[219] | 1 | /**********************************************************************
|
---|
| 2 | *
|
---|
| 3 | * source.cpp --
|
---|
| 4 | * Copyright (C) 1999 The New Zealand Digital Library Project
|
---|
| 5 | *
|
---|
[534] | 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.
|
---|
[219] | 9 | *
|
---|
[534] | 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 | *
|
---|
[219] | 24 | *********************************************************************/
|
---|
| 25 |
|
---|
| 26 | #include "source.h"
|
---|
[15591] | 27 | #include "fileutil.h"
|
---|
| 28 | #include "OIDtools.h"
|
---|
[219] | 29 | #include <assert.h>
|
---|
| 30 |
|
---|
| 31 |
|
---|
[15591] | 32 | sourceclass::sourceclass ()
|
---|
| 33 | {
|
---|
| 34 | db_ptr = NULL;
|
---|
| 35 | textsearchptr = NULL;
|
---|
| 36 | classname = "source";
|
---|
[219] | 37 | }
|
---|
| 38 |
|
---|
[15591] | 39 | sourceclass::~sourceclass ()
|
---|
| 40 | {
|
---|
| 41 | if (db_ptr != NULL) delete db_ptr;
|
---|
| 42 | if (textsearchptr != NULL) delete textsearchptr;
|
---|
[219] | 43 | }
|
---|
| 44 |
|
---|
| 45 | // configure should be called once for each configuration line
|
---|
[15591] | 46 | void sourceclass::configure (const text_t &key, const text_tarray &cfgline)
|
---|
| 47 | {
|
---|
| 48 | if (cfgline.size() >= 1) {
|
---|
| 49 | const text_t &value = cfgline[0];
|
---|
| 50 |
|
---|
| 51 | if (key == "collection") collection = value;
|
---|
| 52 | else if (key == "collectdir") collectdir = value;
|
---|
| 53 | else if (key == "gsdlhome") gsdlhome = value;
|
---|
[16310] | 54 | else if (key == "collecthome") collecthome = value;
|
---|
[15591] | 55 | else if (key == "gdbmhome") dbhome = value;
|
---|
| 56 | }
|
---|
| 57 |
|
---|
| 58 | if (key == "indexmap") {
|
---|
| 59 | indexmap.importmap (cfgline);
|
---|
| 60 |
|
---|
| 61 | } else if (key == "defaultindex") {
|
---|
| 62 | indexmap.from2to (cfgline[0], defaultindex);
|
---|
| 63 |
|
---|
| 64 | } else if (key == "subcollectionmap") {
|
---|
| 65 | subcollectionmap.importmap (cfgline);
|
---|
| 66 |
|
---|
| 67 | } else if (key == "defaultsubcollection") {
|
---|
| 68 | subcollectionmap.from2to (cfgline[0], defaultsubcollection);
|
---|
| 69 |
|
---|
| 70 | } else if (key == "languagemap") {
|
---|
| 71 | languagemap.importmap (cfgline);
|
---|
| 72 |
|
---|
| 73 | } else if (key == "defaultlanguage") {
|
---|
| 74 | languagemap.from2to (cfgline[0], defaultlanguage);
|
---|
| 75 | } else if (key == "indexstem") {
|
---|
| 76 | indexstem = cfgline[0];
|
---|
| 77 | }
|
---|
[219] | 78 | }
|
---|
| 79 |
|
---|
[16310] | 80 | text_t sourceclass::resolve_db_filename(const text_t& idx,
|
---|
| 81 | const text_t& file_ext)
|
---|
| 82 | {
|
---|
| 83 | // This is an exact copy of the method (of the same name) in filterclass
|
---|
| 84 | // Makes sense to merge them, in which either gsdlhome, collecthome,
|
---|
| 85 | // dbhome, and collection need to also be passed in as parameters,
|
---|
| 86 | // or else there is some notion of a shared base class that both
|
---|
| 87 | // filter and source inherit from
|
---|
| 88 |
|
---|
| 89 | // NB: there is an even greater opportunity to share more code in this
|
---|
| 90 | // function if sql_db_ptr/db_ptr and db_filename and sql_db_filename
|
---|
| 91 | // are also drawn from one object
|
---|
| 92 |
|
---|
| 93 | text_t resolved_filename;
|
---|
| 94 |
|
---|
| 95 | if (gsdlhome==dbhome) {
|
---|
| 96 | // dbhome has defaulted to gsdlhome which we take to means the
|
---|
| 97 | // database has been specifically moved out of gsdlhome area.
