[9345] | 1 | /**********************************************************************
|
---|
| 2 | *
|
---|
| 3 | * gdbmsource.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 "gdbmsource.h"
|
---|
| 27 | #include "fileutil.h"
|
---|
| 28 | #include "OIDtools.h"
|
---|
| 29 | #include "gsdltools.h"
|
---|
| 30 |
|
---|
| 31 |
|
---|
| 32 | gdbmsourceclass::gdbmsourceclass () {
|
---|
[15584] | 33 | db_ptr = NULL;
|
---|
[9345] | 34 | textsearchptr = NULL;
|
---|
| 35 | classname = "gdbmsource";
|
---|
| 36 | }
|
---|
| 37 |
|
---|
| 38 | gdbmsourceclass::~gdbmsourceclass () {
|
---|
[15584] | 39 | if (db_ptr != NULL) delete db_ptr;
|
---|
[9345] | 40 | if (textsearchptr != NULL) delete textsearchptr;
|
---|
| 41 | }
|
---|
| 42 |
|
---|
| 43 | void gdbmsourceclass::configure (const text_t &key, const text_tarray &cfgline) {
|
---|
| 44 | if (cfgline.size() >= 1) {
|
---|
| 45 | const text_t &value = cfgline[0];
|
---|
| 46 |
|
---|
| 47 | if (key == "collection") collection = value;
|
---|
| 48 | else if (key == "collectdir") collectdir = value;
|
---|
| 49 | else if (key == "gsdlhome") gsdlhome = value;
|
---|
| 50 | else if (key == "gdbmhome") gdbmhome = value;
|
---|
| 51 | }
|
---|
| 52 |
|
---|
| 53 | if (key == "indexmap") {
|
---|
| 54 | indexmap.importmap (cfgline);
|
---|
| 55 |
|
---|
| 56 | } else if (key == "defaultindex") {
|
---|
| 57 | indexmap.from2to (cfgline[0], defaultindex);
|
---|
| 58 |
|
---|
| 59 | } else if (key == "subcollectionmap") {
|
---|
| 60 | subcollectionmap.importmap (cfgline);
|
---|
| 61 |
|
---|
| 62 | } else if (key == "defaultsubcollection") {
|
---|
| 63 | subcollectionmap.from2to (cfgline[0], defaultsubcollection);
|
---|
| 64 |
|
---|
| 65 | } else if (key == "languagemap") {
|
---|
| 66 | languagemap.importmap (cfgline);
|
---|
| 67 |
|
---|
[9937] | 68 | } else if (key == "defaultlanguage") {
|
---|
[9345] | 69 | languagemap.from2to (cfgline[0], defaultlanguage);
|
---|
[9937] | 70 | } else if (key == "indexstem") {
|
---|
| 71 | indexstem = cfgline[0];
|
---|
| 72 | }
|
---|
| 73 |
|
---|
[9345] | 74 | }
|
---|
| 75 |
|
---|
| 76 | bool gdbmsourceclass::init (ostream &logout) {
|
---|
| 77 | outconvertclass text_t2ascii;
|
---|
| 78 |
|
---|
| 79 | if (gdbmhome.empty()) gdbmhome = gsdlhome;
|
---|
| 80 |
|
---|
| 81 | if (!sourceclass::init (logout)) return false;
|
---|
| 82 |
|
---|
| 83 | if (defaultindex.empty()) {
|
---|
| 84 | // use first index in map as default if no default is set explicitly
|
---|
| 85 | text_tarray toarray;
|
---|
| 86 | indexmap.gettoarray(toarray);
|
---|
| 87 | if (toarray.size()) {
|
---|
| 88 | defaultindex = toarray[0];
|
---|
| 89 | }
|
---|
| 90 | }
|
---|
| 91 |
|
---|
| 92 | if (defaultsubcollection.empty()) {
|
---|
| 93 | // use first subcollection in map as default if no default is set explicitly
|
---|
| 94 | text_tarray toarray;
|
---|
| 95 | subcollectionmap.gettoarray(toarray);
|
---|
| 96 | if (toarray.size()) {
|
---|
| 97 | defaultsubcollection = toarray[0];
|
---|
| 98 | }
|
---|
| 99 | }
|
