1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * mgsearch.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 | * $Id: mgsearch.cpp 633 1999-09-24 02:41:21Z rjmcnab $
|
---|
25 | *
|
---|
26 | *********************************************************************/
|
---|
27 |
|
---|
28 | /*
|
---|
29 | $Log$
|
---|
30 | Revision 1.22 1999/09/24 02:41:21 rjmcnab
|
---|
31 | change to use has_unicode_letdig in text_t
|
---|
32 |
|
---|
33 | Revision 1.21 1999/09/21 21:41:41 sjboddie
|
---|
34 | fixed an error in what I committed last
|
---|
35 |
|
---|
36 | Revision 1.20 1999/09/21 11:59:26 sjboddie
|
---|
37 | added Maxdocs queryfilter option (which may be -1 for 'all)
|
---|
38 |
|
---|
39 | Revision 1.19 1999/09/07 22:52:52 rjmcnab
|
---|
40 | Seems to be an error in mg for retrieving documents using a paragraph
|
---|
41 | based index for some cases. Just added a work around (loads the default
|
---|
42 | index every time).
|
---|
43 |
|
---|
44 | Revision 1.18 1999/09/07 04:57:22 sjboddie
|
---|
45 | added gpl notice
|
---|
46 |
|
---|
47 | Revision 1.17 1999/08/31 22:42:41 rjmcnab
|
---|
48 | A couple of minor things.
|
---|
49 |
|
---|
50 | Revision 1.16 1999/08/25 04:51:06 sjboddie
|
---|
51 | small change to allow for searching using boolean operators
|
---|
52 |
|
---|
53 | Revision 1.15 1999/07/16 08:35:03 rjmcnab
|
---|
54 | Fixed a weird bug to do with a faulty case statement.
|
---|
55 |
|
---|
56 | Revision 1.14 1999/07/16 03:42:22 sjboddie
|
---|
57 | changed isApprox
|
---|
58 |
|
---|
59 | Revision 1.13 1999/07/16 00:12:46 sjboddie
|
---|
60 | removed all the old post-processing stuff
|
---|
61 |
|
---|
62 | Revision 1.12 1999/07/07 06:17:47 rjmcnab
|
---|
63 | broke search_index into index+subcollection+language
|
---|
64 | within mgsearch
|
---|
65 |
|
---|
66 | Revision 1.11 1999/07/05 21:06:43 rjmcnab
|
---|
67 | Disabled quoted strings.
|
---|
68 |
|
---|
69 | Revision 1.10 1999/07/01 09:29:19 rjmcnab
|
---|
70 | Changes for better reporting of number documents which match a query. Changes
|
---|
71 | should still work as before with older versions of mg.
|
---|
72 |
|
---|
73 | Revision 1.9 1999/07/01 03:54:48 rjmcnab
|
---|
74 | Added code to plug in the equivalent terms of each of the query terms.
|
---|
75 | Also added a function to get a raw utf8 encoded mg document (for speeding
|
---|
76 | up a phrase matching function)
|
---|
77 |
|
---|
78 | Revision 1.8 1999/06/30 04:04:12 rjmcnab
|
---|
79 | made stemming functions available from mgsearch and made the stems
|
---|
80 | for the query terms available in queryinfo
|
---|
81 |
|
---|
82 | Revision 1.7 1999/06/27 22:07:27 sjboddie
|
---|
83 | got rid of all the old functions for dealing with dir indexes
|
---|
84 |
|
---|
85 | Revision 1.6 1999/06/09 00:41:32 sjboddie
|
---|
86 | phrase searching now uses case-folding if it's turned on
|
---|
87 |
|
---|
88 | Revision 1.5 1999/02/21 22:31:35 rjmcnab
|
---|
89 |
|
---|
90 | Removed locateinfo.
|
---|
91 |
|
---|
92 | Revision 1.4 1999/02/03 01:13:27 sjboddie
|
---|
93 |
|
---|
94 | Got interface to handle subcollections and language subcollections -
|
---|
95 | committed changes made to some of the collections
|
---|
96 |
|
---|
97 | Revision 1.3 1999/01/19 01:38:17 rjmcnab
|
---|
98 |
|
---|
99 | Made the source more portable.
