[24394] | 1 | /*
|
---|
| 2 | * AbstractGS2TextSearch.java
|
---|
| 3 | * Copyright (C) 2011 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 |
|
---|
[25851] | 20 | import java.util.ArrayList;
|
---|
| 21 |
|
---|
| 22 | import org.apache.log4j.Logger;
|
---|
[24394] | 23 | import org.greenstone.gsdl3.util.GSXML;
|
---|
[26042] | 24 | import org.greenstone.gsdl3.util.BasicDocumentDatabase;
|
---|
[24394] | 25 | import org.w3c.dom.Document;
|
---|
[24857] | 26 | import org.w3c.dom.Element;
|
---|
[24394] | 27 | import org.w3c.dom.NodeList;
|
---|
| 28 |
|
---|
[24857] | 29 | public abstract class AbstractGS2TextSearch extends AbstractTextSearch
|
---|
[24394] | 30 | {
|
---|
[24857] | 31 | protected static final String EQUIV_TERM_ELEM = "equivTerm";
|
---|
[24394] | 32 |
|
---|
[24857] | 33 | protected static final String STEM_ATT = "stem";
|
---|
| 34 | protected static final String NUM_DOCS_MATCH_ATT = "numDocsMatch";
|
---|
| 35 | protected static final String FREQ_ATT = "freq";
|
---|
[24394] | 36 |
|
---|
[24857] | 37 | // Elements used in the config file that are specific to this class
|
---|
| 38 | protected static final String DEFAULT_INDEX_ELEM = "defaultIndex";
|
---|
| 39 | protected static final String INDEX_STEM_ELEM = "indexStem";
|
---|
| 40 | protected static final String INDEX_ELEM = "index";
|
---|
| 41 | protected static final String DEFAULT_INDEX_SUBCOLLECTION_ELEM = "defaultIndexSubcollection";
|
---|
| 42 | protected static final String DEFAULT_INDEX_LANGUAGE_ELEM = "defaultIndexLanguage";
|
---|
| 43 | protected static final String INDEX_SUBCOLLECTION_ELEM = "indexSubcollection";
|
---|
| 44 | protected static final String INDEX_LANGUAGE_ELEM = "indexLanguage";
|
---|
[24394] | 45 |
|
---|
[24857] | 46 | // Some indexing options
|
---|
| 47 | protected static final String STEMINDEX_OPTION = "stemIndexes";
|
---|
| 48 | protected static final String MAXNUMERIC_OPTION = "maxnumeric";
|
---|
[24394] | 49 |
|
---|
[24857] | 50 | /** the stem used for the index files */
|
---|
| 51 | protected String index_stem = null;
|
---|
[24394] | 52 |
|
---|
[24857] | 53 | // stem indexes available
|
---|
[26344] | 54 | protected boolean does_case = false;
|
---|
| 55 | protected boolean does_stem = false;
|
---|
[24857] | 56 | protected boolean does_accent = false;
|
---|
[24394] | 57 |
|
---|
[24857] | 58 | // maxnumeric -
|
---|
| 59 | protected int maxnumeric = 4;
|
---|
[24394] | 60 |
|
---|
[26042] | 61 | BasicDocumentDatabase gs_doc_db = null;
|
---|
[24394] | 62 |
|
---|
[24857] | 63 | static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.AbstractGS2TextSearch.class.getName());
|
---|
[24394] | 64 |
|
---|
[24857] | 65 | /** constructor */
|
---|
| 66 | public AbstractGS2TextSearch()
|
---|
| 67 | {
|
---|
[24394] | 68 |
|
---|
| 69 | }
|
---|
| 70 |
|
---|
[24857] | 71 | public void cleanUp()
|
---|
| 72 | {
|
---|
| 73 | super.cleanUp();
|
---|
| 74 | this.gs_doc_db.cleanUp();
|
---|
[24394] | 75 | }
|
---|
| 76 |
|
---|
[24857] | 77 | /** configure this service */
|
---|
| 78 | public boolean configure(Element info, Element extra_info)
|
---|
| 79 | {
|
---|
| 80 | if (!super.configure(info, extra_info))
|
---|
| 81 | {
|
---|
| 82 | return false;
|
---|
| 83 | }
|
---|
[24394] | 84 |
|
---|
[24857] | 85 | // find out what kind of database we have
|
---|
| 86 | Element database_type_elem = (Element) GSXML.getChildByTagName(info, GSXML.DATABASE_TYPE_ELEM);