---|
| 100 |
|
---|
| 101 | if (defaultlanguage.empty()) {
|
---|
| 102 | // use first language in map as default if no default is set explicitly
|
---|
| 103 | text_tarray toarray;
|
---|
| 104 | languagemap.gettoarray(toarray);
|
---|
| 105 | if (toarray.size()) {
|
---|
| 106 | defaultlanguage = toarray[0];
|
---|
| 107 | }
|
---|
| 108 | }
|
---|
| 109 |
|
---|
| 110 | // get the collection directory name
|
---|
| 111 | if (collectdir.empty()) {
|
---|
| 112 | collectdir = filename_cat (gsdlhome, "collect", collection);
|
---|
| 113 | }
|
---|
| 114 |
|
---|
| 115 | // get the filename for the database and make sure it exists
|
---|
[9937] | 116 | if (indexstem.empty()) {
|
---|
| 117 | indexstem = collection;
|
---|
| 118 | }
|
---|
[15584] | 119 | db_filename = filename_cat(gdbmhome, "collect", collection, "index", "text", indexstem);
|
---|
| 120 | if (littleEndian()) db_filename += ".ldb";
|
---|
| 121 | else db_filename += ".bdb";
|
---|
[9345] | 122 |
|
---|
[15584] | 123 | if (!file_exists(db_filename)) {
|
---|
[9345] | 124 | logout << text_t2ascii
|
---|
[15584] | 125 | << "warning: database \"" << db_filename << "\" does not exist\n\n";
|
---|
| 126 | // return false;
|
---|
[9345] | 127 | }
|
---|
| 128 |
|
---|
| 129 | return true;
|
---|
| 130 | }
|
---|
| 131 |
|
---|
| 132 | bool gdbmsourceclass::translate_OID (const text_t &OIDin, text_t &OIDout,
|
---|
| 133 | comerror_t &err, ostream &logout) {
|
---|
| 134 |
|
---|
| 135 | outconvertclass text_t2ascii;
|
---|
| 136 |
|
---|
| 137 | err = noError;
|
---|
[15584] | 138 | if (db_ptr == NULL) {
|
---|
[9345] | 139 | // most likely a configuration problem
|
---|
| 140 | logout << text_t2ascii
|
---|
[15584] | 141 | << "configuration error: " << classname << " contains a null dbclass\n\n";
|
---|
[9345] | 142 | err = configurationError;
|
---|
| 143 | return true;
|
---|
| 144 | }
|
---|
| 145 |
|
---|
| 146 | // open the database
|
---|
[15584] | 147 | db_ptr->setlogout(&logout);
|
---|
| 148 | if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
|
---|
| 149 | // most likely a system problem (we have already checked that the database exists)
|
---|
[9345] | 150 | logout << text_t2ascii
|
---|
[15584] | 151 | << "system problem: open on database \"" << db_filename << "\" failed\n\n";
|
---|
[9345] | 152 | err = systemProblem;
|
---|
| 153 | return true;
|
---|
| 154 | }
|
---|
| 155 |
|
---|
| 156 | infodbclass info;
|
---|
[15584] | 157 | OIDout = db_ptr->translate_OID (OIDin, info);
|
---|
| 158 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
[9345] | 159 | return true;
|
---|
| 160 | }
|
---|
| 161 |
|
---|
| 162 | bool gdbmsourceclass::get_metadata (const text_t &/*requestParams*/, const text_t &/*refParams*/,
|
---|
| 163 | bool getParents, const text_tset &fields,
|
---|
| 164 | const text_t &OID, MetadataInfo_tmap &metadata,
|
---|
| 165 | comerror_t &err, ostream &logout) {
|
---|
| 166 | outconvertclass text_t2ascii;
|
---|
| 167 |
|
---|
| 168 | metadata.erase(metadata.begin(), metadata.end());
|
---|
