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 | // request.filterResultOptions and request.fields (if required) should
|
---|
31 | // be set from the calling code
|
---|
32 | void set_queryfilter_options (FilterRequest_t &request, const text_t &querystring,
|
---|
33 | cgiargsclass &args) {
|
---|
34 |
|
---|
35 | request.filterName = "QueryFilter";
|
---|
36 |
|
---|
37 | OptionValue_t option;
|
---|
38 |
|
---|
39 | option.name = "Term";
|
---|
40 | option.value = querystring;
|
---|
41 | request.filterOptions.push_back (option);
|
---|
42 |
|
---|
43 | option.name = "QueryType";
|
---|
44 | option.value = (args.getintarg("t")) ? "ranked" : "boolean";
|
---|
45 | request.filterOptions.push_back (option);
|
---|
46 |
|
---|
47 | option.name = "MatchMode";
|
---|
48 | option.value = (args.getintarg("t")) ? "some" : "all";
|
---|
49 | request.filterOptions.push_back (option);
|
---|
50 |
|
---|
51 | option.name = "Casefold";
|
---|
52 | option.value = (args.getintarg("k")) ? "true" : "false";
|
---|
53 | request.filterOptions.push_back (option);
|
---|
54 |
|
---|
55 | option.name = "Stem";
|
---|
56 | option.value = (args.getintarg("s")) ? "true" : "false";
|
---|
57 | request.filterOptions.push_back (option);
|
---|
58 |
|
---|
59 | if (!args["h"].empty()) {
|
---|
60 | option.name = "Index";
|
---|
61 | option.value = args["h"];
|
---|
62 | request.filterOptions.push_back (option);
|
---|
63 | }
|
---|
64 |
|
---|
65 | if (!args["j"].empty()) {
|
---|
66 | option.name = "Subcollection";
|
---|
67 | option.value = args["j"];
|
---|
68 | request.filterOptions.push_back (option);
|
---|
69 | }
|
---|
70 |
|
---|
71 | if (!args["n"].empty()) {
|
---|
72 | option.name = "Language";
|
---|
73 | option.value = args["n"];
|
---|
74 | request.filterOptions.push_back (option);
|
---|
75 | }
|
---|
76 |
|
---|
77 | if (!args["g"].empty()) { // granularity for mgpp
|
---|
78 | option.name = "Level";
|
---|
79 | option.value = args["g"];
|
---|
80 | request.filterOptions.push_back (option);
|
---|
81 | }
|
---|
82 |
|
---|
83 | set_more_queryfilter_options (request, args);
|
---|
84 | }
|
---|
85 |
|
---|
86 | void set_queryfilter_options (FilterRequest_t &request, const text_t &querystring1,
|
---|
87 | const text_t &querystring2, cgiargsclass &args) {
|
---|
88 |
|
---|
89 | set_queryfilter_options (request, querystring1, args);
|
---|
90 |
|
---|
91 | // fill in the second query if needed
|
---|
92 | if (!args["cq2"].empty()) {
|
---|
93 | OptionValue_t option;
|
---|
94 |
|
---|
95 | option.name = "CombineQuery";
|
---|
96 | option.value = args["cq2"];
|
---|
97 | request.filterOptions.push_back (option);
|
---|
98 |
|
---|
99 | option.name = "Term";
|
---|
100 | option.value = querystring2;
|
---|
101 | request.filterOptions.push_back (option);
|
---|
102 |
|
---|
103 | option.name = "QueryType";
|
---|
104 | option.value = (args.getintarg("t")) ? "ranked" : "boolean";
|
---|
105 | request.filterOptions.push_back (option);
|
---|
106 |
|
---|
107 | option.name = "Casefold";
|
---|
