[13242] | 1 | /*
|
---|
| 2 | * AbstractMGSearch.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.mg.*;
|
---|
| 23 | import org.greenstone.gsdl3.util.GSXML;
|
---|
| 24 | import org.greenstone.gsdl3.util.GSFile;
|
---|
| 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 | /** Partially implements a generic MG search service
|
---|
| 42 | *
|
---|
| 43 | * @author <a href="mailto:[email protected]">Katherine Don</a>
|
---|
| 44 | */
|
---|
| 45 |
|
---|
| 46 | abstract public class AbstractMGSearch
|
---|
| 47 | extends AbstractSearch
|
---|
| 48 | {
|
---|
| 49 |
|
---|
| 50 | static Category logger = Category.getInstance(org.greenstone.gsdl3.service.AbstractMGSearch.class.getName());
|
---|
| 51 | protected static final String CASE_PARAM = "case";
|
---|
| 52 | protected static final String STEM_PARAM = "stem";
|
---|
| 53 | protected static final String MATCH_PARAM = "matchMode";
|
---|
| 54 | protected static final String MATCH_PARAM_ALL = "all";
|
---|
| 55 | protected static final String MATCH_PARAM_SOME = "some";
|
---|
| 56 |
|
---|
| 57 | protected static final String BOOLEAN_PARAM_ON = "1";
|
---|
| 58 | protected static final String BOOLEAN_PARAM_OFF = "0";
|
---|
| 59 |
|
---|
| 60 | protected static final String EQUIV_TERM_ELEM = "equivTerm";
|
---|
| 61 |
|
---|
| 62 | protected static final String STEM_ATT = "stem";
|
---|
| 63 | protected static final String NUM_DOCS_MATCH_ATT = "numDocsMatch";
|
---|
| 64 | protected static final String FREQ_ATT = "freq";
|
---|
| 65 |
|
---|
| 66 | // Elements used in the config file that are specific to this class
|
---|
| 67 | protected static final String DEFAULT_INDEX_ELEM = "defaultIndex";
|
---|
| 68 | protected static final String INDEX_STEM_ELEM = "indexStem";
|
---|
| 69 | protected static final String INDEX_ELEM = "index";
|
---|
| 70 |
|
---|
| 71 | /** the stem used for the index files */
|
---|
| 72 | protected String index_stem = null;
|
---|
| 73 | protected MGWrapper mg_src = null;
|
---|
| 74 |
|
---|
| 75 |
|
---|
| 76 | public AbstractMGSearch()
|
---|
| 77 | {
|
---|
| 78 | this.mg_src = new MGWrapper();
|
---|
| 79 | }
|
---|
| 80 |
|
---|
| 81 | public void cleanUp() {
|
---|
| 82 | super.cleanUp();
|
---|
| 83 | this.mg_src.unloadIndexData();
|
---|
| 84 | }
|
---|
| 85 | public boolean configure(Element info, Element extra_info)
|
---|
| 86 | {
|
---|
| 87 | if (!super.configure(info, extra_info)) {
|
---|
| 88 | return false;
|
---|
| 89 | }
|
---|
| 90 | // do we support any of the extended features?
