source: greenstone3/trunk/src/java/org/greenstone/gsdl3/service/AbstractGS2Search.java@ 15326

Last change on this file since 15326 was 15326, checked in by kjdon, 16 years ago

added support for JDBM (or other) in place of GDBM: use SimpleCollectionDatabase instead of GDBMWrapper. new Element in buildConfig file: databaseType, set to gdbm or jdbm. If not present, assume gdbm. Also may be some small style changes to some files
  • Property svn:keywords set to Author Date Id Revision
File size: 12.6 KB
Line 
1/*
2 * GS2MGSearch.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 * the Free Software Foundation; either version 2 of the License, or
7 * (at your option) any later version.
8 *
9 * This program is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 * GNU General Public License for more details.
13 *
14 * You should have received a copy of the GNU General Public License
15 * along with this program; if not, write to the Free Software
16 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
17 */
18package org.greenstone.gsdl3.service;
19
20
21// Greenstone classes
22import org.greenstone.gsdl3.util.OID;
23import org.greenstone.gsdl3.util.DBInfo;
24import org.greenstone.gsdl3.util.GSXML;
25import org.greenstone.gsdl3.util.SimpleCollectionDatabase;
26import org.greenstone.gsdl3.util.GSFile;
27
28// XML classes
29import org.w3c.dom.Document;
30import org.w3c.dom.Element;
31import org.w3c.dom.NodeList;
32
33// java
34import java.util.Vector;
35import java.util.ArrayList;
36import java.util.HashMap;
37import java.util.Map;
38import java.util.Set;
39import java.util.Iterator;
40import java.io.File;
41
42import org.apache.log4j.*;
43
44public abstract class AbstractGS2Search
45 extends AbstractSearch
46{
47
48 protected static final String EQUIV_TERM_ELEM = "equivTerm";
49
50 protected static final String STEM_ATT = "stem";
51 protected static final String NUM_DOCS_MATCH_ATT = "numDocsMatch";
52 protected static final String FREQ_ATT = "freq";
53
54 // Elements used in the config file that are specific to this class
55 protected static final String DEFAULT_INDEX_ELEM = "defaultIndex";
56 protected static final String INDEX_STEM_ELEM = "indexStem";
57 protected static final String INDEX_ELEM = "index";
58 protected static final String DEFAULT_INDEX_SUBCOLLECTION_ELEM = "defaultIndexSubcollection";
59 protected static final String DEFAULT_INDEX_LANGUAGE_ELEM = "defaultIndexLanguage";
60 protected static final String INDEX_SUBCOLLECTION_ELEM = "indexSubcollection";
61 protected static final String INDEX_LANGUAGE_ELEM = "indexLanguage";
62
63
64 // Some indexing options
65 protected static final String STEMINDEX_OPTION = "stemIndexes";
66 protected static final String MAXNUMERIC_OPTION = "maxnumeric";
67
68 /** the stem used for the index files */
69 protected String index_stem = null;
70
71 // stem indexes available
72 protected boolean does_case=true;
73 protected boolean does_stem=true;
74 protected boolean does_accent=false;
75
76 // maxnumeric -
77 protected int maxnumeric = 4;
78
79 // collection database
80 protected SimpleCollectionDatabase coll_db = null;
81
82 static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.AbstractGS2Search.class.getName());
83
84
85 /** constructor */
86 public AbstractGS2Search()
87 {
88
89 }
90 public void cleanUp() {
91 super.cleanUp();
92 this.coll_db.closeDatabase();
93 }
94
95 /** configure this service */
96 public boolean configure(Element info, Element extra_info)
97 {
98 if (!super.configure(info, extra_info)){
99 return false;
100 }
101
102 // find out what kind of database we have
103 Element database_type_elem = (Element) GSXML.getChildByTagName(info, GSXML.DATABASE_TYPE_ELEM);
104 String database_type = null;
105 if (database_type_elem != null) {
106 database_type = database_type_elem.getAttribute(GSXML.NAME_ATT);
107 }
108 if (database_type == null || database_type.equals("")) {
109 database_type = "gdbm"; // the default
110 }
111 coll_db = new SimpleCollectionDatabase(database_type);
112 if (coll_db == null) {
113 logger.error("Couldn't create the collection database of type "+database_type);
114 return false;
115 }
116
117 // the index stem is either the collection name or is specified in the config file
118 Element index_stem_elem = (Element) GSXML.getChildByTagName(info, INDEX_STEM_ELEM);
119 if (index_stem_elem != null) {
120 this.index_stem = index_stem_elem.getAttribute(GSXML.NAME_ATT);
121 }
122 if (this.index_stem == null || this.index_stem.equals("")) {
123 logger.warn("indexStem element not found, stem will default to collection name");
124 this.index_stem = this.cluster_name;
125 }
126
127 // Open database for querying
128 String coll_db_file = GSFile.collectionDatabaseFile(this.site_home, this.cluster_name, this.index_stem, database_type);
129 if (!this.coll_db.openDatabase(coll_db_file, SimpleCollectionDatabase.READ)) {
130 logger.error("Could not open collection database!");
131 return false;
132 }
133
134 // do we support any of the extended features?
