1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * queryfilter.cpp --
|
---|
4 | * Copyright (C) 1999 The New Zealand Digital Library Project
|
---|
5 | *
|
---|
6 | * PUT COPYRIGHT NOTICE HERE
|
---|
7 | *
|
---|
8 | * $Id: queryfilter.cpp 334 1999-07-01 09:29:21Z rjmcnab $
|
---|
9 | *
|
---|
10 | *********************************************************************/
|
---|
11 |
|
---|
12 | /*
|
---|
13 | $Log$
|
---|
14 | Revision 1.9 1999/07/01 09:29:20 rjmcnab
|
---|
15 | Changes for better reporting of number documents which match a query. Changes
|
---|
16 | should still work as before with older versions of mg.
|
---|
17 |
|
---|
18 | Revision 1.8 1999/07/01 03:59:54 rjmcnab
|
---|
19 | reduced MAXDOCS to 200 (more reasonable ???). I also added a virtual
|
---|
20 | method for post-processing the query.
|
---|
21 |
|
---|
22 | Revision 1.7 1999/06/30 04:04:13 rjmcnab
|
---|
23 | made stemming functions available from mgsearch and made the stems
|
---|
24 | for the query terms available in queryinfo
|
---|
25 |
|
---|
26 | Revision 1.6 1999/06/29 22:06:23 rjmcnab
|
---|
27 | Added a couple of fields to queryinfo to handle a special version
|
---|
28 | of mg.
|
---|
29 |
|
---|
30 | Revision 1.5 1999/06/27 22:08:48 sjboddie
|
---|
31 | now check for defaultindex, defaultsubcollection, and defaultlanguage
|
---|
32 | entries in config files
|
---|
33 |
|
---|
34 | Revision 1.4 1999/06/16 02:03:25 sjboddie
|
---|
35 | fixed bug in isApprox and set MAXDOCS to always be 500
|
---|
36 |
|
---|
37 | Revision 1.3 1999/04/19 23:56:09 rjmcnab
|
---|
38 | Finished the gdbm metadata stuff
|
---|
39 |
|
---|
40 | Revision 1.2 1999/04/12 03:45:03 rjmcnab
|
---|
41 | Finished the query filter.
|
---|
42 |
|
---|
43 | Revision 1.1 1999/04/06 22:22:09 rjmcnab
|
---|
44 | Initial revision.
|
---|
45 |
|
---|
46 | */
|
---|
47 |
|
---|
48 |
|
---|
49 | #include "queryfilter.h"
|
---|
50 | #include "fileutil.h"
|
---|
51 | #include "queryinfo.h"
|
---|
52 |
|
---|
53 | #define MAXDOCS 200 // note that maxdocs must be at least as large
|
---|
54 | // as the highest possible value of EndResults
|
---|
55 |
|
---|
56 | // some useful functions
|
---|
57 |
|
---|
58 | // translate will return true if successful
|
---|
59 | static bool translate (gdbmclass *gdbmptr, int docnum, text_t &trans_OID) {
|
---|
60 | infodbclass info;
|
---|
61 |
|
---|
62 | trans_OID.clear();
|
---|
63 |
|
---|
64 | // get the info
|
---|
65 | if (gdbmptr == NULL) return false;
|
---|
66 | if (!gdbmptr->getinfo(docnum, info)) return false;
|
---|
67 |
|
---|
68 | // translate
|
---|
69 | if (info["section"].empty()) return false;
|
---|
70 |
|
---|
71 | trans_OID = info["section"];
|
---|
72 | return true;
|
---|
73 | }
|
---|
74 |
|
---|
75 |
|
---|
76 | // do aditional query processing
|
---|
77 | void queryfilterclass::post_process (const queryparamclass &/*queryparams*/, const text_t &/*index*/,
|
---|
78 | queryresultsclass &/*queryresults*/) {
|
---|
79 | }
|
---|
80 |
|
---|
81 |
|
---|
82 |
|
---|
83 | queryfilterclass::queryfilterclass () {
|
---|
84 | gdbmptr = NULL;
|
---|
85 | mgsearchptr = NULL;
|
---|
86 |
|
---|
87 | // -- onePerQuery StartResults integer
|
---|
88 | FilterOption_t filtopt;
|
---|
89 | filtopt.name = "StartResults";
|
---|
90 | filtopt.type = FilterOption_t::integert;
|
---|