|
---|
100 |
|
---|
101 | Revision 1.2 1999/01/12 01:51:02 rjmcnab
|
---|
102 |
|
---|
103 | Standard header.
|
---|
104 |
|
---|
105 | Revision 1.1 1999/01/08 09:02:16 rjmcnab
|
---|
106 |
|
---|
107 | Moved from src/library.
|
---|
108 |
|
---|
109 | */
|
---|
110 |
|
---|
111 |
|
---|
112 | #include "gsdlconf.h"
|
---|
113 | #include "mgsearch.h"
|
---|
114 | #include "fileutil.h"
|
---|
115 |
|
---|
116 | #include <string.h>
|
---|
117 | #include <stdio.h>
|
---|
118 | #include <stdlib.h>
|
---|
119 | #include <ctype.h>
|
---|
120 |
|
---|
121 | #if defined(GSDL_USE_OBJECTSPACE)
|
---|
122 | # include <ospace\std\iostream>
|
---|
123 | #elif defined(GSDL_USE_IOS_H)
|
---|
124 | # include <iostream.h>
|
---|
125 | #else
|
---|
126 | # include <iostream>
|
---|
127 | #endif
|
---|
128 |
|
---|
129 | #if defined(__WIN32__)
|
---|
130 | // gdbm stuff
|
---|
131 | # include "autoconf.h"
|
---|
132 | # include "systems.h"
|
---|
133 | # include "gdbmconst.h"
|
---|
134 | # include "gdbm.h"
|
---|
135 | #else
|
---|
136 | # include <gdbm.h>
|
---|
137 | #endif
|
---|
138 |
|
---|
139 |
|
---|
140 | #include <assert.h>
|
---|
141 |
|
---|
142 | #include "mgq.h"
|
---|
143 | // #include "locateinfo.h"
|
---|
144 | #include "gsdlunicode.h"
|
---|
145 | #include "unitool.h"
|
---|
146 |
|
---|
147 |
|
---|
148 | /////////////
|
---|
149 | // globals //
|
---|
150 | /////////////
|
---|
151 |
|
---|
152 | static char *tempdoc = NULL;
|
---|
153 | static int templen = 0;
|
---|
154 |
|
---|
155 |
|
---|
156 | //////////////////////
|
---|
157 | // useful functions //
|
---|
158 | //////////////////////
|
---|
159 |
|
---|
160 |
|
---|
161 | // input and output are in utf8
|
---|
162 | text_t mgsearch_stemword (const text_t &word) {
|
---|
163 | // allocate working stem space
|
---|
164 | int maxstemlen = mgq_getmaxstemlen ();
|
---|
165 | unsigned char *word_stem = new unsigned char [maxstemlen + 2];
|
---|
166 | if (word_stem == NULL) return "";
|
---|
167 |
|
---|
168 | // copy word to word_stem
|
---|
169 | int len = 0;
|
---|
170 | text_t::const_iterator here = word.begin();
|
---|
171 | text_t::const_iterator end = word.end();
|
---|
172 | while (len < maxstemlen && here != end) {
|
---|
173 | word_stem[len+1] = (unsigned char)(*here);
|
---|
174 | len++; here++;
|
---|
175 | }
|
---|
176 | word_stem[len+1] = '\0';
|
---|
177 | word_stem[0] = len;
|
---|
178 |
|
---|
179 | mgq_stemword (word_stem);
|
---|
180 |
|
---|
181 | // copy word_stem back to tempstr
|
---|
182 | text_t tempstr;
|
---|
183 | tempstr.setcarr((char *)(&word_stem[1]), word_stem[0]);
|
---|
184 |
|
---|
185 | delete [] word_stem;
|
---|
186 |
|
---|
187 | return tempstr;
|
---|
188 | }
|
---|
189 |
|
---|
190 |
|
---|
191 |
|
---|
192 | ////////////////////////
|
---|
193 | // callback functions //
|
---|
194 | ////////////////////////
|
---|
195 |
|
---|
196 | // This routine is called for each document found in a search
|
---|
197 | // it assumes that cache_num is set up correctly to point to
|
---|
198 | // a suitable result cache
|
---|
199 | int ourquerycallback(char * /*UDoc*/, int /*ULen*/, int DocNum,
|
---|
200 | float Weight, void *info) {
|
---|
201 |
|
---|
202 |
|
---|
203 | queryresultsclass *queryresults = (queryresultsclass * )info;
|
---|
204 |
|
---|
205 | // append this entry to the document results
|
---|
206 | docresultclass docresult;
|
---|
207 | docresult.docnum = DocNum;
|
---|
208 | docresult.num_query_terms_matched = (int)(Weight/100.0); // will always be 0 on some versions of mg...
