1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * collectset.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 |
|
---|
27 | #include "collectset.h"
|
---|
28 | #include "collectserver.h"
|
---|
29 | #include "colservrconfig.h"
|
---|
30 | #include "gsdlsitecfg.h"
|
---|
31 | #include "filter.h"
|
---|
32 | #include "browsefilter.h"
|
---|
33 | #include "queryfilter.h"
|
---|
34 | #include "mgqueryfilter.h"
|
---|
35 | #include "mgppqueryfilter.h"
|
---|
36 | #include "mggdbmsource.h"
|
---|
37 | #include "lucenequeryfilter.h"
|
---|
38 | #include "lucenegdbmsource.h"
|
---|
39 |
|
---|
40 | #include "fileutil.h"
|
---|
41 | #include <assert.h>
|
---|
42 |
|
---|
43 |
|
---|
44 | collectset::collectset (text_t &gsdlhome) {
|
---|
45 |
|
---|
46 | text_tarray collections;
|
---|
47 | text_t collectdir;
|
---|
48 |
|
---|
49 | // get gsdlhome (if we fail the error will be picked up later -- in
|
---|
50 | // cgiwrapper)
|
---|
51 |
|
---|
52 | if (site_cfg_read (gsdlhome, httpdomain, httpprefix)) {
|
---|
53 | if (!gsdlhome.empty() && directory_exists(gsdlhome)) {
|
---|
54 | collectdir = filename_cat (gsdlhome, "collect");
|
---|
55 | if (read_dir (collectdir, collections)) {
|
---|
56 |
|
---|
57 | text_tarray::const_iterator thiscol = collections.begin();
|
---|
58 | text_tarray::const_iterator endcol = collections.end();
|
---|
59 |
|
---|
60 | while (thiscol != endcol) {
|
---|
61 | // ignore the modelcol
|
---|
62 | if (*thiscol == "modelcol") {
|
---|
63 | ++thiscol;
|
---|
64 | continue;
|
---|
65 | }
|
---|
66 |
|
---|
67 | this->add_collection (*thiscol, gsdlhome);
|
---|
68 |
|
---|
69 | ++thiscol;
|
---|
70 | }
|
---|
71 |
|
---|
72 | this->add_all_collection_groups(gsdlhome);
|
---|
73 | }
|
---|
74 | }
|
---|
75 | }
|
---|
76 | }
|
---|
77 |
|
---|
78 | collectset::~collectset () {
|
---|
79 | collectservermapclass::iterator here = cservers.begin();
|
---|
80 | collectservermapclass::iterator end = cservers.end();
|
---|
81 |
|
---|
82 | while (here != end) {
|
---|
83 | if ((*here).second.c != NULL) {
|
---|
84 | delete (*here).second.c;
|
---|
85 | }
|
---|
86 | ++here;
|
---|
87 | }
|
---|
88 | cservers.clear();
|
---|
89 | }
|
---|
90 |
|
---|
91 | bool collectset::init (ostream &logout) {
|
---|
92 | collectservermapclass::iterator here = cservers.begin();
|
---|
93 | collectservermapclass::iterator end = cservers.end();
|
---|
94 |
|
---|
95 | while (here != end) {
|
---|
96 | assert ((*here).second.c != NULL);
|
---|
97 | if ((*here).second.c != NULL) {
|
---|
98 | const colservrconf &configinfo = (*here).second.c->get_configinfo ();
|
---|
99 |
|
---|
100 | // configure this collection server
|
---|
101 |
|
---|
102 | // note that we read build.cfg before collect.cfg so that the indexmaps
|
---|
103 | // are available to decode defaultindex, defaultsubcollection, and
|
---|
104 | // defaultlanguage
|
---|
105 |
|
---|
106 | bool failed_build_cfg = false;
|
---|
107 | if (!build_cfg_read (*((*here).second.c), configinfo.gsdlhome,
|
---|
108 | configinfo.collection)) {
|
---|
109 | failed_build_cfg = true;
|
---|
110 |
|
---|
111 | outconvertclass text_t2ascii;
|
---|
112 | logout << text_t2ascii
|
---|
