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 | */
|
---|
18 | package org.greenstone.gsdl3.service;
|
---|
19 |
|
---|
20 |
|
---|
21 | // Greenstone classes
|
---|
22 | import org.greenstone.mg.*;
|
---|
23 | import org.greenstone.gsdl3.util.*;
|
---|
24 |
|
---|
25 | // XML classes
|
---|
26 | import org.w3c.dom.Document;
|
---|
27 | import org.w3c.dom.Element;
|
---|
28 | import org.w3c.dom.NodeList;
|
---|
29 |
|
---|
30 | // java
|
---|
31 | import java.util.Vector;
|
---|
32 | import java.util.ArrayList;
|
---|
33 | import java.util.HashMap;
|
---|
34 | import java.util.Map;
|
---|
35 | import java.util.Set;
|
---|
36 | import java.util.Iterator;
|
---|
37 | import java.io.File;
|
---|
38 |
|
---|
39 | import org.apache.log4j.*;
|
---|
40 |
|
---|
41 | /**
|
---|
42 | *
|
---|
43 | * @author <a href="mailto:[email protected]">Katherine Don</a>
|
---|
44 | * @author <a href="mailto:[email protected]">Michael Dewsnip</a>
|
---|
45 | */
|
---|
46 |
|
---|
47 | public class GS2MGSearch
|
---|
48 | extends AbstractGS2Search
|
---|
49 | {
|
---|
50 |
|
---|
51 | protected MGWrapper mg_src = null;
|
---|
52 |
|
---|
53 | static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.GS2MGSearch.class.getName());
|
---|
54 |
|
---|
55 |
|
---|
56 | /** constructor */
|
---|
57 | public GS2MGSearch()
|
---|
58 | {
|
---|
59 | this.mg_src = new MGWrapper();
|
---|
60 |
|
---|
61 | }
|
---|
62 | public void cleanUp() {
|
---|
63 | super.cleanUp();
|
---|
64 | this.mg_src.unloadIndexData();
|
---|
65 | }
|
---|
66 |
|
---|
67 | /** configure this service */
|
---|
68 | public boolean configure(Element info, Element extra_info)
|
---|
69 | {
|
---|
70 | if (!super.configure(info, extra_info)){
|
---|
71 | return false;
|
---|
72 | }
|
---|
73 |
|
---|
74 | this.mg_src.setMaxNumeric(this.maxnumeric);
|
---|
75 |
|
---|
76 | return true;
|
---|
77 | }
|
---|
78 |
|
---|
79 |
|
---|
80 |
|
---|
81 | /** do the actual query */
|
---|
82 | protected Element processTextQuery(Element request)
|
---|
83 | {
|
---|
84 |
|
---|
85 | // Create a new (empty) result message
|
---|
86 | Element result = this.doc.createElement(GSXML.RESPONSE_ELEM);
|
---|
87 | result.setAttribute(GSXML.FROM_ATT, TEXT_QUERY_SERVICE);
|
---|
88 | result.setAttribute(GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
|
---|
89 |
|
---|
90 | // Get the parameters of the request
|
---|
91 | Element param_list = (Element) GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
|
---|
92 | if (param_list == null) {
|
---|
93 | logger.error("TextQuery request had no paramList.");
|
---|
94 | return result; // Return the empty result
|
---|
95 | }
|
---|
96 |
|
---|
97 | // Process the request parameters
|
---|
98 | HashMap params = GSXML.extractParams(param_list, false);
|
---|
99 |
|
---|
100 | // Make sure a query has been specified
|
---|
101 | String query = (String) params.get(QUERY_PARAM);
|
---|
102 | if (query == null || query.equals("")) {
|
---|
103 | return result; // Return the empty result
|
---|
104 | }
|
---|
105 |
|
---|
106 | // If an index hasn't been specified, use the default
|
---|
107 | String index = (String) params.get(INDEX_PARAM);
|
---|
108 | if (index == null) {
|
---|
109 | index = this.default_index;
|
---|
110 | }
|
---|
111 |
|
---|
112 | // If a subcollection index has been specified, use it
|
---|
