[8027] | 1 | /**********************************************************************
|
---|
| 2 | *
|
---|
| 3 | * lucenequeryfilter.cpp --
|
---|
| 4 | * Copyright (C) 1999 The New Zealand Digital Library Project
|
---|
| 5 | *
|
---|
| 6 | * A component of the Greenstone digital library software
|
---|
| 7 | * from the New Zealand Digital Library Project at the
|
---|
| 8 | * University of Waikato, New Zealand.
|
---|
| 9 | *
|
---|
| 10 | * This program is free software; you can redistribute it and/or modify
|
---|
| 11 | * it under the terms of the GNU General Public License as published by
|
---|
| 12 | * the Free Software Foundation; either version 2 of the License, or
|
---|
| 13 | * (at your option) any later version.
|
---|
| 14 | *
|
---|
| 15 | * This program is distributed in the hope that it will be useful,
|
---|
| 16 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 17 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 18 | * GNU General Public License for more details.
|
---|
| 19 | *
|
---|
| 20 | * You should have received a copy of the GNU General Public License
|
---|
| 21 | * along with this program; if not, write to the Free Software
|
---|
| 22 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 23 | *
|
---|
| 24 | *********************************************************************/
|
---|
| 25 |
|
---|
| 26 | #include "lucenequeryfilter.h"
|
---|
| 27 | #include "fileutil.h"
|
---|
| 28 | #include "lucenesearch.h"
|
---|
| 29 |
|
---|
| 30 | lucenequeryfilterclass::lucenequeryfilterclass ()
|
---|
[27064] | 31 | : fieldedqueryfilterclass() {
|
---|
[8027] | 32 |
|
---|
| 33 |
|
---|
| 34 | FilterOption_t filtopt;
|
---|
| 35 |
|
---|
[27064] | 36 | // -- onePerQuery SortField, enumerated, used to list available sorting fields
|
---|
[8027] | 37 | filtopt.clear();
|
---|
[27064] | 38 | filtopt.name = "SortField";
|
---|
[8027] | 39 | filtopt.type = FilterOption_t::enumeratedt;
|
---|
[27064] | 40 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
[20727] | 41 | filtopt.defaultValue = "";
|
---|
[27064] | 42 | filterOptions["SortField"] = filtopt;
|
---|
[8027] | 43 |
|
---|
[27064] | 44 | // -- onePerQuery SortOder enumerated (0=ascending, 1=descending)
|
---|
[8027] | 45 | filtopt.clear();
|
---|
[27064] | 46 | filtopt.name = "SortOrder";
|
---|
[8027] | 47 | filtopt.type = FilterOption_t::enumeratedt;
|
---|
[27064] | 48 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
| 49 | filtopt.defaultValue = "ascending";
|
---|
| 50 | filtopt.validValues.push_back("ascending");
|
---|
| 51 | filtopt.validValues.push_back("descending");
|
---|
| 52 | filterOptions["SortOrder"] = filtopt;
|
---|
| 53 |
|
---|
| 54 | // -- onePerQuery Fuzziness string 0.0-1.0
|
---|
| 55 | filtopt.clear();
|
---|
| 56 | filtopt.name = "Fuzziness";
|
---|
| 57 | filtopt.type = FilterOption_t::stringt;
|
---|
| 58 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
[8027] | 59 | filtopt.defaultValue = "";
|
---|
[27064] | 60 | filterOptions["Fuzziness"] = filtopt;
|
---|
[8027] | 61 |
|
---|
[27064] | 62 | // -- onePerQuery FilterString string
|
---|
| 63 | filtopt.clear();
|
---|
| 64 | filtopt.name = "FilterString";
|
---|