|
---|
209 | docresult.docweight = Weight - docresult.num_query_terms_matched*100;
|
---|
210 |
|
---|
211 | queryresults->docs.docset[DocNum] = docresult;
|
---|
212 | queryresults->docs.docorder.push_back(DocNum);
|
---|
213 |
|
---|
214 | return 0;
|
---|
215 | }
|
---|
216 |
|
---|
217 | int termequivcallback(char *Word, int ULen, int /*Freq*/,
|
---|
218 | float /*Weight*/, void *info) {
|
---|
219 | text_tset *equivterms = (text_tset *)info;
|
---|
220 | if (equivterms == NULL) return 0;
|
---|
221 |
|
---|
222 | text_t thisterm;
|
---|
223 | thisterm.setcarr(Word, ULen);
|
---|
224 |
|
---|
225 | equivterms->insert(thisterm);
|
---|
226 |
|
---|
227 | return 0;
|
---|
228 | }
|
---|
229 |
|
---|
230 |
|
---|
231 | void mgsearch_equivterms (const text_t &word, text_tset &equivterms) {
|
---|
232 | // allocate working stem space
|
---|
233 | int maxstemlen = mgq_getmaxstemlen ();
|
---|
234 | unsigned char *word_stem = new unsigned char [maxstemlen + 2];
|
---|
235 | if (word_stem == NULL) return;
|
---|
236 |
|
---|
237 | // copy word to word_stem
|
---|
238 | int len = 0;
|
---|
239 | text_t::const_iterator here = word.begin();
|
---|
240 | text_t::const_iterator end = word.end();
|
---|
241 | while (len < maxstemlen && here != end) {
|
---|
242 | word_stem[len+1] = (unsigned char)(*here);
|
---|
243 | len++; here++;
|
---|
244 | }
|
---|
245 | word_stem[len+1] = '\0';
|
---|
246 | word_stem[0] = len;
|
---|
247 |
|
---|
248 | // get the equivalent terms
|
---|
249 | mgq_equivterms (word_stem, termequivcallback, (void *)(&equivterms));
|
---|
250 |
|
---|
251 | delete [] word_stem;
|
---|
252 |
|
---|
253 | return;
|
---|
254 | }
|
---|
255 |
|
---|
256 | text_tset utf8equivterms; // kept as utf8 string for fast matching
|
---|
257 |
|
---|
258 |
|
---|
259 | // This callback is called once for each term in the query
|
---|
260 | int termfreqcallback(char *Word, int ULen, int Freq,
|
---|
261 | float /*Weight*/, void *info) {
|
---|
262 | queryresultsclass *queryresults = (queryresultsclass *)info;
|
---|
263 | if (queryresults == NULL) return 0;
|
---|
264 |
|
---|
265 | text_t term;
|
---|
266 | term.setcarr(Word, ULen);
|
---|
267 | termfreqclass termfreq;
|
---|
268 |
|
---|
269 | termfreq.termstr = to_uni(term);
|
---|
270 | text_t utf8termstem = mgsearch_stemword (term);
