[3649] | 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
|
---|
[3851] | 6 | * the Free Software Foundation; either version 2 of the License, or
|
---|
[3649] | 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 |
|
---|
[3823] | 20 |
|
---|
| 21 | // Greenstone classes
|
---|
[13238] | 22 | import org.greenstone.mg.*;
|
---|
[3649] | 23 | import org.greenstone.gsdl3.util.*;
|
---|
| 24 |
|
---|
[3823] | 25 | // XML classes
|
---|
[9002] | 26 | import org.w3c.dom.Document;
|
---|
[14428] | 27 | import org.w3c.dom.Element;
|
---|
[9002] | 28 | import org.w3c.dom.NodeList;
|
---|
[3755] | 29 |
|
---|
[13238] | 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;
|
---|
[25635] | 38 | import java.io.Serializable;
|
---|
[13238] | 39 |
|
---|
[13124] | 40 | import org.apache.log4j.*;
|
---|
| 41 |
|
---|
[3649] | 42 |
|
---|
[14428] | 43 | public class GS2MGSearch
|
---|
[24394] | 44 | extends AbstractGS2TextSearch {
|
---|
[14428] | 45 |
|
---|
[14633] | 46 | protected static MGSearchWrapper mg_src = null;
|
---|
[14428] | 47 |
|
---|
| 48 | static Logger logger = Logger.getLogger (org.greenstone.gsdl3.service.GS2MGSearch.class.getName ());
|
---|
| 49 |
|
---|
| 50 |
|
---|
[3755] | 51 | /** constructor */
|
---|
[14428] | 52 | public GS2MGSearch () {
|
---|
[29558] | 53 | does_chunking = true;
|
---|
[14633] | 54 | if(this.mg_src == null){
|
---|
| 55 | this.mg_src = new MGSearchWrapper ();
|
---|
| 56 | }
|
---|
[3649] | 57 | }
|
---|
[14428] | 58 | public void cleanUp () {
|
---|
| 59 | super.cleanUp ();
|
---|
| 60 | this.mg_src.unloadIndexData ();
|
---|
[9874] | 61 | }
|
---|
[14428] | 62 |
|
---|
[13289] | 63 | /** configure this service */
|
---|
[14428] | 64 | public boolean configure (Element info, Element extra_info) {
|
---|
| 65 | if (!super.configure (info, extra_info)){
|
---|
| 66 | return false;
|
---|
| 67 | }
|
---|
[29558] | 68 | // internally mg uses 50, so set this here
|
---|
| 69 | paramDefaults.put(MAXDOCS_PARAM, "50");
|
---|
[29630] | 70 | this.mg_src.setMaxNumeric (this.maxnumeric);
|
---|
[14428] | 71 | return true;
|
---|
[13289] | 72 | }
|
---|
| 73 |
|
---|
[14428] | 74 |
|
---|
| 75 |
|
---|
[13238] | 76 | /** do the actual query */
|
---|
[14428] | 77 | protected Element processTextQuery (Element request) {
|
---|
[14633] | 78 | synchronized(this.mg_src){
|
---|
[14428] | 79 | // Create a new (empty) result message ('doc' is in ServiceRack.java)
|
---|
[28966] | 80 | Document result_doc = XMLConverter.newDOM();
|
---|
| 81 | Element result = result_doc.createElement (GSXML.RESPONSE_ELEM);
|
---|
[24394] | 82 | result.setAttribute (GSXML.FROM_ATT, QUERY_SERVICE);
|
---|
[14428] | 83 | result.setAttribute (GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
|
---|
| 84 |
|
---|
| 85 | // Get the parameters of the request
|
---|
| 86 | Element param_list = (Element) GSXML.getChildByTagName (request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
|
---|
| 87 | if (param_list == null) {
|
---|
| 88 | logger.error ("TextQuery request had no paramList.");
|
---|
| 89 | return result; // Return the empty result
|
---|
| 90 | }
|
---|
| 91 |
|
---|
| 92 | // Process the request parameters
|
---|
[25635] | 93 | HashMap<String, Serializable> params = GSXML.extractParams (param_list, false);
|
---|
[14428] | 94 |
|
---|
| 95 | // Make sure a query has been specified
|
---|
| 96 | String query = (String) params.get (QUERY_PARAM);
|
---|
| 97 | if (query == null || query.equals ("")) {
|
---|
| 98 | return result; // Return the empty result
|
---|
| 99 | }
|
---|
| 100 |
|
---|
| 101 | // If an index hasn't been specified, use the default
|
---|
| 102 | String index = (String) params.get (INDEX_PARAM);
|
---|
| 103 | if (index == null) {
|
---|
| 104 | index = this.default_index;
|
---|
| 105 | }
|