|
---|
| 91 | does_chunking = true;
|
---|
| 92 |
|
---|
| 93 | // Get the default index out of <defaultIndex> (buildConfig.xml)
|
---|
| 94 | Element def = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_ELEM);
|
---|
| 95 | if (def != null) {
|
---|
| 96 | this.default_index = def.getAttribute(GSXML.NAME_ATT);
|
---|
| 97 | } // otherwise will be "", and the first one will be the default
|
---|
| 98 |
|
---|
| 99 | // the index stem is either the collection name or is specified in the config file
|
---|
| 100 | Element index_stem_elem = (Element) GSXML.getChildByTagName(info, INDEX_STEM_ELEM);
|
---|
| 101 | if (index_stem_elem != null) {
|
---|
| 102 | this.index_stem = index_stem_elem.getAttribute(GSXML.NAME_ATT);
|
---|
| 103 | }
|
---|
| 104 | if (this.index_stem == null || this.index_stem.equals("")) {
|
---|
| 105 | logger.error("AbstractMGSearch.configure(): indexStem element not found, stem will default to collection name");
|
---|
| 106 | this.index_stem = this.cluster_name;
|
---|
| 107 | }
|
---|
| 108 |
|
---|
| 109 | // get display info from extra info
|
---|
| 110 | if (extra_info !=null) {
|
---|
| 111 | Document owner = info.getOwnerDocument();
|
---|
| 112 | // so far we have index specific display elements, and global format elements
|
---|
| 113 | NodeList indexes = info.getElementsByTagName(GSXML.INDEX_ELEM);
|
---|
| 114 | Element config_search = (Element)GSXML.getChildByTagName(extra_info, GSXML.SEARCH_ELEM);
|
---|
| 115 |
|
---|
| 116 | for (int i=0; i<indexes.getLength();i++) {
|
---|
| 117 | Element ind = (Element)indexes.item(i);
|
---|
| 118 | String name = ind.getAttribute(GSXML.NAME_ATT);
|
---|
| 119 | Element node_extra = GSXML.getNamedElement(config_search,
|
---|
| 120 | GSXML.INDEX_ELEM,
|
---|
| 121 | GSXML.NAME_ATT,
|
---|
| 122 | name);
|
---|
| 123 | if (node_extra == null) {
|
---|
| 124 | logger.error("haven't found extra info for index named "+name);
|
---|
| 125 | continue;
|
---|
| 126 | }
|
---|
| 127 |
|
---|
| 128 | // get the display elements if any - displayName
|
---|
| 129 | NodeList display_names = node_extra.getElementsByTagName(GSXML.DISPLAY_TEXT_ELEM);
|
---|
| 130 | if (display_names !=null) {
|
---|
| 131 | for (int j=0; j<display_names.getLength(); j++) {
|
---|
| 132 | Element e = (Element)display_names.item(j);
|
---|
| 133 | ind.appendChild(owner.importNode(e, true));
|
---|
| 134 | }
|
---|
| 135 | }
|
---|
| 136 | } // for each index
|
---|
| 137 | }
|
---|
| 138 | return true;
|
---|
| 139 | }
|
---|
| 140 |
|
---|
| 141 | protected void addCustomQueryParams(Element param_list, String lang)
|
---|
| 142 | {
|
---|
| 143 | createParameter(CASE_PARAM, param_list, lang);
|
---|
| 144 | createParameter(STEM_PARAM, param_list, lang);
|
---|
| 145 | createParameter(MATCH_PARAM, param_list, lang);
|
---|
| 146 | }
|
---|
| 147 |
|
---|
| 148 | protected void getIndexData(ArrayList index_ids, ArrayList index_names, String lang) {
|
---|
| 149 | // the index info -
|
---|
| 150 | Element index_list = (Element)GSXML.getChildByTagName(this.config_info, INDEX_ELEM+GSXML.LIST_MODIFIER);
|
---|
| 151 | NodeList indexes = index_list.getElementsByTagName(INDEX_ELEM);
|
---|
| 152 | int len = indexes.getLength();
|
---|
| 153 | // now add even if there is only one
|
---|
| 154 | for (int i=0; i<len; i++) {
|
---|
| 155 | Element index = (Element)indexes.item(i);
|
---|
| 156 | index_ids.add(index.getAttribute(GSXML.NAME_ATT));
|
---|
| 157 | index_names.add(GSXML.getDisplayText(index, GSXML.DISPLAY_TEXT_NAME, lang, "en"));
|
---|
| 158 |
|
---|
| 159 | }
|
---|
| 160 |
|
---|
| 161 | }
|
---|
| 162 | /** do the actual query */
|
---|
| 163 | protected Element processTextQuery(Element request)
|
---|