108 | option.value = (args.getintarg("k")) ? "true" : "false";
|
---|
109 | request.filterOptions.push_back (option);
|
---|
110 |
|
---|
111 | option.name = "Stem";
|
---|
112 | option.value = (args.getintarg("s")) ? "true" : "false";
|
---|
113 | request.filterOptions.push_back (option);
|
---|
114 |
|
---|
115 | if (!args["h2"].empty()) {
|
---|
116 | option.name = "Index";
|
---|
117 | option.value = args["h2"];
|
---|
118 | request.filterOptions.push_back (option);
|
---|
119 | }
|
---|
120 |
|
---|
121 | if (!args["j2"].empty()) {
|
---|
122 | option.name = "Subcollection";
|
---|
123 | option.value = args["j2"];
|
---|
124 | request.filterOptions.push_back (option);
|
---|
125 | }
|
---|
126 |
|
---|
127 | if (!args["n2"].empty()) {
|
---|
128 | option.name = "Language";
|
---|
129 | option.value = args["n2"];
|
---|
130 | request.filterOptions.push_back (option);
|
---|
131 | }
|
---|
132 | }
|
---|
133 | set_more_queryfilter_options (request, args);
|
---|
134 | }
|
---|
135 |
|
---|
136 | void set_more_queryfilter_options (FilterRequest_t &request, cgiargsclass &args) {
|
---|
137 |
|
---|
138 | OptionValue_t option;
|
---|
139 | int arg_m = args.getintarg("m");
|
---|
140 |
|
---|
141 | option.name = "Maxdocs";
|
---|
142 | option.value = arg_m;
|
---|
143 | request.filterOptions.push_back (option);
|
---|
144 |
|
---|
145 | // option.name = "StartResults";
|
---|
146 | // option.value = args["r"];
|
---|
147 | // request.filterOptions.push_back (option);
|
---|
148 |
|
---|
149 | // option.name = "EndResults";
|
---|
150 | // int endresults = args.getintarg("o") + (args.getintarg("r") - 1);
|
---|
151 | // if ((endresults > arg_m) && (arg_m != -1)) endresults = arg_m;
|
---|
152 | // option.value = endresults;
|
---|
153 | // request.filterOptions.push_back (option);
|
---|
154 | }
|
---|
155 |
|
---|
156 | void format_querystring (text_t &querystring, int querymode) {
|
---|
157 | text_t formattedstring;
|
---|
158 |
|
---|
159 | text_t::const_iterator here = querystring.begin();
|
---|
160 | text_t::const_iterator end = querystring.end();
|
---|
161 |
|
---|
162 | // space is used to insert spaces between Chinese
|
---|
163 | // characters. No space is needed before the first
|
---|
164 | // Chinese character.
|
---|
165 | bool space = false;
|
---|
166 |
|
---|
167 | // want to remove ()|!& from querystring so boolean queries are just
|
---|
168 | // "all the words" queries (unless querymode is advanced)
|
---|
169 | while (here != end) {
|
---|
170 | if ((querymode == 0) && (*here == '(' || *here == ')' || *here == '|' ||
|
---|
171 | *here == '!' || *here == '&')) {
|
---|
172 | formattedstring.push_back(' ');
|
---|
173 | } else {
|
---|
174 | if ((*here >= 0x4e00 && *here <= 0x9fa5) ||
|
---|
175 | (*here >= 0xf900 && *here <= 0xfa2d)) {
|
---|
176 | // Chinese character
|
---|
177 | if (space) formattedstring.push_back (0x200b);
|
---|
178 | formattedstring.push_back (*here);
|
---|
179 | formattedstring.push_back (0x200b);
|
---|
180 | space = true;
|
---|
181 | } else {
|
---|