| 65 | filtopt.type = FilterOption_t::stringt;
|
---|
| 66 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
| 67 | filtopt.defaultValue = "";
|
---|
| 68 | filterOptions["FilterString"] = filtopt;
|
---|
[8027] | 69 | }
|
---|
| 70 |
|
---|
| 71 | lucenequeryfilterclass::~lucenequeryfilterclass () {
|
---|
| 72 | }
|
---|
| 73 |
|
---|
| 74 |
|
---|
| 75 |
|
---|
| 76 | void lucenequeryfilterclass::configure (const text_t &key, const text_tarray &cfgline) {
|
---|
[27064] | 77 | fieldedqueryfilterclass::configure(key, cfgline);
|
---|
[8027] | 78 |
|
---|
[27064] | 79 | if (key == "textlevel") {
|
---|
| 80 | ((lucenesearchclass *)textsearchptr)->set_text_level(cfgline[0]);
|
---|
[8027] | 81 | }
|
---|
[28955] | 82 | else if (key == "indexsortfieldmap") {
|
---|
| 83 | sortfieldmap.importmap (cfgline);
|
---|
| 84 | }
|
---|
[27361] | 85 | else if (key == "indexsortfields") {
|
---|
| 86 | filterOptions["SortField"].validValues.erase(filterOptions["SortField"].validValues.begin(), filterOptions["SortField"].validValues.end());
|
---|
| 87 | text_tarray::const_iterator here = cfgline.begin();
|
---|
| 88 | text_tarray::const_iterator end = cfgline.end();
|
---|
| 89 | while (here != end) {
|
---|
| 90 | if (!(*here).empty()) {
|
---|
| 91 | filterOptions["SortField"].validValues.push_back(*here);
|
---|
| 92 | }
|
---|
| 93 | ++here;
|
---|
| 94 | }
|
---|
| 95 | }
|
---|
[28955] | 96 | else if (key == "defaultsortfield") {
|
---|
| 97 | sortfieldmap.from2to (cfgline[0], filterOptions["SortField"].defaultValue);
|
---|
| 98 | }
|
---|
[8027] | 99 | }
|
---|
| 100 |
|
---|
[20727] | 101 | bool lucenequeryfilterclass::init (ostream &logout) {
|
---|
| 102 |
|
---|
[27064] | 103 | if (!fieldedqueryfilterclass::init(logout)) {
|
---|
[20727] | 104 | return false;
|
---|
| 105 | }
|
---|
| 106 |
|
---|
[27361] | 107 | if (filterOptions["SortField"].defaultValue.empty() && filterOptions["SortField"].validValues.begin() != filterOptions["SortField"].validValues.end() && !filterOptions["SortField"].validValues[0].empty()) {
|
---|
| 108 | filterOptions["SortField"].defaultValue = filterOptions["SortField"].validValues[0];
|
---|
[20727] | 109 | }
|
---|
[27361] | 110 |
|
---|
| 111 | return true;
|
---|
[20727] | 112 | }
|
---|
| 113 |
|
---|
[27064] | 114 | void lucenequeryfilterclass::set_queryparam_defaults(queryparamclass &query ) {
|
---|
| 115 |
|
---|
| 116 | fieldedqueryfilterclass::set_queryparam_defaults(query);
|
---|
| 117 | query.filterstring = filterOptions["FilterString"].defaultValue;
|
---|
| 118 | query.sortfield = filterOptions["SortField"].defaultValue;
|
---|
| 119 | query.sortorder = (filterOptions["SortOrder"].defaultValue == "descending");
|
---|
| 120 | query.fuzziness = filterOptions["Fuzziness"].defaultValue;
|
---|
| 121 |
|
---|
| 122 | }
|
---|
| 123 |
|
---|
[27084] | 124 | bool lucenequeryfilterclass::set_queryparam_field(const OptionValue_t &option, queryparamclass &query) {
|
---|
[27064] | 125 |
|
---|
[27084] | 126 | if (option.name == "FilterString") {
|
---|
| 127 | query.filterstring = option.value;
|
---|
| 128 | return true;
|
---|
| 129 | }
|
---|
| 130 | if (option.name == "SortField") {
|
---|
| 131 | query.sortfield = option.value;
|
---|