|
---|
| 87 | String database_type = null;
|
---|
| 88 | if (database_type_elem != null)
|
---|
| 89 | {
|
---|
| 90 | database_type = database_type_elem.getAttribute(GSXML.NAME_ATT);
|
---|
[24394] | 91 | }
|
---|
[24857] | 92 | if (database_type == null || database_type.equals(""))
|
---|
| 93 | {
|
---|
| 94 | database_type = "gdbm"; // the default
|
---|
[24394] | 95 | }
|
---|
| 96 |
|
---|
[24857] | 97 | // the index stem is either the collection name or is specified in the config file
|
---|
| 98 | Element index_stem_elem = (Element) GSXML.getChildByTagName(info, INDEX_STEM_ELEM);
|
---|
| 99 | if (index_stem_elem != null)
|
---|
| 100 | {
|
---|
| 101 | this.index_stem = index_stem_elem.getAttribute(GSXML.NAME_ATT);
|
---|
[24394] | 102 | }
|
---|
[24857] | 103 | if (this.index_stem == null || this.index_stem.equals(""))
|
---|
| 104 | {
|
---|
| 105 | logger.warn("indexStem element not found, stem will default to collection name");
|
---|
| 106 | this.index_stem = this.cluster_name;
|
---|
[24394] | 107 | }
|
---|
| 108 |
|
---|
[24857] | 109 | // replaces default AbstractSearch version with one tied to database
|
---|
[28966] | 110 | gs_doc_db = new BasicDocumentDatabase(database_type, this.site_home, this.cluster_name, this.index_stem);
|
---|
[24857] | 111 | if (!gs_doc_db.isValid())
|
---|
| 112 | {
|
---|
| 113 | logger.error("Failed to open Document Database.");
|
---|
| 114 | return false;
|
---|
[24394] | 115 | }
|
---|
[24857] | 116 | this.gs_doc = gs_doc_db;
|
---|
[24394] | 117 |
|
---|
[24857] | 118 | // do we support any of the extended features?
|
---|
[29558] | 119 | //does_chunking = true;
|
---|
[24857] | 120 |
|
---|
| 121 | // Get the default index out of <defaultIndex> (buildConfig.xml)
|
---|
| 122 | Element def = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_ELEM);
|
---|
| 123 | if (def != null)
|
---|
| 124 | {
|
---|
| 125 | this.default_index = def.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 126 | } // otherwise will be "", and the first one will be the default
|
---|
| 127 |
|
---|
| 128 | //get the default indexSubcollection out of <defaultIndexSubcollection> (buildConfig.xml)
|
---|
| 129 | Element defSub = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_SUBCOLLECTION_ELEM);
|
---|
| 130 | if (defSub != null)
|
---|
| 131 | {
|
---|
| 132 | this.default_index_subcollection = defSub.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
[24394] | 133 | }
|
---|
[24857] | 134 |
|
---|
| 135 | //get the default indexLanguage out of <defaultIndexLanguage> (buildConfig.xml)
|
---|
| 136 | Element defLang = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_LANGUAGE_ELEM);
|
---|
| 137 | if (defLang != null)
|
---|
| 138 | {
|
---|
| 139 | this.default_index_language = defLang.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 140 | } //concate defaultIndex + defaultIndexSubcollection + defaultIndexLanguage
|
---|
| 141 |
|
---|
| 142 | // get index options
|
---|
| 143 | Element index_option_list = (Element) GSXML.getChildByTagName(info, GSXML.INDEX_OPTION_ELEM + GSXML.LIST_MODIFIER);
|
---|
| 144 | if (index_option_list != null)
|
---|
| 145 | {
|
---|
| 146 | NodeList options = index_option_list.getElementsByTagName(GSXML.INDEX_OPTION_ELEM);
|
---|
| 147 | for (int i = 0; i < options.getLength(); i++)
|
---|
| 148 | {
|
---|
| 149 | Element opt = (Element) options.item(i);