| 164 | {
|
---|
| 165 |
|
---|
| 166 | // Create a new (empty) result message
|
---|
| 167 | Element result = this.doc.createElement(GSXML.RESPONSE_ELEM);
|
---|
| 168 | result.setAttribute(GSXML.FROM_ATT, TEXT_QUERY_SERVICE);
|
---|
| 169 | result.setAttribute(GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
|
---|
| 170 |
|
---|
| 171 | // Get the parameters of the request
|
---|
| 172 | Element param_list = (Element) GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
|
---|
| 173 | if (param_list == null) {
|
---|
| 174 | logger.error("TextQuery request had no paramList.");
|
---|
| 175 | return result; // Return the empty result
|
---|
| 176 | }
|
---|
| 177 |
|
---|
| 178 | // Process the request parameters
|
---|
| 179 | HashMap params = GSXML.extractParams(param_list, false);
|
---|
| 180 |
|
---|
| 181 | // Make sure a query has been specified
|
---|
| 182 | String query = (String) params.get(QUERY_PARAM);
|
---|
| 183 | if (query == null || query.equals("")) {
|
---|
| 184 | return result; // Return the empty result
|
---|
| 185 | }
|
---|
| 186 |
|
---|
| 187 | // If an index hasn't been specified, use the default
|
---|
| 188 | String index = (String) params.get(INDEX_PARAM);
|
---|
| 189 | if (index == null) {
|
---|
| 190 | index = this.default_index;
|
---|
| 191 | }
|
---|
| 192 |
|
---|
| 193 | // The location of the MG index and text files
|
---|
| 194 | String basedir = GSFile.collectionBaseDir(this.site_home, this.cluster_name) + File.separatorChar; // Needed for MG
|
---|
| 195 | String textdir = GSFile.collectionTextPath(this.index_stem);
|
---|
| 196 | String indexpath = GSFile.collectionIndexPath(this.index_stem, index);
|
---|
| 197 | this.mg_src.setIndex(indexpath);
|
---|
| 198 |
|
---|
| 199 | // set the mg query parameters to the values the user has specified
|
---|
| 200 | setStandardQueryParams(params);
|
---|
| 201 | this.mg_src.runQuery(basedir, textdir, query);
|
---|
| 202 | MGQueryResult mqr = this.mg_src.getQueryResult();
|
---|
| 203 | if (mqr.isClear()) {
|
---|
| 204 | // something has gone wrong
|
---|
| 205 | GSXML.addError(this.doc, result, "Couldn't query the mg database", GSXML.ERROR_TYPE_SYSTEM);
|
---|
| 206 | return result;
|
---|
| 207 | }
|
---|
| 208 | long totalDocs = mqr.getTotalDocs();
|
---|
| 209 |
|
---|
| 210 | // Get the docnums out, and convert to HASH ids
|
---|
| 211 | Vector docs = mqr.getDocs();
|
---|
| 212 | if (docs.size() == 0) {
|
---|
| 213 | logger.error("No results found...\n");
|
---|
| 214 | }
|
---|
| 215 |
|
---|
| 216 | // Create a metadata list to store information about the query results
|
---|
| 217 | Element metadata_list = this.doc.createElement(GSXML.METADATA_ELEM+GSXML.LIST_MODIFIER);
|
---|
| 218 | result.appendChild(metadata_list);
|
---|
| 219 |
|
---|
| 220 | // Add a metadata element specifying the number of matching documents
|
---|
| 221 | // because teh total number is just the number returned, use numDocsReturned, not numDocsMatched
|
---|
| 222 | GSXML.addMetadata(this.doc, metadata_list, "numDocsReturned", ""+totalDocs);
|
---|
| 223 | // 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
|
---|
| 224 | GSXML.addMetadata(this.doc, metadata_list, "query", query);
|
---|
| 225 |
|
---|
| 226 | if (docs.size() > 0) {
|
---|
| 227 | // Create a document list to store the matching documents, and add them
|
---|
| 228 | Element document_list = this.doc.createElement(GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
|
---|
| 229 | result.appendChild(document_list);
|
---|
| 230 | for (int d = 0; d < docs.size(); d++) {
|
---|
| 231 | long docnum = ((MGDocInfo) docs.elementAt(d)).num_;
|
---|
| 232 | float rank = ((MGDocInfo) docs.elementAt(d)).rank_;