| 132 | return true;
|
---|
| 133 | }
|
---|
| 134 | if (option.name == "SortOrder") {
|
---|
| 135 | query.sortorder = (option.value == "descending");
|
---|
| 136 | return true;
|
---|
| 137 | }
|
---|
| 138 | if (option.name == "Fuzziness") {
|
---|
| 139 | query.fuzziness = option.value;
|
---|
| 140 | return true;
|
---|
| 141 | }
|
---|
| 142 | return fieldedqueryfilterclass::set_queryparam_field(option, query);
|
---|
| 143 | }
|
---|
| 144 |
|
---|
[8027] | 145 | void lucenequeryfilterclass::filter(const FilterRequest_t &request,
|
---|
| 146 | FilterResponse_t &response,
|
---|
| 147 | comerror_t &err, ostream &logout) {
|
---|
| 148 |
|
---|
| 149 | outconvertclass text_t2ascii;
|
---|
| 150 |
|
---|
| 151 | response.clear ();
|
---|
| 152 | err = noError;
|
---|
[15558] | 153 | if (db_ptr == NULL) {
|
---|
[8027] | 154 | // most likely a configuration problem
|
---|
| 155 | logout << text_t2ascii
|
---|
[15558] | 156 | << "configuration error: queryfilter contains a null dbclass\n\n";
|
---|
[8027] | 157 | err = configurationError;
|
---|
| 158 | return;
|
---|
| 159 | }
|
---|
| 160 | if (textsearchptr == NULL) {
|
---|
| 161 | // most likely a configuration problem
|
---|
| 162 | logout << text_t2ascii
|
---|
[9110] | 163 | << "configuration error: queryfilter contains a null textsearchclass (lucene)\n\n";
|
---|
[8027] | 164 | err = configurationError;
|
---|
| 165 | return;
|
---|
| 166 | }
|
---|
| 167 | if (full_text_browse(request.filterResultOptions)) {
|
---|
| 168 | browsefilter(request, response, err, logout);
|
---|
| 169 | return;
|
---|
| 170 | }
|
---|
| 171 | // open the database
|
---|
[15558] | 172 | db_ptr->setlogout(&logout);
|
---|
| 173 | if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
|
---|
| 174 | // most likely a system problem (we have already checked that the database exists)
|
---|
[8027] | 175 | logout << text_t2ascii
|
---|
[15558] | 176 | << "system problem: open on database \"" << db_filename << "\" failed\n\n";
|
---|
[8027] | 177 | err = systemProblem;
|
---|
| 178 | return;
|
---|
| 179 | }
|
---|
| 180 |
|
---|
| 181 |
|
---|
| 182 | // get the query parameters
|
---|
| 183 | int startresults, endresults;
|
---|
| 184 | vector<queryparamclass> queryfilterparams;
|
---|
| 185 | parse_query_params (request, queryfilterparams, startresults,
|
---|
[27064] | 186 | endresults, logout);
|
---|
[8027] | 187 |
|
---|
| 188 |
|
---|
| 189 | // do query
|
---|
| 190 | queryresultsclass queryresults;
|
---|
| 191 | do_multi_query (request, queryfilterparams, queryresults, err, logout);
|
---|
[12421] | 192 | response.error_message = queryresults.error_message;
|
---|
[8027] | 193 | if (err != noError) return;
|
---|
[12421] | 194 |
|
---|
[8027] | 195 | // assemble document results
|
---|
[16947] | 196 | if (need_matching_docs (request.filterResultOptions))
|
---|
| 197 | {
|
---|
| 198 | // Loop through the query results (ordered by ranking)
|
---|
[8027] | 199 | int resultnum = 1;
|
---|
[16947] | 200 | vector<text_t>::iterator docorder_iterator = queryresults.docs.docorder.begin();
|
---|
| 201 | while (docorder_iterator != queryresults.docs.docorder.end())
|
---|
| 202 | {
|
---|
| 203 | text_t doc_OID = (*docorder_iterator);
