source: gsdl/trunk/src/recpt/querytools.cpp@ 15597

Last change on this file since 15597 was 12930, checked in by kjdon, 18 years ago

changed set_query_type_args to make plain the default if no search types have been specified

  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 24.1 KB
Line 
1/**********************************************************************
2 *
3 * querytools.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 "querytools.h"
27#include <ctype.h>
28#include "unitool.h" // for is_unicode_letdig
29
30// sets the ct, qt, qto arguments
31void set_query_type_args(ColInfoResponse_t *cinfo, cgiargsclass &args) {
32
33 if (args["ct"].empty()) {
34 text_t build_type = cinfo->buildType;
35 if (build_type == "mgpp") {
36 args["ct"] = "1";
37 } else if (build_type == "lucene") {
38 args["ct"] = "2";
39 } else {
40 args["ct"] = "0";
41 }
42 }
43 text_t arg_ct = args["ct"];
44 if (arg_ct == "0") {
45 // mg
46 args["qt"] = "0";
47 args["qto"] = "0";
48 return;
49 }
50
51 if (!args["qt"].empty() && !args["qto"].empty()) {
52 return;
53 }
54
55 text_tmap::iterator check = cinfo->format.find("SearchTypes");
56 text_t search_types;
57 if(check != cinfo->format.end() && !(*check).second.empty()){
58 search_types = (*check).second;
59 } else {
60 // assume plain,form
61 if (args["qto"].empty()) args["qto"] = "3";
62 if (args["qt"].empty()) {
63 int arg_qto = args.getintarg("qto");
64 if (arg_qto == 2) {
65 args["qt"] = "1";
66 } else {
67 args["qt"] = "0";
68 }
69 }
70 return;
71 }
72
73
74 if (args["qto"].empty()) {
75 unsigned int type = 0;
76 if (findword(search_types.begin(), search_types.end(), "form") != search_types.end()) {
77 type |= 2;
78 }
79 if (findword(search_types.begin(), search_types.end(), "plain") != search_types.end()) {
80 type |= 1;
81 }
82 args.setintarg("qto", type);
83 }
84
85 if (args["qt"].empty()) {
86 int arg_qto = args.getintarg("qto");
87 if (arg_qto == 2 || (arg_qto == 3 && starts_with(search_types, "form"))) {
88 args["qt"] = "1";
89 } else {
90 args["qt"] = "0";
91 }
92 }
93}
94
95// sets the ks, ss, afs (casesupport, stemsupport, accentfoldsupport) args
96void set_stem_index_args(ColInfoResponse_t *cinfo, cgiargsclass &args) {
97 int stemIndexes = cinfo->stemIndexes;
98
99 if (stemIndexes & SIcasefold) {
100 args["ks"] = 1;
101 }
102 if (stemIndexes & SIstem) {
103 args["ss"] = 1;
104 }
105 if (stemIndexes & SIaccentfold) {
106 args["afs"] = 1;
107 }
108
109}
110
111// request.filterResultOptions and request.fields (if required) should
112// be set from the calling code
113void set_queryfilter_options (FilterRequest_t &request,
114 const text_t &querystring,
115 cgiargsclass &args) {
116
117 request.filterName = "QueryFilter";
118
119 OptionValue_t option;
120
121 option.name = "Term";
122 option.value = querystring;
123 request.filterOptions.push_back (option);
124
125 option.name = "QueryType";
126 option.value = (args.getintarg("t")) ? "ranked" : "boolean";
127 request.filterOptions.push_back (option);
128
129 option.name = "MatchMode";
130 // mgpp in advanced mode, always use some query
131 if (args.getintarg("ct") == 1 && args.getintarg("b") == 1) {
132 option.value = "some";
133 } else {
134 option.value = (args.getintarg("t")) ? "some" : "all";
135 }
136 request.filterOptions.push_back (option);
137
138 option.name = "Casefold";
139 option.value = (args.getintarg("k")) ? "true" : "false";
140 request.filterOptions.push_back (option);
141
142 option.name = "Stem";
143 option.value = (args.getintarg("s")) ? "true" : "false";
144 request.filterOptions.push_back (option);
145
146 option.name = "AccentFold";
147 option.value = (args.getintarg("af")) ? "true" : "false";
148 request.filterOptions.push_back (option);
149
150 if (!args["h"].empty()) {
151 option.name = "Index";
152 option.value = args["h"];
153 request.filterOptions.push_back (option);
154 }
155
156 if (!args["j"].empty()) {
157 option.name = "Subcollection";
158 option.value = args["j"];
159 request.filterOptions.push_back (option);
160 }
161
162 if (!args["n"].empty()) {
163 option.name = "Language";
164 option.value = args["n"];