|
---|
271 | termfreq.termstemstr = to_uni (utf8termstem);
|
---|
272 |
|
---|
273 | mgsearch_equivterms (utf8termstem, termfreq.utf8equivterms);
|
---|
274 |
|
---|
275 | termfreq.termfreq = Freq;
|
---|
276 | queryresults->orgterms.push_back(termfreq);
|
---|
277 |
|
---|
278 | return 0;
|
---|
279 | }
|
---|
280 |
|
---|
281 | // this callback is called once for each variation of each term
|
---|
282 | int termvariantscallback(char *Word, int ULen, int /*Freq*/,
|
---|
283 | float /*Weight*/, void *info) {
|
---|
284 |
|
---|
285 | text_t term;
|
---|
286 | term.setcarr(Word, ULen);
|
---|
287 | queryresultsclass *queryresults = (queryresultsclass *)info;
|
---|
288 | queryresults->termvariants.insert(to_uni(term));
|
---|
289 |
|
---|
290 | return 0;
|
---|
291 | }
|
---|
292 |
|
---|
293 | // This callback is for getting document text
|
---|
294 | int doctextcallback(char *Doc, int ULen, int /*Freq*/,
|
---|
295 | float /*Weight*/, void * /*info*/) {
|
---|
296 | tempdoc = Doc;
|
---|
297 | templen = ULen;
|
---|
298 |
|
---|
299 | return 0;
|
---|
300 | }
|
---|
301 |
|
---|
302 |
|
---|
303 | static text_t getindexsuffix (const text_t &collection,
|
---|
304 | const text_t &index) {
|
---|
305 |
|
---|
306 | text_t indexsuffix = "index";
|
---|
307 | indexsuffix = filename_cat (indexsuffix, index);
|
---|
308 | indexsuffix = filename_cat (indexsuffix, collection);
|
---|
309 | return indexsuffix;
|
---|
310 | }
|
---|
311 |
|
---|
312 |
|
---|
313 |
|
---|
314 |
|
---|
315 | ////////////////////
|
---|
316 | // mgsearch class //
|
---|
317 | ////////////////////
|
---|
318 |
|
---|
319 | mgsearchclass::mgsearchclass ()
|
---|
320 | {
|
---|
321 | cache = new querycache (RESULTCACHESIZE);
|
---|
322 | }
|
---|
323 |
|
---|
324 | mgsearchclass::~mgsearchclass ()
|
---|
325 | {
|
---|
326 | if (cache != NULL)
|
---|
327 | {
|
---|
328 | delete cache;
|
---|
329 | cache = NULL;
|
---|
330 | }
|
---|
331 | }
|
---|
332 |
|
---|
333 |
|
---|
334 | void mgsearchclass::setcollectdir (const text_t &thecollectdir)
|
---|
335 | {
|
---|
336 | collectdir = thecollectdir;
|
---|
337 | }
|
---|
338 |
|
---|
339 | // you only need to use this function before doing any stemming
|
---|
340 | // casefolding and stemming will be set if values for them are
|
---|
341 | // provided (0 or 1).