182 | // non-Chinese character
|
---|
183 | formattedstring.push_back (*here);
|
---|
184 | space = false;
|
---|
185 | }
|
---|
186 | }
|
---|
187 | here ++;
|
---|
188 | }
|
---|
189 | querystring = formattedstring;
|
---|
190 | }
|
---|
191 |
|
---|
192 |
|
---|
193 |
|
---|
194 | void add_dates(text_t &querystring, int startdate, int enddate,
|
---|
195 | int startbc, int endbc)
|
---|
196 | {
|
---|
197 | if(startdate)
|
---|
198 | {
|
---|
199 | int querystringis = 0;
|
---|
200 | text_t::const_iterator here = querystring.begin();
|
---|
201 | text_t::const_iterator end = querystring.end();
|
---|
202 | while(here!=end)
|
---|
203 | {
|
---|
204 | if(!(isspace((*here)))){
|
---|
205 | here = end;
|
---|
206 | querystringis = 1;
|
---|
207 | }
|
---|
208 | else
|
---|
209 | here++;
|
---|
210 | }
|
---|
211 | //converting BCE dates
|
---|
212 | if(startbc && startdate > 0)
|
---|
213 | {
|
---|
214 | startdate *= -1;
|
---|
215 | }
|
---|
216 | if(endbc && enddate > 0)
|
---|
217 | {
|
---|
218 | enddate *= -1;
|
---|
219 | }
|
---|
220 | if(enddate != 0 && enddate<startdate)
|
---|
221 | {
|
---|
222 | cout<<"enddate too small"<<endl;
|
---|
223 | return;
|
---|
224 | }
|
---|
225 | if(querystringis)
|
---|
226 | querystring.appendcstr(" AND");
|
---|
227 | if(!enddate)
|
---|
228 | {
|
---|
229 | querystring.appendcstr(" [");
|
---|
230 | if(startdate<0)
|
---|
231 | {
|
---|
232 | querystring.appendcstr("bc");
|
---|
233 | querystring.appendint((startdate*-1));
|
---|
234 | }
|
---|
235 | else
|
---|
236 | querystring.appendint(startdate);
|
---|
237 | querystring.appendcstr("]:CO");
|
---|
238 | }
|
---|
239 | else{
|
---|
240 | int nextdate = startdate;
|
---|
241 | querystring.appendcstr(" (");
|
---|
242 | while(nextdate<=enddate)
|
---|
243 | {
|
---|
244 | if(nextdate!=0)
|
---|
245 | {
|
---|
246 | querystring.appendcstr(" [");
|
---|
247 | if(nextdate<0)
|
---|
248 | {
|
---|
249 | querystring.appendcstr("bc");
|
---|
250 | querystring.appendint((nextdate*-1));
|
---|
251 | }
|
---|
252 | else
|
---|
253 | querystring.appendint(nextdate);
|
---|
254 | querystring.appendcstr("]:CO");
|
---|
255 | }
|
---|
256 | nextdate++;
|
---|
257 | }
|
---|
258 | querystring.appendcstr(" )");
|
---|
259 | }
|
---|
260 | }
|
---|
261 |
|
---|
262 | }
|
---|
263 |
|
---|
264 | void get_phrases (const text_t &querystring, text_tarray &phrases) {
|
---|
265 |
|
---|
266 | phrases.erase (phrases.begin(), phrases.end());
|
---|
267 | if (!querystring.empty()) {
|
---|
268 |
|
---|
269 | text_t::const_iterator end = querystring.end();
|
---|
270 | text_t::const_iterator here = findchar (querystring.begin(), end, '"');
|
---|
271 | if (here != end) {
|
---|
272 | text_t tmptext;
|
---|
273 | bool foundquote = false;
|
---|
274 | while (here != end) {
|
---|
275 | if (*here == '"') {
|
---|
276 | if (foundquote) {
|
---|
277 | if (!tmptext.empty()) {
|
---|
278 | phrases.push_back(tmptext);
|
---|
279 | tmptext.clear();
|
---|
280 | }
|
---|
281 | foundquote = false;
|
---|
282 | } else foundquote = true;
