source: gsdl/trunk/src/colservr/lucenequeryfilter.cpp@ 15594

Last change on this file since 15594 was 15594, checked in by mdewsnip, 16 years ago

(Adding new DB support) Replaced all "gdbm_level" with "text_level".

  • Property svn:keywords set to Author Date Id Revision
File size: 13.0 KB
Line 
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
27
28#include "lucenequeryfilter.h"
29#include "fileutil.h"
30#include <assert.h>
31#include "lucenesearch.h"
32
33/////////////////////////////////
34// functions for queryfilterclass
35/////////////////////////////////
36
37
38lucenequeryfilterclass::lucenequeryfilterclass ()
39 : queryfilterclass() {
40
41
42 FilterOption_t filtopt;
43
44 // -- onePerTerm Level enumerated
45 // likely to be Doc, Sec, Para, but we dont assume anything now
46 filtopt.clear();
47 filtopt.name = "Level";
48 filtopt.type = FilterOption_t::enumeratedt;
49 filtopt.repeatable = FilterOption_t::onePerTerm;
50 filterOptions["Level"] = filtopt;
51
52 // -- IndexField, enumerated, used to list available fields
53 filtopt.clear();
54 filtopt.name = "IndexField";
55 filtopt.type = FilterOption_t::enumeratedt;
56 filtopt.repeatable = FilterOption_t::onePerTerm;
57 filtopt.defaultValue = "";
58 filterOptions["IndexField"] = filtopt;
59
60}
61
62lucenequeryfilterclass::~lucenequeryfilterclass () {
63}
64
65
66//whether a query is a full text browse
67bool lucenequeryfilterclass::full_text_browse (int filterRequestOptions) {
68 return (filterRequestOptions & FRfullTextBrowse);
69}
70
71void lucenequeryfilterclass::configure (const text_t &key, const text_tarray &cfgline) {
72 queryfilterclass::configure(key, cfgline);
73
74 if (key == "indexfieldmap") {
75 indexfieldmap.importmap (cfgline);
76
77 // update the list of indexes in the filter information
78 text_tarray options;
79 indexfieldmap.gettoarray (options);
80
81 text_tarray::const_iterator here = options.begin();
82 text_tarray::const_iterator end = options.end();
83 bool start = true;
84 while (here !=end) {
85 if (!(*here).empty()) {
86 filterOptions["IndexField"].validValues.push_back(*here);
87 if (start) {
88 filterOptions["IndexField"].defaultValue = *here;
89 start = false;
90 }
91 }
92 ++here;
93 }
94 } else if (key == "indexlevels") {
95 text_tarray::const_iterator here = cfgline.begin();
96 text_tarray::const_iterator end = cfgline.end();
97 bool first=true;
98 while (here != end) {
99 if (!(*here).empty()) {
100 if (first) {
101 first = false;
102 // the default is the first value
103 filterOptions["Level"].defaultValue = *here;
104 }
105 filterOptions["Level"].validValues.push_back(*here);
106 }
107 ++here;
108 }
109 } else if (key == "textlevel") {
110 ((lucenesearchclass *)textsearchptr)->set_text_level(cfgline[0]);
111 }
112
113}
114
115
116void lucenequeryfilterclass::filter(const FilterRequest_t &request,
117 FilterResponse_t &response,
118 comerror_t &err, ostream &logout) {
119
120 outconvertclass text_t2ascii;
121
122 response.clear ();
123 err = noError;
124 if (db_ptr == NULL) {
125 // most likely a configuration problem
126 logout << text_t2ascii
127 << "configuration error: queryfilter contains a null dbclass\n\n";
128 err = configurationError;
129 return;
130 }
131 if (textsearchptr == NULL) {
132 // most likely a configuration problem
133 logout << text_t2ascii
134 << "configuration error: queryfilter contains a null textsearchclass (lucene)\n\n";
135 err = configurationError;
136 return;
137 }
138 if (full_text_browse(request.filterResultOptions)) {
139 browsefilter(request, response, err, logout);
140 return;
141 }
142 // open the database
143 db_ptr->setlogout(&logout);
144 if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
145 // most likely a system problem (we have already checked that the database exists)
146 logout << text_t2ascii
147 << "system problem: open on database \"" << db_filename << "\" failed\n\n";
148 err = systemProblem;
149 return;
150 }
151
152
153 // get the query parameters
154 int startresults, endresults;
155 text_t phrasematch; // not used here any more
156 vector<queryparamclass> queryfilterparams;
157 parse_query_params (request, queryfilterparams, startresults,