|
---|
| 204 | // logout << "Matching doc OID: " << doc_OID << endl;
|
---|
[8027] | 205 |
|
---|
[16947] | 206 | // Make sure this result is in the docset, and either in the request set or the request set is empty
|
---|
| 207 | docresultmap::iterator doc_result = queryresults.docs.docset.find (doc_OID);
|
---|
| 208 | if (doc_result != queryresults.docs.docset.end() && (request.docSet.empty() || in_set(request.docSet, doc_OID)))
|
---|
[12655] | 209 | {
|
---|
[16947] | 210 | // Add the matching document
|
---|
| 211 | ResultDocInfo_t resultdoc;
|
---|
| 212 | resultdoc.OID = doc_OID;
|
---|
| 213 | resultdoc.result_num = resultnum;
|
---|
| 214 | resultdoc.ranking = (int)((*doc_result).second.docweight * 10000.0 + 0.5);
|
---|
| 215 | resultdoc.num_terms_matched = (*doc_result).second.num_query_terms_matched;
|
---|
| 216 | response.docInfo.push_back (resultdoc);
|
---|
[8027] | 217 |
|
---|
[16947] | 218 | resultnum++;
|
---|
| 219 | }
|
---|
[8027] | 220 |
|
---|
[16947] | 221 | docorder_iterator++;
|
---|
| 222 | }
|
---|
| 223 | }
|
---|
[12655] | 224 |
|
---|
[8027] | 225 | // assemble the term results
|
---|
| 226 | if (need_term_info(request.filterResultOptions)) {
|
---|
| 227 | // note: the terms have already been sorted and uniqued - ?? have they??
|
---|
| 228 |
|
---|
| 229 | TermInfo_t terminfo;
|
---|
| 230 | bool terms_first = true;
|
---|
| 231 |
|
---|
| 232 | termfreqclassarray::iterator terms_here = queryresults.terms.begin();
|
---|
| 233 | termfreqclassarray::iterator terms_end = queryresults.terms.end();
|
---|
| 234 |
|
---|
| 235 | while (terms_here != terms_end) {
|
---|
| 236 | terminfo.clear();
|
---|
| 237 | terminfo.term = (*terms_here).termstr;
|
---|
| 238 | terminfo.freq = (*terms_here).termfreq;
|
---|
[13063] | 239 | // lucene doesn't return any termvariants at this stage,
|
---|
| 240 | // so make sure the original term is set
|
---|
| 241 | terminfo.matchTerms.push_back(terminfo.term);
|
---|
| 242 |
|
---|
[8027] | 243 | // this bit gets the matchTerms ie the equivalent (stem/casefold) terms
|
---|
| 244 | if (terms_first) {
|
---|
| 245 | text_tset::iterator termvariants_here = queryresults.termvariants.begin();
|
---|
| 246 | text_tset::iterator termvariants_end = queryresults.termvariants.end();
|
---|
| 247 | while (termvariants_here != termvariants_end) {
|
---|
| 248 | terminfo.matchTerms.push_back (*termvariants_here);
|
---|
[9620] | 249 | ++termvariants_here;
|
---|
[8027] | 250 | }
|
---|
| 251 | }
|
---|
| 252 | terms_first = false;
|
---|
| 253 |
|
---|
| 254 | response.termInfo.push_back (terminfo);
|
---|
| 255 |
|
---|
[9620] | 256 | ++terms_here;
|
---|
[8027] | 257 | }
|
---|
[12380] | 258 |
|
---|
| 259 | // add the stop words
|
---|
| 260 | text_tset::iterator stopwords_here = queryresults.stopwords.begin();
|
---|
| 261 | text_tset::iterator stopwords_end = queryresults.stopwords.end();
|
---|
| 262 | while (stopwords_here != stopwords_end) {
|
---|
| 263 | response.stopwords.insert(*stopwords_here);
|
---|
| 264 | ++stopwords_here;
|
---|
| 265 | }
|
---|
[8027] | 266 | }
|
---|
| 267 |
|
---|
[15558] | 268 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