|
---|
| 98 | // => it should be whereever collecthome is set to
|
---|
| 99 |
|
---|
| 100 | resolved_filename = filename_cat(collecthome, collection, "index", "text", idx);
|
---|
| 101 | }
|
---|
| 102 | else {
|
---|
| 103 | // dbhome is explicitly set to something other than gsdlhome
|
---|
| 104 | // => use dbhome
|
---|
| 105 | resolved_filename = filename_cat(dbhome, "collect", collection, "index", "text", idx);
|
---|
| 106 | }
|
---|
| 107 |
|
---|
| 108 | resolved_filename += file_ext;
|
---|
| 109 |
|
---|
| 110 | return resolved_filename;
|
---|
| 111 | }
|
---|
| 112 |
|
---|
| 113 |
|
---|
| 114 |
|
---|
[219] | 115 | // init should be called after all the configuration is done but
|
---|
| 116 | // before any other methods are called
|
---|
[15591] | 117 | bool sourceclass::init (ostream &logout)
|
---|
| 118 | {
|
---|
| 119 | outconvertclass text_t2ascii;
|
---|
| 120 |
|
---|
[16310] | 121 | if (collecthome.empty()) collecthome = filename_cat(gsdlhome,"collect");
|
---|
[15591] | 122 | if (dbhome.empty()) dbhome = gsdlhome;
|
---|
| 123 |
|
---|
| 124 | if (defaultindex.empty()) {
|
---|
| 125 | // use first index in map as default if no default is set explicitly
|
---|
| 126 | text_tarray toarray;
|
---|
| 127 | indexmap.gettoarray(toarray);
|
---|
| 128 | if (toarray.size()) {
|
---|
| 129 | defaultindex = toarray[0];
|
---|
| 130 | }
|
---|
| 131 | }
|
---|
| 132 |
|
---|
| 133 | if (defaultsubcollection.empty()) {
|
---|
| 134 | // use first subcollection in map as default if no default is set explicitly
|
---|
| 135 | text_tarray toarray;
|
---|
| 136 | subcollectionmap.gettoarray(toarray);
|
---|
| 137 | if (toarray.size()) {
|
---|
| 138 | defaultsubcollection = toarray[0];
|
---|
| 139 | }
|
---|
| 140 | }
|
---|
| 141 |
|
---|
| 142 | if (defaultlanguage.empty()) {
|
---|
| 143 | // use first language in map as default if no default is set explicitly
|
---|
| 144 | text_tarray toarray;
|
---|
| 145 | languagemap.gettoarray(toarray);
|
---|
| 146 | if (toarray.size()) {
|
---|
| 147 | defaultlanguage = toarray[0];
|
---|
| 148 | }
|
---|
| 149 | }
|
---|
| 150 |
|
---|
| 151 | // get the collection directory name
|
---|
| 152 | if (collectdir.empty()) {
|
---|
[16310] | 153 | collectdir = filename_cat (collecthome, collection);
|
---|
[15591] | 154 | }
|
---|
[15680] | 155 |
|
---|
| 156 | if (db_ptr == NULL) {
|
---|
| 157 | // most likely a configuration problem
|
---|
| 158 | logout << text_t2ascii
|
---|
| 159 | << "configuration error: queryfilter contains a null dbclass\n\n";
|
---|
| 160 | return false;
|
---|
| 161 | }
|
---|
| 162 |
|
---|
[15591] | 163 | // get the filename for the database and make sure it exists
|
---|
| 164 | if (indexstem.empty()) {
|
---|
| 165 | indexstem = collection;
|
---|
| 166 | }
|
---|
[16310] | 167 | db_filename = resolve_db_filename(indexstem, db_ptr->getfileextension());
|
---|
[15591] | 168 | if (!file_exists(db_filename)) {
|
---|
| 169 | logout << text_t2ascii
|
---|
| 170 | << "warning: database \"" << db_filename << "\" does not exist\n\n";
|
---|
| 171 | // return false;
|
---|
| 172 | }
|
---|
| 173 |
|
---|
[219] | 174 | return true;
|
---|
| 175 | }
|
---|
| 176 |
|
---|
[249] | 177 |
|
---|
| 178 | // translate_OID translates OIDs using ".pr", ."fc" etc.
