1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * oaiconfig.cpp --
|
---|
4 | *
|
---|
5 | * Copyright (C) 2004-2010 The New Zealand Digital Library Project
|
---|
6 | *
|
---|
7 | * A component of the Greenstone digital library software
|
---|
8 | * from the New Zealand Digital Library Project at the
|
---|
9 | * University of Waikato, New Zealand.
|
---|
10 | *
|
---|
11 | * This program is free software; you can redistribute it and/or modify
|
---|
12 | * it under the terms of the GNU General Public License as published by
|
---|
13 | * the Free Software Foundation; either version 2 of the License, or
|
---|
14 | * (at your option) any later version.
|
---|
15 | *
|
---|
16 | * This program is distributed in the hope that it will be useful,
|
---|
17 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
18 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
19 | * GNU General Public License for more details.
|
---|
20 | *
|
---|
21 | * You should have received a copy of the GNU General Public License
|
---|
22 | * along with this program; if not, write to the Free Software
|
---|
23 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
24 | *
|
---|
25 | *********************************************************************/
|
---|
26 |
|
---|
27 | #include "oaiconfig.h"
|
---|
28 | #include <iostream>
|
---|
29 | #include <stdlib.h>
|
---|
30 | #include "fileutil.h"
|
---|
31 |
|
---|
32 |
|
---|
33 | /**
|
---|
34 | * The mapping works as follows in the oai.cfg or collect.cfg file.
|
---|
35 | *
|
---|
36 | * A line is in the format oaimapping <collection field> <oai field>
|
---|
37 | *
|
---|
38 | * The map here is used to look up the "Greenstone" name which is mapped from
|
---|
39 | * a given OAI field name, the reverse direction to that given in the
|
---|
40 | * Greenstone collect.cfg file. The oairecordaction class instance which
|
---|
41 | * produces output for an OAI record information request thus uses the map
|
---|
42 | * to work from the field in the collection it has on hand which OAI
|
---|
43 | * record name it should use instead.
|
---|
44 | *
|
---|
45 | * An extension is to be used for this in which the OAI field name in the
|
---|
46 | * collect.cfg file can be made specific for a particular record format.
|
---|
47 | * This is done using the OAI field name in the format of
|
---|
48 | * <OAI format>.<OAI field name>
|
---|
49 | * Thus, an rfc1807 Title field would be referred to as rfc1807.Title
|
---|
50 | *
|
---|
51 | * A collection-level mapping can be provided in oai.cfg by prepending
|
---|
52 | * collname to collection field:
|
---|
53 | * <collname:field> <oai field>
|
---|
54 |
|
---|
55 | * In the absence of a particular format name, the mapping is taken to be
|
---|
56 | * universal.
|
---|
57 | */
|
---|
58 |
|
---|
59 | oaiconfig::oaiconfig() : configurable () {
|
---|
60 | this->resumptionSize = -1; // Default = do not use resumption tokens
|
---|
61 | }
|
---|
62 |
|
---|
63 | oaiconfig::oaiconfig(text_t &gsdlhome, text_t &gsdlcollect)
|
---|
64 | {
|
---|
65 | // read main configuration file (oai.cfg) to get oai collections
|
---|
66 | text_t mainconfig = filename_cat(gsdlhome, "etc", "oai.cfg");
|
---|
67 | this->collection = "";
|
---|
68 | this->resumptionSize = -1;
|
---|
69 | this->read_configfile(mainconfig);
|
---|
70 |
|
---|
71 | // then if we've not got a specified collection in the gsdlcollect
|
---|
72 | // parameter, read in all the collection's individual configurations
|
---|
73 | if (gsdlcollect == "") {
|
---|
74 | text_tarray::iterator start = this->collectList.begin();
|
---|
75 | text_tarray::iterator here = this->collectList.end()-1;
|
---|
76 | while (here != start) {
|
---|
77 | if (!this->configureCollection(gsdlhome, *here)) {
|
---|
78 | this->collectList.erase(here);
|
---|
79 | }
|
---|
80 | --here;
|
---|
81 | }
|
---|
82 | // and do the first one
|
---|
83 | if (!this->configureCollection(gsdlhome, *here)) {
|
---|
84 | this->collectList.erase(here);
|
---|
85 | }
|
---|
86 |
|
---|
87 | }
|
---|
88 | else {
|
---|
89 | // what do we do if this fails?
