1 | /*
|
---|
2 | * GS2Browse.java
|
---|
3 | * Copyright (C) 2005 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 | */
|
---|
19 | package org.greenstone.gsdl3.service;
|
---|
20 |
|
---|
21 | // Greenstone classes
|
---|
22 | import org.greenstone.gsdl3.util.OID;
|
---|
23 | import org.greenstone.gsdl3.util.GSXML;
|
---|
24 | import org.greenstone.gsdl3.util.GSFile;
|
---|
25 | import org.greenstone.gsdl3.util.MacroResolver;
|
---|
26 | import org.greenstone.gsdl3.util.GS2MacroResolver;
|
---|
27 | import org.greenstone.gsdl3.util.SimpleCollectionDatabase;
|
---|
28 | import org.greenstone.gsdl3.util.DBInfo;
|
---|
29 | // XML classes
|
---|
30 | import org.w3c.dom.Document;
|
---|
31 | import org.w3c.dom.Element;
|
---|
32 | import org.w3c.dom.NodeList;
|
---|
33 |
|
---|
34 | // General Java classes
|
---|
35 | import java.util.ArrayList;
|
---|
36 | import java.util.StringTokenizer;
|
---|
37 | import java.util.Set;
|
---|
38 | import java.util.Iterator;
|
---|
39 |
|
---|
40 | import org.apache.log4j.*;
|
---|
41 |
|
---|
42 | /** Greenstone 2 collection classifier service
|
---|
43 | *
|
---|
44 | * @author <a href="mailto:[email protected]">Katherine Don</a>
|
---|
45 | * @author <a href="mailto:[email protected]">Michael Dewsnip</a>
|
---|
46 | */
|
---|
47 |
|
---|
48 | public class GS2Browse
|
---|
49 | extends AbstractBrowse
|
---|
50 | {
|
---|
51 |
|
---|
52 | static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.GS2Browse.class.getName());
|
---|
53 |
|
---|
54 | protected SimpleCollectionDatabase coll_db = null;
|
---|
55 |
|
---|
56 | public GS2Browse()
|
---|
57 | {
|
---|
58 | }
|
---|
59 |
|
---|
60 | public void cleanUp() {
|
---|
61 | super.cleanUp();
|
---|
62 | this.coll_db.closeDatabase();
|
---|
63 | }
|
---|
64 |
|
---|
65 | public boolean configure(Element info, Element extra_info)
|
---|
66 | {
|
---|
67 | if (!super.configure(info, extra_info)){
|
---|
68 | return false;
|
---|
69 | }
|
---|
70 |
|
---|
71 | logger.info("Configuring GS2Browse...");
|
---|
72 | // the index stem is either specified in the config file or is the collection name
|
---|
73 | Element index_stem_elem = (Element) GSXML.getChildByTagName(info, GSXML.INDEX_STEM_ELEM);
|
---|
74 | String index_stem = null;
|
---|
75 | if (index_stem_elem != null) {
|
---|
76 | index_stem = index_stem_elem.getAttribute(GSXML.NAME_ATT);
|
---|
77 | }
|
---|
78 | if (index_stem == null || index_stem.equals("")) {
|
---|
79 | index_stem = this.cluster_name;
|
---|
80 | }
|
---|
81 |
|
---|
82 | // find out what kind of database we have
|
---|
83 | Element database_type_elem = (Element) GSXML.getChildByTagName(info, GSXML.DATABASE_TYPE_ELEM);
|
---|
84 | String database_type = null;
|
---|
85 | if (database_type_elem != null) {
|
---|
86 | database_type = database_type_elem.getAttribute(GSXML.NAME_ATT);
|
---|
87 | }
|
---|
88 |
|
---|
89 | if (database_type == null || database_type.equals("")) {
|
---|
90 | database_type = "gdbm"; // the default
|
---|
91 | }
|
---|
92 | coll_db = new SimpleCollectionDatabase(database_type);
|
---|
93 | if (!coll_db.databaseOK()) {
|
---|
94 | logger.error("Couldn't create the collection database of type "+database_type);
|
---|
95 | return false;
|
---|
96 | }
|
---|
97 |
|
---|
98 | // Open database for querying
|
---|
99 | String coll_db_file = GSFile.collectionDatabaseFile(this.site_home, this.cluster_name, index_stem, database_type);
|
---|
100 | if (!this.coll_db.openDatabase(coll_db_file, SimpleCollectionDatabase.READ)) {
|
---|
101 | logger.error("Could not open collection database!");
|
---|
102 | return false;
|
---|
103 | }
|
---|
104 | this.macro_resolver = new GS2MacroResolver(this.coll_db);
|
---|
105 | return true;
|
---|
106 | }
|
---|
107 |
|
---|
108 | /** if id ends in .fc, .pc etc, then translate it to the correct id */
|
---|
