1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * browsefilter.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 "browsefilter.h"
|
---|
27 | #include "colservertools.h"
|
---|
28 | #include "fileutil.h"
|
---|
29 | #include <assert.h>
|
---|
30 |
|
---|
31 |
|
---|
32 | browsefilterclass::browsefilterclass () {
|
---|
33 | db_ptr = NULL;
|
---|
34 | oaidb_ptr = NULL;
|
---|
35 |
|
---|
36 | // -- onePerQuery StartResults integer
|
---|
37 | FilterOption_t filtopt;
|
---|
38 | filtopt.name = "StartResults";
|
---|
39 | filtopt.type = FilterOption_t::integert;
|
---|
40 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
41 | filtopt.defaultValue = "1";
|
---|
42 | filtopt.validValues.push_back("1");
|
---|
43 | filtopt.validValues.push_back("10000");
|
---|
44 | filterOptions["StartResults"] = filtopt;
|
---|
45 |
|
---|
46 | // -- onePerQuery EndResults integer
|
---|
47 | filtopt.clear();
|
---|
48 | filtopt.name = "EndResults";
|
---|
49 | filtopt.type = FilterOption_t::integert;
|
---|
50 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
51 | filtopt.defaultValue = "-1";
|
---|
52 | filtopt.validValues.push_back("-1");
|
---|
53 | filtopt.validValues.push_back("10000");
|
---|
54 | filterOptions["EndResults"] = filtopt;
|
---|
55 |
|
---|
56 | // -- onePerQuery ParentNode string ("" will return the browsing available)
|
---|
57 | filtopt.clear();
|
---|
58 | filtopt.name = "ParentNode";
|
---|
59 | filtopt.type = FilterOption_t::stringt;
|
---|
60 | filtopt.repeatable = FilterOption_t::onePerQuery;
|
---|
61 | filtopt.defaultValue = g_EmptyText;
|
---|
62 | filterOptions["ParentNode"] = filtopt;
|
---|
63 | }
|
---|
64 |
|
---|
65 | browsefilterclass::~browsefilterclass () {}
|
---|
66 |
|
---|
67 | void browsefilterclass::configure (const text_t &key, const text_tarray &cfgline) {
|
---|
68 | filterclass::configure (key, cfgline);
|
---|
69 | if (key == "indexstem") {
|
---|
70 | indexstem = cfgline[0];
|
---|
71 | }
|
---|
72 | }
|
---|
73 |
|
---|
74 | bool browsefilterclass::init (ostream &logout) {
|
---|
75 | outconvertclass text_t2ascii;
|
---|
76 |
|
---|
77 | if (!filterclass::init(logout)) return false;
|
---|
78 |
|
---|
79 | if (db_ptr == NULL || oaidb_ptr == NULL) {
|
---|
80 | // most likely a configuration problem
|
---|
81 | logout << text_t2ascii
|
---|
82 | << "configuration error: browsefilter contains a null dbclass\n\n";
|
---|
83 | return false;
|
---|
84 | }
|
---|
85 |
|
---|
86 | if (indexstem.empty()) {
|
---|
87 | indexstem = collection;
|
---|
88 | }
|
---|
89 |
|
---|
90 | db_filename = resolve_db_filename(gsdlhome, dbhome, collecthome, collection,
|
---|
91 | indexstem,db_ptr->getfileextension());
|
---|
92 | if (!file_exists(db_filename)) {
|
---|
93 | logout << text_t2ascii
|
---|
94 | << "warning: database \"" << db_filename << "\" does not exist\n\n";
|
---|
95 | // return false;
|
---|
96 | }
|
---|
97 |
|
---|
98 | oaidb_filename = resolve_oaidb_filename(gsdlhome, dbhome, collecthome, collection,
|
---|
99 | oaidb_ptr->getfileextension());
|
---|
100 |
|
---|
101 | return true;
|
---|
102 | }
|
---|
103 |
|
---|
104 | void browsefilterclass::filter (const FilterRequest_t &request,