|
---|
342 | // makeindexcurrent returns true if it was able to load the database
|
---|
343 | bool mgsearchclass::makeindexcurrent (const text_t &index,
|
---|
344 | const text_t &subcollection,
|
---|
345 | const text_t &language,
|
---|
346 | const text_t &collection,
|
---|
347 | int casefolding,
|
---|
348 | int stemming) {
|
---|
349 | bool databaseloaded = true;
|
---|
350 |
|
---|
351 | // get the names of the collection, index and text suffixes
|
---|
352 | char *ccollection = collection.getcstr();
|
---|
353 | assert (ccollection != NULL);
|
---|
354 | char *idxsuffix = (getindexsuffix (collection, (index+subcollection+language))).getcstr();
|
---|
355 | assert (idxsuffix != NULL);
|
---|
356 | char *txtsuffix = (getindexsuffix (collection, "text")).getcstr();
|
---|
357 | assert (txtsuffix != NULL);
|
---|
358 |
|
---|
359 | #ifdef __WIN32__
|
---|
360 | char *ccollectdir = (collectdir+"\\").getcstr(); assert (ccollectdir != NULL);
|
---|
361 | #else
|
---|
362 | char *ccollectdir = collectdir.getcstr(); assert (ccollectdir != NULL);
|
---|
363 | #endif
|
---|
364 |
|
---|
365 | if (load_database(ccollection, ccollectdir, idxsuffix, txtsuffix)) {
|
---|
366 | if (casefolding == 0) mgq_ask(".set casefold off");
|
---|
367 | else if (casefolding > 0) mgq_ask(".set casefold on");
|
---|
368 | if (stemming == 0) mgq_ask(".set stem off");
|
---|
369 | else if (stemming > 0) mgq_ask(".set stem on");
|
---|
370 |
|
---|
371 | } else databaseloaded = false;
|
---|
372 |
|
---|
373 | // free up the c strings
|
---|
374 | delete ccollection;
|
---|
375 | delete idxsuffix;
|
---|
376 | delete txtsuffix;
|
---|
377 | delete ccollectdir;
|
---|
378 |
|
---|
379 | return databaseloaded;
|
---|
380 | }
|
---|
381 |
|
---|
382 |
|
---|
383 | // stem word uses the values set in the last call to makeindexcurrent
|
---|
384 | // to stem the word. It is assumed that word is in unicode
|
---|
385 | text_t mgsearchclass::stemword (const text_t &word) {
|
---|
386 | return to_uni (mgsearch_stemword (to_utf8 (word)));
|
---|
387 | }
|
---|
388 |
|
---|
389 | text_t mgsearchclass::stemword (text_t::const_iterator here, text_t::const_iterator end) {
|
---|
390 | return to_uni (mgsearch_stemword (to_utf8 (here, end)));
|
---|
391 | }
|
---|
392 |
|
---|
393 |
|
---|
394 | bool mgsearchclass::search(const queryparamclass &queryparams,
|
---|
395 | queryresultsclass &queryresults) {
|
---|
396 | assert (cache != NULL);
|
---|
397 |
|
---|
398 | queryresults.clear();
|
---|
399 |
|
---|
400 | // first check the cache
|
---|
401 | if (cache->find(queryparams, queryresults)) return true;
|
---|
402 |
|
---|
403 | // make sure there is a query to be processed
|
---|
404 | if (!has_unicode_letdig(queryparams.querystring)) return true;
|
---|
405 |
|
---|
406 | if (makeindexcurrent (queryparams.index, queryparams.subcollection,
|
---|
407 | queryparams.language, queryparams.collection)) {
|
---|
408 | setsearchmode (queryparams);
|
---|
409 | submitquery (queryparams);
|
---|
410 | getresults (queryparams, queryresults);
|
---|
411 | return true;
|
---|
412 | }
|
---|
413 |
|
---|
414 | return false;
|
---|
415 | }
|
---|
416 |
|
---|
417 |
|
---|
418 | void mgsearchclass::setsearchmode (const queryparamclass &queryparams)
|
---|
419 | {
|
---|
420 | mgq_ask(".set expert true");
|
---|
421 | mgq_ask(".set sorted_terms true");
|
---|
422 | mgq_ask(".set accumulator_method list");
|
---|