109 | protected String translateId(String node_id) {
|
---|
110 | return OID.translateOID(this.coll_db, node_id); //return this.coll_db.translateOID(node_id);
|
---|
111 | }
|
---|
112 |
|
---|
113 | /** returns the document type of the doc that the specified node
|
---|
114 | belongs to. should be one of
|
---|
115 | GSXML.DOC_TYPE_SIMPLE,
|
---|
116 | GSXML.DOC_TYPE_PAGED,
|
---|
117 | GSXML.DOC_TYPE_HIERARCHY
|
---|
118 | */
|
---|
119 | protected String getDocType(String node_id) {
|
---|
120 | DBInfo info = this.coll_db.getInfo(node_id);
|
---|
121 | if (info == null) {
|
---|
122 | return GSXML.DOC_TYPE_SIMPLE;
|
---|
123 | }
|
---|
124 | String doc_type = info.getInfo("doctype");
|
---|
125 | if (!doc_type.equals("")&&!doc_type.equals("doc")) {
|
---|
126 | return doc_type;
|
---|
127 | }
|
---|
128 |
|
---|
129 | String top_id = OID.getTop(node_id);
|
---|
130 | boolean is_top = (top_id.equals(node_id) ? true : false);
|
---|
131 |
|
---|
132 | String children = info.getInfo("contains");
|
---|
133 | boolean is_leaf = (children.equals("") ? true : false);
|
---|
134 |
|
---|
135 | if (is_top && is_leaf) { // a single section document
|
---|
136 | return GSXML.DOC_TYPE_SIMPLE;
|
---|
137 | }
|
---|
138 |
|
---|
139 | // now we just check the top node
|
---|
140 | if (!is_top) { // we need to look at the top info
|
---|
141 | info = this.coll_db.getInfo(top_id);
|
---|
142 | }
|
---|
143 | if (info == null) {
|
---|
144 | return GSXML.DOC_TYPE_HIERARCHY;
|
---|
145 | }
|
---|
146 |
|
---|
147 | String childtype = info.getInfo("childtype");
|
---|
148 | if (childtype.equals("Paged")) {
|
---|
149 | return GSXML.DOC_TYPE_PAGED;
|
---|
150 | }
|
---|
151 | return GSXML.DOC_TYPE_HIERARCHY;
|
---|
152 |
|
---|
153 | }
|
---|
154 |
|
---|
155 | /** returns the id of the root node of the document containing node node_id. . may be the same as node_id */
|
---|
156 | protected String getRootId(String node_id) {
|
---|
157 | return OID.getTop(node_id);
|
---|
158 | }
|
---|
159 | /** returns a list of the child ids in order, null if no children */
|
---|
160 | protected ArrayList getChildrenIds(String node_id) {
|
---|
161 | DBInfo info = this.coll_db.getInfo(node_id);
|
---|
162 | if (info == null) {
|
---|
163 | return null;
|
---|
164 | }
|
---|
165 |
|
---|
166 | ArrayList children = new ArrayList();
|
---|
167 |
|
---|
168 | String contains = info.getInfo("contains");
|
---|
169 | StringTokenizer st = new StringTokenizer(contains, ";");
|
---|
170 | while (st.hasMoreTokens()) {
|
---|
171 | String child_id = st.nextToken().replaceAll("\"", node_id);
|
---|
172 | children.add(child_id);
|
---|
173 | }
|
---|
174 | return children;
|
---|
175 |
|
---|
176 | }
|
---|
177 | /** returns the node id of the parent node, null if no parent */
|
---|
178 | protected String getParentId(String node_id){
|
---|
179 | String parent = OID.getParent(node_id);
|
---|
180 | if (parent.equals(node_id)) {
|
---|
181 | return null;
|
---|
182 | }
|
---|
183 | return parent;
|
---|
184 | }
|
---|
185 |
|
---|
186 | protected String getMetadata(String node_id, String key){
|
---|
187 | DBInfo info = this.coll_db.getInfo(node_id);
|
---|
188 | if (info == null) {
|
---|
189 | return "";
|
---|
190 | }
|
---|
191 |
|
---|
192 | Set keys = info.getKeys();
|
---|
193 | Iterator it = keys.iterator();
|
---|
194 | while(it.hasNext()) {
|
---|
195 | String key_in = (String)it.next();
|
---|
196 | String value = info.getInfo(key);
|
---|
197 | if (key_in.equals(key)){
|
---|
198 | return value;
|
---|
199 | }
|
---|
200 | }
|
---|
201 |
|
---|
202 | return "";
|
---|
203 |
|
---|
204 | }
|
---|
205 |
|
---|
206 | /** get the metadata for the classifier node node_id
|
---|
207 | * returns a metadataList element:
|
---|
208 | * <metadataList><metadata name="xxx">value</metadata></metadataList>
|
---|
209 | * if all_metadata is true, returns all available metadata, otherwise just
