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
|
---|
31 | void 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
|
---|
96 | void 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
|
---|
113 | void 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 |
|
---|
195 | void 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 |
|
---|
250 | void 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 |
|
---|
271 | bool 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
|
---|
286 | void 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 >= 0x2e80 && *here <= 0xd7a3) ||
|
---|
308 | ( *here >= 0xf900 && *here <= 0xfa6a)) {
|
---|
309 | /* text_t not big enough to handle these. */
|
---|
310 | /* (*here >= 0x20000 && *here <= 0x2a6d6) ||
|
---|
311 | (*here >= 0x2f800 && *here <= 0x2fa1d)) { */
|
---|
312 |
|
---|
313 | // CJK character
|
---|
314 | if (!space) formattedstring.push_back (0x200b); // zero width space
|
---|
315 | formattedstring.push_back (*here);
|
---|
316 | formattedstring.push_back (0x200b);
|
---|
317 | space = true;
|
---|
318 | } else {
|
---|
319 |
|
---|
320 | // non-Chinese character
|
---|
321 | formattedstring.push_back (*here);
|
---|
322 | space = false;
|
---|
323 |
|
---|
324 | }
|
---|
325 |
|
---|
326 | } else {
|
---|
327 | formattedstring.push_back (*here);
|
---|
328 | }
|
---|
329 | ++here;
|
---|
330 | }
|
---|
331 | querystring = formattedstring;
|
---|
332 | }
|
---|
333 |
|
---|
334 | // turn query string into terms separated by spaces.
|
---|
335 | // still working on this...
|
---|
336 | text_t get_plain_query_terms(const text_t &querystring, const text_t &arg_ct) {
|
---|
337 | text_t::const_iterator here = querystring.begin();
|
---|
338 | text_t::const_iterator end = querystring.end();
|
---|
339 |
|
---|
340 | // lets look for [] and () first - these are a pain.
|
---|
341 | text_t::const_iterator bracket;
|
---|
342 | text_t query_no_brackets = "";
|
---|
343 |
|
---|
344 | // mgpp brackets: [xxx]:TI
|
---|
345 | if (findchar(here, end, '[') != end) {
|
---|
346 | while ((bracket = findchar(here, end, '[')) != end) {
|
---|
347 | // get the first bit
|
---|
348 | query_no_brackets += substr(here, bracket);
|
---|
349 | bracket++;
|
---|
350 | here = bracket;
|
---|
351 | // get the end bracket
|
---|
352 | bracket = findchar(here, end, ']');
|
---|
353 | query_no_brackets += substr(here, bracket);
|
---|
354 | // skip the :TI bits
|
---|
355 | while (*bracket != ' ' && bracket != end) { bracket++;}
|
---|
356 | here = bracket;
|
---|
357 | }
|
---|
358 | if (here != end) {
|
---|
359 | query_no_brackets += substr(here,end);
|
---|
360 | }
|
---|
361 | } else if (findchar(here, end, '(') != end) {
|
---|
362 | // lucene brackets TI:(xxx)
|
---|
363 | while ((bracket = findchar(here, end, '(')) != end) {
|
---|
364 | // back up the field name
|
---|
365 | text_t::const_iterator old_bracket = bracket;
|
---|
366 | while (*bracket != ' ' && bracket != here) {
|
---|
367 | --bracket;
|
---|
368 | }
|
---|
369 | if (bracket != here) {
|
---|
370 | // get the first bit
|
---|
371 | query_no_brackets += substr(here, bracket+1);
|
---|