165 request.filterOptions.push_back (option);
166 }
167
168 if (!args["g"].empty()) { // granularity for mgpp
169 option.name = "Level";
170 option.value = args["g"];
171 request.filterOptions.push_back (option);
172 }
173
174 if (!args["fs"].empty()) { // filter string for lucene
175 option.name = "FilterString";
176 option.value = args["fs"];
177 request.filterOptions.push_back (option);
178 }
179
180 if (!args["sf"].empty()) { // sort field for lucene
181 option.name = "SortField";
182 option.value = args["sf"];
183 request.filterOptions.push_back (option);
184 }
185
186 if (!args["fuzziness"].empty() && args["fuzziness"] != "100") { // fuzziness value for lucene
187 option.name = "Fuzziness";
188 option.value = (text_t) "0." + args["fuzziness"];
189 request.filterOptions.push_back (option);
190 }
191
192 set_more_queryfilter_options (request, args);
193}
194
195void set_queryfilter_options (FilterRequest_t &request,
196 const text_t &querystring1,
197 const text_t &querystring2, cgiargsclass &args) {
198
199 set_queryfilter_options (request, querystring1, args);
200
201 // fill in the second query if needed
202 if (!args["cq2"].empty()) {
203 OptionValue_t option;
204
205 option.name = "CombineQuery";
206 option.value = args["cq2"];
207 request.filterOptions.push_back (option);
208
209 option.name = "Term";
210 option.value = querystring2;
211 request.filterOptions.push_back (option);
212
213 option.name = "QueryType";
214 option.value = (args.getintarg("t")) ? "ranked" : "boolean";
215 request.filterOptions.push_back (option);
216
217 option.name = "Casefold";
218 option.value = (args.getintarg("k")) ? "true" : "false";
219 request.filterOptions.push_back (option);
220
221 option.name = "Stem";
222 option.value = (args.getintarg("s")) ? "true" : "false";
223 request.filterOptions.push_back (option);
224
225 option.name = "AccentFold";
226 option.value = (args.getintarg("af")) ? "true" : "false";
227 request.filterOptions.push_back (option);
228
229 if (!args["h2"].empty()) {
230 option.name = "Index";
231 option.value = args["h2"];
232 request.filterOptions.push_back (option);
233 }
234
235 if (!args["j2"].empty()) {
236 option.name = "Subcollection";
237 option.value = args["j2"];
238 request.filterOptions.push_back (option);
239 }
240
241 if (!args["n2"].empty()) {
242 option.name = "Language";
243 option.value = args["n2"];
244 request.filterOptions.push_back (option);
245 }
246 }
247 set_more_queryfilter_options (request, args);
248}
249
250void set_more_queryfilter_options (FilterRequest_t &request,
251 cgiargsclass &args) {
252
253 OptionValue_t option;
254 int arg_m = args.getintarg("m");
255
256 option.name = "Maxdocs";
257 option.value = arg_m;
258 request.filterOptions.push_back (option);
259
260 // option.name = "StartResults";
261 // option.value = args["r"];
262 // request.filterOptions.push_back (option);
263
264 // option.name = "EndResults";
265 // int endresults = args.getintarg("o") + (args.getintarg("r") - 1);
266 // if ((endresults > arg_m) && (arg_m != -1)) endresults = arg_m;
267 // option.value = endresults;
268 // request.filterOptions.push_back (option);
269}
270
271bool is_special_character(int indexer_type, unsigned short character) {
272 // mgpp
273 if (indexer_type == 1) {
274 return (character == '#' || character == '/' || character == '*');
275 }
276 // lucene
277 else if (indexer_type == 2) {
278 return (character == '?' || character == '*' || character == '~' ||
279 character == '^');
280 }
281 return false;
282}
283
284// This function removes boolean operators from simple searches, and segments
285// chinese characters if segment=true
286void format_querystring (text_t &querystring, int querymode, bool segment) {
287 text_t formattedstring;
288
289 // advanced search, no segmenting, don't need to do anything
290 if (querymode == 1 && !segment) return;
291
292 text_t::const_iterator here = querystring.begin();
293 text_t::const_iterator end = querystring.end();
294
295 // space is used to insert spaces between Chinese
296 // characters. No space is needed before the first
297 // Chinese character.