|
---|
283 | } else {
|
---|
284 | if (foundquote) tmptext.push_back (*here);
|
---|
285 | }
|
---|
286 | here ++;
|
---|
287 | }
|
---|
288 | }
|
---|
289 | }
|
---|
290 | }
|
---|
291 |
|
---|
292 | // search history tool
|
---|
293 | // also used for form query macros
|
---|
294 | text_t escape_quotes(const text_t &querystring) {
|
---|
295 |
|
---|
296 | text_t::const_iterator here = querystring.begin();
|
---|
297 | text_t::const_iterator end = querystring.end();
|
---|
298 |
|
---|
299 | text_t escquery = "";
|
---|
300 | while (here != end) {
|
---|
301 | if (*here != '\'' && *here != '\"' && *here != '\n' && *here != '\r') escquery.push_back(*here);
|
---|
302 | else if (*here == '\n' || *here == '\r') {
|
---|
303 | escquery.push_back(' ');
|
---|
304 | } else {
|
---|
305 | escquery +="\\\\";
|
---|
306 | escquery.push_back(*here);
|
---|
307 | }
|
---|
308 |
|
---|
309 | here++;
|
---|
310 | }
|
---|
311 | return escquery;
|
---|
312 |
|
---|
313 | }
|
---|
314 |
|
---|
315 | // some query form parsing functions for use with mgpp
|
---|
316 |
|
---|
317 | void parse_reg_query_form(text_t &querystring, cgiargsclass &args){
|
---|
318 |
|
---|
319 | querystring.clear();
|
---|
320 | text_t combine;
|
---|
321 | int argt = args.getintarg("t");// t=0 -and, t=1 - or
|
---|
322 | if (argt == 0) combine = "&";
|
---|
323 | else combine = "|";
|
---|
324 |
|
---|
325 | text_t field = args["fqf"];
|
---|
326 | if (field.empty()) return; // no query
|
---|
327 | text_tarray fields;
|
---|
328 | splitchar(field.begin(), field.end(), ',', fields);
|
---|
329 |
|
---|
330 | text_t value = args["fqv"];
|
---|
331 | if (value.empty()) return; // somethings wrong
|
---|
332 | text_tarray values;
|
---|
333 | splitchar(value.begin(), value.end(), ',', values);
|
---|
334 |
|
---|
335 | for (int i=0; i< values.size(); i++) {
|
---|
336 | if (!values[i].empty()) {
|
---|
337 | addqueryelem(querystring, fields[i], values[i], combine);
|
---|
338 | }
|
---|
339 | }
|
---|
340 |
|
---|
341 | }
|
---|
342 |
|
---|
343 |
|
---|
344 | void parse_adv_query_form(text_t &querystring, cgiargsclass &args){
|
---|
345 |
|
---|
346 | querystring.clear();
|
---|
347 | text_t combine = "&";
|
---|
348 |
|
---|
349 | text_t field = args["fqf"];
|
---|
350 | if (field.empty()) return; // no query
|
---|
351 | text_tarray fields;
|
---|
352 | splitchar(field.begin(), field.end(), ',', fields);
|
---|
353 |
|
---|
354 | text_t value = args["fqv"];
|
---|
355 | if (value.empty()) return; // somethings wrong
|
---|
356 | text_tarray values;
|
---|
357 | splitchar(value.begin(), value.end(), ',', values);
|
---|
358 |
|
---|
359 | text_t stem = args["fqs"];
|
---|
360 | if (stem.empty()) return; // somethings wrong
|
---|
361 | text_tarray stems;
|
---|
362 | splitchar(stem.begin(), stem.end(), ',', stems);
|
---|
363 |
|
---|
364 | text_t fold = args["fqk"];
|
---|
365 | if (fold.empty()) return; // somethings wrong
|
---|
366 | text_tarray folds;
|
---|
367 | splitchar(fold.begin(), fold.end(), ',', folds);
|
---|
