KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > org > archive > crawler > deciderules > MatchesFilePatternDecideRule


1 /*
2  * MatchesFilePatternDecideRule
3  *
4  * $Id: MatchesFilePatternDecideRule.java,v 1.4.16.1 2007/01/13 01:31:14 stack-sf Exp $
5  *
6  * Created on Mar 11, 2004
7  *
8  * Copyright (C) 2004 Internet Archive.
9  *
10  * This file is part of the Heritrix web crawler (crawler.archive.org).
11  *
12  * Heritrix is free software; you can redistribute it and/or modify
13  * it under the terms of the GNU Lesser Public License as published by
14  * the Free Software Foundation; either version 2.1 of the License, or
15  * any later version.
16  *
17  * Heritrix is distributed in the hope that it will be useful,
18  * but WITHOUT ANY WARRANTY; without even the implied warranty of
19  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20  * GNU Lesser Public License for more details.
21  *
22  * You should have received a copy of the GNU Lesser Public License
23  * along with Heritrix; if not, write to the Free Software
24  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
25  */
26
27 package org.archive.crawler.deciderules;
28
29 import java.util.logging.Logger JavaDoc;
30
31 import javax.management.AttributeNotFoundException JavaDoc;
32
33 import org.archive.crawler.settings.SimpleType;
34
35 /**
36  * Compares suffix of a passed CrawlURI, UURI, or String against a regular
37  * expression pattern, applying its configured decision to all matches.
38  *
39  * Several predefined patterns are available for convenience. Choosing
40  * 'custom' makes this the same as a regular MatchesRegExpDecideRule.
41  *
42  * @author Igor Ranitovic
43  */
44 public class MatchesFilePatternDecideRule extends MatchesRegExpDecideRule {
45
46     private static final long serialVersionUID = -4182743018517062411L;
47
48     private static final Logger JavaDoc logger =
49         Logger.getLogger(MatchesFilePatternDecideRule.class.getName());
50     public static final String JavaDoc ATTR_USE_PRESET = "use-preset-pattern";
51     public static final String JavaDoc IMAGES_PATTERNS =
52         ".*(?i)(\\.(bmp|gif|jpe?g|png|tiff?))$";
53     public static final String JavaDoc AUDIO_PATTERNS =
54         ".*(?i)(\\.(mid|mp2|mp3|mp4|wav))$";
55     public static final String JavaDoc VIDEO_PATTERNS =
56         ".*(?i)(\\.(avi|mov|mpeg|ram|rm|smil|wmv))$";
57     public static final String JavaDoc MISC_PATTERNS =
58         ".*(?i)(\\.(doc|pdf|ppt|swf))$";
59     public static final String JavaDoc ALL_DEFAULT_PATTERNS =
60         ".*(?i)(\\.(bmp|gif|jpe?g|png|tiff?|mid|mp2|mp3|mp4|wav|avi|mov|mpeg" +
61         "|ram|rm|smil|wmv|doc|pdf|ppt|swf))$";
62
63     public static final String JavaDoc ALL = "All";
64     public static final String JavaDoc IMAGES = "Images";
65     public static final String JavaDoc AUDIO = "Audio";
66     public static final String JavaDoc VIDEO = "Video";
67     public static final String JavaDoc MISC = "Miscellaneous";
68     public static final String JavaDoc CUSTOM = "Custom";
69
70     /**
71      * Usual constructor.
72      * @param name
73      */
74     public MatchesFilePatternDecideRule(String JavaDoc name) {
75         super(name);
76         setDescription("MatchesFilePatternDecideRule. Applies its decision " + "to all URIs that end with the specified pattern(s). Anything " +
77             " that does not match is let PASS. " +
78             " Default file patterns are: .avi, .bmp, " +
79             ".doc, .gif, .jp(e)g, .mid, .mov, .mp2, .mp3, .mp4, .mpeg, " +
80             ".pdf, .png, .ppt, .ram, .rm,.smil, .swf, .tif(f), .wav, .wmv. " +
81             "It is also possible to specify a custom regular expression, " +
82             "in which case this behaves exactly like the " +
83             " MatchesRegExpDecideRule. See also " +
84             "NotMatchesFilePatternDecideRule.");
85
86         String JavaDoc[] options = new String JavaDoc[] {ALL, IMAGES, AUDIO, VIDEO, MISC,
87             CUSTOM};
88
89         addElementToDefinition(
90             new SimpleType(ATTR_USE_PRESET, "URIs that match selected file " +
91                 "patterns will have the decision applied. Default file " +
92                 "patterns are:\n" +
93                 "Images: .bmp, .gif, .jp(e)g, .png, .tif(f)\n" +
94                 "Audio: .mid, mp2, .mp3, .mp4, .wav\n" +
95                 "Video: .avi, .mov, .mpeg, .ram, .rm, .smil, .wmv\n" +
96                 "Miscellaneous: .doc, .pdf, .ppt, .swf\n" +
97                 "All: All above patterns\n" +
98                 "Choose 'Custom' to specify your own pattern. Preset " +
99                 "patterns are case insensitive.",
100                 "All", options));
101
102         addElementToDefinition(
103             new SimpleType(ATTR_REGEXP, "Custom java regular expression. " +
104                     "This regular expression will be used instead of the " +
105                     "supplied pattern groups for matching. An example " +
106                     "of such a regular expression (Miscellaneous): " +
107                     ".*(?i)(\\.(doc|pdf|ppt|swf))$ " +
108                     "Any arbitrary regular expression may be entered and " +
109                     "will be applied to the URI.", ""));
110     }
111
112     /**
113      * Use a preset if configured to do so.
114      * @param o Context
115      * @return Regex to use.
116      *
117      * @see org.archive.crawler.filter.URIRegExpFilter#getRegexp(Object)
118      */
119     protected String JavaDoc getRegexp(Object JavaDoc o) {
120         try {
121             String JavaDoc patternType = (String JavaDoc) getAttribute(o, ATTR_USE_PRESET);
122             if (patternType.equals(ALL)) {
123                 return ALL_DEFAULT_PATTERNS;
124             } else if (patternType.equals(IMAGES)) {
125                 return IMAGES_PATTERNS;
126             } else if (patternType.equals(AUDIO)) {
127                 return AUDIO_PATTERNS;
128             } else if (patternType.equals(VIDEO)) {
129                 return VIDEO_PATTERNS;
130             } else if (patternType.equals(MISC)) {
131                 return MISC_PATTERNS;
132             } else if (patternType.equals(CUSTOM)) {
133                 return super.getRegexp(o);
134             } else {
135                 assert false : "Unrecognized pattern type " + patternType
136                         + ". Should never happen!";
137             }
138         } catch (AttributeNotFoundException JavaDoc e) {
139             logger.severe(e.getMessage());
140         }
141         return null; // Basically the rule is inactive if this occurs.
142 }
143 }
144
Popular Tags