[8027] | 269 | response.numDocs = queryresults.docs_matched;
|
---|
| 270 | response.isApprox = queryresults.is_approx;
|
---|
| 271 | }
|
---|
| 272 |
|
---|
| 273 | void lucenequeryfilterclass::browsefilter(const FilterRequest_t &request,
|
---|
| 274 | FilterResponse_t &response,
|
---|
| 275 | comerror_t &err, ostream &logout) {
|
---|
| 276 |
|
---|
| 277 | outconvertclass text_t2ascii;
|
---|
| 278 |
|
---|
| 279 | // get the query parameters
|
---|
| 280 | int startresults, endresults;
|
---|
| 281 |
|
---|
| 282 | vector<queryparamclass> queryfilterparams;
|
---|
| 283 | parse_query_params (request, queryfilterparams, startresults,
|
---|
[27064] | 284 | endresults, logout);
|
---|
[8027] | 285 |
|
---|
[22050] | 286 | vector<queryparamclass>::const_iterator query_here = queryfilterparams.begin();
|
---|
[8027] | 287 |
|
---|
| 288 | // do query
|
---|
| 289 | queryresultsclass queryresults;
|
---|
| 290 | queryresults.clear();
|
---|
| 291 |
|
---|
| 292 | int numDocs = endresults-startresults;
|
---|
| 293 | textsearchptr->setcollectdir (collectdir);
|
---|
| 294 |
|
---|
| 295 | if (!((lucenesearchclass*)textsearchptr)->browse_search((*query_here), startresults, numDocs, queryresults)) {
|
---|
| 296 | // most likely a system problem
|
---|
| 297 | logout << text_t2ascii
|
---|
| 298 | << "system problem: could not do full text browse with lucene for index \""
|
---|
| 299 | << (*query_here).index << (*query_here).subcollection
|
---|
| 300 | << (*query_here).language << "\".\n\n";
|
---|
| 301 | err = systemProblem;
|
---|
| 302 | return;
|
---|
| 303 | }
|
---|
| 304 |
|
---|
| 305 | // assemble the term results
|
---|
| 306 | TermInfo_t terminfo;
|
---|
| 307 |
|
---|
| 308 | termfreqclassarray::iterator terms_here = queryresults.terms.begin();
|
---|
| 309 | termfreqclassarray::iterator terms_end = queryresults.terms.end();
|
---|
| 310 |
|
---|
| 311 | while (terms_here != terms_end) {
|
---|
| 312 | terminfo.clear();
|
---|
| 313 | terminfo.term = (*terms_here).termstr;
|
---|
| 314 | terminfo.freq = (*terms_here).termfreq;
|
---|
| 315 |
|
---|
| 316 | response.termInfo.push_back (terminfo);
|
---|
| 317 |
|
---|
[9620] | 318 | ++terms_here;
|
---|
[8027] | 319 | }
|
---|
| 320 |
|
---|
| 321 |
|
---|
| 322 | }
|
---|
| 323 |
|
---|
[15558] | 324 | // lucenesearchptr and db_ptr are assumed to be valid
|
---|
[8027] | 325 | void lucenequeryfilterclass::do_multi_query (const FilterRequest_t &request,
|
---|
| 326 | const vector<queryparamclass> &query_params,
|
---|
| 327 | queryresultsclass &multiresults,
|
---|
| 328 | comerror_t &err, ostream &logout) {
|
---|
| 329 | outconvertclass text_t2ascii;
|
---|
| 330 |
|
---|
| 331 | err = noError;
|
---|
| 332 | textsearchptr->setcollectdir (collectdir);
|
---|
| 333 | multiresults.clear();
|
---|
| 334 |
|
---|
| 335 | vector<queryparamclass>::const_iterator query_here = query_params.begin();
|
---|
| 336 | vector<queryparamclass>::const_iterator query_end = query_params.end();
|
---|
| 337 | while (query_here != query_end) {
|
---|
| 338 | queryresultsclass thisqueryresults;
|
---|
| 339 | if (!textsearchptr->search((*query_here), thisqueryresults)) {
|
---|
| 340 | // most likely a system problem
|
---|
| 341 | logout << text_t2ascii
