[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 |
|
---|
| 20 | // Greenstone classes
|
---|
[25851] | 21 | import java.util.ArrayList;
|
---|
| 22 |
|
---|
| 23 | import org.apache.log4j.Logger;
|
---|
[24394] | 24 | import org.greenstone.gsdl3.util.GSXML;
|
---|
| 25 | import org.greenstone.gsdl3.util.SimpleDocumentDatabase;
|
---|
| 26 | import org.w3c.dom.Document;
|
---|
[24857] | 27 | import org.w3c.dom.Element;
|
---|
[24394] | 28 | import org.w3c.dom.NodeList;
|
---|
| 29 |
|
---|
[24857] | 30 | public abstract class AbstractGS2TextSearch extends AbstractTextSearch
|
---|
[24394] | 31 | {
|
---|
[24857] | 32 | protected static final String EQUIV_TERM_ELEM = "equivTerm";
|
---|
[24394] | 33 |
|
---|
[24857] | 34 | protected static final String STEM_ATT = "stem";
|
---|
| 35 | protected static final String NUM_DOCS_MATCH_ATT = "numDocsMatch";
|
---|
| 36 | protected static final String FREQ_ATT = "freq";
|
---|
[24394] | 37 |
|
---|
[24857] | 38 | // Elements used in the config file that are specific to this class
|
---|
| 39 | protected static final String DEFAULT_INDEX_ELEM = "defaultIndex";
|
---|
| 40 | protected static final String INDEX_STEM_ELEM = "indexStem";
|
---|
| 41 | protected static final String INDEX_ELEM = "index";
|
---|
| 42 | protected static final String DEFAULT_INDEX_SUBCOLLECTION_ELEM = "defaultIndexSubcollection";
|
---|
| 43 | protected static final String DEFAULT_INDEX_LANGUAGE_ELEM = "defaultIndexLanguage";
|
---|
| 44 | protected static final String INDEX_SUBCOLLECTION_ELEM = "indexSubcollection";
|
---|
| 45 | protected static final String INDEX_LANGUAGE_ELEM = "indexLanguage";
|
---|
[24394] | 46 |
|
---|
[24857] | 47 | // Some indexing options
|
---|
| 48 | protected static final String STEMINDEX_OPTION = "stemIndexes";
|
---|
| 49 | protected static final String MAXNUMERIC_OPTION = "maxnumeric";
|
---|
[24394] | 50 |
|
---|
[24857] | 51 | /** the stem used for the index files */
|
---|
| 52 | protected String index_stem = null;
|
---|
[24394] | 53 |
|
---|
[24857] | 54 | // stem indexes available
|
---|
| 55 | protected boolean does_case = true;
|
---|
| 56 | protected boolean does_stem = true;
|
---|
| 57 | protected boolean does_accent = false;
|
---|
[24394] | 58 |
|
---|
[24857] | 59 | // maxnumeric -
|
---|
| 60 | protected int maxnumeric = 4;
|
---|
[24394] | 61 |
|
---|
[24857] | 62 | SimpleDocumentDatabase gs_doc_db = null;
|
---|
[24394] | 63 |
|
---|
[24857] | 64 | static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.AbstractGS2TextSearch.class.getName());
|
---|
[24394] | 65 |
|
---|
[24857] | 66 | /** constructor */
|
---|
| 67 | public AbstractGS2TextSearch()
|
---|
| 68 | {
|
---|
[24394] | 69 |
|
---|
| 70 | }
|
---|
| 71 |
|
---|
[24857] | 72 | public void cleanUp()
|
---|
| 73 | {
|
---|
| 74 | super.cleanUp();
|
---|
| 75 | this.gs_doc_db.cleanUp();
|
---|
[24394] | 76 | }
|
---|
| 77 |
|
---|
[24857] | 78 | /** configure this service */
|
---|
| 79 | public boolean configure(Element info, Element extra_info)
|
---|
| 80 | {
|
---|
| 81 | if (!super.configure(info, extra_info))
|
---|
| 82 | {
|
---|
| 83 | return false;
|
---|
| 84 | }
|
---|
[24394] | 85 |
|
---|
[24857] | 86 | // find out what kind of database we have
|
---|
| 87 | Element database_type_elem = (Element) GSXML.getChildByTagName(info, GSXML.DATABASE_TYPE_ELEM);
|
---|
| 88 | String database_type = null;
|
---|
| 89 | if (database_type_elem != null)
|
---|
| 90 | {
|
---|
| 91 | database_type = database_type_elem.getAttribute(GSXML.NAME_ATT);
|
---|
[24394] | 92 | }
|
---|
[24857] | 93 | if (database_type == null || database_type.equals(""))
|
---|
| 94 | {
|
---|