298 bool space = false;
299
300 // want to remove ()|!& from querystring so boolean queries are just
301 // "all the words" queries (unless querymode is advanced)
302 while (here != end) {
303 if ((querymode == 0) && (*here == '(' || *here == ')' || *here == '|' ||
304 *here == '!' || *here == '&')) {
305 formattedstring.push_back(' ');
306 } else if (segment) {
307 if ((*here >= 0x4e00 && *here <= 0x9fa5) ||
308 (*here >= 0xf900 && *here <= 0xfa2d)) {
309 // Chinese character
310 if (!space) formattedstring.push_back (0x200b); // zero width space
311 formattedstring.push_back (*here);
312 formattedstring.push_back (0x200b);
313 space = true;
314 } else {
315
316 // non-Chinese character
317 formattedstring.push_back (*here);
318 space = false;
319
320 }
321
322 } else {
323 formattedstring.push_back (*here);
324 }
325 ++here;
326 }
327 querystring = formattedstring;
328}
329
330
331
332
333// search history tool
334// also used for form query macros
335text_t escape_quotes(const text_t &querystring) {
336
337 text_t::const_iterator here = querystring.begin();
338 text_t::const_iterator end = querystring.end();
339
340 text_t escquery = "";
341 while (here != end) {
342 if (*here != '\'' && *here != '\"' && *here != '\n' && *here != '\r') escquery.push_back(*here);
343 else if (*here == '\n' || *here == '\r') {
344 escquery.push_back(' ');
345 } else {
346 escquery +="\\\\";
347 escquery.push_back(*here);
348 }
349
350 ++here;
351 }
352 return escquery;
353
354}
355
356// Parses the terms into words, and adds #si if necessary
357text_t addstemcase(const text_t &terms, const text_t &stem, const text_t &fold,
358 const int indexer_type) {
359
360 // the default stem and case are set to 0 if this is being used, so we are only adding on qualifiers if stem or fold is 1.
