifier;

         }

         document.terms = null;

         document.metadata = new HashMap<String, String>();

         document.tags = new ArrayList<Tag>();

         // parse the text into pieces

         tokenizer.process(document);

         // send it on to the next stage

         processor.process(document);

         lastDocument = null;

         links.clear();

     }

     @Override

     public void close() throws IOException {

         flush();

         super.close();

     }

     public Class<IdentifiedLink> getInputClass() {

         return IdentifiedLink.class;

     }

     public Class<Document> getOutputClass() {

         return Document.class;

     }

 }