| 169 |
|
---|
| 170 | err = noError;
|
---|
[15584] | 171 | if (db_ptr == NULL) {
|
---|
[9345] | 172 | // most likely a configuration problem
|
---|
| 173 | logout << text_t2ascii
|
---|
[15584] | 174 | << "configuration error: " << classname <<" contains a null dbclass\n\n";
|
---|
[9345] | 175 | err = configurationError;
|
---|
| 176 | return true;
|
---|
| 177 | }
|
---|
| 178 |
|
---|
| 179 | // open the database
|
---|
[15584] | 180 | db_ptr->setlogout(&logout);
|
---|
| 181 | if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
|
---|
| 182 | // most likely a system problem (we have already checked that the database exists)
|
---|
[9345] | 183 | logout << text_t2ascii
|
---|
[15584] | 184 | << "system problem: open on database \"" << db_filename << "\" failed\n\n";
|
---|
[9345] | 185 | err = systemProblem;
|
---|
| 186 | return true;
|
---|
| 187 | }
|
---|
| 188 |
|
---|
| 189 | // get the metadata - if getParents is set we need to get
|
---|
| 190 | // info for all parents of OID as well as OID
|
---|
| 191 | vector<infodbclass> info_array;
|
---|
| 192 | text_tarray OIDs;
|
---|
| 193 | if (getParents) get_parents_array (OID, OIDs);
|
---|
| 194 | OIDs.push_back (OID);
|
---|
| 195 |
|
---|
| 196 | text_tarray::const_iterator this_OID = OIDs.begin();
|
---|
| 197 | text_tarray::const_iterator end_OID = OIDs.end();
|
---|
| 198 |
|
---|
| 199 | while (this_OID != end_OID) {
|
---|
| 200 | infodbclass info;
|
---|
[15584] | 201 | if (!db_ptr->getinfo(*this_OID, info)) return false;
|
---|
[9345] | 202 |
|
---|
| 203 | // adjust the metadata
|
---|
| 204 | text_t &contains = info["contains"];
|
---|
| 205 | if (contains.empty()) info["haschildren"] = 0;
|
---|
| 206 | else info["haschildren"] = 1;
|
---|
[10414] | 207 | //contains.clear();
|
---|
[9345] | 208 |
|
---|
| 209 | info_array.push_back(info);
|
---|
[9620] | 210 | ++this_OID;
|
---|
[9345] | 211 | }
|
---|
| 212 |
|
---|
| 213 | // if fields set is empty we want to get all available metadata
|
---|
| 214 | text_tset tfields = fields;
|
---|
| 215 | if (tfields.empty() && !info_array.empty()) {
|
---|
| 216 | infodbclass::iterator t_info = info_array[0].begin();
|
---|
| 217 | infodbclass::iterator e_info = info_array[0].end();
|
---|
| 218 | while (t_info != e_info) {
|
---|
| 219 | if ((*t_info).first != "contains")
|
---|
| 220 | tfields.insert ((*t_info).first);
|
---|
[9620] | 221 | ++t_info;
|
---|
[9345] | 222 | }
|
---|
| 223 | tfields.insert ("hasnext");
|
---|
| 224 | tfields.insert ("hasprevious");
|
---|
| 225 | }
|
---|
| 226 |
|
---|
| 227 | // collect together the metadata
|
---|
| 228 | bool donenextprevtest = false;
|
---|
| 229 | bool hasnext=false, hasprevious=false;
|
---|
| 230 | MetadataInfo_t this_metadata;
|
---|
| 231 | text_tarray *pos_metadata;
|
---|
| 232 | text_tset::const_iterator fields_here = tfields.begin();
|
---|
| 233 | text_tset::const_iterator fields_end = tfields.end();
|
---|
| 234 |
|
---|
| 235 | while (fields_here != fields_end) {
|
---|
| 236 | this_metadata.clear();
|
---|