113 | String indexSub = (String) params.get(INDEX_SUBCOLLECTION_PARAM);
|
---|
114 | if (indexSub != null) {
|
---|
115 | index += indexSub;
|
---|
116 | }
|
---|
117 | else{
|
---|
118 | if (!this.default_index_subcollection.equals("")){
|
---|
119 | index += this.default_index_subcollection;
|
---|
120 | }
|
---|
121 | }
|
---|
122 |
|
---|
123 | // If a subcollection index has been specified, use it
|
---|
124 | String indexLang = (String) params.get(INDEX_LANGUAGE_PARAM);
|
---|
125 | if (indexLang != null) {
|
---|
126 | index += indexLang;
|
---|
127 | }
|
---|
128 | else{
|
---|
129 | if (!this.default_index_language.equals("")){
|
---|
130 | index += this.default_index_language;
|
---|
131 | }
|
---|
132 | }
|
---|
133 |
|
---|
134 | // The location of the MG index and text files
|
---|
135 | String basedir = GSFile.collectionBaseDir(this.site_home, this.cluster_name) + File.separatorChar; // Needed for MG
|
---|
136 | String textdir = GSFile.collectionTextPath(this.index_stem);
|
---|
137 | String indexpath = GSFile.collectionIndexPath(this.index_stem, index);
|
---|
138 | this.mg_src.setIndex(indexpath);
|
---|
139 | System.err.println("index path = "+indexpath);
|
---|
140 | // set the mg query parameters to the values the user has specified
|
---|
141 | setStandardQueryParams(params);
|
---|
142 | this.mg_src.runQuery(basedir, textdir, query);
|
---|
143 | MGQueryResult mqr = this.mg_src.getQueryResult();
|
---|
144 | if (mqr.isClear()) {
|
---|
145 | // something has gone wrong
|
---|
146 | GSXML.addError(this.doc, result, "Couldn't query the mg database", GSXML.ERROR_TYPE_SYSTEM);
|
---|
147 | return result;
|
---|
148 | }
|
---|
149 | long totalDocs = mqr.getTotalDocs();
|
---|
150 |
|
---|
151 | // Get the docnums out, and convert to HASH ids
|
---|
152 | Vector docs = mqr.getDocs();
|
---|
153 | if (docs.size() == 0) {
|
---|
154 | logger.error("No results found...\n");
|
---|
155 | }
|
---|
156 |
|
---|
157 | // Create a metadata list to store information about the query results
|
---|
158 | Element metadata_list = this.doc.createElement(GSXML.METADATA_ELEM+GSXML.LIST_MODIFIER);
|
---|
159 | result.appendChild(metadata_list);
|
---|
160 |
|
---|
161 | // Add a metadata element specifying the number of matching documents
|
---|
162 | // because teh total number is just the number returned, use numDocsReturned, not numDocsMatched
|
---|
163 | GSXML.addMetadata(this.doc, metadata_list, "numDocsReturned", ""+totalDocs);
|
---|
164 | // add a metadata item to specify what actual query was done - eg if stuff was stripped out etc. and then we can use the query later, cos we don't know which parameter was the query
|
---|
165 | GSXML.addMetadata(this.doc, metadata_list, "query", query);
|
---|
166 |
|
---|
167 | if (docs.size() > 0) {
|
---|
168 | // Create a document list to store the matching documents, and add them
|
---|
169 | Element document_list = this.doc.createElement(GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
|
---|
170 | result.appendChild(document_list);
|
---|
171 | for (int d = 0; d < docs.size(); d++) {
|
---|
172 | long docnum = ((MGDocInfo) docs.elementAt(d)).num_;
|
---|
173 | float rank = ((MGDocInfo) docs.elementAt(d)).rank_;
|
---|
174 | String doc_id = internalNum2OID(docnum);
|
---|
175 | Element doc_node = createDocNode(doc_id, Float.toString(rank));