372 | }
|
---|
373 | here = old_bracket +1;
|
---|
374 | // get the end bracket
|
---|
375 | bracket = findchar(here, end, ')');
|
---|
376 | query_no_brackets += substr(here, bracket);
|
---|
377 | if (bracket != end) {
|
---|
378 | here = bracket+1;
|
---|
379 | }
|
---|
380 | }
|
---|
381 | if (here != end) {
|
---|
382 | query_no_brackets += substr(here,end);
|
---|
383 | }
|
---|
384 | } else {
|
---|
385 | // was no brackets
|
---|
386 | query_no_brackets = querystring;
|
---|
387 | }
|
---|
388 |
|
---|
389 |
|
---|
390 | if (arg_ct == "2") { // lucene
|
---|
391 | // look for AND OR NOT and remove
|
---|
392 | here = query_no_brackets.begin();
|
---|
393 | end = query_no_brackets.end();
|
---|
394 | text_tlist terms;
|
---|
395 | splitword(here, end, "AND", terms);
|
---|
396 | joinchar(terms, ' ', query_no_brackets);
|
---|
397 | here = query_no_brackets.begin();
|
---|
398 | end = query_no_brackets.end();
|
---|
399 | splitword(here, end, "OR", terms);
|
---|
400 | joinchar(terms, ' ', query_no_brackets);
|
---|
401 | here = query_no_brackets.begin();
|
---|
402 | end = query_no_brackets.end();
|
---|
403 | splitword(here, end, "NOT", terms);
|
---|
404 | joinchar(terms, ' ', query_no_brackets);
|
---|
405 |
|
---|
406 | }
|
---|
407 | text_t terms = "";
|
---|
408 | bool space = false;
|
---|
409 | here = query_no_brackets.begin();
|
---|
410 | end = query_no_brackets.end();
|
---|
411 |
|
---|
412 | while (here != end) {
|
---|
413 | if (*here == '#' || *here == '/') {
|
---|
414 | // skip over #is /10 etc
|
---|
415 | ++here;
|
---|
416 | while (here != end && *here != ' ') {
|
---|
417 | ++here;
|
---|
418 | }
|
---|
419 | if (here == end) break;
|
---|
420 | }
|
---|
421 | if (is_unicode_letdig(*here)) {
|
---|
422 | terms.push_back(*here);
|
---|
423 | space = false;
|
---|
424 | } else {
|
---|
425 | if (!space) {
|
---|
426 | terms.push_back(' ');
|
---|
427 | space = true;
|
---|
428 | }
|
---|
429 | }
|
---|
430 | ++here;
|
---|
431 | }
|
---|
432 | return terms;
|
---|
433 |
|
---|
434 | }
|
---|
435 |
|
---|
436 | // search history tool
|
---|
437 | // also used for form query macros
|
---|
438 | text_t escape_quotes(const text_t &querystring) {
|
---|
439 |
|
---|
440 | text_t::const_iterator here = querystring.begin();
|
---|
441 | text_t::const_iterator end = querystring.end();
|
---|
442 |
|
---|
443 | text_t escquery = "";
|
---|
444 | while (here != end) {
|
---|
445 | if (*here != '\'' && *here != '\"' && *here != '\n' && *here != '\r') escquery.push_back(*here);
|
---|
446 | else if (*here == '\n' || *here == '\r') {
|
---|
447 | escquery.push_back(' ');
|
---|
448 | } else {
|
---|
449 | escquery +="\\\\";
|
---|
450 | escquery.push_back(*here);
|
---|
451 | }
|
---|
452 |
|
---|
453 | ++here;
|
---|
454 | }
|
---|
455 | return escquery;
|
---|
456 |
|
---|
457 | }
|
---|
458 |
|
---|
459 | // Parses the terms into words, and adds #si if necessary
|
---|
460 | text_t addstemcase(const text_t &terms, const text_t &stem, const text_t &fold,
|
---|
461 | const int indexer_type) {
|
---|
462 |
|
---|
463 | // 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.