423 | mgq_ask(".set max_accumulators 500000");
|
---|
424 | mgq_ask(".set maxparas 500000");
|
---|
425 | mgq_ask(".set verbatim true");
|
---|
426 | mgq_ask(".unset skip_dump");
|
---|
427 | mgq_ask(".set mode docnums");
|
---|
428 |
|
---|
429 | switch (queryparams.search_type)
|
---|
430 | {
|
---|
431 | case 0: mgq_ask(".set query boolean"); break;
|
---|
432 | case 1: mgq_ask(".set query ranked"); break;
|
---|
433 | }
|
---|
434 | switch (queryparams.casefolding)
|
---|
435 | {
|
---|
436 | case 1: mgq_ask(".set casefold on"); break;
|
---|
437 | case 0: mgq_ask(".set casefold off"); break;
|
---|
438 | }
|
---|
439 | switch (queryparams.stemming)
|
---|
440 | {
|
---|
441 | case 1: mgq_ask(".set stem on"); break;
|
---|
442 | case 0: mgq_ask(".set stem off"); break;
|
---|
443 | }
|
---|
444 | mgq_ask(".set heads_length 150");
|
---|
445 |
|
---|
446 | if (queryparams.maxdocs == -1) {
|
---|
447 | mgq_ask(".set maxdocs all");
|
---|
448 | } else {
|
---|
449 | char maxdocstr[32];
|
---|
450 | sprintf(maxdocstr, ".set maxdocs %i", queryparams.maxdocs);
|
---|
451 | mgq_ask(maxdocstr);
|
---|
452 | }
|
---|
453 | }
|
---|
454 |
|
---|
455 |
|
---|
456 | void mgsearchclass::submitquery (const queryparamclass &queryparams)
|
---|
457 | {
|
---|
458 | // sort out the query string
|
---|
459 | text_t ttquerystring = queryparams.querystring;
|
---|
460 | filterquery (ttquerystring);
|
---|
461 | char *querystring = to_utf8(ttquerystring).getcstr();
|
---|
462 |
|
---|
463 | // submit the query
|
---|
464 | mgq_ask(querystring);
|
---|
465 |
|
---|
466 | delete querystring;
|
---|
467 | }
|
---|
468 |
|
---|
469 |
|
---|
470 | void mgsearchclass::getresults (const queryparamclass &queryparams,
|
---|
471 | queryresultsclass &queryresults) {
|
---|
472 |
|
---|
473 | int howmany = queryparams.maxdocs;
|
---|
474 | if (howmany == -1) howmany = MAXNUMDOCS;
|
---|
475 | mgq_results(result_docnums, 0, howmany,
|
---|
476 | ourquerycallback, (void *)(&queryresults));
|
---|
477 |
|
---|
478 | // get the term frequencies
|
---|
479 | mgq_results(result_termfreqs, 0, MAXNUMTERMS,
|
---|
480 | termfreqcallback, (void *)(&queryresults));
|
---|
481 | queryresults.sortuniqqueryterms();
|
---|
482 |
|
---|
483 | // get term variants
|
---|
484 | mgq_results(result_terms, 0, MAXNUMTERMS,
|
---|
485 | termvariantscallback, (void *)(&queryresults));
|
---|
486 |
|
---|
487 | // get the number of documents retrieved
|
---|
488 | int total_retrieved = 0, is_approx = 0;
|
---|
489 | mgq_docsretrieved (&total_retrieved, &is_approx);
|
---|
490 |
|
---|
491 | if (total_retrieved == 0) {
|
---|
492 | // not available (or really was zero)
|
---|
493 | queryresults.docs_matched = queryresults.docs.docset.size();
|
---|
494 | if ((queryparams.maxdocs == -1) ||
|
---|
495 | (queryresults.docs_matched < queryparams.maxdocs))
|
---|
496 | queryresults.is_approx = Exact;
|
---|
497 | else
|
---|
498 | queryresults.is_approx = MoreThan;
|
---|
499 | } else {
|
---|
500 | queryresults.docs_matched = total_retrieved;
|
---|
501 | if (is_approx) queryresults.is_approx = Approximate;
|
---|
502 | else queryresults.is_approx = Exact;
|
---|
503 | }
|
---|
504 | }
|
---|
505 |
|
---|
506 | void mgsearchclass::filterquery (text_t &ttquerystring) {
|
---|
507 | text_t::iterator ithere = ttquerystring.begin ();
|
---|
508 | text_t::iterator itend = ttquerystring.end ();
|
---|
509 |
|
---|
510 | // remove all non alphanumeric characters (except