|
---|
105 | FilterResponse_t &response,
|
---|
106 | comerror_t &err, ostream &logout) {
|
---|
107 | int numDocs = 0;
|
---|
108 | outconvertclass text_t2ascii;
|
---|
109 |
|
---|
110 | response.clear ();
|
---|
111 | err = noError;
|
---|
112 |
|
---|
113 | // get the browse parameters
|
---|
114 | int startresults = filterOptions["StartResults"].defaultValue.getint();
|
---|
115 | int endresults = filterOptions["EndResults"].defaultValue.getint();
|
---|
116 | text_t parentnode = filterOptions["ParentNode"].defaultValue;
|
---|
117 | OptionValue_tarray::const_iterator options_here = request.filterOptions.begin();
|
---|
118 | OptionValue_tarray::const_iterator options_end = request.filterOptions.end();
|
---|
119 | while (options_here != options_end) {
|
---|
120 | if ((*options_here).name == "StartResults")
|
---|
121 | startresults = (*options_here).value.getint();
|
---|
122 | else if ((*options_here).name == "EndResults")
|
---|
123 | endresults = (*options_here).value.getint();
|
---|
124 | else if ((*options_here).name == "ParentNode")
|
---|
125 | parentnode = (*options_here).value;
|
---|
126 | else {
|
---|
127 | logout << text_t2ascii
|
---|
128 | << "warning: unknown browsefilter option \""
|
---|
129 | << (*options_here).name
|
---|
130 | << "\" ignored.\n\n";
|
---|
131 | }
|
---|
132 |
|
---|
133 | ++options_here;
|
---|
134 | }
|
---|
135 |
|
---|
136 | // if we're only working on oai, open oai_db, no need to work with index_db in browsefilter.cpp
|
---|
137 | // (but source.cpp uses both oai-inf.db and index.db to get metadata for OAI request)
|
---|
138 | // If we can't open the oai-inf db, this can be because it didn't exist in older versions of GS
|
---|
139 | // in that case, proceed as usual, using the index db.
|
---|
140 | if((request.filterResultOptions & FROAI)) { // OAI request
|
---|
141 | bool success = false;
|
---|
142 |
|
---|
143 | if(parentnode == "oai") { // doing an OAI listidentifiers request
|
---|
144 |
|
---|
145 | // open up the oai-inf db, if it exists, and return all IDs
|
---|
146 | // if it doesn't exist, proceed as usual
|
---|
147 | success = get_oaiinf_db_entries(response, err, logout); //adds any stuff in oai-inf db for the current OID to resultdoc.metadata
|
---|
148 |
|
---|
149 | response.numDocs = response.docInfo.size();
|
---|
150 | response.isApprox = Exact;
|
---|
151 | }
|
---|
152 | if (success) return; // oai request successfully completed with oai-inf.db, no need to open index_db
|
---|
153 | }
|
---|
154 |
|
---|
155 | // Since we're here, it means we're not doing anything oai (or oai-inf.db did not exist/open)
|
---|
156 | // So we don't need to work with oai_db. Instead, work with index_db:
|
---|
157 |
|
---|
158 | if (db_ptr == NULL) {
|
---|
159 | // most likely a configuration problem
|
---|
160 | logout << text_t2ascii
|
---|
161 | << "configuration error: browsefilter contains a null index dbclass\n\n";
|
---|
162 | err = configurationError;
|
---|
163 | return;
|
---|
164 | }
|
---|
165 |
|
---|
166 | // open the database
|
---|
167 | db_ptr->setlogout(&logout);
|
---|
168 | if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
|
---|