|
---|
464 | if (stem == "0" && fold == "0") {
|
---|
465 | return terms;
|
---|
466 | }
|
---|
467 | // this is only for mgpp collections, shouldn't be called for anything else
|
---|
468 | if (indexer_type != 1) {
|
---|
469 | return terms;
|
---|
470 | }
|
---|
471 |
|
---|
472 | text_t outtext;
|
---|
473 | text_t word;
|
---|
474 |
|
---|
475 | text_t::const_iterator here = terms.begin();
|
---|
476 | text_t::const_iterator end = terms.end();
|
---|
477 |
|
---|
478 | while (here !=end) {
|
---|
479 |
|
---|
480 | if (is_unicode_letdig(*here) || is_special_character(indexer_type, *here)) {
|
---|
481 | // not word boundary
|
---|
482 | word.push_back(*here);
|
---|
483 | ++here;
|
---|
484 | }
|
---|
485 | else {
|
---|
486 | // found word boundary
|
---|
487 | if (!word.empty() ) {
|
---|
488 | if (starts_with(word, "NEAR") || starts_with(word, "WITHIN")) {
|
---|
489 | outtext += word;
|
---|
490 | word.clear();
|
---|
491 | }
|
---|
492 | else {
|
---|
493 | word += "#";
|
---|
494 | if (stem == "1") word += "s";
|
---|
495 | if (fold == "1") word += "i";
|
---|
496 | outtext += word;
|
---|
497 | word.clear();
|
---|
498 | }
|
---|
499 | }
|
---|
500 | // this only used in advanced form, so we leave in boolean operators
|
---|
501 | if (*here == '\"' || *here == '&' || *here == '|' || *here == '!' ||
|
---|
502 | *here == '(' || *here == ')' || is_unicode_space(*here)) {
|
---|
503 | outtext.push_back(*here);
|
---|
504 | }
|
---|
505 | ++here;
|
---|
506 | }
|
---|
507 | }
|
---|
508 |
|
---|
509 | // get last word
|
---|
510 | if (!word.empty()) {
|
---|
511 | word += "#";
|
---|
512 | if (stem == "1") word += "s";
|
---|
513 | if (fold == "1") word += "i";
|
---|
514 | word += " ";
|
---|
515 | outtext += word;
|
---|
516 | }
|
---|
517 | return outtext;
|
---|
518 | }
|
---|
519 |
|
---|
520 |
|
---|
521 | // some query form parsing functions for use with mgpp & lucene
|
---|
522 |
|
---|
523 | void parse_reg_query_form(text_t &querystring, cgiargsclass &args, bool segment)
|
---|
524 | {
|
---|
525 | querystring.clear();
|
---|
526 |
|
---|
527 | int argct = args.getintarg("ct");
|
---|
528 | int argt = args.getintarg("t");// t=0 -and, t=1 - or
|
---|
529 | int argb = args.getintarg("b");
|
---|
530 |
|
---|
531 | text_t combine;
|
---|
532 |
|
---|
533 | // lucene uses global combine, so only need this for mgpp
|
---|
534 | if (argct==1) {
|
---|
535 | if (argt == 0) combine = "&";
|
---|
536 | else combine = "|";
|
---|
537 | }
|
---|
538 |
|
---|
539 | text_t field = args["fqf"];
|
---|
540 | if (field.empty()) return; // no query
|
---|
541 | text_tarray fields;
|
---|
542 | splitchar(field.begin(), field.end(), ',', fields);
|
---|
543 |
|
---|
544 | text_t value = args["fqv"];
|
---|
545 | if (value.empty()) return; // somethings wrong
|
---|
546 | text_tarray values;
|
---|
547 | splitchar(value.begin(), value.end(), ',', values);
|
---|
548 |
|
---|
549 |
|
---|
550 | for (int i=0; i< values.size(); ++i) {
|
---|
551 | if (!values[i].empty()) {
|
---|
552 | text_t this_value = values[i];
|
---|
553 | // remove operators for simple search, segments text if necessary
|
---|
554 | format_querystring(this_value, argb, segment);
|
---|
555 | // add tag info for this field (and other processing)
|
---|
556 | format_field_info(this_value, fields[i], argct, argt, argb);
|
---|
557 | // add into query string
|
---|
558 | if (argct == 2) {
|
---|
559 | // lucene
|
---|
560 | // we don't worry about AND/OR, cos this is done by defaultcombineoperator
|
---|
561 | querystring += this_value+" ";
|
---|
562 | } else {
|
---|
563 | // mgpp
|
---|
564 | if (!querystring.empty()) {
|
---|
565 | querystring += " "+ combine+ " ";
|
---|
566 | }
|
---|
567 | querystring += this_value;
|
---|
568 | }
|
---|
569 | }
|
---|
570 | }
|
---|
571 | }
|
---|
572 |
|
---|
573 |
|
---|
574 | void parse_adv_query_form(text_t &querystring, cgiargsclass &args, bool segment){
|
---|
575 | querystring.clear();
|
---|
576 |
|
---|
577 | const int argct = args.getintarg("ct");
|
---|
578 | 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
|
---|
579 | int argb = args.getintarg("b");
|
---|
580 | text_t combine;
|
---|
581 | if (argct==1) {
|
---|
582 | combine = "&";
|
---|
583 | }
|
---|
584 | else { // lucene
|
---|
585 | combine = "AND";
|
---|
586 | }
|
---|
587 |
|
---|
588 | text_t field = args["fqf"];
|
---|
589 | if (field.empty()) return; // no query
|
---|
590 | text_tarray fields;
|
---|
591 | splitchar(field.begin(), field.end(), ',', fields);
|
---|
592 |
|
---|
593 | text_t value = args["fqv"];
|
---|
594 | if (value.empty()) return; // somethings wrong
|
---|
595 | text_tarray values;
|
---|
596 | splitchar(value.begin(), value.end(), ',', values);
|
---|
597 |
|
---|
598 | text_t comb = args["fqc"];
|
---|
599 | if (comb.empty()) return; //somethings wrong
|
---|
600 | text_tarray combs;
|
---|
601 | splitchar(comb.begin(), comb.end(), ',', combs);
|
---|
602 |
|
---|
603 | text_tarray stems;
|
---|
604 | text_tarray folds;
|
---|
605 | if (argct == 1) {// mgpp - lucene doesn't do stem/case
|
---|
606 | text_t stem = args["fqs"];
|
---|
607 | if (stem.empty()) return; // somethings wrong
|
---|
608 | splitchar(stem.begin(), stem.end(), ',', stems);
|
---|
609 |
|
---|
610 | text_t fold = args["fqk"];
|
---|
611 | if (fold.empty()) return; // somethings wrong
|
---|
612 | splitchar(fold.begin(), fold.end(), ',', folds);
|
---|
613 | }
|
---|
614 |
|
---|
615 | for(int i=0; i< values.size(); ++i) {
|
---|
616 | if (!values[i].empty()) {
|
---|
617 | if (i!=0) {
|
---|
618 | if (argct==1) {
|
---|
619 | if (combs[i-1]=="and") combine = "&";
|
---|
620 | else if (combs[i-1]=="or")combine = "|";
|
---|
621 | else if (combs[i-1]=="not")combine = "!";
|
---|
622 | }
|
---|
623 | else { // lucene
|
---|
624 | if (combs[i-1]=="and") combine = "AND";
|
---|
625 | else if (combs[i-1]=="or")combine = "OR";
|
---|
626 | else if (combs[i-1]=="not")combine = "NOT";
|
---|
627 | }
|
---|
628 | }
|
---|
629 | text_t this_value = values[i];
|
---|
630 | // remove operators for simple search, segments text if necessary
|
---|
631 | format_querystring(this_value, argb, segment);
|
---|
632 | if (argct == 1) { // mgpp only
|
---|
633 | this_value = addstemcase(this_value, stems[i], folds[i], argct);
|
---|
634 | }
|
---|
635 | // add tag info for this field (and other processing)
|
---|
636 | format_field_info(this_value, fields[i], argct, argt, argb);
|
---|
637 | // add into query string
|
---|
638 | if (!querystring.empty()) {
|
---|
639 | querystring += " "+ combine+ " ";
|
---|
640 | }
|
---|
641 | querystring += this_value;
|
---|
642 |
|
---|
643 | }
|
---|
644 | }
|
---|
645 | }
|
---|
646 |
|
---|
647 | // Extended addqueryelem for Human Info project
|
---|
648 | void addqueryelem_ex(text_t &querystring, const text_t &tag,
|
---|
649 | const text_t &terms, const text_t &stem,
|
---|
650 | const text_t &fold,
|
---|
651 | const text_t& combine, const text_t& word_combine) {
|
---|
652 |
|
---|
653 | if (!querystring.empty()) { // have to put and/or
|
---|
654 | querystring += " " + combine + " ";
|
---|
655 | }
|
---|
656 | text_t outtext; outtext.reserve(512);
|
---|
657 | text_t word; word.reserve(100);
|
---|
658 | //unsigned short c;
|
---|
659 | text_t::const_iterator here = terms.begin();
|
---|
660 | text_t::const_iterator end = terms.end();
|
---|
661 | bool inquote = false, firstword = true;
|
---|
662 |
|
---|
663 | text_t word2; word2.reserve(256);
|
---|
664 |
|
---|
665 | while (here !=end) {
|
---|
666 | if (is_unicode_space(*here)) {
|
---|
667 | if (word2 == "AND") { word2.clear(); word2.push_back(7527); word2.appendcarr("AND", 3); word2.push_back(7527); }
|
---|
668 | else if (word2 == "OR") { word2.clear(); word2.push_back(7527); word2.appendcarr("OR", 2); word2.push_back(7527); }
|
---|
669 | else if (word2 == "NOT") { word2.clear(); word2.push_back(7527); word2.appendcarr("NOT", 3); word2.push_back(7527); }
|
---|
670 | else if (word2 == "NEAR") { word2.clear(); word2.push_back(7527); word2.appendcarr("NEAR", 4); word2.push_back(7527); }
|
---|
671 | else if (word2 == "WITHIN") { word2.clear(); word2.push_back(7527); word2.appendcarr("WITHIN", 6); word2.push_back(7527); }
|
---|
672 | if (inquote) {
|
---|
673 | word2.push_back(*here);
|
---|
674 | }
|
---|
675 | word.append(word2); word2.clear();
|
---|
676 |
|
---|
677 | if (!inquote && !word.empty() ) {
|
---|
678 | // found word boundary
|
---|
679 |
|
---|
680 | if (stem == "1" || fold =="1") {
|
---|
681 | word += "#";
|
---|
682 | if (stem == "1") word += "s";
|
---|
683 | //else word += "u";
|
---|
684 |
|
---|
685 | if (fold == "1") word += "i";
|
---|
686 | //else word += "c";
|
---|
687 | }
|
---|
688 | if (firstword) {
|
---|
689 | firstword = false;
|
---|
690 | } else {
|
---|
691 | outtext += " " + word_combine + " ";
|
---|
692 | }
|
---|
693 | outtext += "[" + word + "]:"+tag;
|
---|
694 | word.clear();
|
---|
695 | }
|
---|
696 | ++here;
|
---|
697 | } else if (*here == '\"') {
|
---|
698 | word2.push_back(*here);
|
---|
699 | inquote = !inquote;
|
---|
700 | ++here;
|
---|
701 | } else {
|
---|
702 | // not word boundary
|
---|
703 | word2.push_back(*here);
|
---|
704 | ++here;
|
---|
705 | }
|
---|
706 | }
|
---|
707 |
|
---|
708 | // get last word
|
---|
709 | if (!word2.empty()) {
|
---|
710 | if (word2 == "AND") { word2.clear(); word2.push_back(7527); word2.appendcarr("AND", 3); word2.push_back(7527); }
|
---|
711 | else if (word2 == "OR") { word2.clear(); word2.push_back(7527); word2.appendcarr("OR", 2); word2.push_back(7527); }
|
---|
712 | else if (word2 == "NOT") { word2.clear(); word2.push_back(7527); word2.appendcarr("NOT", 3); word2.push_back(7527); }
|
---|
713 | else if (word2 == "NEAR") { word2.clear(); word2.push_back(7527); word2.appendcarr("NEAR", 4); word2.push_back(7527); }
|
---|
714 | else if (word2 == "WITHIN") { word2.clear(); word2.push_back(7527); word2.appendcarr("WITHIN", 6); word2.push_back(7527); }
|
---|
715 | word.append(word2); word2.clear();
|
---|
716 |
|
---|
717 | if (stem == "1"|| fold == "1") {
|
---|
718 | word += "#";
|
---|
719 | if (stem == "1") word += "s";
|
---|
720 | //else word += "u";
|
---|
721 |
|
---|
722 | if (fold == "1") word += "i";
|
---|
723 | //else word += "c";
|
---|
724 | }
|
---|
725 | if (!outtext.empty()) outtext += " " + word_combine + " ";
|
---|
726 | outtext += "[" + word + "]:"+tag;
|
---|
727 | }
|
---|
728 | querystring += "(" + outtext + ")";
|
---|
729 | }
|
---|
730 |
|
---|
731 | void add_field_info(text_t &querystring, const text_t &tag, int type) {
|
---|
732 |
|
---|
733 | if (tag == "") return; // do nothing
|
---|
734 | if (tag == "ZZ" && type == 1) return; // mgpp doesn't use ZZ tag internally
|
---|
735 | if (type == 1) { //mgpp
|
---|
736 | querystring = "["+querystring+"]:"+tag;
|
---|
737 | } else if (type == 2) { // lucene
|
---|
738 | querystring = tag+":("+querystring+")";
|
---|
739 | }
|
---|
740 |
|
---|
741 | }
|
---|
742 |
|
---|
743 |
|
---|
744 | void format_field_info_lucene(text_t &querystring, text_t &tag, int argt, int argb) {
|
---|
745 |
|
---|
746 | int type = 2; //lucene
|
---|
747 |
|
---|
748 | if (argb==0) { // simple
|
---|
749 | // there will be no & or | as they should have already been removed
|
---|
750 | // just tag the entire thing
|
---|
751 | if (tag != "") {
|
---|
752 | add_field_info(querystring, tag, type);
|
---|
753 | }
|
---|
754 | return;
|
---|
755 | }
|
---|
756 |
|
---|
757 | // need to replace & with &&, | with ||
|
---|
758 | text_t::const_iterator here = querystring.begin();
|
---|
759 | text_t::const_iterator end = querystring.end();
|
---|
760 |
|
---|
761 | text_t finalquery = "";
|
---|
762 | while (here != end) {
|
---|
763 | if (*here == '&') {
|
---|
764 | finalquery.push_back('&');
|
---|
765 | finalquery.push_back('&');
|
---|
766 | while (*(here+1) == '&') {
|
---|
767 | ++here;
|
---|
768 | }
|
---|
769 | }
|
---|
770 | else if (*here == '|') {
|
---|
771 | finalquery.push_back('|');
|
---|
772 | finalquery.push_back('|');
|
---|
773 | while (*(here+1) == '|') {
|
---|
774 | ++here;
|
---|
775 | }
|
---|
776 | }
|
---|
777 | else {
|
---|
778 | finalquery.push_back(*here);
|
---|
779 | }
|
---|
780 | ++here;
|
---|
781 | }
|
---|
782 | querystring = finalquery;
|
---|
783 | add_field_info(querystring, tag, type);
|
---|
784 | }
|
---|
785 |
|
---|
786 |
|
---|
787 | void format_field_info_mgpp(text_t &querystring, text_t tag, int argt, int argb) {
|
---|
788 |
|
---|
789 | if (tag == "ZZ") tag = ""; // ZZ is a special tag meaning no tag (all fields)
|
---|
790 | if (tag == "" && argb == 1) {
|
---|
791 | return; // no field specifier, advanced mode, the query stays as written
|
---|
792 | }
|
---|
793 |
|
---|
794 | int type = 1; // mgpp
|
---|
795 |
|
---|
796 | bool simple_and = (argb==0 && argt==0);
|
---|
797 | text_t finalquery = "";
|
---|
798 | text_t fieldpart ="";
|
---|
799 | text_t queryelem = "";
|
---|
800 | bool in_phrase = false;
|
---|
801 | bool in_field = false;
|
---|
802 |
|
---|
803 | text_t::const_iterator here = querystring.begin();
|
---|
804 | text_t::const_iterator end = querystring.end();
|
---|
805 | while (here != end) {
|
---|
806 | if (is_unicode_letdig(*here) || *here == '&' || is_special_character(type, *here)) {
|
---|
807 | queryelem.push_back(*here);
|
---|
808 | }
|
---|
809 | else if (*here == '|') {
|
---|
810 | in_field = false;
|
---|
811 | }
|
---|
812 | else if (*here == '!' || *here == '(' || *here == ')') {
|
---|
813 | if (!in_phrase) { // ignore these if in_phrase
|
---|
814 | // output field, then output operator
|
---|
815 | in_field = false;
|
---|
816 | if (!queryelem.empty()) {
|
---|
817 | if (!simple_and && !fieldpart.empty()) {
|
---|
818 | add_field_info(fieldpart, tag, type);
|
---|
819 | finalquery += fieldpart;
|
---|
820 | finalquery.push_back(' ');
|
---|
821 | fieldpart.clear();
|
---|
822 | }
|
---|
823 | fieldpart += queryelem;
|
---|
824 | }
|
---|
825 | if (!fieldpart.empty()) {
|
---|
826 | add_field_info(fieldpart, tag, type);
|
---|
827 | finalquery += fieldpart;
|
---|
828 | finalquery.push_back(' ');
|
---|
829 | }
|
---|
830 | fieldpart.clear();
|
---|
831 | queryelem.clear();
|
---|
832 | finalquery.push_back(*here);
|
---|
833 | finalquery.push_back(' ');
|
---|
834 | }
|
---|
835 | }
|
---|
836 | else if (*here == '"') {
|
---|
837 | queryelem.push_back(*here);
|
---|
838 | if (in_phrase == false) in_phrase = true;
|
---|
839 | else {
|
---|
840 | in_phrase = false;
|
---|
841 | }
|
---|
842 | }
|
---|
843 |
|
---|
844 | // Found word boundary, in a phrase
|
---|
845 | else if (in_phrase) {
|
---|
846 | queryelem.push_back(*here);
|
---|
847 | }
|
---|
848 | // Found a word boundary
|
---|
849 | else {
|
---|
850 | if (!queryelem.empty()) {
|
---|
851 | if (queryelem == "&") {
|
---|
852 | in_field = true;
|
---|
853 | queryelem.clear();
|
---|
854 | }
|
---|
855 | else if (starts_with(queryelem, "NEAR") || starts_with(queryelem, "WITHIN")) {
|
---|
856 |
|
---|
857 | if (argb==1) {
|
---|
858 | // simple search, these not allowed
|
---|
859 | in_field = true;
|
---|
860 | fieldpart += queryelem;
|
---|
861 | fieldpart.push_back(' ');
|
---|
862 | }
|
---|
863 | queryelem.clear();
|
---|
864 |
|
---|
865 | }
|
---|
866 | else {
|
---|
867 | if (!simple_and && !in_field) {
|
---|
868 | if (!fieldpart.empty()) {
|
---|
869 | add_field_info(fieldpart, tag, type);
|
---|
870 | finalquery += fieldpart;
|
---|
871 | finalquery.push_back(' ');
|
---|
872 | fieldpart.clear();
|
---|
873 | }
|
---|
874 | }
|
---|
875 |
|
---|
876 | fieldpart += queryelem;
|
---|
877 | fieldpart.push_back(' ');
|
---|
878 | queryelem.clear();
|
---|
879 | }
|
---|
880 | }
|
---|
881 | }
|
---|
882 | ++here;
|
---|
883 | }
|
---|
884 | // at the end
|
---|
885 | if (!queryelem.empty()) {
|
---|
886 | if (!simple_and && !in_field && !fieldpart.empty()) {
|
---|
887 | add_field_info(fieldpart, tag, type);
|
---|
888 | finalquery += fieldpart;
|
---|
889 | finalquery.push_back(' ');
|
---|
890 | fieldpart.clear();
|
---|
891 | }
|
---|
892 | fieldpart += queryelem;
|
---|
893 | }
|
---|
894 | if (!fieldpart.empty()) {
|
---|
895 | add_field_info(fieldpart, tag, type);
|
---|
896 | finalquery += fieldpart;
|
---|
897 | fieldpart.clear();
|
---|
898 |
|
---|
899 | // doesn't the following just leave a dangling space at the end ?? (used to make mgpp crash)
|
---|
900 | // consider cutting this line
|
---|
901 | finalquery.push_back(' ');
|
---|
902 | }
|
---|
903 |
|
---|
904 | querystring = finalquery;
|
---|
905 | }
|
---|
906 |
|
---|
907 |
|
---|
908 | void format_field_info(text_t &querystring, text_t tag, int argct, int argt, int argb) {
|
---|
909 | if (argct == 1) {
|
---|
910 | format_field_info_mgpp(querystring, tag, argt, argb);
|
---|
911 | } else if (argct == 2) {
|
---|
912 | format_field_info_lucene(querystring, tag, argt, argb);
|
---|
913 | }
|
---|
914 | }
|
---|
915 |
|
---|
916 | void mgpp_adddateelem(text_t& querystring, const int date)
|
---|
917 | {
|
---|
918 | querystring.appendcstr(" [");
|
---|
919 | if(date<0) {
|
---|
920 | querystring.appendcstr("bc");
|
---|
921 | querystring.appendint((date*-1));
|
---|
922 | }
|
---|
923 | else {
|
---|
924 | querystring.appendint(date);
|
---|
925 | }
|
---|
926 | querystring.appendcstr("]:CV");
|
---|
927 | }
|
---|
928 |
|
---|
929 | void lucene_adddateelem(text_t& querystring, const int date)
|
---|
930 | {
|
---|
931 | querystring.appendcstr(" CV:(");
|
---|
932 | if(date<0) {
|
---|
933 | querystring.appendcstr("bc");
|
---|
934 | querystring.appendint((date*-1));
|
---|
935 | }
|
---|
936 | else {
|
---|
937 | querystring.appendint(date);
|
---|
938 | }
|
---|
939 | querystring.appendcstr(")");
|
---|
940 | }
|
---|
941 |
|
---|
942 |
|
---|
943 | void add_dates(text_t &querystring, int startdate, int enddate,
|
---|
944 | int startbc, int endbc, int ct)
|
---|
945 | {
|
---|
946 | if(startdate)
|
---|
947 | {
|
---|
948 | int querystringis = 0;
|
---|
949 | text_t::const_iterator here = querystring.begin();
|
---|
950 | text_t::const_iterator end = querystring.end();
|
---|
951 | while(here!=end)
|
---|
952 | {
|
---|
953 | if(!(isspace((*here)))){
|
---|
954 | here = end;
|
---|
955 | querystringis = 1;
|
---|
956 | }
|
---|
957 | else
|
---|
958 | ++here;
|
---|
959 | }
|
---|
960 | //converting BCE dates
|
---|
961 | if(startbc && startdate > 0)
|
---|
962 | {
|
---|
963 | startdate *= -1;
|
---|
964 | }
|
---|
965 | if(endbc && enddate > 0)
|
---|
966 | {
|
---|
967 | enddate *= -1;
|
---|
968 | }
|
---|
969 | if(enddate != 0 && enddate<startdate)
|
---|
970 | {
|
---|
971 | cout<<"enddate too small"<<endl;
|
---|
972 | return;
|
---|
973 | }
|
---|
974 | if(querystringis)
|
---|
975 | querystring.appendcstr(" AND");
|
---|
976 | if(!enddate)
|
---|
977 | {
|
---|
978 | if (ct==1) {
|
---|
979 | mgpp_adddateelem(querystring,startdate);
|
---|
980 | }
|
---|
981 | else { // lucene
|
---|
982 | lucene_adddateelem(querystring,startdate);
|
---|
983 | }
|
---|
984 | }
|
---|
985 | else{
|
---|
986 | int nextdate = startdate;
|
---|
987 | querystring.appendcstr(" (");
|
---|
988 | while(nextdate<=enddate)
|
---|
989 | {
|
---|
990 | if(nextdate!=0) {
|
---|
991 | if (ct==1) {
|
---|
992 | mgpp_adddateelem(querystring,nextdate);
|
---|
993 | }
|
---|
994 | else { // lucene
|
---|
995 | lucene_adddateelem(querystring,nextdate);
|
---|
996 | }
|
---|
997 | }
|
---|
998 | ++nextdate;
|
---|
999 | }
|
---|
1000 | querystring.appendcstr(" )");
|
---|
1001 | }
|
---|
1002 | }
|
---|
1003 |
|
---|
1004 | }
|
---|