|
---|
| 342 | << "system problem: could not do search with lucene for index \""
|
---|
[9090] | 343 | << (*query_here).index << (*query_here).level
|
---|
| 344 | << (*query_here).subcollection
|
---|
[8027] | 345 | << (*query_here).language << "\".\n\n";
|
---|
| 346 | err = systemProblem;
|
---|
| 347 | return;
|
---|
| 348 | }
|
---|
| 349 |
|
---|
| 350 | // check for syntax error
|
---|
| 351 | if (thisqueryresults.syntax_error==true) {
|
---|
| 352 | logout << text_t2ascii
|
---|
| 353 | << "syntax problem: invalid query string \""
|
---|
| 354 | << (*query_here).querystring<<"\".\n";
|
---|
| 355 | err = syntaxError;
|
---|
| 356 | return;
|
---|
| 357 | }
|
---|
| 358 | // combine the results
|
---|
| 359 | if (need_matching_docs (request.filterResultOptions)) {
|
---|
| 360 |
|
---|
| 361 | if (query_params.size() == 1) {
|
---|
[12421] | 362 | multiresults.error_message = thisqueryresults.error_message;
|
---|
[8027] | 363 | multiresults.docs = thisqueryresults.docs; // just one set of results
|
---|
| 364 | multiresults.docs_matched = thisqueryresults.docs_matched;
|
---|
| 365 | multiresults.is_approx = thisqueryresults.is_approx;
|
---|
| 366 |
|
---|
| 367 | } else {
|
---|
| 368 | if ((*query_here).combinequery == "and") {
|
---|
| 369 | multiresults.docs.combine_and (thisqueryresults.docs);
|
---|
| 370 | } else if ((*query_here).combinequery == "or") {
|
---|
| 371 | multiresults.docs.combine_or (thisqueryresults.docs);
|
---|
| 372 | } else if ((*query_here).combinequery == "not") {
|
---|
| 373 | multiresults.docs.combine_not (thisqueryresults.docs);
|
---|
| 374 | }
|
---|
| 375 | multiresults.docs_matched = multiresults.docs.docset.size();
|
---|
| 376 | multiresults.is_approx = Exact;
|
---|
| 377 | }
|
---|
| 378 | }
|
---|
| 379 |
|
---|
| 380 | // combine the term information
|
---|
| 381 | if (need_term_info (request.filterResultOptions)) {
|
---|
| 382 | // append the terms
|
---|
| 383 | multiresults.orgterms.insert(multiresults.orgterms.end(),
|
---|
| 384 | thisqueryresults.orgterms.begin(),
|
---|
| 385 | thisqueryresults.orgterms.end());
|
---|
| 386 |
|
---|
| 387 |
|
---|
| 388 | // add the term variants -
|
---|
| 389 | text_tset::iterator termvar_here = thisqueryresults.termvariants.begin();
|
---|
| 390 | text_tset::iterator termvar_end = thisqueryresults.termvariants.end();
|
---|
| 391 | while (termvar_here != termvar_end) {
|
---|
| 392 | multiresults.termvariants.insert(*termvar_here);
|
---|
[9620] | 393 | ++termvar_here;
|
---|
[8027] | 394 | }
|
---|
[12380] | 395 |
|
---|
| 396 | // add the stop words
|
---|
| 397 | text_tset::iterator stopwords_here = thisqueryresults.stopwords.begin();
|
---|
| 398 | text_tset::iterator stopwords_end = thisqueryresults.stopwords.end();
|
---|
| 399 | while (stopwords_here != stopwords_end) {
|
---|
| 400 | multiresults.stopwords.insert(*stopwords_here);
|
---|
| 401 | ++stopwords_here;
|
---|
| 402 | }
|
---|
[8027] | 403 | }
|
---|
| 404 |
|
---|
[9620] | 405 | ++query_here;
|
---|
[8027] | 406 | }
|
---|
| 407 |
|
---|
| 408 | // sort and unique the query terms
|
---|
| 409 | multiresults.sortuniqqueryterms ();
|
---|
| 410 | }
|
---|
| 411 |
|
---|
| 412 |
|
---|