|
---|
| 150 | String name = opt.getAttribute(GSXML.NAME_ATT);
|
---|
| 151 | String value = opt.getAttribute(GSXML.VALUE_ATT);
|
---|
| 152 | if (name.equals(MAXNUMERIC_OPTION))
|
---|
| 153 | {
|
---|
| 154 | int maxnum = Integer.parseInt(value);
|
---|
| 155 | if (4 <= maxnum && maxnum < 512)
|
---|
| 156 | {
|
---|
| 157 | maxnumeric = maxnum;
|
---|
| 158 | }
|
---|
| 159 | }
|
---|
| 160 | else if (name.equals(STEMINDEX_OPTION))
|
---|
| 161 | {
|
---|
| 162 | int stemindex = Integer.parseInt(value);
|
---|
[26344] | 163 | if ((stemindex & 1) != 0)
|
---|
[24857] | 164 | {
|
---|
[26344] | 165 | does_case = true;
|
---|
[24857] | 166 | }
|
---|
[26344] | 167 | if ((stemindex & 2) != 0)
|
---|
[24857] | 168 | {
|
---|
[26344] | 169 | does_stem = true;
|
---|
[24857] | 170 | }
|
---|
| 171 | if ((stemindex & 4) != 0)
|
---|
| 172 | {
|
---|
| 173 | does_accent = true;
|
---|
| 174 | }
|
---|
| 175 | }
|
---|
| 176 | }
|
---|
[24394] | 177 | }
|
---|
| 178 |
|
---|
[24857] | 179 | // get display info from extra info
|
---|
| 180 | if (extra_info != null)
|
---|
| 181 | {
|
---|
| 182 | Document owner = info.getOwnerDocument();
|
---|
| 183 | Element config_search = (Element) GSXML.getChildByTagName(extra_info, GSXML.SEARCH_ELEM);
|
---|
[24394] | 184 |
|
---|
[26130] | 185 | // so far we have index and indexSubcollection specific display elements, and global format elements
|
---|
| 186 |
|
---|
| 187 | NodeList indexes = info.getElementsByTagName(GSXML.INDEX_ELEM);
|
---|
[24857] | 188 | for (int i = 0; i < indexes.getLength(); i++)
|
---|
| 189 | {
|
---|
| 190 | Element ind = (Element) indexes.item(i);
|
---|
| 191 | String name = ind.getAttribute(GSXML.NAME_ATT);
|
---|
| 192 | Element node_extra = GSXML.getNamedElement(config_search, GSXML.INDEX_ELEM, GSXML.NAME_ATT, name);
|
---|
| 193 | if (node_extra == null)
|
---|
| 194 | {
|
---|
| 195 | logger.error("haven't found extra info for index named " + name);
|
---|
| 196 | continue;
|
---|
| 197 | }
|
---|
[24394] | 198 |
|
---|
[24857] | 199 | // get the display elements if any - displayName
|
---|
| 200 | NodeList display_names = node_extra.getElementsByTagName(GSXML.DISPLAY_TEXT_ELEM);
|
---|
| 201 | if (display_names != null)
|
---|
| 202 | {
|
---|
| 203 | for (int j = 0; j < display_names.getLength(); j++)
|
---|
| 204 | {
|
---|
| 205 | Element e = (Element) display_names.item(j);
|
---|
| 206 | ind.appendChild(owner.importNode(e, true));
|
---|
| 207 | }
|
---|
| 208 | }
|
---|
| 209 | } // for each index
|
---|
[26130] | 210 |
|
---|
| 211 | NodeList indexSubcollections = info.getElementsByTagName(INDEX_SUBCOLLECTION_ELEM); // buildConfig.xml
|
---|
| 212 |
|
---|
| 213 | for (int i = 0; i < indexSubcollections.getLength(); i++)
|
---|
| 214 | {
|
---|
| 215 | Element indexSubcollection = (Element) indexSubcollections.item(i);
|
---|
| 216 | String name = indexSubcollection.getAttribute(GSXML.NAME_ATT);
|
---|
| 217 | Element node_extra = GSXML.getNamedElement(config_search, INDEX_SUBCOLLECTION_ELEM, GSXML.NAME_ATT, name); // collectionConfig.xml
|
---|
| 218 | if (node_extra == null)
|
---|
| 219 | {
|
---|
| 220 | logger.error("haven't found extra info for indexSubCollection named " + name);
|
---|
| 221 | continue;
|
---|
| 222 | }
|
---|
| 223 |
|
---|
| 224 | // get the display elements if any - displayName
|
---|
| 225 | NodeList display_names = node_extra.getElementsByTagName(GSXML.DISPLAY_TEXT_ELEM);