| 237 | this_metadata.isRef = false;
|
---|
| 238 |
|
---|
| 239 | vector<infodbclass>::reverse_iterator this_info = info_array.rbegin();
|
---|
| 240 | vector<infodbclass>::reverse_iterator end_info = info_array.rend();
|
---|
| 241 | MetadataInfo_t *tmetaptr = &this_metadata;
|
---|
| 242 | while (this_info != end_info) {
|
---|
| 243 |
|
---|
| 244 | pos_metadata = (*this_info).getmultinfo(*fields_here);
|
---|
| 245 | if ((*fields_here == "hasnext" || *fields_here == "hasprevious")) {
|
---|
| 246 |
|
---|
| 247 | // collect metadata
|
---|
| 248 | if (!donenextprevtest) {
|
---|
| 249 | donenextprevtest = true;
|
---|
| 250 |
|
---|
| 251 | // cache parent contents array
|
---|
| 252 | text_t thisparent = get_parent (OID);
|
---|
| 253 | if (!thisparent.empty()) {
|
---|
| 254 | if (thisparent != parentOID) {
|
---|
| 255 | parentOID = thisparent;
|
---|
| 256 | parentcontents.erase(parentcontents.begin(), parentcontents.end());
|
---|
[15584] | 257 | if (db_ptr->getinfo(parentOID, parentinfo)) {
|
---|
[9345] | 258 | text_t &parentinfocontains = parentinfo["contains"];
|
---|
| 259 | if (!parentinfocontains.empty())
|
---|
| 260 | splitchar (parentinfocontains.begin(), parentinfocontains.end(),
|
---|
| 261 | ';', parentcontents);
|
---|
| 262 | }
|
---|
| 263 | }
|
---|
| 264 |
|
---|
| 265 | // do tests
|
---|
| 266 | text_tarray::const_iterator parentcontents_here = parentcontents.begin();
|
---|
| 267 | text_tarray::const_iterator parentcontents_end = parentcontents.end();
|
---|
| 268 | text_t shrunk_OID = OID;
|
---|
| 269 | shrink_parent (shrunk_OID);
|
---|
| 270 | while (parentcontents_here != parentcontents_end) {
|
---|
| 271 | if (*parentcontents_here == shrunk_OID) {
|
---|
| 272 | if (parentcontents_here == parentcontents.begin()) hasprevious = false;
|
---|
| 273 | else hasprevious = true;
|
---|
| 274 |
|
---|
[9620] | 275 | ++parentcontents_here;
|
---|
[9345] | 276 |
|
---|
| 277 | if (parentcontents_here == parentcontents.end()) hasnext = false;
|
---|
| 278 | else hasnext = true;
|
---|
| 279 |
|
---|
| 280 | break;
|
---|
| 281 | }
|
---|
| 282 |
|
---|
[9620] | 283 | ++parentcontents_here;
|
---|
[9345] | 284 | }
|
---|
| 285 |
|
---|
| 286 | // fill in metadata
|
---|
| 287 | if ((*fields_here == "hasnext" && hasnext) ||
|
---|
| 288 | (*fields_here == "hasprevious" && hasprevious))
|
---|
| 289 | tmetaptr->values.push_back("1");
|
---|
| 290 | else
|
---|
| 291 | tmetaptr->values.push_back("0");
|
---|
| 292 | } else
|
---|
| 293 | tmetaptr->values.push_back("0");
|
---|
| 294 | }
|
---|
| 295 | }
|
---|
[10414] | 296 | //else if (pos_metadata != NULL && *fields_here != "contains") {
|
---|
| 297 | else if (pos_metadata != NULL) {
|
---|
[9345] | 298 | tmetaptr->values = *pos_metadata;
|
---|
| 299 | }
|
---|
| 300 | else
|
---|
| 301 | tmetaptr->values.push_back("");
|
---|
| 302 |
|
---|
[9620] | 303 | ++this_info;
|
---|
[9345] | 304 | if (this_info != end_info) {
|
---|
| 305 | tmetaptr->parent = new MetadataInfo_t();
|
---|
| 306 | tmetaptr = tmetaptr->parent;