361 if (stem == "0" && fold == "0") {
362 return terms;
363 }
364 // this is only for mgpp collections, shouldn't be called for anything else
365 if (indexer_type != 1) {
366 return terms;
367 }
368
369 text_t outtext;
370 text_t word;
371
372 text_t::const_iterator here = terms.begin();
373 text_t::const_iterator end = terms.end();
374
375 while (here !=end) {
376
377 if (is_unicode_letdig(*here) || is_special_character(indexer_type, *here)) {
378 // not word boundary
379 word.push_back(*here);
380 ++here;
381 }
382 else {
383 // found word boundary
384 if (!word.empty() ) {
385 if (starts_with(word, "NEAR") || starts_with(word, "WITHIN")) {
386 outtext += word;
387 word.clear();
388 }
389 else {
390 word += "#";
391 if (stem == "1") word += "s";
392 if (fold == "1") word += "i";
393 outtext += word;
394 word.clear();
395 }
396 }
397 // this only used in advanced form, so we leave in boolean operators
398 if (*here == '\"' || *here == '&' || *here == '|' || *here == '!' ||
399 *here == '(' || *here == ')' || is_unicode_space(*here)) {
400 outtext.push_back(*here);
401 }
402 ++here;
403 }
404 }
405
406 // get last word
407 if (!word.empty()) {
408 word += "#";
409 if (stem == "1") word += "s";
410 if (fold == "1") word += "i";
411 word += " ";
412 outtext += word;
413 }
414 return outtext;
415}
416
417
418// some query form parsing functions for use with mgpp & lucene
419
420void parse_reg_query_form(text_t &querystring, cgiargsclass &args, bool segment)
421{
422 querystring.clear();
423
424 int argct = args.getintarg("ct");
425 int argt = args.getintarg("t");// t=0 -and, t=1 - or
426 int argb = args.getintarg("b");
427
428 text_t combine;
429
430 // lucene uses global combine, so only need this for mgpp
431 if (argct==1) {
432 if (argt == 0) combine = "&";
433 else combine = "|";
434 }
435
436 text_t field = args["fqf"];
437 if (field.empty()) return; // no query
438 text_tarray fields;
439 splitchar(field.begin(), field.end(), ',', fields);
440
441 text_t value = args["fqv"];
442 if (value.empty()) return; // somethings wrong
443 text_tarray values;
444 splitchar(value.begin(), value.end(), ',', values);
445
446
447 for (int i=0; i< values.size(); ++i) {
448 if (!values[i].empty()) {
449 text_t this_value = values[i];
450 // remove operators for simple search, segments text if necessary
451 format_querystring(this_value, argb, segment);
452 // add tag info for this field (and other processing)
453 format_field_info(this_value, fields[i], argct, argt, argb);
454 // add into query string
455 if (argct == 2) {
456 // lucene
457 // we don't worry about AND/OR, cos this is done by defaultcombineoperator
458 querystring += this_value+" ";
459 } else {
460 // mgpp
461 if (!querystring.empty()) {
462 querystring += " "+ combine+ " ";
463 }
464 querystring += this_value;
465 }
466 }
467 }
468}
469
470
471void parse_adv_query_form(text_t &querystring, cgiargsclass &args, bool segment){
472 querystring.clear();
473
474 const int argct = args.getintarg("ct");
475 int argt = 0;// arg t is either not used (lucene) or used for natural/ranked (mgpp), so we set it to 0 = AND, by default
476 int argb = args.getintarg("b");
477 text_t combine;
478 if (argct==1) {
479 combine = "&";
480 }
481 else { // lucene
482 combine = "AND";
483 }
484
485 text_t field = args["fqf"];
486 if (field.empty()) return; // no query
487 text_tarray fields;
488 splitchar(field.begin(), field.end(), ',', fields);
489
490 text_t value = args["fqv"];
491 if (value.empty()) return; // somethings wrong
492 text_tarray values;
493 splitchar(value.begin(), value.end(), ',', values);
494
495 text_t comb = args["fqc"];
496 if (comb.empty()) return; //somethings wrong