91 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
92 | filtopt.defaultValue = "1";
|
---|
93 | filtopt.validValues.push_back("1");
|
---|
94 | filtopt.validValues.push_back("1000");
|
---|
95 | filterOptions["StartResults"] = filtopt;
|
---|
96 |
|
---|
97 | // -- onePerQuery EndResults integer
|
---|
98 | filtopt.clear();
|
---|
99 | filtopt.name = "EndResults";
|
---|
100 | filtopt.type = FilterOption_t::integert;
|
---|
101 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
102 | filtopt.defaultValue = "10";
|
---|
103 | filtopt.validValues.push_back("1");
|
---|
104 | filtopt.validValues.push_back("1000");
|
---|
105 | filterOptions["EndResults"] = filtopt;
|
---|
106 |
|
---|
107 | // -- onePerQuery QueryType enumerated (boolean, ranked)
|
---|
108 | filtopt.clear();
|
---|
109 | filtopt.name = "QueryType";
|
---|
110 | filtopt.type = FilterOption_t::enumeratedt;
|
---|
111 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
112 | filtopt.defaultValue = "ranked";
|
---|
113 | filtopt.validValues.push_back("boolean");
|
---|
114 | filtopt.validValues.push_back("ranked");
|
---|
115 | filterOptions["QueryType"] = filtopt;
|
---|
116 |
|
---|
117 | // -- onePerTerm Term string ???
|
---|
118 | filtopt.clear();
|
---|
119 | filtopt.name = "Term";
|
---|
120 | filtopt.type = FilterOption_t::stringt;
|
---|
121 | filtopt.repeatable = FilterOption_t::onePerTerm;
|
---|
122 | filtopt.defaultValue = "";
|
---|
123 | filterOptions["Term"] = filtopt;
|
---|
124 |
|
---|
125 | // -- onePerTerm Casefold boolean
|
---|
126 | filtopt.clear();
|
---|
127 | filtopt.name = "Casefold";
|
---|
128 | filtopt.type = FilterOption_t::booleant;
|
---|
129 | filtopt.repeatable = FilterOption_t::onePerTerm;
|
---|
130 | filtopt.defaultValue = "true";
|
---|
131 | filtopt.validValues.push_back("false");
|
---|
132 | filtopt.validValues.push_back("true");
|
---|
133 | filterOptions["Casefold"] = filtopt;
|
---|
134 |
|
---|
135 | // -- onePerTerm Stem boolean
|
---|
136 | filtopt.clear();
|
---|
137 | filtopt.name = "Stem";
|
---|
138 | filtopt.type = FilterOption_t::booleant;
|
---|
139 | filtopt.repeatable = FilterOption_t::onePerTerm;
|
---|
140 | filtopt.defaultValue = "false";
|
---|
141 | filtopt.validValues.push_back("false");
|
---|
142 | filtopt.validValues.push_back("true");
|
---|
143 | filterOptions["Stem"] = filtopt;
|
---|
144 |
|
---|
145 | // -- onePerTerm Index enumerated
|
---|
146 | filtopt.clear();
|
---|
147 | filtopt.name = "Index";
|
---|
148 | filtopt.type = FilterOption_t::enumeratedt;
|
---|
149 | filtopt.repeatable = FilterOption_t::onePerTerm;
|
---|
150 | filtopt.defaultValue = "";
|
---|
151 | filterOptions["Index"] = filtopt;
|
---|
152 |
|
---|
153 | // -- onePerTerm Subcollection enumerated
|
---|
154 | filtopt.clear();
|
---|
155 | filtopt.name = "Subcollection";
|
---|
156 | filtopt.type = FilterOption_t::enumeratedt;
|
---|
157 | filtopt.repeatable = FilterOption_t::onePerTerm;
|
---|
158 | filtopt.defaultValue = "";
|
---|
159 | filterOptions["Subcollection"] = filtopt;
|
---|
160 |
|
---|
161 | // -- onePerTerm Language enumerated
|
---|
162 | filtopt.clear();
|
---|
163 | filtopt.name = "Language";
|
---|
164 | filtopt.type = FilterOption_t::enumeratedt;
|
---|
165 | filtopt.repeatable = FilterOption_t::onePerTerm;
|
---|
166 | filtopt.defaultValue = "";
|
---|