|
---|
| 307 | }
|
---|
| 308 | }
|
---|
| 309 | metadata[*fields_here] = this_metadata;
|
---|
[9620] | 310 | ++fields_here;
|
---|
[9345] | 311 | }
|
---|
[13780] | 312 |
|
---|
[15584] | 313 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
[9345] | 314 | return true;
|
---|
| 315 | }
|
---|
| 316 |
|
---|
| 317 |
|
---|
| 318 | bool gdbmsourceclass::get_document (const text_t &OID, text_t &doc,
|
---|
| 319 | comerror_t &err, ostream &logout) {
|
---|
| 320 |
|
---|
| 321 | outconvertclass text_t2ascii;
|
---|
| 322 |
|
---|
| 323 | err = noError;
|
---|
[15584] | 324 | if (db_ptr == NULL) {
|
---|
[9345] | 325 | // most likely a configuration problem
|
---|
| 326 | logout << text_t2ascii
|
---|
[15584] | 327 | << "configuration error: " << classname << " contains a null dbclass\n\n";
|
---|
[9345] | 328 | err = configurationError;
|
---|
| 329 | return true;
|
---|
| 330 | }
|
---|
| 331 |
|
---|
| 332 | // open the database
|
---|
[15584] | 333 | db_ptr->setlogout(&logout);
|
---|
| 334 | if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
|
---|
| 335 | // most likely a system problem (we have already checked that the database exists)
|
---|
[9345] | 336 | logout << text_t2ascii
|
---|
[15584] | 337 | << "system problem: open on database \"" << db_filename << "\" failed\n\n";
|
---|
[9345] | 338 | err = systemProblem;
|
---|
| 339 | return true;
|
---|
| 340 | }
|
---|
| 341 |
|
---|
| 342 | text_t tOID = OID;
|
---|
| 343 | if (needs_translating (OID))
|
---|
| 344 | translate_OID (OID, tOID, err, logout);
|
---|
| 345 | infodbclass info;
|
---|
[15584] | 346 | if (!db_ptr->getinfo(tOID, info)) {
|
---|
| 347 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
[13780] | 348 | return false;
|
---|
| 349 | }
|
---|
[9345] | 350 |
|
---|
| 351 | if (info["hastxt"].getint() == 1) {
|
---|
| 352 | int docnum = info["docnum"].getint();
|
---|
| 353 |
|
---|
| 354 | // set the collection directory
|
---|
| 355 | textsearchptr->setcollectdir (collectdir);
|
---|
| 356 |
|
---|
| 357 | // get the text
|
---|
| 358 | textsearchptr->docTargetDocument(defaultindex, defaultsubcollection,
|
---|
| 359 | defaultlanguage, collection, docnum, doc);
|
---|
| 360 | }
|
---|
[13780] | 361 |
|
---|
[15584] | 362 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
[9345] | 363 | return true;
|
---|
| 364 | }
|
---|
| 365 |
|
---|
| 366 | bool gdbmsourceclass::is_searchable(bool &issearchable, comerror_t &err, ostream &logout) {
|
---|
| 367 | err = noError;
|
---|
| 368 | issearchable = false;
|
---|
| 369 |
|
---|
| 370 | text_tarray fromarray;
|
---|
| 371 | indexmap.getfromarray(fromarray);
|
---|
| 372 | if (fromarray.size() == 0) {
|
---|
| 373 | return true;
|
---|
| 374 | } else if (fromarray.size() == 1) {
|
---|
| 375 | if (fromarray[0] == "dummy:text") {
|
---|
| 376 | // always return true - issearchable is false here though
|
---|
| 377 | return true;
|
---|
| 378 | }
|
---|
| 379 | }
|
---|
| 380 | issearchable = true;
|
---|
| 381 | return true;
|
---|
| 382 | }
|
---|