169 | // most likely a system problem (we have already checked that the database exists)
|
---|
170 | logout << text_t2ascii
|
---|
171 | << "system problem: open on database \""
|
---|
172 | << db_filename << "\" failed\n\n";
|
---|
173 | err = systemProblem;
|
---|
174 | return;
|
---|
175 | }
|
---|
176 |
|
---|
177 | infodbclass info;
|
---|
178 |
|
---|
179 | // translate any ".fc", ".pr" etc. stuff in the parentnode
|
---|
180 | parentnode = db_ptr->translate_OID (parentnode, info);
|
---|
181 |
|
---|
182 | // adjust topmost browsing node
|
---|
183 | if (parentnode.empty()) parentnode = "browse";
|
---|
184 |
|
---|
185 | // get the node
|
---|
186 | if ((request.filterResultOptions & FROID) ||
|
---|
187 | (request.filterResultOptions & FRmetadata)) {
|
---|
188 | if (!db_ptr->getinfo(parentnode, info)) {
|
---|
189 | // didn't find the node in index db
|
---|
190 | logout << text_t2ascii
|
---|
191 | << "warning: lookup for node \"" << parentnode
|
---|
192 | << "\" failed for browsefilter.\n\n";
|
---|
193 | } else {
|
---|
194 | // found the node
|
---|
195 |
|
---|
196 | // replace " with the parent node name and split the contains string
|
---|
197 | // into the result set
|
---|
198 | text_tarray resultset;
|
---|
199 | text_t tmptext;
|
---|
200 | text_t &contains = info["contains"];
|
---|
201 | text_t::iterator contains_here = contains.begin();
|
---|
202 | text_t::iterator contains_end = contains.end();
|
---|
203 | while (contains_here != contains_end) {
|
---|
204 | if (*contains_here == '"') tmptext += parentnode;
|
---|
205 | else if (*contains_here == ';') {
|
---|
206 | if (!tmptext.empty()) resultset.push_back (tmptext);
|
---|
207 | tmptext.clear();
|
---|
208 | } else tmptext.push_back(*contains_here);
|
---|
209 |
|
---|
210 | ++contains_here;
|
---|
211 | }
|
---|
212 | // insert the last result in the set
|
---|
213 | if (!tmptext.empty()) resultset.push_back (tmptext);
|
---|
214 |
|
---|
215 | text_tarray offset_resultset;
|
---|
216 | text_t &md_type = info["mdtype"];
|
---|
217 | if (!md_type.empty())
|
---|
218 | {
|
---|
219 | text_t &md_offset = info["mdoffset"];
|
---|
220 | if (!md_offset.empty())
|
---|
221 | {
|
---|
222 | text_t offsettext;
|
---|
223 |
|
---|
224 | text_t::iterator offset_here = md_offset.begin();
|
---|
225 | text_t::iterator offset_end = md_offset.end();
|
---|
226 | while (offset_here != offset_end)
|
---|
227 | {
|
---|
228 | if (*offset_here == ';')
|
---|
229 | {
|
---|
230 | if (offsettext.empty())
|
---|
231 | {
|
---|
232 | offset_resultset.push_back ("0");
|
---|
233 | }
|
---|
234 | else
|
---|
235 | {
|
---|
236 | offset_resultset.push_back (offsettext);
|
---|
237 | }
|
---|
238 | offsettext.clear();
|
---|
239 | }
|
---|
240 | else
|
---|
241 | {
|
---|
242 | offsettext.push_back(*offset_here);
|
---|
243 | }
|
---|
244 |
|
---|
245 | ++offset_here;
|
---|
246 | }
|
---|
247 | // insert the last result in the set
|
---|
248 | if (offsettext.empty())
|
---|
249 | {
|
---|
250 | offset_resultset.push_back ("0");
|
---|
251 | }
|
---|
252 | else
|
---|
253 | {
|
---|
254 | offset_resultset.push_back (offsettext);
|
---|
255 | }
|
---|
256 | }
|
---|
257 | else
|
---|
258 | {
|
---|
