highlight.java

关于Ultraseek的一些用法,刚初学,所以都是比较简单

/* -*- mode:java; indent-tabs-mode:nil; c-basic-offset:2 -*-
 *  $RCSFile$ $Revision: 1.10 $ $Date: 2006/02/01 00:20:30 $
 *  Copyright (c) 2003 Autonomy Corp.  All Rights Reserved.
 */

import java.io.*;
import java.util.*;

import com.ultraseek.xpa.search.*;
import com.ultraseek.xpa.server.*;


/**
 * Highlighting of Titles, Description, and documents.
 * <p>
 * This application demonstrates how to use Ultraseek to
 * highlight search terms appearing in a title or description.
 * <p>
 * It also demonstrates which documents can be highlighted
 * by the Ultraseek server.
 * <p>
 * To use this application, edit the file "Sample.properties" 
 * to point to your Ultraseek server.
 * 
 * @see UltraseekSearchResult#getTitleForHighlight
 * @see UltraseekSearchResult#getDescriptionForHighlight
 * @see UltraseekSearchResult#isDocumentHighlightable
 * @see UltraseekSearchResult#getMimeType
 * @since XPA2.1
 * @requires Ultraseek 5.1.1
 */
public class Highlight {

  /**
   * Loads parameters from the file Sample.properties
   */
  static String getString(String key, String def) {
    try {
      String resourceName = System.getProperty("Sample");
      if (resourceName==null) resourceName = "Sample";
      ResourceBundle settings = ResourceBundle.getBundle( resourceName );
      return settings.getString(key);
    } catch (MissingResourceException e) {
      return def;
    } catch (Exception e) {
      System.out.println( "Property file problem: "  + e );
      return def;
    }
  }

  static final String serverName  = getString("UltraseekServer.host", 
                                              "software-demo.ultraseek.com");
  static final int    serverPort  = Integer.valueOf(getString("UltraseekServer.port", 
                                                              "80")).intValue();
  static final String serverProto = getString("UltraseekServer.protocol", 
                                              "http");
  static UltraseekServer server = new UltraseekServer(serverName, serverPort, serverProto);

  /**
   * Highlight terms with * for display at the terminal.
   */
  static String highlightResult(String split[]) {
    if (split==null) return "";
    StringBuffer sb = new StringBuffer();
    for (int i = 0; i < split.length; i++) {
      if (i % 2 == 1) sb.append( "***" );
      sb.append(split[i]);
      if (i % 2 == 1) sb.append( "***" );
    }
    return sb.toString();
  }

  /**
   * HTMLQuotes a String.
   * @return the HTML quoted string.
   * @param s the String to quote.
   */
  public static String htmlquote(String s) {
    StringBuffer sb = new StringBuffer(s.length() * 2);
    for (int i = 0; i < s.length(); i++) {
      char ch = s.charAt(i);
      if (ch == '&') sb.append("&amp;");
      else if (ch == '"') sb.append("&quot;");
      else if (ch == '<') sb.append("&lt;");
      else if (ch == '>') sb.append("&gt;");
      else if (ch > '\u00FF') {
        sb.append( "&#" );
        sb.append( Integer.toHexString(ch) );
        sb.append( ";" );
      } else
        sb.append( ch );
    }
    return sb.toString();
  }

  public static String HIGHLIGHT_START = "<b>";
  public static String HIGHLIGHT_END   = "</b>";
  public static String PASSAGE_SEP     = "&nbsp;...&nbsp;";

  /**
   * Prepares HTML for a search result's description.
   * <p>
   * The DESCRIPTION_SEPARATOR (Unicode Horizontal Ellipsis) does not
   * always display very well in a web page.  This function remaps it to
   * three ASCII dots.
   * <p>
   * The result of this function is HTML quoted.
   * @param text the description of a search result.
   * @return the HTML-Quoted and prepared description.
   */
  public static String quoteDescription(String text) {
    /* find each passage, replace the separator with ... */
    StringBuffer sb = new StringBuffer(text.length() * 2);
    while (text.length() > 0) {
        int next_separator = text.indexOf(UltraseekSearchResult.DESCRIPTION_SEPARATOR);
        if (next_separator == -1) {
          sb.append( htmlquote(text) );
          break;
        } else {
          String before = text.substring(0,next_separator);
          sb.append( htmlquote(before) );
          sb.append( PASSAGE_SEP );
          text = text.substring(next_separator+1);
        }
      }
    return sb.toString();
  }