|
---|
| 233 | String doc_id = MGNum2OID(docnum);
|
---|
| 234 | Element doc_node = createDocNode(doc_id, Float.toString(rank));
|
---|
| 235 | document_list.appendChild(doc_node);
|
---|
| 236 | }
|
---|
| 237 | }
|
---|
| 238 |
|
---|
| 239 | // Create a term list to store the term information, and add it
|
---|
| 240 | Element term_list = this.doc.createElement(GSXML.TERM_ELEM+GSXML.LIST_MODIFIER);
|
---|
| 241 | result.appendChild(term_list);
|
---|
| 242 | Vector terms = mqr.getTerms();
|
---|
| 243 | for (int t = 0; t < terms.size(); t++) {
|
---|
| 244 | MGTermInfo term_info = (MGTermInfo) terms.get(t);
|
---|
| 245 |
|
---|
| 246 | String term = term_info.term_;
|
---|
| 247 | int stem_method = term_info.stem_method_;
|
---|
| 248 | Vector equiv_terms = term_info.equiv_terms_;
|
---|
| 249 |
|
---|
| 250 | Element term_elem = this.doc.createElement(GSXML.TERM_ELEM);
|
---|
| 251 | term_elem.setAttribute(GSXML.NAME_ATT, term);
|
---|
| 252 | term_elem.setAttribute(STEM_ATT, "" + stem_method);
|
---|
| 253 |
|
---|
| 254 | Element equiv_term_list = this.doc.createElement(EQUIV_TERM_ELEM+GSXML.LIST_MODIFIER);
|
---|
| 255 | term_elem.appendChild(equiv_term_list);
|
---|
| 256 |
|
---|
| 257 | long total_term_freq = 0;
|
---|
| 258 | for (int et = 0; et < equiv_terms.size(); et++) {
|
---|
| 259 | MGEquivTermInfo equiv_term_info = (MGEquivTermInfo) equiv_terms.get(et);
|
---|
| 260 |
|
---|
| 261 | Element equiv_term_elem = this.doc.createElement(GSXML.TERM_ELEM);
|
---|
| 262 | equiv_term_elem.setAttribute(GSXML.NAME_ATT, equiv_term_info.term_);
|
---|
| 263 | equiv_term_elem.setAttribute(NUM_DOCS_MATCH_ATT, "" + equiv_term_info.match_docs_);
|
---|
| 264 | equiv_term_elem.setAttribute(FREQ_ATT, "" + equiv_term_info.term_freq_);
|
---|
| 265 | equiv_term_list.appendChild(equiv_term_elem);
|
---|
| 266 |
|
---|
| 267 | total_term_freq += equiv_term_info.term_freq_;
|
---|
| 268 | }
|
---|
| 269 |
|
---|
| 270 | term_elem.setAttribute(FREQ_ATT, "" + total_term_freq);
|
---|
| 271 | term_list.appendChild(term_elem);
|
---|
| 272 | }
|
---|
| 273 | return result;
|
---|
| 274 | }
|
---|
| 275 |
|
---|
| 276 | // should probably use a list rather than map
|
---|
| 277 | protected boolean setStandardQueryParams(HashMap params)
|
---|
| 278 | {
|
---|
| 279 | // set the default ones
|
---|
| 280 | this.mg_src.setReturnTerms(true);
|
---|
| 281 | this.mg_src.setCase(true); // turn casefolding on by default
|
---|
| 282 | Set entries = params.entrySet();
|
---|
| 283 | Iterator i = entries.iterator();
|
---|
| 284 | while (i.hasNext()) {
|
---|
| 285 | Map.Entry m = (Map.Entry)i.next();
|
---|
| 286 | String name = (String)m.getKey();
|
---|
| 287 | String value = (String)m.getValue();
|
---|
| 288 |
|
---|
| 289 | if (name.equals(CASE_PARAM)) {
|
---|
| 290 | boolean val = (value.equals(BOOLEAN_PARAM_ON) ? true : false);
|
---|
| 291 | this.mg_src.setCase(val);
|
---|
| 292 | }
|
---|
| 293 | else if (name.equals(STEM_PARAM)) {
|
---|
| 294 | boolean val = (value.equals(BOOLEAN_PARAM_ON) ? true : false);
|
---|
| 295 | this.mg_src.setStem(val);
|
---|
| 296 | }
|
---|
| 297 | else if (name.equals(MATCH_PARAM)) {
|
---|
| 298 | int mode = (value.equals(MATCH_PARAM_ALL) ? 1 : 0);
|
---|
| 299 | this.mg_src.setMatchMode(mode);
|
---|
| 300 | }
|
---|
| 301 | else if (name.equals(MAXDOCS_PARAM)) {
|
---|
| 302 | int docs = Integer.parseInt(value);
|
---|
| 303 | this.mg_src.setMaxDocs(docs);
|
---|
| 304 | } // ignore any others
|
---|
| 305 | }
|
---|
| 306 | return true;
|
---|
| 307 | }
|
---|
| 308 |
|
---|
| 309 | /** convert MG internal id to Greenstone oid */
|
---|
| 310 | abstract protected String MGNum2OID(long docnum);
|
---|
| 311 | }
|
---|