source: main/trunk/greenstone3/src/java/org/greenstone/gsdl3/service/AbstractGS2DocumentRetrieve.java@ 26198

Last change on this file since 26198 was 26198, checked in by kjdon, 12 years ago

getMetadataList needs a language param, as we may need to look up a dictionary for macro resolving. and set a class loader for the macro resolver so we can have collection specific dictionaries
  • Property svn:keywords set to Author Date Id Revision
File size: 9.5 KB
Line 
1/*
2 * AbstractGS2DocumentRetrieve.java
3 * Copyright (C) 2002 New Zealand Digital Library, http://www.nzdl.org
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program 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 General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
18 */
19package org.greenstone.gsdl3.service;
20
21// Greenstone classes
22import org.greenstone.gsdl3.core.GSException;
23import org.greenstone.gsdl3.util.BasicDocumentDatabase;
24import org.greenstone.gsdl3.util.GSXML;
25import org.greenstone.gsdl3.util.GSFile;
26import org.greenstone.gsdl3.util.OID;
27import org.greenstone.gsdl3.util.MacroResolver;
28import org.greenstone.gsdl3.util.GS2MacroResolver;
29import org.greenstone.gsdl3.util.GSConstants;
30import org.greenstone.gsdl3.util.SimpleCollectionDatabase;
31import org.greenstone.gsdl3.util.DBInfo;
32// XML classes
33import org.w3c.dom.Document;
34import org.w3c.dom.Element;
35import org.w3c.dom.NodeList;
36
37// General Java classes
38import java.io.File;
39import java.util.StringTokenizer;
40import java.util.Vector;
41import java.util.Set;
42import java.util.Iterator;
43import java.util.ArrayList;
44
45import org.apache.log4j.*;
46
47// Apache Commons
48import org.apache.commons.lang3.*;
49
50/**
51 * Implements the generic retrieval and classifier services for GS2 collections.
52 *
53 * @author Katherine Don
54 * @author Michael Dewsnip
55 */
56
57public abstract class AbstractGS2DocumentRetrieve extends AbstractDocumentRetrieve
58{
59
60 static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.AbstractGS2DocumentRetrieve.class.getName());
61
62 // protected static final String EXTLINK_PARAM = "ext"; here or in base??
63 protected String index_stem = null;
64
65 protected SimpleCollectionDatabase coll_db = null;
66 BasicDocumentDatabase gs_doc_db = null;
67 /** constructor */
68 protected AbstractGS2DocumentRetrieve()
69 {
70 this.macro_resolver = new GS2MacroResolver();
71 }
72
73 public void cleanUp()
74 {
75 super.cleanUp();
76 this.coll_db.closeDatabase();
77 this.gs_doc_db.cleanUp();
78 }
79
80 /** configure this service */
81 public boolean configure(Element info, Element extra_info)
82 {
83 if (!super.configure(info, extra_info))
84 {
85 return false;
86 }
87
88 logger.info("Configuring AbstractGS2DocumentRetrieve...");
89 //this.config_info = info;
90
91 // the index stem is either specified in the config file or is the collection name
92 Element index_stem_elem = (Element) GSXML.getChildByTagName(info, GSXML.INDEX_STEM_ELEM);
93 if (index_stem_elem != null)
94 {
95 this.index_stem = index_stem_elem.getAttribute(GSXML.NAME_ATT);
96 }
97 if (this.index_stem == null || this.index_stem.equals(""))
98 {
99 logger.error("AbstractGS2DocumentRetrieve.configure(): indexStem element not found, stem will default to collection name");
100 this.index_stem = this.cluster_name;
101 }
102
103 // find out what kind of database we have
104 Element database_type_elem = (Element) GSXML.getChildByTagName(info, GSXML.DATABASE_TYPE_ELEM);
105 String database_type = null;
106 if (database_type_elem != null)
107 {
108 database_type = database_type_elem.getAttribute(GSXML.NAME_ATT);
109 }
110 if (database_type == null || database_type.equals(""))
111 {
112 database_type = "gdbm"; // the default
113 }
114 coll_db = new SimpleCollectionDatabase(database_type);
115 if (!coll_db.databaseOK())
116 {
117 logger.error("Couldn't create the collection database of type " + database_type);
118 return false;
119 }
120
121 // Open database for querying
122 String coll_db_file = GSFile.collectionDatabaseFile(this.site_home, this.cluster_name, this.index_stem, database_type);
123 if (!this.coll_db.openDatabase(coll_db_file, SimpleCollectionDatabase.READ))
124 {
125 logger.error("Could not open collection database!");
126 return false;
127 }
128
129 gs_doc_db = new BasicDocumentDatabase(this.doc, database_type, this.site_home, this.cluster_name, this.index_stem);
130 if (!gs_doc_db.isValid())
131 {
132 logger.error("Failed to open Document Database.");
133 return false;
134 }
135 this.gs_doc = gs_doc_db;
136
137 // we need to set the database for our GS2 macro resolver
138 GS2MacroResolver gs2_macro_resolver = (GS2MacroResolver) this.macro_resolver;
139 gs2_macro_resolver.setDB(this.coll_db);
140 // set the class loader in case we have collection specific properties files
141 gs2_macro_resolver.setClassLoader(this.class_loader);
142 return true;
143 }
144
145 /** if id ends in .fc, .pc etc, then translate it to the correct id */
146 protected String translateId(String node_id)
147 {
148 return OID.translateOID(this.coll_db, node_id); //return this.coll_db.translateOID(node_id);
149 }
150
151 /**
152 * if an id is not a greenstone id (an external id) then translate it to a
153 * greenstone one
154 */
155 protected String translateExternalId(String node_id)
156 {
157 return this.coll_db.externalId2OID(node_id);
158 }
159
160 /**
161 * returns the id of the root node of the document containing node node_id.