167 | filterOptions["Language"] = filtopt;
|
---|
168 | }
|
---|
169 |
|
---|
170 | queryfilterclass::~queryfilterclass () {
|
---|
171 | }
|
---|
172 |
|
---|
173 | void queryfilterclass::configure (const text_t &key, const text_tarray &cfgline) {
|
---|
174 | filterclass::configure (key, cfgline);
|
---|
175 |
|
---|
176 | if (key == "indexmap") {
|
---|
177 | indexmap.importmap (cfgline);
|
---|
178 |
|
---|
179 | // update the list of indexes in the filter information
|
---|
180 | text_tarray options;
|
---|
181 | indexmap.gettoarray (options);
|
---|
182 | filterOptions["Index"].validValues = options;
|
---|
183 |
|
---|
184 | } else if (key == "defaultindex") {
|
---|
185 | indexmap.from2to (cfgline[0], filterOptions["Index"].defaultValue);
|
---|
186 |
|
---|
187 | } else if (key == "subcollectionmap") {
|
---|
188 | subcollectionmap.importmap (cfgline);
|
---|
189 |
|
---|
190 | // update the list of subcollections in the filter information
|
---|
191 | text_tarray options;
|
---|
192 | subcollectionmap.gettoarray (options);
|
---|
193 | filterOptions["Subcollection"].validValues = options;
|
---|
194 |
|
---|
195 | } else if (key == "defaultsubcollection") {
|
---|
196 | subcollectionmap.from2to (cfgline[0], filterOptions["Subcollection"].defaultValue);
|
---|
197 |
|
---|
198 | } else if (key == "languagemap") {
|
---|
199 | languagemap.importmap (cfgline);
|
---|
200 |
|
---|
201 | // update the list of languages in the filter information
|
---|
202 | text_tarray options;
|
---|
203 | languagemap.gettoarray (options);
|
---|
204 | filterOptions["Language"].validValues = options;
|
---|
205 |
|
---|
206 | } else if (key == "defaultlanguage")
|
---|
207 | languagemap.from2to (cfgline[0], filterOptions["Language"].defaultValue);
|
---|
208 | }
|
---|
209 |
|
---|
210 | bool queryfilterclass::init (ostream &logout) {
|
---|
211 | outconvertclass text_t2ascii;
|
---|
212 |
|
---|
213 | if (!filterclass::init(logout)) return false;
|
---|
214 |
|
---|
215 | // get the filename for the database and make sure it exists
|
---|
216 | gdbm_filename = filename_cat(collectdir,"index","text",collection);
|
---|
217 | #ifdef _LITTLE_ENDIAN
|
---|
218 | gdbm_filename += ".ldb";
|
---|
219 | #else
|
---|
220 | gdbm_filename += ".bdb";
|
---|
221 | #endif
|
---|
222 | if (!file_exists(gdbm_filename)) {
|
---|
223 | logout << text_t2ascii
|
---|
224 | << "error: gdbm database \""
|
---|
225 | << gdbm_filename << "\" does not exist\n\n";
|
---|
226 | return false;
|
---|
227 | }
|
---|
228 |
|
---|
229 | return true;
|
---|
230 | }
|
---|
231 |
|
---|
232 | void queryfilterclass::filter (const FilterRequest_t &request,
|
---|
233 | FilterResponse_t &response,
|
---|
234 | comerror_t &err, ostream &logout) {
|
---|
235 | outconvertclass text_t2ascii;
|
---|
236 |
|
---|
237 | response.clear ();
|
---|
238 | err = noError;
|
---|
239 | if (gdbmptr == NULL) {
|
---|
240 | // most likely a configuration problem
|
---|
241 | logout << text_t2ascii
|
---|
242 | << "configuration error: queryfilter contains a null gdbmclass\n\n";
|
---|
243 | err = configurationError;
|
---|
244 | return;
|
---|
245 | }
|
---|
246 | if (mgsearchptr == NULL) {
|
---|
247 | // most likely a configuration problem
|
---|
248 | logout << text_t2ascii
|
---|
249 | << "configuration error: queryfilter contains a null mgsearchclass\n\n";
|
---|