|
---|
90 | this->configureCollection(gsdlhome, gsdlcollect);
|
---|
91 | }
|
---|
92 | }
|
---|
93 |
|
---|
94 | oaiconfig::~oaiconfig()
|
---|
95 | {
|
---|
96 | oaicollectmap::iterator here = this->collectMap.begin();
|
---|
97 | oaicollectmap::iterator end = this->collectMap.end();
|
---|
98 | while (here != end) {
|
---|
99 | delete here->second;
|
---|
100 | ++here;
|
---|
101 | }
|
---|
102 | }
|
---|
103 |
|
---|
104 | int oaiconfig::resumeAfter()
|
---|
105 | { return this->resumptionSize;
|
---|
106 | }
|
---|
107 |
|
---|
108 | int oaiconfig::getOAIVersion()
|
---|
109 | {
|
---|
110 | if (this->oaiVersion == "1.1") {
|
---|
111 | return 110;
|
---|
112 | }
|
---|
113 | return 200;
|
---|
114 | }
|
---|
115 |
|
---|
116 | bool oaiconfig::configureCollection(const text_t &gsdlhome, const text_t &gsdlcollect)
|
---|
117 | {
|
---|
118 | text_t cnfgfile = filename_cat(gsdlhome, "collect", gsdlcollect, "etc", "collect.cfg");
|
---|
119 | if (!file_exists(cnfgfile)) {
|
---|
120 | return false;
|
---|
121 | }
|
---|
122 | this->collection = gsdlcollect;
|
---|
123 | this->read_configfile(cnfgfile);
|
---|
124 |
|
---|
125 | return true;
|
---|
126 | }
|
---|
127 | void oaiconfig::configure (const text_t &key, const text_tarray &cfgline)
|
---|
128 | {
|
---|
129 | // we've got an oai mapping item, and at least two fields
|
---|
130 | if (key == "oaimapping" && cfgline.size() > 1) {
|
---|
131 | text_t::const_iterator colonAt;
|
---|
132 | text_t index, name, configCollection;
|
---|
133 |
|
---|
134 | // Take a default collection as being whatever the collection being configured is...
|
---|
135 | configCollection = this->collection;
|
---|
136 |
|
---|
137 | // get the name of the (collection) field to map; this may actually
|
---|
138 | // be in a colon separated format of the type
|
---|
139 | // <collection name>:<field name>
|
---|
140 | index = cfgline[0];
|
---|
141 | if ((colonAt = findchar(index.begin(), index.end(), ':')) != index.end()) {
|
---|
142 | configCollection = substr(index.begin(), colonAt);
|
---|
143 |
|
---|
144 | if (this->collection != "" && configCollection != this->collection) {
|
---|
145 | cerr << "Attempt to configure OAI mappings for " << configCollection << " in " << this->collection << endl;
|
---|
146 | }
|
---|
147 |
|
---|
148 | colonAt += 1;
|
---|
149 | index = substr(colonAt, index.end());
|
---|
150 | }
|
---|
151 |
|
---|
152 | // the second parameter is the metadata field to map the collection
|
---|
153 | // field onto. It may be provided with a metadata protocol (which
|
---|
154 | // will be given first and separated by a period or full stop). In
|
---|
155 | // the case of format.field name, the splitting is done here.