368 |
|
---|
369 | text_t comb = args["fqc"];
|
---|
370 | if (comb.empty()) return; //somethings wrong
|
---|
371 | text_tarray combs;
|
---|
372 | splitchar(comb.begin(), comb.end(), ',', combs);
|
---|
373 |
|
---|
374 | for(int i=0; i< values.size(); i++) {
|
---|
375 | if (!values[i].empty()) {
|
---|
376 | if (i!=0) {
|
---|
377 | if (combs[i-1]=="and") combine = "&";
|
---|
378 | else if (combs[i-1]=="or")combine = "|";
|
---|
379 | else if (combs[i-1]=="not")combine = "!";
|
---|
380 | }
|
---|
381 | text_t term = addstemcase(values[i], stems[i], folds[i]);
|
---|
382 | addqueryelem(querystring, fields[i], term, combine);
|
---|
383 | }
|
---|
384 |
|
---|
385 | }
|
---|
386 | }
|
---|
387 |
|
---|
388 | text_t addstemcase(text_t &terms, text_t &stem, text_t &fold) {
|
---|
389 |
|
---|
390 | text_t outtext;
|
---|
391 | text_t word;
|
---|
392 | //unsigned short c;
|
---|
393 | text_t::iterator here = terms.begin();
|
---|
394 | text_t::iterator end = terms.end();
|
---|
395 |
|
---|
396 | while (here !=end) {
|
---|
397 |
|
---|
398 | if (is_unicode_letdig(*here)) {
|
---|
399 | // not word boundary
|
---|
400 | word.push_back(*here);
|
---|
401 | here++;
|
---|
402 | }
|
---|
403 | else {
|
---|
404 | // found word boundary
|
---|
405 | if (!word.empty() ) {
|
---|
406 | if (stem == "1" || fold =="1") {
|
---|
407 | word += "#";
|
---|
408 | if (stem == "1") word += "s";
|
---|
409 | //else word += "u";
|
---|
410 |
|
---|
411 | if (fold == "1") word += "i";
|
---|
412 | //else word += "c";
|
---|
413 | }
|
---|
414 |
|
---|
415 | word += " ";
|
---|
416 | outtext += word;
|
---|
417 | word.clear();
|
---|
418 | }
|
---|
419 | if (*here == '\"') {
|
---|
420 | outtext.push_back(*here);
|
---|
421 | }
|
---|
422 | here++;
|
---|
423 | }
|
---|
424 | }
|
---|
425 |
|
---|
426 | // get last word
|
---|
427 | if (!word.empty()) {
|
---|
428 | if (stem == "1"|| fold == "1") {
|
---|
429 | word += "#";
|
---|
430 | if (stem == "1") word += "s";
|
---|
431 | //else word += "u";
|
---|
432 |
|
---|
433 | if (fold == "1") word += "i";
|
---|
434 | //else word += "c";
|
---|
435 | }
|
---|
436 | word += " ";
|
---|
437 | outtext += word;
|
---|
438 | }
|
---|
439 | return outtext;
|
---|
440 | }
|
---|
441 |
|
---|
442 |
|
---|
443 |
|
---|
444 | void addqueryelem(text_t &querystring, text_t &tag,
|
---|
445 | text_t &query, text_t combine) {
|
---|
446 | if (!querystring.empty()) { // have to put and/or
|
---|
447 | querystring += " "+combine + " ";
|
---|
448 |
|
---|
449 | }
|
---|
450 | if (tag=="ZZ" || tag=="") { // just add onto querystring
|
---|
451 | querystring += query;
|
---|
452 | }
|
---|
453 | else {
|
---|
454 | querystring += "["+query+"]:"+tag;
|
---|
455 | }
|
---|
456 |
|
---|
457 | }
|
---|
458 |
|
---|
459 |
|
---|
460 | void format_field_info(text_t & querystring, text_t &tag) {
|
---|
461 |
|
---|
462 | if (tag == "ZZ" || tag == "") {
|
---|
463 | return; // do nothing
|
---|
464 | }
|
---|
465 |
|
---|
466 | querystring = "["+querystring+"]:"+tag;
|
---|
467 | }
|
---|