259 | // add 0 offset for each 'contains' entry
|
---|
260 | text_tarray::iterator result_here = resultset.begin();
|
---|
261 | text_tarray::iterator result_end = resultset.end();
|
---|
262 | while (result_here != result_end) {
|
---|
263 | offset_resultset.push_back("0");
|
---|
264 | ++result_here;
|
---|
265 | }
|
---|
266 | }
|
---|
267 |
|
---|
268 | // do an intersection with the input set
|
---|
269 | if (!request.docSet.empty()) {
|
---|
270 |
|
---|
271 | text_tarray intersect_resultset;
|
---|
272 | text_tarray intersect_offset_resultset;
|
---|
273 |
|
---|
274 | text_tarray::const_iterator resultset_here = resultset.begin();
|
---|
275 | text_tarray::const_iterator resultset_end = resultset.end();
|
---|
276 | text_tarray::const_iterator offset_resultset_here = offset_resultset.begin();
|
---|
277 |
|
---|
278 | while (resultset_here != resultset_end) {
|
---|
279 | if (in_set (request.docSet, *resultset_here))
|
---|
280 | {
|
---|
281 | intersect_resultset.push_back (*resultset_here);
|
---|
282 | intersect_offset_resultset.push_back (*offset_resultset_here);
|
---|
283 | }
|
---|
284 | ++resultset_here;
|
---|
285 | ++offset_resultset_here;
|
---|
286 | }
|
---|
287 | resultset = intersect_resultset;
|
---|
288 | offset_resultset = intersect_offset_resultset;
|
---|
289 | }
|
---|
290 | }
|
---|
291 | else
|
---|
292 | {
|
---|
293 | // do an intersection with the input set
|
---|
294 | if (!request.docSet.empty()) {
|
---|
295 | intersect (resultset, request.docSet);
|
---|
296 | }
|
---|
297 |
|
---|
298 | // add 0 offset for each 'contains' entry
|
---|
299 | text_tarray::iterator result_here = resultset.begin();
|
---|
300 | text_tarray::iterator result_end = resultset.end();
|
---|
301 | while (result_here != result_end) {
|
---|
302 | offset_resultset.push_back("0");
|
---|
303 | ++result_here;
|
---|
304 | }
|
---|
305 | }
|
---|
306 |
|
---|
307 | // create the response
|
---|
308 | numDocs = resultset.size();
|
---|
309 | int resultnum = 1;
|
---|
310 | ResultDocInfo_t resultdoc;
|
---|
311 | text_tarray::iterator result_here = resultset.begin();
|
---|
312 | text_tarray::iterator result_end = resultset.end();
|
---|
313 | text_tarray::iterator offset_result_here = offset_resultset.begin();
|
---|
314 |
|
---|
315 | while (result_here != result_end) {
|
---|
316 | // if endresults is -1 get all results
|
---|
317 | if ((endresults != -1) && (resultnum > endresults)) break;
|
---|
318 | if (resultnum >= startresults) {
|
---|
319 | resultdoc.OID = (*result_here);
|
---|
320 | if (!md_type.empty())
|
---|
321 | {
|
---|
322 | resultdoc.classifier_metadata_type = md_type;
|
---|
323 | resultdoc.classifier_metadata_offset = offset_result_here->getint();
|
---|
324 | }
|
---|
325 | response.docInfo.push_back(resultdoc);
|
---|
326 | }
|
---|
327 |
|
---|
328 | ++resultnum;
|
---|
329 | ++result_here;
|
---|
330 | if (!md_type.empty()) ++offset_result_here;
|
---|
331 | }
|
---|
332 | }
|
---|
333 | }
|
---|
334 |
|
---|
335 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
336 | response.numDocs = numDocs;
|
---|
337 | response.isApprox = Exact;
|
---|
338 | }
|
---|
339 |
|
---|