  /**
   * Prepares HTML for a search result's description.
   * <p>
   * The DESCRIPTION_SEPARATOR (Unicode Horizontal Ellipsis) does not
   * always display very well in a web page.  This function remaps it to
   * three ASCII dots.
   * <p>
   * Highlighting of search terms uses HIGHLIGHT_START and 
   * HIGHLIGHT_END around portions to be highlighted.
   * <p>
   * The result of this function is HTML quoted.
   * @param sr The UltraseekSearchResult to prepare a description.
   * @return the HTML-Quoted and prepared description.
   * @exception IOException There was a problem fetching the description.
   */
  public static String makeDescription(UltraseekSearchResult sr) 
    throws IOException {
    String parts[] = sr.getDescriptionForHighlight();
    StringBuffer sb = new StringBuffer(1024);
    for (int i = 0; i < parts.length; i++) {
      if (i % 2 == 1) {
        /* Odd numbered array element - highlight this part */
        sb.append( HIGHLIGHT_START );
        sb.append( quoteDescription( parts[i] ));
        sb.append( HIGHLIGHT_END );
      } else {
        /* Even numbered array element - no highlighting */
        sb.append( quoteDescription( parts[i] ));
      }
    }
    return sb.toString();
  }


  /**
   * Simple display of a search result.
   */
  static void showResult(SearchResult sr) 
    throws IOException {
    System.out.println("---------------------------------");
    if (sr instanceof UltraseekSearchResult) {
      UltraseekSearchResult usr = (UltraseekSearchResult) sr;
      System.out.print("Title             : ");
      System.out.println(highlightResult(usr.getTitleForHighlight()));
      String description = makeDescription(usr);
      System.out.print("Description       : ");
      System.out.println( description );
      System.out.print("URL               : ");
      System.out.println(usr.getURLString());
      if (usr.getMimeType()==null)
        System.out.println("Document MIME type: unknown");
      else
        System.out.println("Document MIME type: " + usr.getMimeType());
      if (usr.isDocumentHighlightable()) {
        System.out.println("This document can be highlighted by Ultraseek.");
        System.out.print("Terms: ");
        Iterator it = usr.getTermTFs().keySet().iterator();
        while (it.hasNext()) {
          String term = (String) it.next();
          System.out.print( "'" + term + "' " );
        }
        System.out.println();
      } else
        System.out.println("This document cannot be highlighted by Ultraseek.");
    } else {
      System.out.println(sr.getTitle());
      System.out.println(sr.getDescription());
      System.out.println(sr.getURLString());
      System.out.println("Not an UltraseekSearchResult.");
    }
    System.out.println();
  }

  static void demoHighlight(String query)
    throws IOException {
    Query q = Query.parse(query);
    final int SHOWCOUNT = 4;

    try {
      SearchResultList srl = server.search(q);

      // Show summary of search results.
      System.out.println();
      System.out.println( "" + srl.getDocCount() + " documents were searched.");
      System.out.println( "" + srl.getResultCount() + " document(s) match your query.");
      System.out.println( "" + srl.size() + " search result(s) have been prioritized.");
      System.out.println( "" + Math.min(SHOWCOUNT,srl.size()) + 
                          " search result(s) will be displayed." );
      Iterator it = srl.getTermDFs().keySet().iterator();
      while (it.hasNext()) {
        String term = (String) it.next();
        Integer count = (Integer) srl.getTermDFs().get(term);
        System.out.println( term + ": " + count.intValue() + " documents." );
      }
      
      int count = 0;
      try {
        // Only show a few results for this demo.
        while (count < SHOWCOUNT) {
          SearchResult sr = (SearchResult) srl.get(count);
          showResult(sr);
          count++;
        }
      } catch (IndexOutOfBoundsException e) {
        // Normal end of list - not an error.
      }
    } catch (QueryNotSupportedException e) {
      System.out.println( "Query problem: " + e );
    }
  }

  static public void main(String args[]) 
    throws IOException {
    System.out.println("Highlight Demo");
    System.out.println("Using Ultraseek at " + server );
    System.out.println(" (version " + server.getVersionString() +")" );
    if (!server.versionAtLeast(5,1,1)) {
      System.out.println("Ultraseek version 5.1.1 (or newer) is required for this demo." );
      return;
    }

    InputStreamReader inputStreamReader = new InputStreamReader(System.in);
    BufferedReader bufferedReader = new BufferedReader(inputStreamReader);
    while (true) {
      System.out.print("\n\nEnter a query (EOF to end): ");
      String line = bufferedReader.readLine();
      if (line==null) break;
      try {
        demoHighlight( line );
      } catch (IOException e) {
        System.out.println( "Problem communicating with server: " + e );
      } catch (XPARuntimeException e) {
        System.out.println( "Problem communicating with server: " + e.getCause() );
      }
    }
  }
}