135 does_chunking = true;
136
137 // Get the default index out of <defaultIndex> (buildConfig.xml)
138 Element def = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_ELEM);
139 if (def != null) {
140 this.default_index = def.getAttribute(GSXML.SHORTNAME_ATT);
141 } // otherwise will be "", and the first one will be the default
142
143 //get the default indexSubcollection out of <defaultIndexSubcollection> (buildConfig.xml)
144 Element defSub = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_SUBCOLLECTION_ELEM);
145 if (defSub != null) {
146 this.default_index_subcollection = defSub.getAttribute(GSXML.SHORTNAME_ATT);
147 }
148
149 //get the default indexLanguage out of <defaultIndexLanguage> (buildConfig.xml)
150 Element defLang = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_LANGUAGE_ELEM);
151 if (defLang != null) {
152 this.default_index_language = defLang.getAttribute(GSXML.SHORTNAME_ATT);
153 } //concate defaultIndex + defaultIndexSubcollection + defaultIndexLanguage
154
155 // get index options
156 Element index_option_list = (Element) GSXML.getChildByTagName(info, GSXML.INDEX_OPTION_ELEM + GSXML.LIST_MODIFIER);
157 if (index_option_list != null) {
158 NodeList options = index_option_list.getElementsByTagName(GSXML.INDEX_OPTION_ELEM);
159 for (int i=0; i<options.getLength(); i++) {
160 Element opt = (Element)options.item(i);
161 String name = opt.getAttribute(GSXML.NAME_ATT);
162 String value = opt.getAttribute(GSXML.VALUE_ATT);
163 if (name.equals(MAXNUMERIC_OPTION)) {
164 int maxnum = Integer.parseInt(value);
165 if (4 <= maxnum && maxnum < 512) {
166 maxnumeric = maxnum;
167 }
168 }
169 else if (name.equals(STEMINDEX_OPTION)) {
170 int stemindex = Integer.parseInt(value);
171 // stem and case are true by default, accent folding false by default
172 if ((stemindex & 1) == 0) {
173 does_case = false;
174 }
175 if ((stemindex & 2) == 0) {
176 does_stem = false;
177 }
178 if ((stemindex & 4) != 0) {
179 does_accent = true;
180 }
181 }
182 }
183 }
184
185 // get display info from extra info
186 if (extra_info !=null) {
187 Document owner = info.getOwnerDocument();
188 // so far we have index specific display elements, and global format elements
189 NodeList indexes = info.getElementsByTagName(GSXML.INDEX_ELEM);
190 Element config_search = (Element)GSXML.getChildByTagName(extra_info, GSXML.SEARCH_ELEM);
191
192 for (int i=0; i<indexes.getLength();i++) {
193 Element ind = (Element)indexes.item(i);
194 String name = ind.getAttribute(GSXML.NAME_ATT);
195 Element node_extra = GSXML.getNamedElement(config_search,
196 GSXML.INDEX_ELEM,
197 GSXML.NAME_ATT,
198 name);
199 if (node_extra == null) {
200 logger.error("haven't found extra info for index named "+name);
201 continue;
202 }
203
204 // get the display elements if any - displayName
205 NodeList display_names = node_extra.getElementsByTagName(GSXML.DISPLAY_TEXT_ELEM);
206 if (display_names !=null) {
207 for (int j=0; j<display_names.getLength(); j++) {
208 Element e = (Element)display_names.item(j);
209 ind.appendChild(owner.importNode(e, true));
210 }
211 }
212 } // for each index
213 }
214 return true;
215 }
216
217 protected void getIndexData(ArrayList index_ids, ArrayList index_names, String lang) {
218 // the index info -
219 Element index_list = (Element)GSXML.getChildByTagName(this.config_info, INDEX_ELEM+GSXML.LIST_MODIFIER);
220 NodeList indexes = index_list.getElementsByTagName(INDEX_ELEM);
221 int len = indexes.getLength();
222 // now add even if there is only one
223 for (int i=0; i<len; i++) {
224 Element index = (Element)indexes.item(i);
225 String shortname = index.getAttribute(GSXML.SHORTNAME_ATT);
226 if (shortname.equals("")) {
227 continue;
228 }
229 index_ids.add(shortname);
230 String display_name = GSXML.getDisplayText(index, GSXML.DISPLAY_TEXT_NAME, lang, "en");
231 if (display_name.equals("")) {
232 display_name = index.getAttribute(GSXML.NAME_ATT);
233 if (display_name.equals("")) {
234 display_name = shortname;
235 }
236 }
237 index_names.add(display_name);
238 }
239 }
240
241 protected void getIndexSubcollectionData(ArrayList index_sub_ids, ArrayList index_sub_names, String lang){
242 // the index info -
243 Element index_sub_list = (Element)GSXML.getChildByTagName(this.config_info, INDEX_SUBCOLLECTION_ELEM+GSXML.LIST_MODIFIER);