497 text_tarray combs;
498 splitchar(comb.begin(), comb.end(), ',', combs);
499
500 text_tarray stems;
501 text_tarray folds;
502 if (argct == 1) {// mgpp - lucene doesn't do stem/case
503 text_t stem = args["fqs"];
504 if (stem.empty()) return; // somethings wrong
505 splitchar(stem.begin(), stem.end(), ',', stems);
506
507 text_t fold = args["fqk"];
508 if (fold.empty()) return; // somethings wrong
509 splitchar(fold.begin(), fold.end(), ',', folds);
510 }
511
512 for(int i=0; i< values.size(); ++i) {
513 if (!values[i].empty()) {
514 if (i!=0) {
515 if (argct==1) {
516 if (combs[i-1]=="and") combine = "&";
517 else if (combs[i-1]=="or")combine = "|";
518 else if (combs[i-1]=="not")combine = "!";
519 }
520 else { // lucene
521 if (combs[i-1]=="and") combine = "AND";
522 else if (combs[i-1]=="or")combine = "OR";
523 else if (combs[i-1]=="not")combine = "NOT";
524 }
525 }
526 text_t this_value = values[i];
527 // remove operators for simple search, segments text if necessary
528 format_querystring(this_value, argb, segment);
529 if (argct == 1) { // mgpp only
530 this_value = addstemcase(this_value, stems[i], folds[i], argct);
531 }
532 // add tag info for this field (and other processing)
533 format_field_info(this_value, fields[i], argct, argt, argb);
534 // add into query string
535 if (!querystring.empty()) {
536 querystring += " "+ combine+ " ";
537 }
538 querystring += this_value;
539
540 }
541 }
542}
543
544// Extended addqueryelem for Human Info project
545void addqueryelem_ex(text_t &querystring, const text_t &tag,
546 const text_t &terms, const text_t &stem,
547 const text_t &fold,
548 const text_t& combine, const text_t& word_combine) {
549
550 if (!querystring.empty()) { // have to put and/or
551 querystring += " " + combine + " ";
552 }
553 text_t outtext; outtext.reserve(512);
554 text_t word; word.reserve(100);
555 //unsigned short c;
556 text_t::const_iterator here = terms.begin();
557 text_t::const_iterator end = terms.end();
558 bool inquote = false, firstword = true;
559
560 text_t word2; word2.reserve(256);
561
562 while (here !=end) {
563 if (is_unicode_space(*here)) {
564 if (word2 == "AND") { word2.clear(); word2.push_back(7527); word2.appendcarr("AND", 3); word2.push_back(7527); }
565 else if (word2 == "OR") { word2.clear(); word2.push_back(7527); word2.appendcarr("OR", 2); word2.push_back(7527); }
566 else if (word2 == "NOT") { word2.clear(); word2.push_back(7527); word2.appendcarr("NOT", 3); word2.push_back(7527); }
567 else if (word2 == "NEAR") { word2.clear(); word2.push_back(7527); word2.appendcarr("NEAR", 4); word2.push_back(7527); }
568 else if (word2 == "WITHIN") { word2.clear(); word2.push_back(7527); word2.appendcarr("WITHIN", 6); word2.push_back(7527); }
569 if (inquote) {
570 word2.push_back(*here);
571 }
572 word.append(word2); word2.clear();
573
574 if (!inquote && !word.empty() ) {
575 // found word boundary
576
577 if (stem == "1" || fold =="1") {
578 word += "#";
579 if (stem == "1") word += "s";
580 //else word += "u";
581
582 if (fold == "1") word += "i";
583 //else word += "c";
584 }
585 if (firstword) {
586 firstword = false;
587 } else {
588 outtext += " " + word_combine + " ";
589 }
590 outtext += "[" + word + "]:"+tag;
591 word.clear();
592 }
593 ++here;
594 } else if (*here == '\"') {
595 word2.push_back(*here);
596 inquote = !inquote;
597 ++here;
598 } else {
599 // not word boundary
600 word2.push_back(*here);
601 ++here;
602 }
603 }
604
605 // get last word
606 if (!word2.empty()) {
607 if (word2 == "AND") { word2.clear(); word2.push_back(7527); word2.appendcarr("AND", 3); word2.push_back(7527); }