250 | err = configurationError;
|
---|
251 | return;
|
---|
252 | }
|
---|
253 |
|
---|
254 | // open the database
|
---|
255 | gdbmptr->setlogout(&logout);
|
---|
256 | if (!gdbmptr->opendatabase (gdbm_filename)) {
|
---|
257 | // most likely a system problem (we have already checked that the
|
---|
258 | // gdbm database exists)
|
---|
259 | logout << text_t2ascii
|
---|
260 | << "system problem: open on gdbm database \""
|
---|
261 | << gdbm_filename << "\" failed\n\n";
|
---|
262 | err = systemProblem;
|
---|
263 | return;
|
---|
264 | }
|
---|
265 |
|
---|
266 | // get the query parameters
|
---|
267 | int startresults = filterOptions["StartResults"].defaultValue.getint();
|
---|
268 | int endresults = filterOptions["EndResults"].defaultValue.getint();
|
---|
269 | text_t index = filterOptions["Index"].defaultValue;
|
---|
270 | text_t subcollection = filterOptions["Subcollection"].defaultValue;
|
---|
271 | text_t language = filterOptions["Language"].defaultValue;
|
---|
272 | queryparamclass queryparams;
|
---|
273 | queryparams.collection = collection;
|
---|
274 | queryparams.search_type = (filterOptions["QueryType"].defaultValue == "ranked");
|
---|
275 | queryparams.casefolding = (filterOptions["Casefold"].defaultValue == "true");
|
---|
276 | queryparams.stemming = (filterOptions["Stem"].defaultValue == "true");
|
---|
277 |
|
---|
278 | OptionValue_tarray::const_iterator options_here = request.filterOptions.begin();
|
---|
279 | OptionValue_tarray::const_iterator options_end = request.filterOptions.end();
|
---|
280 | while (options_here != options_end) {
|
---|
281 | if ((*options_here).name == "StartResults") {
|
---|
282 | startresults = (*options_here).value.getint();
|
---|
283 | } else if ((*options_here).name == "EndResults") {
|
---|
284 | endresults = (*options_here).value.getint();
|
---|
285 | } else if ((*options_here).name == "QueryType") {
|
---|
286 | queryparams.search_type = ((*options_here).value == "ranked");
|
---|
287 | } else if ((*options_here).name == "Term") {
|
---|
288 | queryparams.querystring = (*options_here).value;
|
---|
289 | } else if ((*options_here).name == "Casefold") {
|
---|
290 | queryparams.casefolding = ((*options_here).value == "true");
|
---|
291 | } else if ((*options_here).name == "Stem") {
|
---|
292 | queryparams.stemming = ((*options_here).value == "true");
|
---|
293 | } else if ((*options_here).name == "Index") {
|
---|
294 | index = (*options_here).value;
|
---|
295 | } else if ((*options_here).name == "Subcollection") {
|
---|
296 | subcollection = (*options_here).value;
|
---|
297 | } else if ((*options_here).name == "Language") {
|
---|
298 | language = (*options_here).value;
|
---|
299 | } else {
|
---|
300 | logout << text_t2ascii
|
---|
301 | << "warning: unknown queryfilter option \""
|
---|
302 | << (*options_here).name
|
---|
303 | << "\" ignored.\n\n";
|
---|
304 | }
|
---|
305 |
|
---|
306 | options_here++;
|
---|
307 | }
|
---|
308 |
|
---|
309 | queryparams.search_index = index+subcollection+language;
|
---|
310 | queryparams.maxdocs = MAXDOCS;
|
---|
311 |
|
---|
312 | // do query
|
---|
313 | queryresultsclass queryresults;
|
---|
314 | mgsearchptr->setcollectdir (collectdir);
|
---|
315 | if (!mgsearchptr->search(queryparams, queryresults)) {
|
---|
316 | // most likely a system problem
|
---|
317 | logout << text_t2ascii