| 95 | database_type = "gdbm"; // the default
|
---|
[24394] | 96 | }
|
---|
| 97 |
|
---|
[24857] | 98 | // the index stem is either the collection name or is specified in the config file
|
---|
| 99 | Element index_stem_elem = (Element) GSXML.getChildByTagName(info, INDEX_STEM_ELEM);
|
---|
| 100 | if (index_stem_elem != null)
|
---|
| 101 | {
|
---|
| 102 | this.index_stem = index_stem_elem.getAttribute(GSXML.NAME_ATT);
|
---|
[24394] | 103 | }
|
---|
[24857] | 104 | if (this.index_stem == null || this.index_stem.equals(""))
|
---|
| 105 | {
|
---|
| 106 | logger.warn("indexStem element not found, stem will default to collection name");
|
---|
| 107 | this.index_stem = this.cluster_name;
|
---|
[24394] | 108 | }
|
---|
| 109 |
|
---|
[24857] | 110 | // replaces default AbstractSearch version with one tied to database
|
---|
| 111 | gs_doc_db = new SimpleDocumentDatabase(this.doc, database_type, this.site_home, this.cluster_name, this.index_stem);
|
---|
| 112 | if (!gs_doc_db.isValid())
|
---|
| 113 | {
|
---|
| 114 | logger.error("Failed to open Document Database.");
|
---|
| 115 | return false;
|
---|
[24394] | 116 | }
|
---|
[24857] | 117 | this.gs_doc = gs_doc_db;
|
---|
[24394] | 118 |
|
---|
[24857] | 119 | // do we support any of the extended features?
|
---|
| 120 | does_chunking = true;
|
---|
| 121 |
|
---|
| 122 | // Get the default index out of <defaultIndex> (buildConfig.xml)
|
---|
| 123 | Element def = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_ELEM);
|
---|
| 124 | if (def != null)
|
---|
| 125 | {
|
---|
| 126 | this.default_index = def.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 127 | } // otherwise will be "", and the first one will be the default
|
---|
| 128 |
|
---|
| 129 | //get the default indexSubcollection out of <defaultIndexSubcollection> (buildConfig.xml)
|
---|
| 130 | Element defSub = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_SUBCOLLECTION_ELEM);
|
---|
| 131 | if (defSub != null)
|
---|
| 132 | {
|
---|
| 133 | this.default_index_subcollection = defSub.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
[24394] | 134 | }
|
---|
[24857] | 135 |
|
---|
| 136 | //get the default indexLanguage out of <defaultIndexLanguage> (buildConfig.xml)
|
---|
| 137 | Element defLang = (Element) GSXML.getChildByTagName(info, DEFAULT_INDEX_LANGUAGE_ELEM);
|
---|
| 138 | if (defLang != null)
|
---|
| 139 | {
|
---|
| 140 | this.default_index_language = defLang.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 141 | } //concate defaultIndex + defaultIndexSubcollection + defaultIndexLanguage
|
---|
| 142 |
|
---|
| 143 | // get index options
|
---|
| 144 | Element index_option_list = (Element) GSXML.getChildByTagName(info, GSXML.INDEX_OPTION_ELEM + GSXML.LIST_MODIFIER);
|
---|
| 145 | if (index_option_list != null)
|
---|
| 146 | {
|
---|
| 147 | NodeList options = index_option_list.getElementsByTagName(GSXML.INDEX_OPTION_ELEM);
|
---|
| 148 | for (int i = 0; i < options.getLength(); i++)
|
---|
| 149 | {
|
---|
| 150 | Element opt = (Element) options.item(i);
|
---|
| 151 | String name = opt.getAttribute(GSXML.NAME_ATT);
|
---|
| 152 | String value = opt.getAttribute(GSXML.VALUE_ATT);
|
---|
| 153 | if (name.equals(MAXNUMERIC_OPTION))
|
---|
| 154 | {
|
---|
| 155 | int maxnum = Integer.parseInt(value);
|
---|
| 156 | if (4 <= maxnum && maxnum < 512)
|
---|
| 157 | {
|
---|
| 158 | maxnumeric = maxnum;
|
---|
| 159 | }
|
---|
| 160 | }
|
---|
| 161 | else if (name.equals(STEMINDEX_OPTION))
|
---|
| 162 | {
|
---|
| 163 | int stemindex = Integer.parseInt(value);
|
---|
| 164 | // stem and case are true by default, accent folding false by default
|
---|