113 | << "Warning: couldn't read build.cfg file for collection \"" //****
|
---|
114 | << configinfo.collection << "\", gsdlhome=\""
|
---|
115 | << configinfo.gsdlhome << "\"\n";
|
---|
116 | }
|
---|
117 |
|
---|
118 | bool failed_collect_cfg = false;
|
---|
119 | if (!collect_cfg_read (*((*here).second.c), configinfo.gsdlhome,
|
---|
120 | configinfo.collection)) {
|
---|
121 | failed_collect_cfg = true;
|
---|
122 | outconvertclass text_t2ascii;
|
---|
123 | logout << text_t2ascii
|
---|
124 | << "Warning: couldn't read collect.cfg file for collection \""
|
---|
125 | << configinfo.collection << "\", gsdlhome=\""
|
---|
126 | << configinfo.gsdlhome << "\"\n";
|
---|
127 | }
|
---|
128 |
|
---|
129 |
|
---|
130 | bool is_colgroup = (*here).second.c->is_collection_group();
|
---|
131 |
|
---|
132 | if (failed_collect_cfg) {
|
---|
133 | ++here;
|
---|
134 | continue;
|
---|
135 | }
|
---|
136 |
|
---|
137 | if (failed_build_cfg && (!is_colgroup)) {
|
---|
138 | ++here;
|
---|
139 | continue;
|
---|
140 | }
|
---|
141 | // let a failed build.cfg through if its 'collect.cfg' marks it as 'collectgroup true'
|
---|
142 |
|
---|
143 | if (!(*here).second.c->init (logout)) return false;
|
---|
144 |
|
---|
145 | (*here).second.c->configure("httpdomain",httpdomain);
|
---|
146 | (*here).second.c->configure("httpprefix",httpprefix);
|
---|
147 | }
|
---|
148 | ++here;
|
---|
149 | }
|
---|
150 |
|
---|
151 | return true;
|
---|
152 | }
|
---|
153 |
|
---|
154 | collectservermapclass collectset::servers()
|
---|
155 | { return cservers;
|
---|
156 | }
|
---|
157 |
|
---|
158 |
|
---|
159 | void collectset::add_all_collections(const text_t &gsdlhome) {
|
---|
160 |
|
---|
161 | text_tarray collections;
|
---|
162 | text_t collectdir = filename_cat(gsdlhome, "collect");
|
---|
163 | if (read_dir(collectdir, collections)) {
|
---|
164 |
|
---|
165 | text_tarray::const_iterator thiscol = collections.begin();
|
---|
166 | text_tarray::const_iterator endcol = collections.end();
|
---|
167 |
|
---|
168 | while (thiscol != endcol) {
|
---|
169 |
|
---|
170 | // ignore the modelcol
|
---|
171 | if (*thiscol == "modelcol") {
|
---|
172 | ++thiscol;
|
---|
173 | continue;
|
---|
174 | }
|
---|
175 |
|
---|
176 | // create collection server for this collection
|
---|
177 | this->add_collection (*thiscol, gsdlhome);
|
---|
178 |
|
---|
179 | ++thiscol;
|
---|
180 | }
|
---|
181 |
|
---|
182 | this->add_all_collection_groups(gsdlhome);
|
---|
183 | }
|
---|
184 | }
|
---|
185 |
|
---|
186 | // add_collection sets up the collectionserver and calls
|
---|
187 | // add_collectserver
|
---|
188 | void collectset::add_collection (const text_t &collection,
|
---|
189 | const text_t &gsdlhome) {
|
---|
190 |
|
---|
191 | this->remove_collection(collection);
|
---|
192 |
|
---|
193 | // read config file to see if built with mg, mgpp, or lucene
|
---|
194 | text_t buildtype = "mg"; // mg is default
|
---|
195 |
|
---|
196 | text_tarray cfgline;
|
---|
197 | text_t key;
|
---|
198 | text_t build_cfg = filename_cat(gsdlhome, "collect", collection, "index", "build.cfg");
|
---|
199 | char *build_cfgc = build_cfg.getcstr();
|
---|
200 | ifstream confin(build_cfgc);
|
---|
201 |
|
---|
202 | if (confin) {