|
---|
210 | * returns requested metadata
|
---|
211 | */
|
---|
212 | // assumes only one value per metadata
|
---|
213 | protected Element getMetadataList(String node_id, boolean all_metadata,
|
---|
214 | ArrayList metadata_names) {
|
---|
215 | String lang = "en";
|
---|
216 | Element metadata_list = this.doc.createElement(GSXML.METADATA_ELEM+GSXML.LIST_MODIFIER);
|
---|
217 | DBInfo info = this.coll_db.getInfo(node_id);
|
---|
218 | if (info == null) {
|
---|
219 | return null;
|
---|
220 | }
|
---|
221 | if (all_metadata) {
|
---|
222 | // return everything out of the database
|
---|
223 | Set keys = info.getKeys();
|
---|
224 | Iterator it = keys.iterator();
|
---|
225 | while(it.hasNext()) {
|
---|
226 | String key = (String)it.next();
|
---|
227 | String value = info.getInfo(key);
|
---|
228 | GSXML.addMetadata(this.doc, metadata_list, key, this.macro_resolver.resolve(value, lang, GS2MacroResolver.SCOPE_META, node_id));
|
---|
229 | }
|
---|
230 |
|
---|
231 | } else {
|
---|
232 | for (int i=0; i<metadata_names.size(); i++) {
|
---|
233 | String meta_name = (String) metadata_names.get(i);
|
---|
234 | String value = (String)info.getInfo(meta_name);
|
---|
235 | GSXML.addMetadata(this.doc, metadata_list, meta_name, value);
|
---|
236 | }
|
---|
237 | }
|
---|
238 | return metadata_list;
|
---|
239 | }
|
---|
240 |
|
---|
241 | /** returns the structural information asked for.
|
---|
242 | * info_type may be one of
|
---|
243 | * INFO_NUM_SIBS, INFO_NUM_CHILDREN, INFO_SIB_POS
|
---|
244 | */
|
---|
245 | protected String getStructureInfo(String doc_id, String info_type) {
|
---|
246 | String value="";
|
---|
247 | if (info_type.equals(INFO_NUM_SIBS)) {
|
---|
248 | String parent_id = OID.getParent(doc_id);
|
---|
249 | if (parent_id.equals(doc_id)) {
|
---|
250 | value="0";
|
---|
251 | } else {
|
---|
252 | value = String.valueOf(getNumChildren(parent_id));
|
---|
253 | }
|
---|
254 | return value;
|
---|
255 | }
|
---|
256 |
|
---|
257 | if (info_type.equals(INFO_NUM_CHILDREN)) {
|
---|
258 | return String.valueOf(getNumChildren(doc_id));
|
---|
259 | }
|
---|
260 |
|
---|
261 |
|
---|
262 | if (info_type.equals(INFO_SIB_POS)) {
|
---|
263 | String parent_id = OID.getParent(doc_id);
|
---|
264 | if (parent_id.equals(doc_id)) {
|
---|
265 | return "-1";
|
---|
266 | }
|
---|
267 |
|
---|
268 | DBInfo info = this.coll_db.getInfo(parent_id);
|
---|
269 | if (info==null) {
|
---|
270 | return "-1";
|
---|
271 | }
|
---|
272 |
|
---|
273 | String contains = info.getInfo("contains");
|
---|
274 | contains = contains.replaceAll("\"", parent_id);
|
---|
275 | String [] children = contains.split(";");
|
---|
276 | for (int i=0;i<children.length;i++) {
|
---|
277 | String child_id = children[i];
|
---|
278 | if (child_id.equals(doc_id)) {
|
---|
279 | return String.valueOf(i+1); // make it from 1 to length
|
---|
280 |
|
---|
281 | }
|
---|
282 | }
|
---|
283 |
|
---|
284 | return "-1";
|
---|
285 | } else {
|
---|
286 | return null;
|
---|
287 | }
|
---|
288 |
|
---|
289 | }
|
---|
290 |
|
---|
291 | protected int getNumChildren(String node_id) {
|
---|
292 | DBInfo info = this.coll_db.getInfo(node_id);
|
---|
293 | if (info == null) {
|
---|
294 | return 0;
|
---|
295 | }
|
---|
296 | String contains = info.getInfo("contains");
|
---|
297 | if (contains.equals("")) {
|
---|
298 | return 0;
|
---|
299 | }
|
---|
300 | String [] children = contains.split(";");
|
---|
301 | return children.length;
|
---|
302 | }
|
---|
303 |
|
---|
304 | /** returns true if the id refers to a document (rather than
|
---|
305 | * a classifier node)
|
---|
306 | */
|
---|
307 | protected boolean isDocumentId(String node_id){
|
---|
308 | if (node_id.startsWith("CL")) {
|
---|
309 | return false;
|
---|
310 | }
|
---|
311 | return true;
|
---|
312 | }
|
---|
313 |
|
---|
314 | }
|
---|