|
---|
156 | if ((colonAt = findchar(cfgline[1].begin(), cfgline[1].end(), '.')) != cfgline[1].end()) {
|
---|
157 | text_t stub = substr(cfgline[1].begin(), colonAt);
|
---|
158 | colonAt += 1;
|
---|
159 | name = substr(colonAt, cfgline[1].end());
|
---|
160 | index.append(":");
|
---|
161 | index.append(stub);
|
---|
162 | }
|
---|
163 | else {
|
---|
164 | name = cfgline[1];
|
---|
165 | }
|
---|
166 |
|
---|
167 | // now 'index' is in the form <collectionfield>:(formatname)
|
---|
168 | // 'name' is simply the fieldname within the format
|
---|
169 | // 'configCollection' is the collection to be configured
|
---|
170 |
|
---|
171 | // now simply map the field name (index) onto the collection name (name)
|
---|
172 | if (this->collectMap[configCollection] == NULL) {
|
---|
173 | this->collectMap[configCollection] = new oaicollectconfig(configCollection);
|
---|
174 | }
|
---|
175 | this->collectMap[configCollection]->fieldMap[index] = name;
|
---|
176 |
|
---|
177 | // cerr << "Mapping " << index << " to " << name << " in " << configCollection << endl;
|
---|
178 |
|
---|
179 | // TODO: check that the mapped field is actually in use
|
---|
180 | }
|
---|
181 | else if (key == "oaicollection" && cfgline.size() >= 1) {
|
---|
182 | // Configure a collection to be used as part of the OAI archive.
|
---|
183 | // This line should read:
|
---|
184 | //
|
---|
185 | // oaicollection <collectionname>
|
---|
186 | //
|
---|
187 | // Where <collectionname> is the name of the directory inside the
|
---|
188 | // gsdl/collect folder which contains the collection.
|
---|
189 | //
|
---|
190 | // To configure several collections, merely repeat this line,
|
---|
191 | // or alternatively use additional collection names after the
|
---|
192 | // first one.
|
---|
193 | //
|
---|
194 | // This configuration should only appear in oai.cfg
|
---|
195 | //
|
---|
196 | if (this->collection != "") {
|
---|
197 | cerr << "Attempt to configure an oai collection outside of oai.cfg" << endl;
|
---|
198 | cerr << "Configuration attempted in " << this->collection << " collection." << endl;
|
---|
199 | exit(1);
|
---|
200 | }
|
---|
201 | for (int c = 0; c < cfgline.size(); ++c) {
|
---|
202 | this->collectList.push_back(cfgline[c]);
|
---|
203 | }
|
---|
204 | }
|
---|
205 | else if (key == "oaimetadata" && cfgline.size() >= 1) {
|
---|
206 | // List of metadata prefixes to suuport
|
---|
207 | // This line should read:
|
---|
208 | //
|
---|
209 | // oaicollection <metadataname> <metadataname>...
|
---|
210 | //
|
---|
211 | //
|
---|
212 | // This configuration should only appear in oai.cfg
|
---|
213 | //
|
---|
214 | if (this->collection != "") {
|
---|
215 | cerr << "Attempt to configure oai metadata outside of oai.cfg" << endl;
|
---|
216 | cerr << "Configuration attempted in " << this->collection << " collection." << endl;
|
---|
217 | exit(1);
|
---|
218 | }
|
---|
219 | for (int c = 0; c < cfgline.size(); ++c) {
|
---|
220 | // todo: check that the set name is valid
|
---|
221 | this->metadataSet.insert(cfgline[c]);
|
---|
222 | }
|
---|
223 | }
|
---|
224 | else if (key == "oaiinfo" && cfgline.size() >= 1) {
|
---|
225 | // Get a piece of information for the oai repository information
|
---|
226 | // request. The line should read:
|
---|
227 | //
|
---|
228 | // oaiinfo <information field name> <value>
|
---|
229 | //
|
---|