---|
| 106 |
|
---|
| 107 | // If a subcollection index has been specified, use it
|
---|
| 108 | String indexSub = (String) params.get (INDEX_SUBCOLLECTION_PARAM);
|
---|
| 109 | if (indexSub != null) {
|
---|
| 110 | index += indexSub;
|
---|
| 111 | }
|
---|
| 112 | else{
|
---|
| 113 | if (!this.default_index_subcollection.equals ("")){
|
---|
| 114 | index += this.default_index_subcollection;
|
---|
| 115 | }
|
---|
| 116 | }
|
---|
| 117 |
|
---|
| 118 | // If a subcollection index has been specified, use it
|
---|
| 119 | String indexLang = (String) params.get (INDEX_LANGUAGE_PARAM);
|
---|
| 120 | if (indexLang != null) {
|
---|
| 121 | index += indexLang;
|
---|
| 122 | }
|
---|
| 123 | else{
|
---|
| 124 | if (!this.default_index_language.equals ("")){
|
---|
| 125 | index += this.default_index_language;
|
---|
| 126 | }
|
---|
| 127 | }
|
---|
| 128 |
|
---|
| 129 | // The location of the MG index and text files
|
---|
| 130 | String basedir = GSFile.collectionBaseDir (this.site_home, this.cluster_name) + File.separatorChar; // Needed for MG
|
---|
| 131 | String textdir = GSFile.collectionTextPath (this.index_stem);
|
---|
| 132 | String indexpath = GSFile.collectionIndexPath (this.index_stem, index);
|
---|
| 133 | this.mg_src.setIndex (indexpath);
|
---|
[20296] | 134 |
|
---|
[14428] | 135 | // set the mg query parameters to the values the user has specified
|
---|
| 136 | setStandardQueryParams (params);
|
---|
| 137 | this.mg_src.runQuery (basedir, textdir, query);
|
---|
| 138 | MGQueryResult mqr = this.mg_src.getQueryResult ();
|
---|
| 139 | if (mqr.isClear ()) {
|
---|
| 140 | // something has gone wrong
|
---|
[28966] | 141 | GSXML.addError (result, "Couldn't query the mg database", GSXML.ERROR_TYPE_SYSTEM);
|
---|
[14428] | 142 | return result;
|
---|
| 143 | }
|
---|
| 144 | long totalDocs = mqr.getTotalDocs ();
|
---|
| 145 |
|
---|
| 146 | // Get the docnums out, and convert to HASH ids
|
---|
| 147 | Vector docs = mqr.getDocs ();
|
---|
| 148 | if (docs.size () == 0) {
|
---|
| 149 | logger.error ("No results found...\n");
|
---|
| 150 | }
|
---|
| 151 |
|
---|
| 152 | // Create a metadata list to store information about the query results
|
---|
[28966] | 153 | Element metadata_list = result_doc.createElement (GSXML.METADATA_ELEM+GSXML.LIST_MODIFIER);
|
---|
[14428] | 154 | result.appendChild (metadata_list);
|
---|
| 155 |
|
---|
| 156 | // Add a metadata element specifying the number of matching documents
|
---|
| 157 | // because teh total number is just the number returned, use numDocsReturned, not numDocsMatched
|
---|
[28966] | 158 | GSXML.addMetadata (metadata_list, "numDocsReturned", ""+totalDocs);
|
---|
[14428] | 159 | // 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
|
---|
[28966] | 160 | GSXML.addMetadata (metadata_list, "query", query);
|
---|
[14428] | 161 |
|
---|
| 162 | if (docs.size () > 0) {
|
---|
| 163 | // Create a document list to store the matching documents, and add them
|
---|
[28966] | 164 | Element document_list = result_doc.createElement (GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
|
---|
[14428] | 165 | result.appendChild (document_list);
|
---|
| 166 | for (int d = 0; d < docs.size (); d++) {
|
---|
| 167 | long docnum = ((MGDocInfo) docs.elementAt (d)).num_;
|
---|
| 168 | float rank = ((MGDocInfo) docs.elementAt (d)).rank_;
|
---|
| 169 | String doc_id = internalNum2OID (docnum);
|
---|
[28966] | 170 | Element doc_node = createDocNode (result_doc, doc_id, Float.toString (rank));
|
---|
[14428] | 171 | document_list.appendChild (doc_node);
|
---|
| 172 | }
|
---|
| 173 | }
|
---|
| 174 |
|
---|
| 175 | // Create a term list to store the term information, and add it
|
---|
[28966] | 176 | Element term_list = result_doc.createElement (GSXML.TERM_ELEM+GSXML.LIST_MODIFIER);