158 endresults, phrasematch, logout);
159
160
161 // do query
162 queryresultsclass queryresults;
163 do_multi_query (request, queryfilterparams, queryresults, err, logout);
164 response.error_message = queryresults.error_message;
165 if (err != noError) return;
166
167 // assemble document results
168 if (need_matching_docs (request.filterResultOptions)) {
169
170 int resultnum = 1;
171 ResultDocInfo_t resultdoc;
172 text_t trans_OID;
173 vector<int>::iterator docorder_here = queryresults.docs.docorder.begin();
174 vector<int>::iterator docorder_end = queryresults.docs.docorder.end();
175
176 // Now handled by Lucene directly
177 //if (endresults == -1) endresults = MAXNUMDOCS;
178
179 while (docorder_here != docorder_end)
180 {
181 // Now handled by Lucene directly
182 //if (resultnum > endresults) break;
183
184 // translate the document number
185 if (!translate(db_ptr, *docorder_here, trans_OID))
186 {
187 logout << text_t2ascii
188 << "warning: could not translate lucene document number \""
189 << *docorder_here << "\" to OID.\n\n";
190
191 }
192 else
193 {
194 docresultmap::iterator docset_here = queryresults.docs.docset.find (*docorder_here);
195
196 // see if there is a result for this number,
197 // if it is in the request set (or the request set is empty)
198 if (docset_here != queryresults.docs.docset.end() && (request.docSet.empty() || in_set(request.docSet, trans_OID)))
199 {
200 // Now handled by Lucene directly
201 //if (resultnum >= startresults) {
202
203 // add this document
204 resultdoc.OID = trans_OID;
205 resultdoc.result_num = resultnum;
206 resultdoc.ranking = (int)((*docset_here).second.docweight * 10000.0 + 0.5);
207
208 response.docInfo.push_back (resultdoc);
209 //}
210 ++resultnum;
211 }
212 } // else
213
214 ++docorder_here;
215 }
216 } // if need matching docs
217
218 // assemble the term results
219 if (need_term_info(request.filterResultOptions)) {
220 // note: the terms have already been sorted and uniqued - ?? have they??
221
222 TermInfo_t terminfo;
223 bool terms_first = true;
224
225 termfreqclassarray::iterator terms_here = queryresults.terms.begin();
226 termfreqclassarray::iterator terms_end = queryresults.terms.end();
227
228 while (terms_here != terms_end) {
229 terminfo.clear();
230 terminfo.term = (*terms_here).termstr;
231 terminfo.freq = (*terms_here).termfreq;
232 // lucene doesn't return any termvariants at this stage,
233 // so make sure the original term is set
234 terminfo.matchTerms.push_back(terminfo.term);
235
236 // this bit gets the matchTerms ie the equivalent (stem/casefold) terms
237 if (terms_first) {
238 text_tset::iterator termvariants_here = queryresults.termvariants.begin();
239 text_tset::iterator termvariants_end = queryresults.termvariants.end();
240 while (termvariants_here != termvariants_end) {
241 terminfo.matchTerms.push_back (*termvariants_here);
242 ++termvariants_here;
243 }
244 }
245 terms_first = false;
246
247 response.termInfo.push_back (terminfo);
248
249 ++terms_here;
250 }
251
252 // add the stop words
253 text_tset::iterator stopwords_here = queryresults.stopwords.begin();
254 text_tset::iterator stopwords_end = queryresults.stopwords.end();
255 while (stopwords_here != stopwords_end) {
256 response.stopwords.insert(*stopwords_here);
257 ++stopwords_here;
258 }
259 }
260
261 db_ptr->closedatabase(); // Important that local library doesn't leave any files open
262 response.numDocs = queryresults.docs_matched;
263 response.isApprox = queryresults.is_approx;
264}
265
266void lucenequeryfilterclass::browsefilter(const FilterRequest_t &request,
267 FilterResponse_t &response,
268 comerror_t &err, ostream &logout) {
269
270 outconvertclass text_t2ascii;
271
272 // get the query parameters
273 int startresults, endresults;
274 text_t phrasematch; // not used here any more, just have it so can use
275 // parse_query_params function
276
277 vector<queryparamclass> queryfilterparams;
278 parse_query_params (request, queryfilterparams, startresults,
279 endresults, phrasematch, logout);
280
281 vector<queryparamclass>::const_iterator query_here = queryfilterparams.begin();
282
283 // do query
284 queryresultsclass queryresults;