230 | // This configuration should only be attempted in oai.cfg
|
---|
231 | //
|
---|
232 | if (this->collection != "") {
|
---|
233 | cerr << "Attempt to set oai information outside of oai.cfg" << endl;
|
---|
234 | cerr << "Configuration attempted in " << this->collection << " collection." << endl;
|
---|
235 | exit(1);
|
---|
236 | }
|
---|
237 |
|
---|
238 | // if no second parameter is given, then the first parameter
|
---|
239 | if (cfgline.size() == 1) {
|
---|
240 | this->infoMap[cfgline[0]] = cfgline[0];
|
---|
241 | }
|
---|
242 | else {
|
---|
243 | this->infoMap[cfgline[0]] = cfgline[1];
|
---|
244 | }
|
---|
245 | }
|
---|
246 | else if ( key == "oaisetname" || key == "oaisetdescription") {
|
---|
247 | text_t coll_name;
|
---|
248 | text_t value = "";
|
---|
249 | if (this->collection != "") {
|
---|
250 | // we are in collect.cfg
|
---|
251 | coll_name = this->collection;
|
---|
252 | if (cfgline.size() == 1) {
|
---|
253 | // just the collection value
|
---|
254 | value = cfgline[0];
|
---|
255 | }
|
---|
256 | else if (cfgline.size() == 2) {
|
---|
257 | // we have a subset name (eg for classifier)
|
---|
258 | coll_name.append(":");
|
---|
259 | coll_name.append(cfgline[0]);
|
---|
260 | value = cfgline[1];
|
---|
261 | }
|
---|
262 | } else if (cfgline.size() == 2) {
|
---|
263 | // oai.cfg, line should be collname, setName
|
---|
264 | coll_name = cfgline[0];
|
---|
265 | value = cfgline[1];
|
---|
266 | }
|
---|
267 | if (value != "") {
|
---|
268 | if (this->collectMap[coll_name] == NULL) {
|
---|
269 | this->collectMap[coll_name] = new oaicollectconfig(coll_name);
|
---|
270 | }
|
---|
271 | if (key == "oaisetname") {
|
---|
272 | this->collectMap[coll_name]->setName = value;
|
---|
273 | } else if (key == "oaisetdescription") {
|
---|
274 | this->collectMap[coll_name]->setDescription = value;
|
---|
275 | }
|
---|
276 | }
|
---|
277 | }
|
---|
278 |
|
---|
279 | else if (key == "resumeafter" && cfgline.size() >= 1) {
|
---|
280 | this->resumptionSize = cfgline[0].getint();
|
---|
281 | }
|
---|
282 |
|
---|
283 | else if (key == "maintainer") {
|
---|
284 | this->maintainer = cfgline[0];
|
---|
285 | }
|
---|
286 | else if (key == "repositoryName") {
|
---|
287 | this->repositoryName = cfgline[0];
|
---|
288 | }
|
---|
289 | else if (key == "repositoryId") {
|
---|
290 | this->repositoryId = cfgline[0];
|
---|
291 | }
|
---|
292 | else if (key == "repositoryIdVersion") {
|
---|
293 | this->repositoryIdVersion = cfgline[0];
|
---|
294 | }
|
---|
295 | else if (key == "baseURL") {
|
---|
296 | this->baseURL = cfgline[0];
|
---|
297 | }
|
---|
298 | else if (key == "baseLibraryURL") {
|
---|
299 | this->baseLibraryURL = cfgline[0];
|
---|
300 | }
|
---|
301 | else if (key == "baseDocRoot") {
|
---|
302 | this->baseDocRoot = cfgline[0];
|
---|
303 | }
|
---|
304 | else if (key == "oaiversion") {
|
---|
305 | this->oaiVersion = cfgline[0];
|
---|
306 | }
|
---|
307 |
|
---|
308 | }
|
---|
309 |
|
---|
310 | text_t oaiconfig::getMapping(const text_t &collection, const text_t &collectfield)
|
---|
311 | {
|
---|
312 | if (this->collectMap[collection] == NULL) {
|
---|
313 | return "";
|
---|
314 | }
|
---|
315 | return this->collectMap[collection]->fieldMap[collectfield];
|
---|
316 | }
|
---|
317 |
|
---|
318 | /**
|
---|
319 | * Get the mapping for a field in a given collection; if no mapping
|
---|
320 | * exists, the result will be a blank string.