608 else if (word2 == "OR") { word2.clear(); word2.push_back(7527); word2.appendcarr("OR", 2); word2.push_back(7527); }
609 else if (word2 == "NOT") { word2.clear(); word2.push_back(7527); word2.appendcarr("NOT", 3); word2.push_back(7527); }
610 else if (word2 == "NEAR") { word2.clear(); word2.push_back(7527); word2.appendcarr("NEAR", 4); word2.push_back(7527); }
611 else if (word2 == "WITHIN") { word2.clear(); word2.push_back(7527); word2.appendcarr("WITHIN", 6); word2.push_back(7527); }
612 word.append(word2); word2.clear();
613
614 if (stem == "1"|| fold == "1") {
615 word += "#";
616 if (stem == "1") word += "s";
617 //else word += "u";
618
619 if (fold == "1") word += "i";
620 //else word += "c";
621 }
622 if (!outtext.empty()) outtext += " " + word_combine + " ";
623 outtext += "[" + word + "]:"+tag;
624 }
625 querystring += "(" + outtext + ")";
626}
627
628void add_field_info(text_t &querystring, const text_t &tag, int type) {
629
630 if (tag == "" || tag == "ZZ") return; // do nothing
631 if (type == 1) { //mgpp
632 querystring = "["+querystring+"]:"+tag;
633 } else if (type == 2) { // lucene
634 querystring = tag+":("+querystring+")";
635 }
636
637}
638
639
640void format_field_info_lucene(text_t &querystring, text_t tag, int argt, int argb) {
641 if (tag == "ZZ") tag = ""; // ZZ is a special tag meaning no tag (all fields)
642 int type = 2; //lucene
643
644 if (argb==0) { // simple
645 // there will be no & or | as they should have already been removed
646 // just tag the entire thing
647 if (tag != "") {
648 add_field_info(querystring, tag, type);
649 }
650 return;
651 }
652
653 // need to replace & with &&, | with ||
654 text_t::const_iterator here = querystring.begin();
655 text_t::const_iterator end = querystring.end();
656
657 text_t finalquery = "";
658 while (here != end) {
659 if (*here == '&') {
660 finalquery.push_back('&');
661 finalquery.push_back('&');
662 while (*(here+1) == '&') {
663 ++here;
664 }
665 }
666 else if (*here == '|') {
667 finalquery.push_back('|');
668 finalquery.push_back('|');
669 while (*(here+1) == '|') {
670 ++here;
671 }
672 }
673 else {
674 finalquery.push_back(*here);
675 }
676 ++here;
677 }
678 querystring = finalquery;
679 add_field_info(querystring, tag, type);
680}
681
682
683void format_field_info_mgpp(text_t &querystring, text_t tag, int argt, int argb) {
684
685 if (tag == "ZZ") tag = ""; // ZZ is a special tag meaning no tag (all fields)
686 if (tag == "" && argb == 1) {
687 return; // no field specifier, advanced mode, the query stays as written
688 }
689
690 int type = 1; // mgpp
691
692 bool simple_and = (argb==0 && argt==0);
693 text_t finalquery = "";
694 text_t fieldpart ="";
695 text_t queryelem = "";
696 bool in_phrase = false;
697 bool in_field = false;
698
699 text_t::const_iterator here = querystring.begin();
700 text_t::const_iterator end = querystring.end();
701 while (here != end) {
702 if (is_unicode_letdig(*here) || *here == '&' || is_special_character(type, *here)) {
703 queryelem.push_back(*here);
704 }
705 else if (*here == '|') {
706 in_field = false;
707 }
708 else if (*here == '!' || *here == '(' || *here == ')') {
709 if (!in_phrase) { // ignore these if in_phrase
710 // output field, then output operator
711 in_field = false;
712 if (!queryelem.empty()) {
713 if (!simple_and && !fieldpart.empty()) {
714 add_field_info(fieldpart, tag, type);
715 finalquery += fieldpart;
716 finalquery.push_back(' ');
717 fieldpart.clear();
718 }
719 fieldpart += queryelem;
720 }
721 if (!fieldpart.empty()) {
722 add_field_info(fieldpart, tag, type);
723 finalquery += fieldpart;
724 finalquery.push_back(' ');
725 }
726 fieldpart.clear();
727 queryelem.clear();
728 finalquery.push_back(*here);