|
---|
203 | while (read_cfg_line(confin, cfgline) >= 0) {
|
---|
204 | if (cfgline.size() == 2) {
|
---|
205 | key = cfgline[0];
|
---|
206 | cfgline.erase(cfgline.begin());
|
---|
207 | if (key == "buildtype") {
|
---|
208 | buildtype = cfgline[0];
|
---|
209 | break;
|
---|
210 | }
|
---|
211 | }
|
---|
212 | }
|
---|
213 | confin.close();
|
---|
214 | }
|
---|
215 | delete []build_cfgc;
|
---|
216 |
|
---|
217 | collectserver *cserver = new collectserver();
|
---|
218 | gdbmclass *gdbmhandler = new gdbmclass();
|
---|
219 |
|
---|
220 | // add a null filter
|
---|
221 | filterclass *filter = new filterclass ();
|
---|
222 | cserver->add_filter (filter);
|
---|
223 |
|
---|
224 | // add a browse filter
|
---|
225 | browsefilterclass *browsefilter = new browsefilterclass();
|
---|
226 | browsefilter->set_gdbmptr (gdbmhandler);
|
---|
227 |
|
---|
228 | cserver->add_filter (browsefilter);
|
---|
229 |
|
---|
230 | if (buildtype == "mg") {
|
---|
231 | mgsearch = new mgsearchclass();
|
---|
232 |
|
---|
233 | // add a query filter
|
---|
234 | mgqueryfilterclass *queryfilter = new mgqueryfilterclass();
|
---|
235 | queryfilter->set_gdbmptr (gdbmhandler);
|
---|
236 | queryfilter->set_textsearchptr (mgsearch);
|
---|
237 | cserver->add_filter (queryfilter);
|
---|
238 |
|
---|
239 | // add a mg and gdbm source
|
---|
240 | mggdbmsourceclass *mggdbmsource = new mggdbmsourceclass ();
|
---|
241 | mggdbmsource->set_gdbmptr (gdbmhandler);
|
---|
242 | mggdbmsource->set_textsearchptr (mgsearch);
|
---|
243 | cserver->add_source (mggdbmsource);
|
---|
244 | }
|
---|
245 | else if (buildtype == "mgpp") {
|
---|
246 | mgppsearch = new mgppsearchclass();
|
---|
247 |
|
---|
248 | // add a query filter
|
---|
249 | mgppqueryfilterclass *queryfilter = new mgppqueryfilterclass();
|
---|
250 | queryfilter->set_gdbmptr (gdbmhandler);
|
---|
251 | queryfilter->set_textsearchptr (mgppsearch);
|
---|
252 | cserver->add_filter (queryfilter);
|
---|
253 |
|
---|
254 | // add a mg and gdbm source
|
---|
255 | mggdbmsourceclass *mggdbmsource = new mggdbmsourceclass ();
|
---|
256 | mggdbmsource->set_gdbmptr (gdbmhandler);
|
---|
257 | mggdbmsource->set_textsearchptr (mgppsearch);
|
---|
258 | cserver->add_source (mggdbmsource);
|
---|
259 | }
|
---|
260 | else if (buildtype == "lucene") {
|
---|
261 | lucenesearch = new lucenesearchclass();
|
---|
262 | lucenesearch->set_gsdlhome(gsdlhome);
|
---|
263 |
|
---|
264 | // add a query filter
|
---|
265 | lucenequeryfilterclass *queryfilter = new lucenequeryfilterclass();
|
---|
266 | queryfilter->set_gdbmptr (gdbmhandler);
|
---|
267 | queryfilter->set_textsearchptr (lucenesearch);
|
---|
268 | cserver->add_filter (queryfilter);
|
---|
269 |
|
---|
270 | // add a lucene and gdbm source
|
---|
271 | lucenegdbmsourceclass *lucenegdbmsource = new lucenegdbmsourceclass ();
|
---|
272 | lucenegdbmsource->set_gdbmptr (gdbmhandler);
|
---|
273 | lucenegdbmsource->set_textsearchptr (lucenesearch);
|
---|
274 | cserver->add_source (lucenegdbmsource);
|
---|
275 | }
|
---|
276 |
|
---|
277 | // inform collection server and everything it contains about its
|
---|
278 | // collection name
|
---|
279 | cserver->configure ("collection", collection);