162 * . may be the same as node_id
163 */
164 protected String getRootId(String node_id)
165 {
166 return this.gs_doc.getRootId(node_id);
167 }
168
169
170
171 /**
172 * get the metadata for the classifier node node_id returns a metadataList
173 * element: <metadataList><metadata
174 * name="xxx">value</metadata></metadataList>
175 */
176 protected Element getMetadataList(String node_id, boolean all_metadata, ArrayList<String> metadata_names, String lang) throws GSException
177 {
178 Element metadata_list = this.doc.createElement(GSXML.METADATA_ELEM + GSXML.LIST_MODIFIER);
179 DBInfo info = this.coll_db.getInfo(node_id);
180 if (info == null)
181 {
182 return null;
183 }
184
185 if (all_metadata) // this will get all metadata for current node
186 {
187 // return everything out of the database
188 Set<String> keys = info.getKeys();
189 Iterator<String> it = keys.iterator();
190 while (it.hasNext())
191 {
192 String key = it.next();
193 //String value = info.getInfo(key);
194 Vector<String> values = info.getMultiInfo(key);
195 for (int i = 0; i < values.size(); i++)
196 {
197 GSXML.addMetadata(this.doc, metadata_list, key, this.macro_resolver.resolve(values.elementAt(i), lang, MacroResolver.SCOPE_META, node_id));
198 }
199 }
200
201 }
202 // now we go through the list of names. If we have specified
203 // all_metadata, then here we only get the ones like
204 // parent_Title, that are not the current node.
205 for (int i = 0; i < metadata_names.size(); i++)
206 {
207 String meta_name = metadata_names.get(i);
208
209 if (!all_metadata || meta_name.indexOf(GSConstants.META_RELATION_SEP)!=-1) {
210 Vector <String> values = getMetadata(node_id, info, meta_name, lang);
211 if (values != null) {
212 for (int j = 0; j < values.size(); j++)
213 {
214 // some of these may be parent/ancestor. does resolve need a different id???
215 GSXML.addMetadata(this.doc, metadata_list, meta_name, this.macro_resolver.resolve(values.elementAt(j), lang, MacroResolver.SCOPE_META, node_id));
216 }
217 }
218 }
219 }
220
221 return metadata_list;
222 }
223
224 protected Vector<String> getMetadata(String node_id, DBInfo info, String metadata, String lang) {
225
226 DBInfo current_info = info;
227
228 int index = metadata.indexOf(GSConstants.META_RELATION_SEP);
229 if (index == -1) {
230 // metadata is for this node
231 Vector<String> values = info.getMultiInfo(metadata);
232 return values;
233 }
234 // we need to get metadata for one or more different nodes
235 String relation = metadata.substring(0, index);
236 String relation_id="";
237 metadata = metadata.substring(index + 1);
238 if (relation.equals("parent") || relation.equals("ancestors")) {
239 relation_id = OID.getParent(node_id);
240 if (relation_id.equals(node_id)) {
241 return null;
242 }
243 } else if (relation.equals("root")) {
244 relation_id = OID.getTop(node_id);
245 }
246
247 DBInfo relation_info;
248 if (relation_id.equals(node_id)) {
249 relation_info = info;
250 } else {
251 relation_info = this.coll_db.getInfo(relation_id);
252 }
253 if (relation_info == null)
254 {
255 return null;
256 }
257
258 Vector<String> values = relation_info.getMultiInfo(metadata);
259 // do resolving
260 if (!relation.equals("ancestors")){
261 return values;
262 }
263
264 // ancestors: go up the chain
265
266 String current_id = relation_id;
267 relation_id = OID.getParent(current_id);
268 while (!relation_id.equals(current_id))
269 {
270 relation_info = this.coll_db.getInfo(relation_id);
271 if (relation_info == null)
272 return values;
273
274 Vector<String> more_values = relation_info.getMultiInfo(metadata);
275 if (more_values != null)
276 {
277 values.addAll(0, more_values);
278 }
279
280
281 current_id = relation_id;
282 relation_id = OID.getParent(current_id);
283 }
284 return values; // for now
285 }
286
287 protected int getNumChildren(String node_id)
288 {
289 return this.gs_doc.getNumChildren(node_id);
290 }
291
292
293 /**
294 * returns the content of a node should return a nodeContent element:
295 * <nodeContent>text content or other elements</nodeContent>
296 */
297 abstract protected Element getNodeContent(String doc_id, String lang) throws GSException;
298
299
300 /**
301 * needs to get info from collection database - if the calling code gets it
302 * already it may pay to pass it in instead
303 */
304 protected String resolveTextMacros(String doc_content, String doc_id, String lang)
305 {
306 // resolve any collection specific macros
307 doc_content = macro_resolver.resolve(doc_content, lang, MacroResolver.SCOPE_TEXT, doc_id);
308 return doc_content;
309 }
310
311
312
313}
Note: See TracBrowser for help on using the repository browser.