|
---|
| 226 | if (display_names != null)
|
---|
| 227 | {
|
---|
| 228 | for (int j = 0; j < display_names.getLength(); j++)
|
---|
| 229 | {
|
---|
| 230 | Element e = (Element) display_names.item(j);
|
---|
| 231 | indexSubcollection.appendChild(owner.importNode(e, true));
|
---|
| 232 | }
|
---|
| 233 | }
|
---|
| 234 | } // for each indexSubCollection
|
---|
[24857] | 235 | }
|
---|
| 236 | return true;
|
---|
[24394] | 237 | }
|
---|
[24857] | 238 |
|
---|
[25635] | 239 | protected void getIndexData(ArrayList<String> index_ids, ArrayList<String> index_names, String lang)
|
---|
[24857] | 240 | {
|
---|
| 241 | // the index info -
|
---|
| 242 | Element index_list = (Element) GSXML.getChildByTagName(this.config_info, INDEX_ELEM + GSXML.LIST_MODIFIER);
|
---|
| 243 | NodeList indexes = index_list.getElementsByTagName(INDEX_ELEM);
|
---|
| 244 | int len = indexes.getLength();
|
---|
| 245 | // now add even if there is only one
|
---|
| 246 | for (int i = 0; i < len; i++)
|
---|
| 247 | {
|
---|
| 248 | Element index = (Element) indexes.item(i);
|
---|
| 249 | String shortname = index.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 250 | if (shortname.equals(""))
|
---|
| 251 | {
|
---|
| 252 | continue;
|
---|
| 253 | }
|
---|
| 254 | index_ids.add(shortname);
|
---|
[30634] | 255 | String display_name = getDisplayText(index, GSXML.DISPLAY_TEXT_NAME, lang, "en", "metadata_names");
|
---|
[24857] | 256 | if (display_name.equals(""))
|
---|
| 257 | {
|
---|
| 258 | display_name = index.getAttribute(GSXML.NAME_ATT);
|
---|
| 259 | if (display_name.equals(""))
|
---|
| 260 | {
|
---|
| 261 | display_name = shortname;
|
---|
| 262 | }
|
---|
[30634] | 263 | String d_name = getTextString(display_name+".buttonname", lang, "metadata_names");
|
---|
| 264 | if (d_name == null) {
|
---|
| 265 | d_name = getTextString(cleanUpMetadata(display_name)+".buttonname", lang, "metadata_names");
|
---|
| 266 | }
|
---|
| 267 | if (d_name == null) {
|
---|
| 268 | d_name = getTextString(display_name, lang, "metadata_names");
|
---|
| 269 | }
|
---|
| 270 | if (d_name != null) {
|
---|
| 271 |
|
---|
| 272 | display_name = d_name;
|
---|
| 273 | }
|
---|
[24857] | 274 | }
|
---|
| 275 | index_names.add(display_name);
|
---|
| 276 | }
|
---|
[24394] | 277 | }
|
---|
[24857] | 278 |
|
---|
[30634] | 279 | protected String cleanUpMetadata(String meta) {
|
---|
| 280 | // remove namespace, and only take the first item if there is a list of them.
|
---|
| 281 | String[] parts = meta.split("[,;]");
|
---|
| 282 | String cleaned = parts[0];
|
---|
| 283 | if (cleaned.lastIndexOf('.') != -1) {
|
---|
| 284 | cleaned = cleaned.substring(cleaned.lastIndexOf('.')+1);
|
---|
| 285 | }
|
---|
| 286 | return cleaned;
|
---|
| 287 |
|
---|
| 288 | }
|
---|
[25635] | 289 | protected void getIndexSubcollectionData(ArrayList<String> index_sub_ids, ArrayList<String> index_sub_names, String lang)
|
---|
[24857] | 290 | {
|
---|
| 291 | // the index info -
|
---|
| 292 | Element index_sub_list = (Element) GSXML.getChildByTagName(this.config_info, INDEX_SUBCOLLECTION_ELEM + GSXML.LIST_MODIFIER);
|
---|
| 293 | NodeList index_subs = index_sub_list.getElementsByTagName(INDEX_SUBCOLLECTION_ELEM);
|
---|
| 294 | int len = index_subs.getLength();
|
---|
| 295 | // now add even if there is only one
|
---|
| 296 | for (int i = 0; i < len; i++)
|
---|
