HotBot


1   /*
2    * WebSphinx web-crawling toolkit
3    *
4    * Copyright (c) 1998-2002 Carnegie Mellon University.  All rights
5    * reserved.
6    *
7    * Redistribution and use in source and binary forms, with or without
8    * modification, are permitted provided that the following conditions
9    * are met:
10   *
11   * 1. Redistributions of source code must retain the above copyright
12   *    notice, this list of conditions and the following disclaimer.
13   *
14   * 2. Redistributions in binary form must reproduce the above copyright
15   *    notice, this list of conditions and the following disclaimer in
16   *    the documentation and/or other materials provided with the
17   *    distribution.
18   *
19   * THIS SOFTWARE IS PROVIDED BY CARNEGIE MELLON UNIVERSITY ``AS IS'' AND
20   * ANY EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
21   * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
22   * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL CARNEGIE MELLON UNIVERSITY
23   * NOR ITS EMPLOYEES BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24   * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25   * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26   * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27   * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28   * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29   * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30   *
31   */
32  
33  package websphinx.searchengine;
34  
35  import websphinx.*;
36  import java.net.URL  ;
37  import java.net.URLEncoder  ;
38  import java.net.MalformedURLException  ;
39  
40  /**
41   * <A HREF="http://www.hotbot.com/">HotBot</a> search engine.
42   */
43  public class HotBot implements SearchEngine {
44  
45      static Pattern patTitle = new Regexp ("^");
46  
47      static Pattern patCount = new Regexp (
48          "Returned <B>(\\d+)</b> matches"
49      );
50      static Pattern patNoHits = new Regexp (
51          "Sorry -- your search yielded no results"
52      );
53  
54      // FIX: works only for Netscape
55      static Pattern patResult = new Tagexp (
56        "<td><b>(?{rank})</b></td>" // rank
57      + "<td>(?:<a><img></a>)?"     // optional icon
58      + "(?{link}<a>(?{title})</a>)</td>" // title and main link
59      + "</tr><tr><td><font>(?{score})</font></td><td>(?{description})<br></td>" // description
60      );
61  
62      static Pattern patMoreLink = new Tagexp (
63          "<input type=image name=act.next>"
64      );
65  
66      /**
67       * Classify a page.  Sets the following labels:
68       * <TABLE>
69       * <TR><TH>Name <TH>Type  <TH>Meaning
70       * <TR><TD>searchengine.source <TD>Page label <TD>HotBot object that labeled this page
71       * <TR><TD>searchengine.count <TD>Page field <TD>Number of results on page
72       * <TR><TD>searchengine.results <TD>Page fields <TD>Array of results.  Each result region
73       * contains subfields: rank, title, description, and link.
74       * <TR><TD>searchengine.more-results <TD>Link label <TD>Link to a page containing more results.
75       * </TABLE>
76       */
77      public void classify (Page page) {
78          String   title = page.getTitle ();
79          if (title != null && title.startsWith ("HotBot results:")) {
80              page.setObjectLabel ("searchengine.source", this);
81  
82              Region count = patCount.oneMatch (page);
83              if (count != null)
84                  page.setField ("searchengine.count", count.getField ("0"));
85              
86              Region[] results = patResult.allMatches (page);
87              //System.err.println ("found " + results.length + " results");
88              SearchEngineResult[] ser = new SearchEngineResult[results.length];
89              for (int i=0; i<results.length; ++i) {
90                  ser[i] = new SearchEngineResult (results[i]);
91                  //System.out.println (ser[i]);
92              }
93              page.setFields ("searchengine.results", ser);
94  
95              PatternMatcher m = patMoreLink.match (page);
96              while (m.hasMoreElements ()) {
97                  Link link = (Link)m.nextMatch ();
98                  link.setLabel ("searchengine.more-results");
99                  link.setLabel ("hyperlink");             
100             }
101         }
102         else System.err.println ("not a HotBot page");
103 
104     }
105 
106     /**
107      * Priority of this classifier.
108      */
109     public static final float priority = 0.0F;
110     
111     /**
112      * Get priority of this classifier.
113      * @return priority.
114      */
115     public float getPriority () {
116         return priority;
117     }
118 
119     /**
120      * Make a query URL for HotBot.
121      * @param keywords list of keywords, separated by spaces
122      * @return URL that submits the keywords to HotBot.
123      */
124     public URL   makeQuery (String   keywords) {
125         try {
126             return new URL  ("http://www.search.hotbot.com/hResult.html/?SM=MC&MT="
127                            + URLEncoder.encode(keywords)
128                            + "&DV=7&RG=.com&DC=10&DE=2&OPs=MDRTP&_v=2&DU=days&SW=web");
129         } catch (MalformedURLException   e) {
130             throw new RuntimeException   ("internal error");
131         }
132     }
133 
134     /**
135      * Get number of results per page for this search engine.
136      * @return typical number of results per page
137      */
138     public int getResultsPerPage () {
139         return 10;
140     }
141 
142     /**
143      * Search HotBot.
144      * @param keywords list of keywords, separated by spaces
145      * @return enumeration of SearchEngineResults returned by a HotBot query constructed from the keywords.
146      */
147     public static Search search (String   keywords) {
148         return new Search (new HotBot(), keywords);
149     }
150 
151     /**
152      * Search HotBot.
153      * @param keywords list of keywords, separated by spaces
154      * @param maxResults maximum number of results to return
155      * @return enumeration of SearchEngineResults returned by an HotBot query constructed from the keywords.
156      * The enumeration yields at most maxResults objects.
157      */
158     public static Search search (String   keywords, int maxResults) {
159         return new Search (new HotBot(), keywords, maxResults);
160     }
161 } 
162
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags