[270] | 1 | /**********************************************************************
|
---|
| 2 | *
|
---|
| 3 | * querytools.cpp --
|
---|
| 4 | * Copyright (C) 1999 The New Zealand Digital Library Project
|
---|
| 5 | *
|
---|
[533] | 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.
|
---|
[270] | 9 | *
|
---|
[533] | 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 | *
|
---|
[270] | 24 | *********************************************************************/
|
---|
| 25 |
|
---|
| 26 | #include "querytools.h"
|
---|
[1373] | 27 | #include <ctype.h>
|
---|
[1914] | 28 | #include "unitool.h" // for is_unicode_letdig
|
---|
[270] | 29 |
|
---|
[759] | 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) {
|
---|
[270] | 34 |
|
---|
| 35 | request.filterName = "QueryFilter";
|
---|
| 36 |
|
---|
| 37 | OptionValue_t option;
|
---|
[470] | 38 |
|
---|
[270] | 39 | option.name = "Term";
|
---|
[759] | 40 | option.value = querystring;
|
---|
[270] | 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 |
|
---|
[1774] | 47 | option.name = "MatchMode";
|
---|
| 48 | option.value = (args.getintarg("t")) ? "some" : "all";
|
---|
| 49 | request.filterOptions.push_back (option);
|
---|
| 50 |
|
---|
[270] | 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 | }
|
---|
[1329] | 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 | }
|
---|
[270] | 82 |
|
---|
[759] | 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 |
|
---|
[349] | 91 | // fill in the second query if needed
|
---|
| 92 | if (!args["cq2"].empty()) {
|
---|
[759] | 93 | OptionValue_t option;
|
---|
| 94 |
|
---|
[349] | 95 | option.name = "CombineQuery";
|
---|
| 96 | option.value = args["cq2"];
|
---|
| 97 | request.filterOptions.push_back (option);
|
---|
| 98 |
|
---|
| 99 | option.name = "Term";
|
---|
[759] | 100 | option.value = querystring2;
|
---|
[349] | 101 | request.filterOptions.push_back (option);
|
---|
[759] | 102 |
|
---|
[349] | 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 | }
|
---|
[759] | 133 | set_more_queryfilter_options (request, args);
|
---|
| 134 | }
|
---|
[608] | 135 |
|
---|
[759] | 136 | void set_more_queryfilter_options (FilterRequest_t &request, cgiargsclass &args) {
|
---|
| 137 |
|
---|
| 138 | OptionValue_t option;
|
---|
[608] | 139 | int arg_m = args.getintarg("m");
|
---|
[759] | 140 |
|
---|
[608] | 141 | option.name = "Maxdocs";
|
---|
| 142 | option.value = arg_m;
|
---|
| 143 | request.filterOptions.push_back (option);
|
---|
[1329] | 144 |
|
---|
[759] | 145 | // option.name = "StartResults";
|
---|
| 146 | // option.value = args["r"];
|
---|
| 147 | // request.filterOptions.push_back (option);
|
---|
[270] | 148 |
|
---|
[759] | 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);
|
---|
[270] | 154 | }
|
---|
| 155 |
|
---|
[470] | 156 | void format_querystring (text_t &querystring, int querymode) {
|
---|
[270] | 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
|
---|
[470] | 168 | // "all the words" queries (unless querymode is advanced)
|
---|
[270] | 169 | while (here != end) {
|
---|
[470] | 170 | if ((querymode == 0) && (*here == '(' || *here == ')' || *here == '|' ||
|
---|
| 171 | *here == '!' || *here == '&')) {
|
---|
[270] | 172 | formattedstring.push_back(' ');
|
---|
| 173 | } else {
|
---|
[397] | 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;
|
---|
[270] | 181 | } else {
|
---|
[397] | 182 | // non-Chinese character
|
---|
| 183 | formattedstring.push_back (*here);
|
---|
| 184 | space = false;
|
---|
[270] | 185 | }
|
---|
| 186 | }
|
---|
| 187 | here ++;
|
---|
| 188 | }
|
---|
[397] | 189 | querystring = formattedstring;
|
---|
[270] | 190 | }
|
---|
| 191 |
|
---|
[1373] | 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(" [");
|
---|
[1467] | 230 | if(startdate<0)
|
---|
| 231 | {
|
---|
| 232 | querystring.appendcstr("bc");
|
---|
| 233 | querystring.appendint((startdate*-1));
|
---|
| 234 | }
|
---|
| 235 | else
|
---|
| 236 | querystring.appendint(startdate);
|
---|
[1889] | 237 | querystring.appendcstr("]:CO");
|
---|
[1373] | 238 | }
|
---|
| 239 | else{
|
---|
| 240 | int nextdate = startdate;
|
---|
| 241 | querystring.appendcstr(" (");
|
---|
| 242 | while(nextdate<=enddate)
|
---|
| 243 | {
|
---|
| 244 | if(nextdate!=0)
|
---|
| 245 | {
|
---|
[1467] | 246 | querystring.appendcstr(" [");
|
---|
| 247 | if(nextdate<0)
|
---|
| 248 | {
|
---|
| 249 | querystring.appendcstr("bc");
|
---|
| 250 | querystring.appendint((nextdate*-1));
|
---|
| 251 | }
|
---|
| 252 | else
|
---|
| 253 | querystring.appendint(nextdate);
|
---|
[2745] | 254 | querystring.appendcstr("]:CO");
|
---|
[1373] | 255 | }
|
---|
| 256 | nextdate++;
|
---|
| 257 | }
|
---|
| 258 | querystring.appendcstr(" )");
|
---|
| 259 | }
|
---|
| 260 | }
|
---|
[1467] | 261 |
|
---|
[1373] | 262 | }
|
---|
[1467] | 263 |
|
---|
[403] | 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 | }
|
---|
[1914] | 291 |
|
---|
| 292 | // search history tool
|
---|
| 293 | text_t escape_quotes(const text_t &querystring) {
|
---|
| 294 |
|
---|
| 295 | text_t::const_iterator here = querystring.begin();
|
---|
| 296 | text_t::const_iterator end = querystring.end();
|
---|
| 297 |
|
---|
| 298 | text_t escquery = "";
|
---|
| 299 | while (here != end) {
|
---|
[1988] | 300 | if (*here != '\'' && *here != '\"' && *here != '\n' && *here != '\r') escquery.push_back(*here);
|
---|
| 301 | else if (*here == '\n' || *here == '\r') {
|
---|
| 302 | escquery.push_back(' ');
|
---|
| 303 | } else {
|
---|
[1914] | 304 | escquery +="\\\\";
|
---|
| 305 | escquery.push_back(*here);
|
---|
| 306 | }
|
---|
| 307 |
|
---|
| 308 | here++;
|
---|
| 309 | }
|
---|
| 310 | return escquery;
|
---|
| 311 |
|
---|
| 312 | }
|
---|
| 313 |
|
---|
| 314 | // some query form parsing functions for use with mgpp
|
---|
| 315 |
|
---|
| 316 | void parse_reg_query_form(text_t &querystring, cgiargsclass &args){
|
---|
| 317 |
|
---|
| 318 | querystring.clear();
|
---|
| 319 | text_t combine;
|
---|
| 320 | int argt = args.getintarg("t");// t=0 -and, t=1 - or
|
---|
| 321 | if (argt == 0) combine = "&";
|
---|
| 322 | else combine = "|";
|
---|
| 323 |
|
---|
| 324 | text_t field = args["fqf"];
|
---|
| 325 | if (field.empty()) return; // no query
|
---|
| 326 | text_tarray fields;
|
---|
| 327 | splitchar(field.begin(), field.end(), ',', fields);
|
---|
| 328 |
|
---|
| 329 | text_t value = args["fqv"];
|
---|
| 330 | if (value.empty()) return; // somethings wrong
|
---|
| 331 | text_tarray values;
|
---|
| 332 | splitchar(value.begin(), value.end(), ',', values);
|
---|
| 333 |
|
---|
| 334 | for (int i=0; i< values.size(); i++) {
|
---|
| 335 | if (!values[i].empty()) {
|
---|
| 336 | text_t text = formatelem(values[i]);
|
---|
[2745] | 337 | addqueryelem(querystring, fields[i], text, combine);
|
---|
[1914] | 338 | }
|
---|
| 339 | }
|
---|
| 340 |
|
---|
| 341 | }
|
---|
| 342 |
|
---|
| 343 | text_t formatelem(text_t &text) {
|
---|
| 344 |
|
---|
| 345 | text_t::iterator here = text.begin();
|
---|
| 346 | text_t::iterator end = text.end();
|
---|
| 347 |
|
---|
| 348 | bool space = false;
|
---|
[2745] | 349 | int words = 0;
|
---|
[1914] | 350 | text_t newtext = "";
|
---|
| 351 | while (here != end) {
|
---|
| 352 | if (is_unicode_letdig(*here)) {
|
---|
| 353 | newtext.push_back(*here);
|
---|
[2745] | 354 | if (space==false) words++;
|
---|
[1914] | 355 | space = true;
|
---|
| 356 | }
|
---|
| 357 | else {
|
---|
| 358 | if (space) {
|
---|
| 359 | newtext.push_back(' ');
|
---|
| 360 | space = false;
|
---|
| 361 | }
|
---|
| 362 | }
|
---|
| 363 | here++;
|
---|
| 364 | }
|
---|
| 365 |
|
---|
[2745] | 366 | if (words > 1) { // have a phrase, put it in quotes
|
---|
| 367 | newtext = "\""+newtext+"\"";
|
---|
| 368 | cout <<"new text="<<newtext.getcstr()<<endl;
|
---|
| 369 | }
|
---|
[1914] | 370 | return newtext;
|
---|
| 371 | }
|
---|
| 372 |
|
---|
| 373 | void parse_adv_query_form(text_t &querystring, cgiargsclass &args){
|
---|
| 374 |
|
---|
| 375 | querystring.clear();
|
---|
| 376 | text_t combine = "&";
|
---|
| 377 |
|
---|
| 378 | text_t field = args["fqf"];
|
---|
| 379 | if (field.empty()) return; // no query
|
---|
| 380 | text_tarray fields;
|
---|
| 381 | splitchar(field.begin(), field.end(), ',', fields);
|
---|
| 382 |
|
---|
| 383 | text_t value = args["fqv"];
|
---|
| 384 | if (value.empty()) return; // somethings wrong
|
---|
| 385 | text_tarray values;
|
---|
| 386 | splitchar(value.begin(), value.end(), ',', values);
|
---|
| 387 |
|
---|
| 388 | text_t stem = args["fqs"];
|
---|
| 389 | if (stem.empty()) return; // somethings wrong
|
---|
| 390 | text_tarray stems;
|
---|
| 391 | splitchar(stem.begin(), stem.end(), ',', stems);
|
---|
| 392 |
|
---|
| 393 | text_t fold = args["fqk"];
|
---|
| 394 | if (fold.empty()) return; // somethings wrong
|
---|
| 395 | text_tarray folds;
|
---|
| 396 | splitchar(fold.begin(), fold.end(), ',', folds);
|
---|
| 397 |
|
---|
| 398 | text_t comb = args["fqc"];
|
---|
| 399 | if (comb.empty()) return; //somethings wrong
|
---|
| 400 | text_tarray combs;
|
---|
| 401 | splitchar(comb.begin(), comb.end(), ',', combs);
|
---|
| 402 |
|
---|
| 403 | for(int i=0; i< values.size(); i++) {
|
---|
| 404 | if (!values[i].empty()) {
|
---|
| 405 | if (i!=0) {
|
---|
| 406 | if (combs[i-1]=="and") combine = "&";
|
---|
| 407 | else if (combs[i-1]=="or")combine = "|";
|
---|
| 408 | else if (combs[i-1]=="not")combine = "!";
|
---|
| 409 | }
|
---|
[2745] | 410 | text_t term = formatelem(values[i]);
|
---|
[1914] | 411 | term = addstemcase(term, stems[i], folds[i]);
|
---|
| 412 | addqueryelem(querystring, fields[i], term, combine);
|
---|
| 413 | }
|
---|
| 414 |
|
---|
| 415 | }
|
---|
| 416 | }
|
---|
| 417 |
|
---|
| 418 | text_t addstemcase(text_t &terms, text_t &stem, text_t &fold) {
|
---|
| 419 |
|
---|
| 420 | text_t outtext;
|
---|
| 421 | text_t word;
|
---|
| 422 | //unsigned short c;
|
---|
| 423 | text_t::iterator here = terms.begin();
|
---|
| 424 | text_t::iterator end = terms.end();
|
---|
| 425 |
|
---|
| 426 | while (here !=end) {
|
---|
[2745] | 427 |
|
---|
[1914] | 428 | if (is_unicode_letdig(*here)) {
|
---|
| 429 | // not word boundary
|
---|
| 430 | word.push_back(*here);
|
---|
| 431 | here++;
|
---|
| 432 | }
|
---|
| 433 | else {
|
---|
| 434 | // found word boundary
|
---|
| 435 | if (!word.empty() ) {
|
---|
| 436 | if (stem == "1" || fold =="1") {
|
---|
| 437 | word += "#";
|
---|
| 438 | if (stem == "1") word += "s";
|
---|
| 439 | //else word += "u";
|
---|
| 440 |
|
---|
| 441 | if (fold == "1") word += "i";
|
---|
| 442 | //else word += "c";
|
---|
| 443 | }
|
---|
[2745] | 444 |
|
---|
[1914] | 445 | word += " ";
|
---|
| 446 | outtext += word;
|
---|
| 447 | word.clear();
|
---|
| 448 | }
|
---|
[2745] | 449 | if (*here == '\"') {
|
---|
| 450 | outtext.push_back(*here);
|
---|
| 451 | }
|
---|
[1914] | 452 | here++;
|
---|
| 453 | }
|
---|
| 454 | }
|
---|
| 455 |
|
---|
| 456 | // get last word
|
---|
| 457 | if (!word.empty()) {
|
---|
| 458 | if (stem == "1"|| fold == "1") {
|
---|
| 459 | word += "#";
|
---|
| 460 | if (stem == "1") word += "s";
|
---|
| 461 | //else word += "u";
|
---|
| 462 |
|
---|
| 463 | if (fold == "1") word += "i";
|
---|
| 464 | //else word += "c";
|
---|
| 465 | }
|
---|
| 466 | word += " ";
|
---|
| 467 | outtext += word;
|
---|
| 468 | }
|
---|
| 469 | return outtext;
|
---|
| 470 | }
|
---|
| 471 |
|
---|
| 472 |
|
---|
| 473 |
|
---|
| 474 | void addqueryelem(text_t &querystring, text_t &tag,
|
---|
| 475 | text_t &query, text_t combine) {
|
---|
| 476 | if (!querystring.empty()) { // have to put and/or
|
---|
| 477 | querystring += " "+combine + " ";
|
---|
| 478 |
|
---|
| 479 | }
|
---|
| 480 | if (tag=="ZZ") { // just add onto querystring
|
---|
| 481 | querystring += query;
|
---|
| 482 | }
|
---|
| 483 | else {
|
---|
| 484 | querystring += "["+query+"]:"+tag;
|
---|
| 485 | }
|
---|
| 486 |
|
---|
| 487 | }
|
---|
| 488 |
|
---|
| 489 |
|
---|
| 490 |
|
---|