244 NodeList index_subs = index_sub_list.getElementsByTagName(INDEX_SUBCOLLECTION_ELEM);
245 int len = index_subs.getLength();
246 // now add even if there is only one
247 for (int i=0; i<len; i++) {
248 Element indexsub = (Element)index_subs.item(i);
249 String shortname = indexsub.getAttribute(GSXML.SHORTNAME_ATT);
250 if (shortname.equals("")) {
251 continue;
252 }
253 index_sub_ids.add(shortname);
254 String display_name = GSXML.getDisplayText(indexsub, GSXML.DISPLAY_TEXT_NAME, lang, "en");
255 if (display_name.equals("")) {
256 display_name = indexsub.getAttribute(GSXML.NAME_ATT);
257 if (display_name.equals("")) {
258 display_name = shortname;
259 }
260 }
261 index_sub_names.add(display_name);
262 }
263 }
264
265 protected void getIndexLanguageData(ArrayList index_lang_ids, ArrayList index_lang_names, String lang){
266 // the index info -
267 Element index_lang_list = (Element)GSXML.getChildByTagName(this.config_info, INDEX_LANGUAGE_ELEM+GSXML.LIST_MODIFIER);
268 NodeList index_langs = index_lang_list.getElementsByTagName(INDEX_LANGUAGE_ELEM);
269 int len = index_langs.getLength();
270 // now add even if there is only one
271 for (int i=0; i<len; i++) {
272 Element indexlang = (Element)index_langs.item(i);
273 String shortname = indexlang.getAttribute(GSXML.SHORTNAME_ATT);
274 if (shortname.equals("")) {
275 continue;
276 }
277 index_lang_ids.add(shortname);
278 String display_name = GSXML.getDisplayText(indexlang, GSXML.DISPLAY_TEXT_NAME, lang, "en");
279 if (display_name.equals("")) {
280 display_name = indexlang.getAttribute(GSXML.NAME_ATT);
281 if (display_name.equals("")) {
282 display_name = shortname;
283 }
284 }
285 index_lang_names.add(display_name);
286 }
287
288
289 }
290
291
292 protected void addCustomQueryParams(Element param_list, String lang)
293 {
294 if (this.does_case){
295 createParameter(CASE_PARAM, param_list, lang);
296 }
297 if (this.does_stem){
298 createParameter(STEM_PARAM, param_list, lang);
299 }
300 if (this.does_accent){
301 createParameter(ACCENT_PARAM, param_list, lang);
302 }
303 createParameter(MATCH_PARAM, param_list, lang);
304 }
305
306
307
308 /** returns the document type of the doc that the specified node
309 belongs to. should be one of
310 GSXML.DOC_TYPE_SIMPLE,
311 GSXML.DOC_TYPE_PAGED,
312 GSXML.DOC_TYPE_HIERARCHY
313 */
314 protected String getDocType(String node_id){
315 DBInfo info = this.coll_db.getInfo(node_id);
316 if (info == null) {
317 return GSXML.DOC_TYPE_SIMPLE;
318 }
319 String doc_type = info.getInfo("doctype");
320 if (!doc_type.equals("")&&!doc_type.equals("doc")) {
321 return doc_type;
322 }
323
324 String top_id = OID.getTop(node_id);
325 boolean is_top = (top_id.equals(node_id) ? true : false);
326
327 String children = info.getInfo("contains");
328 boolean is_leaf = (children.equals("") ? true : false);
329
330 if (is_top && is_leaf) { // a single section document
331 return GSXML.DOC_TYPE_SIMPLE;
332 }
333
334 // now we just check the top node
335 if (!is_top) { // we need to look at the top info
336 info = this.coll_db.getInfo(top_id);
337 }
338 if (info == null) {
339 return GSXML.DOC_TYPE_HIERARCHY;
340 }
341
342 String childtype = info.getInfo("childtype");
343 if (childtype.equals("Paged")) {
344 return GSXML.DOC_TYPE_PAGED;
345 }
346 return GSXML.DOC_TYPE_HIERARCHY;
347
348 }
349
350 /** returns true if the node has child nodes */
351 protected boolean hasChildren(String node_id){
352 DBInfo info = this.coll_db.getInfo(node_id);
353 if (info == null) {
354 return false;
355 }
356 String contains = info.getInfo("contains");
357 if (contains.equals("")) {
358 return false;
359 }
360 return true;
361 }
362
363 /** returns true if the node has a parent */
364 protected boolean hasParent(String node_id){
365 String parent = OID.getParent(node_id);
366 if (parent.equals(node_id)) {
367 return false;
368 }
369 return true;
370 }
371
372 /** convert indexer internal id to Greenstone oid */
373 protected String internalNum2OID(long docnum)
374 {
375 return this.coll_db.docnum2OID(docnum);
376
377 }
378 protected String internalNum2OID(String docnum)
379 {
380 return this.coll_db.docnum2OID(docnum);
381
382 }
383
384}
385
386
Note: See TracBrowser for help on using the repository browser.