|
---|
511 | // boolean operators
|
---|
512 | while (ithere != itend) {
|
---|
513 | if ((!is_unicode_letdig(*ithere)) && (*ithere != '!') &&
|
---|
514 | (*ithere != '&') && (*ithere != '|') && (*ithere != '(') &&
|
---|
515 | (*ithere != ')')) (*ithere) = ' ';
|
---|
516 | ithere++;
|
---|
517 | }
|
---|
518 | }
|
---|
519 |
|
---|
520 |
|
---|
521 | // the document text for 'docnum' is placed in 'output'
|
---|
522 | // docTargetDocument returns 'true' if it was able to
|
---|
523 | // try to get a document
|
---|
524 | // collection is needed to see if an index from the
|
---|
525 | // collection is loaded. If no index has been loaded
|
---|
526 | // defaultindex is needed to load one
|
---|
527 | bool mgsearchclass::docTargetDocument(const text_t &defaultindex,
|
---|
528 | const text_t &defaultsubcollection,
|
---|
529 | const text_t &defaultlanguage,
|
---|
530 | const text_t &collection,
|
---|
531 | int docnum,
|
---|
532 | text_t &output) {
|
---|
533 | output.clear();
|
---|
534 |
|
---|
535 | // get the mg version of the document
|
---|
536 | char *mgdoc = NULL;
|
---|
537 | int doclen = 0;
|
---|
538 | if (!mgdocument (defaultindex, defaultsubcollection, defaultlanguage,
|
---|
539 | collection, docnum, mgdoc, doclen)) return false;
|
---|
540 | if (mgdoc == NULL) return false;
|
---|
541 |
|
---|
542 | // replace all control-Cs with spaces
|
---|
543 | char *mgdoc_here = mgdoc;
|
---|
544 | char *mgdoc_end = mgdoc + doclen;
|
---|
545 | while (mgdoc_here < mgdoc_end) {
|
---|
546 | if (*mgdoc_here == '\x3') *mgdoc_here = ' ';
|
---|
547 | mgdoc_here++;
|
---|
548 | }
|
---|
549 |
|
---|
550 | // convert this document to unicode
|
---|
551 | utf8inconvertclass inconvert;
|
---|
552 | convertclass::status_t status;
|
---|
553 | inconvert.reset ();
|
---|
554 | inconvert.setinput (mgdoc, doclen);
|
---|
555 | inconvert.convert (output, status);
|
---|
556 |
|
---|
557 | return true;
|
---|
558 | }
|
---|
559 |
|
---|
560 |
|
---|
561 | bool mgsearchclass::mgdocument (const text_t &defaultindex,
|
---|
562 | const text_t &defaultsubcollection,
|
---|
563 | const text_t &defaultlanguage,
|
---|
564 | const text_t &collection,
|
---|
565 | int docnum,
|
---|
566 | char *&UDoc, int &ULen) {
|
---|
567 | int databaseloaded = 0;
|
---|
568 |
|
---|
569 | UDoc = NULL; ULen = 0;
|
---|
570 |
|
---|
571 | // see if we can make an appropriate database current
|
---|
572 | // char *ccollection = collection.getcstr();
|
---|
573 | // assert (ccollection != NULL);
|
---|
574 | // databaseloaded = load_text_database (ccollection);
|
---|
575 | // delete ccollection;
|
---|
576 |
|
---|
577 | // try and load the database
|
---|
578 | // if (!databaseloaded)
|
---|
579 | databaseloaded = makeindexcurrent (defaultindex, defaultsubcollection,
|
---|
580 | defaultlanguage, collection);
|
---|
581 |
|
---|
582 | if (databaseloaded) {
|
---|
583 | // retrieve the document from mg
|
---|
584 | char docstr[32];
|
---|
585 | sprintf(docstr, "%i", docnum);
|
---|
586 |
|
---|
587 | mgq_ask(".set mode text");
|
---|
588 | mgq_ask(".set query docnums");
|
---|
589 | mgq_ask(docstr);
|
---|
590 |
|
---|
591 | tempdoc = NULL;
|
---|
592 | templen = 0;
|
---|
593 | mgq_results (result_docs, 0, 1, doctextcallback, (void *)NULL);
|
---|
594 | UDoc = tempdoc;
|
---|
595 | ULen = templen;
|
---|
596 | }
|
---|
597 |
|
---|
598 | return (bool)databaseloaded;
|
---|
599 | }
|
---|
600 |
|
---|