729 finalquery.push_back(' ');
730 }
731 }
732 else if (*here == '"') {
733 queryelem.push_back(*here);
734 if (in_phrase == false) in_phrase = true;
735 else {
736 in_phrase = false;
737 }
738 }
739
740 // Found word boundary, in a phrase
741 else if (in_phrase) {
742 queryelem.push_back(*here);
743 }
744 // Found a word boundary
745 else {
746 if (!queryelem.empty()) {
747 if (queryelem == "&") {
748 in_field = true;
749 queryelem.clear();
750 }
751 else if (starts_with(queryelem, "NEAR") || starts_with(queryelem, "WITHIN")) {
752
753 if (argb==1) {
754 // simple search, these not allowed
755 in_field = true;
756 fieldpart += queryelem;
757 fieldpart.push_back(' ');
758 }
759 queryelem.clear();
760
761 }
762 else {
763 if (!simple_and && !in_field) {
764 if (!fieldpart.empty()) {
765 add_field_info(fieldpart, tag, type);
766 finalquery += fieldpart;
767 finalquery.push_back(' ');
768 fieldpart.clear();
769 }
770 }
771
772 fieldpart += queryelem;
773 fieldpart.push_back(' ');
774 queryelem.clear();
775 }
776 }
777 }
778 ++here;
779 }
780 // at the end
781 if (!queryelem.empty()) {
782 if (!simple_and && !in_field && !fieldpart.empty()) {
783 add_field_info(fieldpart, tag, type);
784 finalquery += fieldpart;
785 finalquery.push_back(' ');
786 fieldpart.clear();
787 }
788 fieldpart += queryelem;
789 }
790 if (!fieldpart.empty()) {
791 add_field_info(fieldpart, tag, type);
792 finalquery += fieldpart;
793 fieldpart.clear();
794 finalquery.push_back(' ');
795 }
796
797 querystring = finalquery;
798}
799
800
801void format_field_info(text_t &querystring, text_t tag, int argct, int argt, int argb) {
802 if (argct == 1) {
803 format_field_info_mgpp(querystring, tag, argt, argb);
804 } else if (argct == 2) {
805 format_field_info_lucene(querystring, tag, argt, argb);
806 }
807}
808
809void mgpp_adddateelem(text_t& querystring, const int date)
810{
811 querystring.appendcstr(" [");
812 if(date<0) {
813 querystring.appendcstr("bc");
814 querystring.appendint((date*-1));
815 }
816 else {
817 querystring.appendint(date);
818 }
819 querystring.appendcstr("]:CV");
820}
821
822void lucene_adddateelem(text_t& querystring, const int date)
823{
824 querystring.appendcstr(" CV:(");
825 if(date<0) {
826 querystring.appendcstr("bc");
827 querystring.appendint((date*-1));
828 }
829 else {
830 querystring.appendint(date);
831 }
832 querystring.appendcstr(")");
833}
834
835
836void add_dates(text_t &querystring, int startdate, int enddate,
837 int startbc, int endbc, int ct)
838{
839 if(startdate)
840 {
841 int querystringis = 0;
842 text_t::const_iterator here = querystring.begin();
843 text_t::const_iterator end = querystring.end();
844 while(here!=end)
845 {
846 if(!(isspace((*here)))){
847 here = end;
848 querystringis = 1;
849 }
850 else
851 ++here;
852 }
853 //converting BCE dates
854 if(startbc && startdate > 0)
855 {
856 startdate *= -1;
857 }
858 if(endbc && enddate > 0)
859 {
860 enddate *= -1;
861 }
862 if(enddate != 0 && enddate<startdate)
863 {
864 cout<<"enddate too small"<<endl;
865 return;
866 }
867 if(querystringis)
868 querystring.appendcstr(" AND");
869 if(!enddate)
870 {
871 if (ct==1) {
872 mgpp_adddateelem(querystring,startdate);
873 }
874 else { // lucene
875 lucene_adddateelem(querystring,startdate);
876 }
877 }
878 else{
879 int nextdate = startdate;
880 querystring.appendcstr(" (");
881 while(nextdate<=enddate)
882 {
883 if(nextdate!=0) {
884 if (ct==1) {
885 mgpp_adddateelem(querystring,nextdate);
886 }
887 else { // lucene
888 lucene_adddateelem(querystring,nextdate);
889 }
890 }
891 ++nextdate;
892 }
893 querystring.appendcstr(" )");
894 }
895 }
896
897}
Note: See TracBrowser for help on using the repository browser.