340 | bool browsefilterclass::get_oaiinf_db_entries(FilterResponse_t &response,
|
---|
341 | comerror_t &err, ostream &logout)
|
---|
342 | {
|
---|
343 | outconvertclass text_t2ascii;
|
---|
344 |
|
---|
345 | //logout << text_t2ascii << "browsefilterclass::get_oaiinf_db_entries\n";
|
---|
346 |
|
---|
347 | // ONLY if we're doing any OAI stuff (FROAI will be set then) will we be here
|
---|
348 | // So next try to open the oai-inf db if it exists for this collection
|
---|
349 |
|
---|
350 |
|
---|
351 | if (!file_exists(oaidb_filename)) { // if the oaidb file doesn't even exist, let's not bother with oaidb
|
---|
352 |
|
---|
353 | logout << text_t2ascii
|
---|
354 | << "warning: collection's oai-inf database \"" << oaidb_filename << "\" does not exist\n\n";
|
---|
355 | return false;
|
---|
356 |
|
---|
357 | } else { // let's try opening the oaidb file
|
---|
358 | oaidb_ptr->setlogout(&logout);
|
---|
359 | if (!oaidb_ptr->opendatabase (oaidb_filename, DB_READER, 100, false)) {
|
---|
360 | // most likely a system problem (we have already checked that the database exists just above)
|
---|
361 | logout << text_t2ascii
|
---|
362 | << "system problem: open on database \""
|
---|
363 | << oaidb_filename << "\" failed\n\n";
|
---|
364 | err = systemProblem;
|
---|
365 | return false;
|
---|
366 | } // now we've opened the oai-inf db file successfully
|
---|
367 | }
|
---|
368 |
|
---|
369 | infodbclass oai_info;
|
---|
370 | ResultDocInfo_t resultdoc;
|
---|
371 |
|
---|
372 | text_tarray keys = oaidb_ptr->getkeys();
|
---|
373 |
|
---|
374 | text_tarray::iterator key_here = keys.begin();
|
---|
375 | text_tarray::iterator key_end = keys.end();
|
---|
376 | while (key_here != key_end) {
|
---|
377 |
|
---|
378 | resultdoc.OID = (*key_here);
|
---|
379 |
|
---|
380 | if(!oaidb_ptr->getinfo(resultdoc.OID, oai_info)) {
|
---|
381 | logout << text_t2ascii
|
---|
382 | << "warning: lookup for node \"" << resultdoc.OID
|
---|
383 | << "\" in etc/oai-inf db failed for browsefilter.\n\n";
|
---|
384 | }
|
---|
385 | // We don't need to get the oai metadata from oai-inf.db at this stage. That will be
|
---|
386 | // handled by a separate metadata request. See collectserver::filter() and source.cpp's get_oai_metadata().
|
---|
387 | /*
|
---|
388 | else {
|
---|
389 | //logout << text_t2ascii << "@@@@ found node \"" << resultdoc.OID << "\" in etc/oai-inf db.\n\n";
|
---|
390 |
|
---|
391 | resultdoc.metadata["oaiinf.status"].isRef = false;
|
---|
392 | resultdoc.metadata["oaiinf.status"].values.push_back(oai_info["status"]);
|
---|
393 | resultdoc.metadata["oaiinf.timestamp"].isRef = false;
|
---|
394 | resultdoc.metadata["oaiinf.timestamp"].values.push_back(oai_info["timestamp"]);
|
---|
395 | resultdoc.metadata["oaiinf.datestamp"].isRef = false;
|
---|
396 | resultdoc.metadata["oaiinf.datestamp"].values.push_back(oai_info["datestamp"]);
|
---|
397 | }
|
---|
398 | */
|
---|
399 |
|
---|
400 | response.docInfo.push_back(resultdoc);
|
---|
401 | ++key_here;
|
---|
402 |
|
---|
403 | }
|
---|
404 |
|
---|
405 | // we're done with oai-inf db
|
---|
406 |
|
---|
407 | oaidb_ptr->closedatabase(); // don't leave files open
|
---|
408 |
|
---|
409 | return true;
|
---|
410 | }
|
---|