| 297 | {
|
---|
| 298 | Element indexsub = (Element) index_subs.item(i);
|
---|
| 299 | String shortname = indexsub.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 300 | if (shortname.equals(""))
|
---|
| 301 | {
|
---|
| 302 | continue;
|
---|
| 303 | }
|
---|
| 304 | index_sub_ids.add(shortname);
|
---|
[30634] | 305 | String display_name = getDisplayText(indexsub, GSXML.DISPLAY_TEXT_NAME, lang, "en");
|
---|
[24857] | 306 | if (display_name.equals(""))
|
---|
| 307 | {
|
---|
| 308 | display_name = indexsub.getAttribute(GSXML.NAME_ATT);
|
---|
| 309 | if (display_name.equals(""))
|
---|
| 310 | {
|
---|
| 311 | display_name = shortname;
|
---|
| 312 | }
|
---|
| 313 | }
|
---|
| 314 | index_sub_names.add(display_name);
|
---|
| 315 | }
|
---|
[24394] | 316 | }
|
---|
| 317 |
|
---|
[25635] | 318 | protected void getIndexLanguageData(ArrayList<String> index_lang_ids, ArrayList<String> index_lang_names, String lang)
|
---|
[24857] | 319 | {
|
---|
| 320 | // the index info -
|
---|
| 321 | Element index_lang_list = (Element) GSXML.getChildByTagName(this.config_info, INDEX_LANGUAGE_ELEM + GSXML.LIST_MODIFIER);
|
---|
| 322 | NodeList index_langs = index_lang_list.getElementsByTagName(INDEX_LANGUAGE_ELEM);
|
---|
| 323 | int len = index_langs.getLength();
|
---|
| 324 | // now add even if there is only one
|
---|
| 325 | for (int i = 0; i < len; i++)
|
---|
| 326 | {
|
---|
| 327 | Element indexlang = (Element) index_langs.item(i);
|
---|
| 328 | String shortname = indexlang.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 329 | if (shortname.equals(""))
|
---|
| 330 | {
|
---|
| 331 | continue;
|
---|
| 332 | }
|
---|
| 333 | index_lang_ids.add(shortname);
|
---|
[30634] | 334 | String display_name = getDisplayText(indexlang, GSXML.DISPLAY_TEXT_NAME, lang, "en", "metadata_names");
|
---|
[24857] | 335 | if (display_name.equals(""))
|
---|
| 336 | {
|
---|
| 337 | display_name = indexlang.getAttribute(GSXML.NAME_ATT);
|
---|
| 338 | if (display_name.equals(""))
|
---|
| 339 | {
|
---|
| 340 | display_name = shortname;
|
---|
| 341 | }
|
---|
| 342 | }
|
---|
| 343 | index_lang_names.add(display_name);
|
---|
| 344 | }
|
---|
[24394] | 345 |
|
---|
[24857] | 346 | }
|
---|
[24394] | 347 |
|
---|
[24857] | 348 | protected void addCustomQueryParams(Element param_list, String lang)
|
---|
| 349 | {
|
---|
| 350 | if (this.does_case)
|
---|
| 351 | {
|
---|
[28181] | 352 | createParameter(CASE_PARAM, param_list, lang);//, case_default);
|
---|
[24857] | 353 | }
|
---|
| 354 | if (this.does_stem)
|
---|
| 355 | {
|
---|
[28181] | 356 | createParameter(STEM_PARAM, param_list, lang);//, stem_default);
|
---|
[24857] | 357 | }
|
---|
| 358 | if (this.does_accent)
|
---|
| 359 | {
|
---|
[28181] | 360 | createParameter(ACCENT_PARAM, param_list, lang);//, accent_default);
|
---|
[24857] | 361 | }
|
---|
| 362 | createParameter(MATCH_PARAM, param_list, lang);
|
---|
| 363 | }
|
---|
[24394] | 364 |
|
---|
[24857] | 365 | /** convert indexer internal id to Greenstone oid */
|
---|
| 366 | protected String internalNum2OID(long docnum)
|
---|
| 367 | {
|
---|
| 368 | return this.gs_doc_db.internalNum2OID(docnum);
|
---|
| 369 | }
|
---|
[24394] | 370 |
|
---|
[24857] | 371 | protected String internalNum2OID(String docnum)
|
---|
| 372 | {
|
---|
| 373 | return this.gs_doc_db.internalNum2OID(docnum);
|
---|
| 374 | }
|
---|
| 375 |
|
---|
| 376 | }
|
---|