| 165 | if ((stemindex & 1) == 0)
|
---|
| 166 | {
|
---|
| 167 | does_case = false;
|
---|
| 168 | }
|
---|
| 169 | if ((stemindex & 2) == 0)
|
---|
| 170 | {
|
---|
| 171 | does_stem = false;
|
---|
| 172 | }
|
---|
| 173 | if ((stemindex & 4) != 0)
|
---|
| 174 | {
|
---|
| 175 | does_accent = true;
|
---|
| 176 | }
|
---|
| 177 | }
|
---|
| 178 | }
|
---|
[24394] | 179 | }
|
---|
| 180 |
|
---|
[24857] | 181 | // get display info from extra info
|
---|
| 182 | if (extra_info != null)
|
---|
| 183 | {
|
---|
| 184 | Document owner = info.getOwnerDocument();
|
---|
| 185 | // so far we have index specific display elements, and global format elements
|
---|
| 186 | NodeList indexes = info.getElementsByTagName(GSXML.INDEX_ELEM);
|
---|
| 187 | Element config_search = (Element) GSXML.getChildByTagName(extra_info, GSXML.SEARCH_ELEM);
|
---|
[24394] | 188 |
|
---|
[24857] | 189 | for (int i = 0; i < indexes.getLength(); i++)
|
---|
| 190 | {
|
---|
| 191 | Element ind = (Element) indexes.item(i);
|
---|
| 192 | String name = ind.getAttribute(GSXML.NAME_ATT);
|
---|
| 193 | Element node_extra = GSXML.getNamedElement(config_search, GSXML.INDEX_ELEM, GSXML.NAME_ATT, name);
|
---|
| 194 | if (node_extra == null)
|
---|
| 195 | {
|
---|
| 196 | logger.error("haven't found extra info for index named " + name);
|
---|
| 197 | continue;
|
---|
| 198 | }
|
---|
[24394] | 199 |
|
---|
[24857] | 200 | // get the display elements if any - displayName
|
---|
| 201 | NodeList display_names = node_extra.getElementsByTagName(GSXML.DISPLAY_TEXT_ELEM);
|
---|
| 202 | if (display_names != null)
|
---|
| 203 | {
|
---|
| 204 | for (int j = 0; j < display_names.getLength(); j++)
|
---|
| 205 | {
|
---|
| 206 | Element e = (Element) display_names.item(j);
|
---|
| 207 | ind.appendChild(owner.importNode(e, true));
|
---|
| 208 | }
|
---|
| 209 | }
|
---|
| 210 | } // for each index
|
---|
| 211 | }
|
---|
| 212 | return true;
|
---|
[24394] | 213 | }
|
---|
[24857] | 214 |
|
---|
[25635] | 215 | protected void getIndexData(ArrayList<String> index_ids, ArrayList<String> index_names, String lang)
|
---|
[24857] | 216 | {
|
---|
| 217 | // the index info -
|
---|
| 218 | Element index_list = (Element) GSXML.getChildByTagName(this.config_info, INDEX_ELEM + GSXML.LIST_MODIFIER);
|
---|
| 219 | NodeList indexes = index_list.getElementsByTagName(INDEX_ELEM);
|
---|
| 220 | int len = indexes.getLength();
|
---|
| 221 | // now add even if there is only one
|
---|
| 222 | for (int i = 0; i < len; i++)
|
---|
| 223 | {
|
---|
| 224 | Element index = (Element) indexes.item(i);
|
---|
| 225 | String shortname = index.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 226 | if (shortname.equals(""))
|
---|
| 227 | {
|
---|
| 228 | continue;
|
---|
| 229 | }
|
---|
| 230 | index_ids.add(shortname);
|
---|
| 231 | String display_name = GSXML.getDisplayText(index, GSXML.DISPLAY_TEXT_NAME, lang, "en");
|
---|
| 232 | if (display_name.equals(""))
|
---|
| 233 | {
|
---|
| 234 | display_name = index.getAttribute(GSXML.NAME_ATT);
|
---|
| 235 | if (display_name.equals(""))
|
---|
| 236 | {
|
---|
| 237 | display_name = shortname;
|
---|
| 238 | }
|
---|
| 239 | }
|
---|
| 240 | index_names.add(display_name);
|
---|
| 241 | }
|
---|
[24394] | 242 | }
|
---|
[24857] | 243 |
|
---|
[25635] | 244 | protected void getIndexSubcollectionData(ArrayList<String> index_sub_ids, ArrayList<String> index_sub_names, String lang)
|
---|
[24857] | 245 | {
|
---|
| 246 | // the index info -
|
---|
| 247 | Element index_sub_list = (Element) GSXML.getChildByTagName(this.config_info, INDEX_SUBCOLLECTION_ELEM + GSXML.LIST_MODIFIER);
|
---|