|
---|
321 | */
|
---|
322 | text_t oaiconfig::getMapping(const text_t &collection, const text_t &collectfield, const text_t &formatname)
|
---|
323 | {
|
---|
324 | text_t fullName = collectfield;
|
---|
325 | fullName.append(":");
|
---|
326 | fullName.append(formatname);
|
---|
327 |
|
---|
328 | // try the collection-specific options first
|
---|
329 | if (this->collectMap[collection] != NULL) {
|
---|
330 | // first try the most specific item - this collection, and given that protocol
|
---|
331 | if (this->collectMap[collection]->fieldMap.count(fullName) >= 1) {
|
---|
332 | return this->collectMap[collection]->fieldMap[fullName];
|
---|
333 | }
|
---|
334 | // otherwise, fall back to this collection, and all protocols
|
---|
335 | else if (this->collectMap[collection]->fieldMap.count(collectfield) >= 1) {
|
---|
336 | return this->collectMap[collection]->fieldMap[collectfield];
|
---|
337 | }
|
---|
338 | }
|
---|
339 |
|
---|
340 | // if no mappings exist, return an empty item
|
---|
341 | if (this->collectMap[""] == NULL) {
|
---|
342 | return "";
|
---|
343 | }
|
---|
344 |
|
---|
345 | // then try generic rules
|
---|
346 | if (this->collectMap[""]->fieldMap.count(fullName) >= 1) {
|
---|
347 | return this->collectMap[""]->fieldMap[fullName];
|
---|
348 | }
|
---|
349 | else {
|
---|
350 | return this->collectMap[""]->fieldMap[collectfield];
|
---|
351 | }
|
---|
352 | }
|
---|
353 |
|
---|
354 | text_t oaiconfig::getBaseURL()
|
---|
355 | {
|
---|
356 | return this->baseURL;
|
---|
357 | }
|
---|
358 | text_t oaiconfig::getBaseLibraryURL()
|
---|
359 | {
|
---|
360 | return this->baseLibraryURL;
|
---|
361 | }
|
---|
362 | text_t oaiconfig::getBaseDocRoot()
|
---|
363 | {
|
---|
364 | return this->baseDocRoot;
|
---|
365 | }
|
---|
366 | text_t oaiconfig::getRepositoryName()
|
---|
367 | {
|
---|
368 | return this->repositoryName;
|
---|
369 | }
|
---|
370 | text_t oaiconfig::getRepositoryId()
|
---|
371 | {
|
---|
372 | return this->repositoryId;
|
---|
373 | }
|
---|
374 | text_t oaiconfig::getRepositoryIdVersion()
|
---|
375 | {
|
---|
376 | return this->repositoryIdVersion;
|
---|
377 | }
|
---|
378 | text_t oaiconfig::getMaintainer()
|
---|
379 | {
|
---|
380 | return this->maintainer;
|
---|
381 | }
|
---|
382 | text_t oaiconfig::getSetName(const text_t &setSpec)
|
---|
383 | {
|
---|
384 | if (this->collectMap[setSpec] == NULL) {
|
---|
385 | return "" ;
|
---|
386 | }
|
---|
387 |
|
---|
388 | return this->collectMap[setSpec]->setName;
|
---|
389 |
|
---|
390 | }
|
---|
391 |
|
---|
392 | text_t oaiconfig::getSetDescription(const text_t &setSpec)
|
---|
393 | {
|
---|
394 | if (this->collectMap[setSpec] == NULL) {
|
---|
395 | return "" ;
|
---|
396 | }
|
---|
397 |
|
---|
398 | return this->collectMap[setSpec]->setDescription;
|
---|
399 | }
|
---|
400 |
|
---|