|
---|
[15591] | 179 | bool sourceclass::translate_OID (const text_t &OIDin, text_t &OIDout,
|
---|
| 180 | comerror_t &err, ostream &logout)
|
---|
| 181 | {
|
---|
| 182 | outconvertclass text_t2ascii;
|
---|
| 183 |
|
---|
[249] | 184 | err = noError;
|
---|
[15591] | 185 | if (db_ptr == NULL) {
|
---|
| 186 | // most likely a configuration problem
|
---|
| 187 | logout << text_t2ascii
|
---|
| 188 | << "configuration error: " << classname << " contains a null dbclass\n\n";
|
---|
| 189 | err = configurationError;
|
---|
| 190 | return true;
|
---|
| 191 | }
|
---|
[249] | 192 |
|
---|
[15591] | 193 | // open the database
|
---|
| 194 | db_ptr->setlogout(&logout);
|
---|
| 195 | if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
|
---|
| 196 | // most likely a system problem (we have already checked that the database exists)
|
---|
| 197 | logout << text_t2ascii
|
---|
| 198 | << "system problem: open on database \"" << db_filename << "\" failed\n\n";
|
---|
| 199 | err = systemProblem;
|
---|
| 200 | return true;
|
---|
| 201 | }
|
---|
| 202 |
|
---|
| 203 | infodbclass info;
|
---|
| 204 | OIDout = db_ptr->translate_OID (OIDin, info);
|
---|
| 205 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
| 206 | return true;
|
---|
[249] | 207 | }
|
---|
| 208 |
|
---|
| 209 |
|
---|
[21772] | 210 | // get_metadata fills out the metadata if possible, if it is not responsible
|
---|
[226] | 211 | // for the given OID then it will return false.
|
---|
[15591] | 212 | bool sourceclass::get_metadata (const text_t &requestParams, const text_t &refParams,
|
---|
| 213 | bool getParents, const text_tset &fields,
|
---|
| 214 | const text_t &OID, MetadataInfo_tmap &metadata,
|
---|
| 215 | comerror_t &err, ostream &logout)
|
---|
| 216 | {
|
---|
| 217 | outconvertclass text_t2ascii;
|
---|
| 218 |
|
---|
[226] | 219 | metadata.erase(metadata.begin(), metadata.end());
|
---|
[15591] | 220 |
|
---|
[226] | 221 | err = noError;
|
---|
[15591] | 222 | if (db_ptr == NULL) {
|
---|
| 223 | // most likely a configuration problem
|
---|
| 224 | logout << text_t2ascii
|
---|
| 225 | << "configuration error: " << classname <<" contains a null dbclass\n\n";
|
---|
| 226 | err = configurationError;
|
---|
| 227 | return true;
|
---|
| 228 | }
|
---|
[219] | 229 |
|
---|
[15591] | 230 | // open the database
|
---|
| 231 | db_ptr->setlogout(&logout);
|
---|
| 232 | if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
|
---|
| 233 | // most likely a system problem (we have already checked that the database exists)
|
---|
| 234 | logout << text_t2ascii
|
---|
| 235 | << "system problem: open on database \"" << db_filename << "\" failed\n\n";
|
---|
| 236 | err = systemProblem;
|
---|
| 237 | return true;
|
---|
| 238 | }
|
---|
| 239 |
|
---|
| 240 | // get the metadata - if getParents is set we need to get
|
---|
| 241 | // info for all parents of OID as well as OID
|
---|
| 242 | vector<infodbclass> info_array;
|
---|
| 243 | text_tarray OIDs;
|
---|
| 244 | if (getParents) get_parents_array (OID, OIDs);
|
---|
| 245 | OIDs.push_back (OID);
|
---|
| 246 |
|
---|
| 247 | text_tarray::const_iterator this_OID = OIDs.begin();
|
---|
| 248 | text_tarray::const_iterator end_OID = OIDs.end();
|
---|
| 249 |
|
---|
| 250 | while (this_OID != end_OID) {
|
---|
| 251 | infodbclass info;
|
---|
| 252 | if (!db_ptr->getinfo(*this_OID, info)) return false;
|
---|
| 253 |
|
---|
| 254 | // adjust the metadata
|
---|
| 255 | text_t &contains = info["contains"];
|
---|
| 256 | if (contains.empty()) info["haschildren"] = 0;
|
---|
