KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > org > archive > crawler > datamodel > CoreAttributeConstants


1 /* Copyright (C) 2003 Internet Archive.
2  *
3  * This file is part of the Heritrix web crawler (crawler.archive.org).
4  *
5  * Heritrix is free software; you can redistribute it and/or modify
6  * it under the terms of the GNU Lesser Public License as published by
7  * the Free Software Foundation; either version 2.1 of the License, or
8  * any later version.
9  *
10  * Heritrix is distributed in the hope that it will be useful,
11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13  * GNU Lesser Public License for more details.
14  *
15  * You should have received a copy of the GNU Lesser Public License
16  * along with Heritrix; if not, write to the Free Software
17  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
18  *
19  * CoreAttributeConstants.java
20  * Created on Jun 17, 2003
21  *
22  * $Header: /cvsroot/archive-crawler/ArchiveOpenCrawler/src//**
23  * @author gojomo
24  *
25  */
26 package org.archive.crawler.datamodel;
27
28 /**
29  * CrawlURI attribute keys used by the core crawler
30  * classes.
31  *
32  * @author gojomo
33  *
34  */
35 public interface CoreAttributeConstants {
36
37     /**
38      * Extracted MIME type of fetched content; should be
39      * set immediately by fetching module if possible
40      * (rather than waiting for a later analyzer)
41      */
42     public static String JavaDoc A_CONTENT_TYPE = "content-type";
43
44     /**
45      * Multiplier of last fetch duration to wait before
46      * fetching another item of the same class (eg host)
47      */
48     public static String JavaDoc A_DELAY_FACTOR = "delay-factor";
49     /**
50      * Minimum delay before fetching another item of th
51      * same class (eg host). Even if lastFetchTime*delayFactor
52      * is less than this, this period will be waited.
53      */
54     public static String JavaDoc A_MINIMUM_DELAY = "minimum-delay";
55
56     public static String JavaDoc A_RRECORD_SET_LABEL = "dns-records";
57     public static String JavaDoc A_DNS_FETCH_TIME = "dns-fetch-time";
58     public static String JavaDoc A_DNS_SERVER_IP_LABEL = "dns-server-ip";
59     public static String JavaDoc A_FETCH_BEGAN_TIME= "fetch-began-time";
60     public static String JavaDoc A_FETCH_COMPLETED_TIME = "fetch-completed-time";
61     public static String JavaDoc A_HTTP_TRANSACTION = "http-transaction";
62
63     public static String JavaDoc A_RUNTIME_EXCEPTION = "runtime-exception";
64     public static String JavaDoc A_LOCALIZED_ERRORS = "localized-errors";
65
66     /** shorthand string tokens indicating notable occurences,
67      * separated by commas */
68     public static String JavaDoc A_ANNOTATIONS = "annotations";
69
70     public static String JavaDoc A_PREREQUISITE_URI = "prerequisite-uri";
71     public static String JavaDoc A_DISTANCE_FROM_SEED = "distance-from-seed";
72     public static String JavaDoc A_HTML_BASE = "html-base-href";
73     public static String JavaDoc A_RETRY_DELAY = "retry-delay";
74
75     public static String JavaDoc A_META_ROBOTS = "meta-robots";
76     /**
77      * Define for org.archive.crawler.writer.MirrorWriterProcessor.
78      */
79     public static String JavaDoc A_MIRROR_PATH = "mirror-path";
80
81     /**
82      * Key to get credential avatars from A_LIST.
83      */
84     public static final String JavaDoc A_CREDENTIAL_AVATARS_KEY =
85         "credential-avatars";
86     
87     /** a 'source' (usu. URI) that's inherited by discovered URIs */
88     public static String JavaDoc A_SOURCE_TAG = "source";
89     
90     /**
91      * Key to (optional) attribute specifying a list of keys that
92      * are passed to CandidateURIs that 'descend' (are discovered
93      * via) this URI.
94      */
95     public static final String JavaDoc A_HERITABLE_KEYS = "heritable";
96     
97     /** flag indicating the containing queue should be retired */
98     public static final String JavaDoc A_FORCE_RETIRE = "force-retire";
99     
100     /** local override of proxy host */
101     public static final String JavaDoc A_HTTP_PROXY_HOST = "http-proxy-host";
102     /** local override of proxy port */
103     public static final String JavaDoc A_HTTP_PROXY_PORT = "http-proxy-port";
104
105     /**
106      * Fetch truncation codes present in {@link CrawlURI} annotations.
107      * All truncation annotations have a <code>TRUNC_SUFFIX</code> suffix (TODO:
108      * Make for-sure unique or redo truncation so definitive flag marked
109      * against {@link CrawlURI}).
110      */
111     public static final String JavaDoc TRUNC_SUFFIX = "Trunc";
112     // headerTrunc
113 public static final String JavaDoc HEADER_TRUNC = "header" + TRUNC_SUFFIX;
114     // timeTrunc
115 public static final String JavaDoc TIMER_TRUNC = "time" + TRUNC_SUFFIX;
116     // lenTrunc
117 public static final String JavaDoc LENGTH_TRUNC = "len" + TRUNC_SUFFIX;
118 }
119
Popular Tags