|
---|
280 | cserver->configure ("gsdlhome", gsdlhome);
|
---|
281 |
|
---|
282 | cservers.addcollectserver (cserver);
|
---|
283 | }
|
---|
284 |
|
---|
285 | void collectset::remove_all_collections () {
|
---|
286 |
|
---|
287 | // first unload any cached mg databases
|
---|
288 | if (mgsearch != NULL) {
|
---|
289 | mgsearch->unload_database();
|
---|
290 | }
|
---|
291 |
|
---|
292 | // now delete the collection server objects
|
---|
293 | collectservermapclass::iterator here = cservers.begin();
|
---|
294 | collectservermapclass::iterator end = cservers.end();
|
---|
295 |
|
---|
296 | while (here != end) {
|
---|
297 | if ((*here).second.c != NULL) {
|
---|
298 | delete (*here).second.c;
|
---|
299 | }
|
---|
300 | ++here;
|
---|
301 | }
|
---|
302 | cservers.clear();
|
---|
303 | }
|
---|
304 |
|
---|
305 | void collectset::add_collection_group(const text_t& collection,
|
---|
306 | const text_t& gsdlhome)
|
---|
307 | {
|
---|
308 | text_tarray group;
|
---|
309 | text_t collect_group_dir = filename_cat (gsdlhome, "collect", collection);
|
---|
310 |
|
---|
311 | // need to read collect.cfg for 'collectgroup' as class hasn't been initialised through 'init' yet
|
---|
312 | text_t is_collect_group;
|
---|
313 | text_tarray cfgline;
|
---|
314 | text_t key;
|
---|
315 | text_t build_cfg = filename_cat(gsdlhome, "collect", collection, "etc", "collect.cfg");
|
---|
316 | char *collect_cfgc = build_cfg.getcstr();
|
---|
317 | ifstream confin(collect_cfgc);
|
---|
318 |
|
---|
319 | if (confin) {
|
---|
320 | while (read_cfg_line(confin, cfgline) >= 0) {
|
---|
321 | if (cfgline.size() == 2) {
|
---|
322 | key = cfgline[0];
|
---|
323 | cfgline.erase(cfgline.begin());
|
---|
324 | if (key == "collectgroup") {
|
---|
325 | is_collect_group = cfgline[0];
|
---|
326 | break;
|
---|
327 | }
|
---|
328 | }
|
---|
329 | }
|
---|
330 | confin.close();
|
---|
331 | }
|
---|
332 | delete []collect_cfgc;
|
---|
333 |
|
---|
334 | if (is_collect_group == "true") {
|
---|
335 |
|
---|
336 | if (read_dir (collect_group_dir, group)) {
|
---|
337 |
|
---|
338 | text_tarray::const_iterator thiscol = group.begin();
|
---|
339 | text_tarray::const_iterator endcol = group.end();
|
---|
340 |
|
---|
341 | while (thiscol != endcol) {
|
---|
342 | // ignore the modelcol
|
---|
343 | if (*thiscol == "etc") {
|
---|
344 | ++thiscol;
|
---|
345 | continue;
|
---|
346 | }
|
---|
347 |
|
---|
348 | this->add_collection (collection + "/" + *thiscol, gsdlhome);
|
---|
349 |
|
---|
350 | ++thiscol;
|
---|
351 | }
|
---|
352 | }
|
---|
353 | }
|
---|
354 | }
|
---|
355 |
|
---|
356 | void collectset::add_all_collection_groups (const text_t& gsdlhome)
|
---|
357 |
|
---|
358 | {
|
---|
359 | collectservermapclass::iterator here = cservers.begin();
|
---|
360 | collectservermapclass::iterator end = cservers.end();
|
---|
361 |
|
---|
362 | while (here != end) {
|
---|
363 | text_t collection = (*here).second.c->get_collection_name();
|
---|
364 | this->add_collection_group(collection,gsdlhome);
|
---|
365 |
|
---|
366 | ++here;
|
---|
367 | }
|
---|
368 | }
|
---|
369 |
|
---|
370 |
|
---|
371 | // remove_collection deletes the collection server of collection.