|
---|
176 | document_list.appendChild(doc_node);
|
---|
177 | }
|
---|
178 | }
|
---|
179 |
|
---|
180 | // Create a term list to store the term information, and add it
|
---|
181 | Element term_list = this.doc.createElement(GSXML.TERM_ELEM+GSXML.LIST_MODIFIER);
|
---|
182 | result.appendChild(term_list);
|
---|
183 | Vector terms = mqr.getTerms();
|
---|
184 | for (int t = 0; t < terms.size(); t++) {
|
---|
185 | MGTermInfo term_info = (MGTermInfo) terms.get(t);
|
---|
186 |
|
---|
187 | String term = term_info.term_;
|
---|
188 | int stem_method = term_info.stem_method_;
|
---|
189 | Vector equiv_terms = term_info.equiv_terms_;
|
---|
190 |
|
---|
191 | Element term_elem = this.doc.createElement(GSXML.TERM_ELEM);
|
---|
192 | term_elem.setAttribute(GSXML.NAME_ATT, term);
|
---|
193 | term_elem.setAttribute(STEM_ATT, "" + stem_method);
|
---|
194 |
|
---|
195 | Element equiv_term_list = this.doc.createElement(EQUIV_TERM_ELEM+GSXML.LIST_MODIFIER);
|
---|
196 | term_elem.appendChild(equiv_term_list);
|
---|
197 |
|
---|
198 | long total_term_freq = 0;
|
---|
199 | for (int et = 0; et < equiv_terms.size(); et++) {
|
---|
200 | MGEquivTermInfo equiv_term_info = (MGEquivTermInfo) equiv_terms.get(et);
|
---|
201 |
|
---|
202 | Element equiv_term_elem = this.doc.createElement(GSXML.TERM_ELEM);
|
---|
203 | equiv_term_elem.setAttribute(GSXML.NAME_ATT, equiv_term_info.term_);
|
---|
204 | equiv_term_elem.setAttribute(NUM_DOCS_MATCH_ATT, "" + equiv_term_info.match_docs_);
|
---|
205 | equiv_term_elem.setAttribute(FREQ_ATT, "" + equiv_term_info.term_freq_);
|
---|
206 | equiv_term_list.appendChild(equiv_term_elem);
|
---|
207 |
|
---|
208 | total_term_freq += equiv_term_info.term_freq_;
|
---|
209 | }
|
---|
210 |
|
---|
211 | term_elem.setAttribute(FREQ_ATT, "" + total_term_freq);
|
---|
212 | term_list.appendChild(term_elem);
|
---|
213 | }
|
---|
214 | return result;
|
---|
215 | }
|
---|
216 |
|
---|
217 | // should probably use a list rather than map
|
---|
218 | protected boolean setStandardQueryParams(HashMap params)
|
---|
219 | {
|
---|
220 | // set the default ones
|
---|
221 | this.mg_src.setReturnTerms(true);
|
---|
222 | this.mg_src.setCase(true); // turn casefolding on by default
|
---|
223 | Set entries = params.entrySet();
|
---|
224 | Iterator i = entries.iterator();
|
---|
225 | while (i.hasNext()) {
|
---|
226 | Map.Entry m = (Map.Entry)i.next();
|
---|
227 | String name = (String)m.getKey();
|
---|
228 | String value = (String)m.getValue();
|
---|
229 |
|
---|
230 | if (name.equals(CASE_PARAM)) {
|
---|
231 | boolean val = (value.equals(BOOLEAN_PARAM_ON) ? true : false);
|
---|
232 | this.mg_src.setCase(val);
|
---|
233 | }
|
---|
234 | else if (name.equals(STEM_PARAM)) {
|
---|
235 | boolean val = (value.equals(BOOLEAN_PARAM_ON) ? true : false);
|
---|
236 | this.mg_src.setStem(val);
|
---|
237 | }
|
---|
238 | else if (name.equals(MATCH_PARAM)) {
|
---|
239 | int mode = (value.equals(MATCH_PARAM_ALL) ? 1 : 0);
|
---|
240 | this.mg_src.setMatchMode(mode);
|
---|
241 | }
|
---|
242 | else if (name.equals(MAXDOCS_PARAM)) {
|
---|
243 | int docs = Integer.parseInt(value);
|
---|
244 | this.mg_src.setMaxDocs(docs);
|
---|
245 | } // ignore any others
|
---|
246 | }
|
---|
247 | return true;
|
---|
248 | }
|
---|
249 |
|
---|
250 |
|
---|
251 | }
|
---|
252 |
|
---|
253 |
|
---|