285 queryresults.clear();
286
287 int numDocs = endresults-startresults;
288 textsearchptr->setcollectdir (collectdir);
289
290 if (!((lucenesearchclass*)textsearchptr)->browse_search((*query_here), startresults, numDocs, queryresults)) {
291 // most likely a system problem
292 logout << text_t2ascii
293 << "system problem: could not do full text browse with lucene for index \""
294 << (*query_here).index << (*query_here).subcollection
295 << (*query_here).language << "\".\n\n";
296 err = systemProblem;
297 return;
298 }
299
300 // assemble the term results
301 TermInfo_t terminfo;
302
303 termfreqclassarray::iterator terms_here = queryresults.terms.begin();
304 termfreqclassarray::iterator terms_end = queryresults.terms.end();
305
306 while (terms_here != terms_end) {
307 terminfo.clear();
308 terminfo.term = (*terms_here).termstr;
309 terminfo.freq = (*terms_here).termfreq;
310
311 response.termInfo.push_back (terminfo);
312
313 ++terms_here;
314 }
315
316
317}
318
319// lucenesearchptr and db_ptr are assumed to be valid
320void lucenequeryfilterclass::do_multi_query (const FilterRequest_t &request,
321 const vector<queryparamclass> &query_params,
322 queryresultsclass &multiresults,
323 comerror_t &err, ostream &logout) {
324 outconvertclass text_t2ascii;
325
326 err = noError;
327 textsearchptr->setcollectdir (collectdir);
328 multiresults.clear();
329
330 vector<queryparamclass>::const_iterator query_here = query_params.begin();
331 vector<queryparamclass>::const_iterator query_end = query_params.end();
332 while (query_here != query_end) {
333 queryresultsclass thisqueryresults;
334 if (!textsearchptr->search((*query_here), thisqueryresults)) {
335 // most likely a system problem
336 logout << text_t2ascii
337 << "system problem: could not do search with lucene for index \""
338 << (*query_here).index << (*query_here).level
339 << (*query_here).subcollection
340 << (*query_here).language << "\".\n\n";
341 err = systemProblem;
342 return;
343 }
344
345 // check for syntax error
346 if (thisqueryresults.syntax_error==true) {
347 logout << text_t2ascii
348 << "syntax problem: invalid query string \""
349 << (*query_here).querystring<<"\".\n";
350 err = syntaxError;
351 return;
352 }
353 // combine the results
354 if (need_matching_docs (request.filterResultOptions)) {
355
356 if (query_params.size() == 1) {
357 multiresults.error_message = thisqueryresults.error_message;
358 multiresults.docs = thisqueryresults.docs; // just one set of results
359 multiresults.docs_matched = thisqueryresults.docs_matched;
360 multiresults.is_approx = thisqueryresults.is_approx;
361
362 } else {
363 if ((*query_here).combinequery == "and") {
364 multiresults.docs.combine_and (thisqueryresults.docs);
365 } else if ((*query_here).combinequery == "or") {
366 multiresults.docs.combine_or (thisqueryresults.docs);
367 } else if ((*query_here).combinequery == "not") {
368 multiresults.docs.combine_not (thisqueryresults.docs);
369 }
370 multiresults.docs_matched = multiresults.docs.docset.size();
371 multiresults.is_approx = Exact;
372 }
373 }
374
375 // combine the term information
376 if (need_term_info (request.filterResultOptions)) {
377 // append the terms
378 multiresults.orgterms.insert(multiresults.orgterms.end(),
379 thisqueryresults.orgterms.begin(),
380 thisqueryresults.orgterms.end());
381
382
383 // add the term variants -
384 text_tset::iterator termvar_here = thisqueryresults.termvariants.begin();
385 text_tset::iterator termvar_end = thisqueryresults.termvariants.end();
386 while (termvar_here != termvar_end) {
387 multiresults.termvariants.insert(*termvar_here);
388 ++termvar_here;
389 }
390
391 // add the stop words
392 text_tset::iterator stopwords_here = thisqueryresults.stopwords.begin();
393 text_tset::iterator stopwords_end = thisqueryresults.stopwords.end();
394 while (stopwords_here != stopwords_end) {
395 multiresults.stopwords.insert(*stopwords_here);
396 ++stopwords_here;
397 }
398 }
399
400 ++query_here;
401 }
402
403 // sort and unique the query terms
404 multiresults.sortuniqqueryterms ();
405}
406
407
408
Note: See TracBrowser for help on using the repository browser.