|
---|
318 | << "system problem: could not do search with mg for index \""
|
---|
319 | << queryparams.search_index << "\".\n\n";
|
---|
320 | err = systemProblem;
|
---|
321 | return;
|
---|
322 | }
|
---|
323 |
|
---|
324 | // assemble document results
|
---|
325 | if ((request.filterResultOptions & FROID) || (request.filterResultOptions & FRranking) ||
|
---|
326 | (request.filterResultOptions & FRmetadata)) {
|
---|
327 |
|
---|
328 | // post-process the results if needed
|
---|
329 | if (!queryresults.postprocessed && queryresults.orgterms.size() > 1 &&
|
---|
330 | !queryresults.docs.empty()) {
|
---|
331 | post_process (queryparams, index, queryresults);
|
---|
332 | queryresults.postprocessed = true;
|
---|
333 | }
|
---|
334 |
|
---|
335 | int resultnum = 1;
|
---|
336 | ResultDocInfo_t resultdoc;
|
---|
337 | text_t trans_OID;
|
---|
338 | vector<docresultclass>::iterator docs_here = queryresults.docs.begin();
|
---|
339 | vector<docresultclass>::iterator docs_end = queryresults.docs.end();
|
---|
340 |
|
---|
341 | while (docs_here != docs_end) {
|
---|
342 | if (resultnum > endresults) break;
|
---|
343 |
|
---|
344 | // translate the document number
|
---|
345 | if (!translate(gdbmptr, (*docs_here).docnum, trans_OID)) {
|
---|
346 | logout << text_t2ascii
|
---|
347 | << "warning: could not translate mg document number \""
|
---|
348 | << (*docs_here).docnum << "\"to OID.\n\n";
|
---|
349 |
|
---|
350 | } else {
|
---|
351 | // see if it is in the set (or the set is empty)
|
---|
352 | if (request.docSet.empty() || in_set(request.docSet, trans_OID)) {
|
---|
353 | if (resultnum >= startresults) {
|
---|
354 | // add this document
|
---|
355 | resultdoc.OID = trans_OID;
|
---|
356 | resultdoc.ranking = (int)((*docs_here).docweight * 10000.0 + 0.5);
|
---|
357 |
|
---|
358 | // these next two are not available on all versions of mg
|
---|
359 | resultdoc.num_terms_matched = (*docs_here).num_query_terms_matched;
|
---|
360 | resultdoc.query_phrase_match = (*docs_here).query_phrase_match;
|
---|
361 |
|
---|
362 | response.docInfo.push_back (resultdoc);
|
---|
363 | }
|
---|
364 |
|
---|
365 | resultnum++;
|
---|
366 | }
|
---|
367 | }
|
---|
368 |
|
---|
369 | docs_here++;
|
---|
370 | }
|
---|
371 | }
|
---|
372 |
|
---|
373 | // assemble the term results
|
---|
374 | if ((request.filterResultOptions & FRtermFreq) || (request.filterResultOptions & FRmatchTerms)) {
|
---|
375 | // note: the terms have already been sorted and uniqued
|
---|
376 |
|
---|
377 | TermInfo_t terminfo;
|
---|
378 | bool terms_first = true;
|
---|
379 | vector<termfreqclass>::iterator terms_here = queryresults.terms.begin();
|
---|
380 | vector<termfreqclass>::iterator terms_end = queryresults.terms.end();
|
---|
381 |
|
---|
382 | while (terms_here != terms_end) {
|
---|
383 | terminfo.clear();
|
---|
384 | terminfo.term = (*terms_here).termstr;
|
---|
385 | terminfo.freq = (*terms_here).termfreq;
|
---|
386 | if (terms_first) terminfo.matchTerms = queryresults.termvariants;
|
---|
387 | terms_first = false;
|
---|
388 |
|
---|
389 | response.termInfo.push_back (terminfo);
|
---|
390 |
|
---|
391 | terms_here++;
|
---|
392 | }
|
---|
393 | }
|
---|
394 |
|
---|
395 | response.numDocs = queryresults.docs_matched;
|
---|
396 | response.isApprox = queryresults.is_approx;
|
---|
397 | }
|
---|