| 248 | NodeList index_subs = index_sub_list.getElementsByTagName(INDEX_SUBCOLLECTION_ELEM);
|
---|
| 249 | int len = index_subs.getLength();
|
---|
| 250 | // now add even if there is only one
|
---|
| 251 | for (int i = 0; i < len; i++)
|
---|
| 252 | {
|
---|
| 253 | Element indexsub = (Element) index_subs.item(i);
|
---|
| 254 | String shortname = indexsub.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 255 | if (shortname.equals(""))
|
---|
| 256 | {
|
---|
| 257 | continue;
|
---|
| 258 | }
|
---|
| 259 | index_sub_ids.add(shortname);
|
---|
| 260 | String display_name = GSXML.getDisplayText(indexsub, GSXML.DISPLAY_TEXT_NAME, lang, "en");
|
---|
| 261 | if (display_name.equals(""))
|
---|
| 262 | {
|
---|
| 263 | display_name = indexsub.getAttribute(GSXML.NAME_ATT);
|
---|
| 264 | if (display_name.equals(""))
|
---|
| 265 | {
|
---|
| 266 | display_name = shortname;
|
---|
| 267 | }
|
---|
| 268 | }
|
---|
| 269 | index_sub_names.add(display_name);
|
---|
| 270 | }
|
---|
[24394] | 271 | }
|
---|
| 272 |
|
---|
[25635] | 273 | protected void getIndexLanguageData(ArrayList<String> index_lang_ids, ArrayList<String> index_lang_names, String lang)
|
---|
[24857] | 274 | {
|
---|
| 275 | // the index info -
|
---|
| 276 | Element index_lang_list = (Element) GSXML.getChildByTagName(this.config_info, INDEX_LANGUAGE_ELEM + GSXML.LIST_MODIFIER);
|
---|
| 277 | NodeList index_langs = index_lang_list.getElementsByTagName(INDEX_LANGUAGE_ELEM);
|
---|
| 278 | int len = index_langs.getLength();
|
---|
| 279 | // now add even if there is only one
|
---|
| 280 | for (int i = 0; i < len; i++)
|
---|
| 281 | {
|
---|
| 282 | Element indexlang = (Element) index_langs.item(i);
|
---|
| 283 | String shortname = indexlang.getAttribute(GSXML.SHORTNAME_ATT);
|
---|
| 284 | if (shortname.equals(""))
|
---|
| 285 | {
|
---|
| 286 | continue;
|
---|
| 287 | }
|
---|
| 288 | index_lang_ids.add(shortname);
|
---|
| 289 | String display_name = GSXML.getDisplayText(indexlang, GSXML.DISPLAY_TEXT_NAME, lang, "en");
|
---|
| 290 | if (display_name.equals(""))
|
---|
| 291 | {
|
---|
| 292 | display_name = indexlang.getAttribute(GSXML.NAME_ATT);
|
---|
| 293 | if (display_name.equals(""))
|
---|
| 294 | {
|
---|
| 295 | display_name = shortname;
|
---|
| 296 | }
|
---|
| 297 | }
|
---|
| 298 | index_lang_names.add(display_name);
|
---|
| 299 | }
|
---|
[24394] | 300 |
|
---|
[24857] | 301 | }
|
---|
[24394] | 302 |
|
---|
[24857] | 303 | protected void addCustomQueryParams(Element param_list, String lang)
|
---|
| 304 | {
|
---|
| 305 | if (this.does_case)
|
---|
| 306 | {
|
---|
| 307 | // gs2 has case on by default
|
---|
| 308 | createParameter(CASE_PARAM, param_list, lang, BOOLEAN_PARAM_ON);
|
---|
| 309 | }
|
---|
| 310 | if (this.does_stem)
|
---|
| 311 | {
|
---|
| 312 | // but stem is off by default
|
---|
| 313 | createParameter(STEM_PARAM, param_list, lang, BOOLEAN_PARAM_OFF);
|
---|
| 314 | }
|
---|
| 315 | if (this.does_accent)
|
---|
| 316 | {
|
---|
| 317 | // and so is accent folding
|
---|
| 318 | createParameter(ACCENT_PARAM, param_list, lang, BOOLEAN_PARAM_OFF);
|
---|
| 319 | }
|
---|
| 320 | createParameter(MATCH_PARAM, param_list, lang);
|
---|
| 321 | }
|
---|
[24394] | 322 |
|
---|
[24857] | 323 | /** convert indexer internal id to Greenstone oid */
|
---|
| 324 | protected String internalNum2OID(long docnum)
|
---|
| 325 | {
|
---|
| 326 | return this.gs_doc_db.internalNum2OID(docnum);
|
---|
| 327 | }
|
---|
[24394] | 328 |
|
---|
[24857] | 329 | protected String internalNum2OID(String docnum)
|
---|
| 330 | {
|
---|
| 331 | return this.gs_doc_db.internalNum2OID(docnum);
|
---|
| 332 | }
|
---|
| 333 |
|
---|
| 334 | }
|
---|