|
---|
372 | // This only needs to be called if a collectionserver is to be
|
---|
373 | // removed while the library is running. The destructor function
|
---|
374 | // cleans up all collectservers when the program exits.
|
---|
375 | void collectset::remove_collection (const text_t &collection) {
|
---|
376 |
|
---|
377 | // do nothing if no collection server exists for this collection
|
---|
378 | if (cservers.getcollectserver(collection) == NULL) return;
|
---|
379 |
|
---|
380 | // first unload any cached mg databases - we may need to do something
|
---|
381 | // similar to this for mgpp and lucene too
|
---|
382 | if (mgsearch != NULL) {
|
---|
383 | mgsearch->unload_database();
|
---|
384 | }
|
---|
385 |
|
---|
386 | // now delete the collection server object
|
---|
387 | collectservermapclass::iterator here = cservers.begin();
|
---|
388 | collectservermapclass::iterator end = cservers.end();
|
---|
389 |
|
---|
390 | while (here != end) {
|
---|
391 | if ((*here).second.c != NULL && (*here).first == collection) {
|
---|
392 | delete (*here).second.c;
|
---|
393 | cservers.erase (here);
|
---|
394 | return;
|
---|
395 | }
|
---|
396 | ++here;
|
---|
397 | }
|
---|
398 | }
|
---|
399 |
|
---|
400 |
|
---|
401 | // remove_collection deletes the collection server of collection.
|
---|
402 | // This only needs to be called if a collectionserver is to be
|
---|
403 | // removed while the library is running. The destructor function
|
---|
404 | // cleans up all collectservers when the program exits.
|
---|
405 | void collectset::remove_collection (const text_t &collection, ostream &logout) {
|
---|
406 |
|
---|
407 | remove_collection(collection);
|
---|
408 |
|
---|
409 | outconvertclass text_t2ascii;
|
---|
410 | logout << text_t2ascii << "collectset::remove_collection: failed to remove collectserver for "
|
---|
411 | << collection << "\n";
|
---|
412 | }
|
---|
413 |
|
---|
414 | void collectset::configure(const text_t &key, const text_tarray &cfgline)
|
---|
415 | {
|
---|
416 | if (key == "collection" || key == "collectdir") return;
|
---|
417 |
|
---|
418 | collectservermapclass::iterator here = cservers.begin();
|
---|
419 | collectservermapclass::iterator end = cservers.end();
|
---|
420 |
|
---|
421 | while (here != end) {
|
---|
422 | assert ((*here).second.c != NULL);
|
---|
423 | if ((*here).second.c != NULL) {
|
---|
424 | if (key == "collectinfo") {
|
---|
425 | if ((*here).first == cfgline[0]) {
|
---|
426 | (*here).second.c->configure ("gsdlhome", cfgline[1]);
|
---|
427 | (*here).second.c->configure ("gdbmhome", cfgline[2]);
|
---|
428 | }
|
---|
429 | } else {
|
---|
430 | (*here).second.c->configure (key, cfgline);
|
---|
431 | }
|
---|
432 | }
|
---|
433 |
|
---|
434 | ++here;
|
---|
435 | }
|
---|
436 | }
|
---|
437 |
|
---|
438 | void collectset::getCollectionList (text_tarray &collist)
|
---|
439 | {
|
---|
440 | collist.erase(collist.begin(),collist.end());
|
---|
441 |
|
---|
442 | collectservermapclass::iterator here = cservers.begin();
|
---|
443 | collectservermapclass::iterator end = cservers.end();
|
---|
444 | while (here != end) {
|
---|
445 | assert ((*here).second.c != NULL);
|
---|
446 | if ((*here).second.c != NULL) {
|
---|
447 | collist.push_back ((*here).second.c->get_collection_name());
|
---|
448 | }
|
---|
449 | ++here;
|
---|
450 | }
|
---|
451 | }
|
---|