|
---|
[14428] | 177 | result.appendChild (term_list);
|
---|
| 178 | Vector terms = mqr.getTerms ();
|
---|
| 179 | for (int t = 0; t < terms.size (); t++) {
|
---|
| 180 | MGTermInfo term_info = (MGTermInfo) terms.get (t);
|
---|
| 181 |
|
---|
| 182 | String term = term_info.term_;
|
---|
| 183 | int stem_method = term_info.stem_method_;
|
---|
| 184 | Vector equiv_terms = term_info.equiv_terms_;
|
---|
| 185 |
|
---|
[28966] | 186 | Element term_elem = result_doc.createElement (GSXML.TERM_ELEM);
|
---|
[14428] | 187 | term_elem.setAttribute (GSXML.NAME_ATT, term);
|
---|
| 188 | term_elem.setAttribute (STEM_ATT, "" + stem_method);
|
---|
| 189 |
|
---|
[28966] | 190 | Element equiv_term_list = result_doc.createElement (EQUIV_TERM_ELEM+GSXML.LIST_MODIFIER);
|
---|
[14428] | 191 | term_elem.appendChild (equiv_term_list);
|
---|
| 192 |
|
---|
| 193 | long total_term_freq = 0;
|
---|
| 194 | for (int et = 0; et < equiv_terms.size (); et++) {
|
---|
| 195 | MGEquivTermInfo equiv_term_info = (MGEquivTermInfo) equiv_terms.get (et);
|
---|
| 196 |
|
---|
[28966] | 197 | Element equiv_term_elem = result_doc.createElement (GSXML.TERM_ELEM);
|
---|
[14428] | 198 | equiv_term_elem.setAttribute (GSXML.NAME_ATT, equiv_term_info.term_);
|
---|
| 199 | equiv_term_elem.setAttribute (NUM_DOCS_MATCH_ATT, "" + equiv_term_info.match_docs_);
|
---|
| 200 | equiv_term_elem.setAttribute (FREQ_ATT, "" + equiv_term_info.term_freq_);
|
---|
| 201 | equiv_term_list.appendChild (equiv_term_elem);
|
---|
| 202 |
|
---|
| 203 | total_term_freq += equiv_term_info.term_freq_;
|
---|
| 204 | }
|
---|
| 205 |
|
---|
| 206 | term_elem.setAttribute (FREQ_ATT, "" + total_term_freq);
|
---|
| 207 | term_list.appendChild (term_elem);
|
---|
| 208 | }
|
---|
| 209 | return result;
|
---|
[14633] | 210 | }//end of synchronized
|
---|
[3755] | 211 | }
|
---|
[14428] | 212 |
|
---|
[13238] | 213 | // should probably use a list rather than map
|
---|
[25635] | 214 | protected boolean setStandardQueryParams(HashMap<String, Serializable> params)
|
---|
[13238] | 215 | {
|
---|
[20238] | 216 | // set the default settings that gs uses
|
---|
[29637] | 217 | this.mg_src.setMaxDocs(Integer.parseInt(paramDefaults.get(MAXDOCS_PARAM)));
|
---|
[13238] | 218 | this.mg_src.setReturnTerms(true);
|
---|
[29630] | 219 | this.mg_src.setCase(paramDefaults.get(CASE_PARAM).equals(BOOLEAN_PARAM_ON) ? true : false);
|
---|
| 220 | this.mg_src.setStem(paramDefaults.get(STEM_PARAM).equals(BOOLEAN_PARAM_ON) ? true : false);
|
---|
[13238] | 221 | Set entries = params.entrySet();
|
---|
| 222 | Iterator i = entries.iterator();
|
---|
| 223 | while (i.hasNext()) {
|
---|
| 224 | Map.Entry m = (Map.Entry)i.next();
|
---|
| 225 | String name = (String)m.getKey();
|
---|
| 226 | String value = (String)m.getValue();
|
---|
| 227 |
|
---|
[14378] | 228 | if (name.equals(CASE_PARAM) && this.does_case) {
|
---|
[13238] | 229 | boolean val = (value.equals(BOOLEAN_PARAM_ON) ? true : false);
|
---|
| 230 | this.mg_src.setCase(val);
|
---|
| 231 | }
|
---|
[14378] | 232 | else if (name.equals(STEM_PARAM) && this.does_stem) {
|
---|
[13238] | 233 | boolean val = (value.equals(BOOLEAN_PARAM_ON) ? true : false);
|
---|
| 234 | this.mg_src.setStem(val);
|
---|
| 235 | }
|
---|
| 236 | else if (name.equals(MATCH_PARAM)) {
|
---|
| 237 | int mode = (value.equals(MATCH_PARAM_ALL) ? 1 : 0);
|
---|
| 238 | this.mg_src.setMatchMode(mode);
|
---|
| 239 | }
|
---|
| 240 | else if (name.equals(MAXDOCS_PARAM)) {
|
---|
| 241 | int docs = Integer.parseInt(value);
|
---|
| 242 | this.mg_src.setMaxDocs(docs);
|
---|
| 243 | } // ignore any others
|
---|
[9002] | 244 | }
|
---|
| 245 | return true;
|
---|
| 246 | }
|
---|
[14428] | 247 |
|
---|
| 248 |
|
---|
[3755] | 249 | }
|
---|
[9002] | 250 |
|
---|
| 251 |
|
---|