| 257 | else info["haschildren"] = 1;
|
---|
| 258 | //contains.clear();
|
---|
| 259 |
|
---|
| 260 | info_array.push_back(info);
|
---|
| 261 | ++this_OID;
|
---|
| 262 | }
|
---|
| 263 |
|
---|
| 264 | // if fields set is empty we want to get all available metadata
|
---|
| 265 | text_tset tfields = fields;
|
---|
| 266 | if (tfields.empty() && !info_array.empty()) {
|
---|
| 267 | infodbclass::iterator t_info = info_array[0].begin();
|
---|
| 268 | infodbclass::iterator e_info = info_array[0].end();
|
---|
| 269 | while (t_info != e_info) {
|
---|
| 270 | if ((*t_info).first != "contains")
|
---|
| 271 | tfields.insert ((*t_info).first);
|
---|
| 272 | ++t_info;
|
---|
| 273 | }
|
---|
| 274 | tfields.insert ("hasnext");
|
---|
| 275 | tfields.insert ("hasprevious");
|
---|
| 276 | }
|
---|
| 277 |
|
---|
| 278 | // collect together the metadata
|
---|
| 279 | bool donenextprevtest = false;
|
---|
| 280 | bool hasnext=false, hasprevious=false;
|
---|
| 281 | MetadataInfo_t this_metadata;
|
---|
| 282 | text_tarray *pos_metadata;
|
---|
| 283 | text_tset::const_iterator fields_here = tfields.begin();
|
---|
| 284 | text_tset::const_iterator fields_end = tfields.end();
|
---|
| 285 |
|
---|
| 286 | while (fields_here != fields_end) {
|
---|
| 287 | this_metadata.clear();
|
---|
| 288 | this_metadata.isRef = false;
|
---|
| 289 |
|
---|
| 290 | vector<infodbclass>::reverse_iterator this_info = info_array.rbegin();
|
---|
| 291 | vector<infodbclass>::reverse_iterator end_info = info_array.rend();
|
---|
| 292 | MetadataInfo_t *tmetaptr = &this_metadata;
|
---|
| 293 | while (this_info != end_info) {
|
---|
| 294 |
|
---|
| 295 | pos_metadata = (*this_info).getmultinfo(*fields_here);
|
---|
| 296 | if ((*fields_here == "hasnext" || *fields_here == "hasprevious")) {
|
---|
| 297 |
|
---|
| 298 | // collect metadata
|
---|
| 299 | if (!donenextprevtest) {
|
---|
| 300 | donenextprevtest = true;
|
---|
| 301 |
|
---|
| 302 | // cache parent contents array
|
---|
| 303 | text_t thisparent = get_parent (OID);
|
---|
| 304 | if (!thisparent.empty()) {
|
---|
| 305 | if (thisparent != parentOID) {
|
---|
| 306 | parentOID = thisparent;
|
---|
| 307 | parentcontents.erase(parentcontents.begin(), parentcontents.end());
|
---|
| 308 | if (db_ptr->getinfo(parentOID, parentinfo)) {
|
---|
| 309 | text_t &parentinfocontains = parentinfo["contains"];
|
---|
| 310 | if (!parentinfocontains.empty())
|
---|
| 311 | splitchar (parentinfocontains.begin(), parentinfocontains.end(),
|
---|
| 312 | ';', parentcontents);
|
---|
| 313 | }
|
---|
| 314 | }
|
---|
| 315 |
|
---|
| 316 | // do tests
|
---|
| 317 | text_tarray::const_iterator parentcontents_here = parentcontents.begin();
|
---|
| 318 | text_tarray::const_iterator parentcontents_end = parentcontents.end();
|
---|
| 319 | text_t shrunk_OID = OID;
|
---|
| 320 | shrink_parent (shrunk_OID);
|
---|
| 321 | while (parentcontents_here != parentcontents_end) {
|
---|
| 322 | if (*parentcontents_here == shrunk_OID) {
|
---|
| 323 | if (parentcontents_here == parentcontents.begin()) hasprevious = false;
|
---|
| 324 | else hasprevious = true;
|
---|
| 325 |
|
---|
| 326 | ++parentcontents_here;
|
---|
| 327 |
|
---|
| 328 | if (parentcontents_here == parentcontents.end()) hasnext = false;
|
---|
| 329 | else hasnext = true;
|
---|
| 330 |
|
---|
| 331 | break;
|
---|
| 332 | }
|
---|
| 333 |
|
---|
| 334 | ++parentcontents_here;
|
---|
| 335 | }
|
---|
| 336 |
|
---|
| 337 | // fill in metadata
|
---|
| 338 | if ((*fields_here == "hasnext" && hasnext) ||
|
---|
| 339 | (*fields_here == "hasprevious" && hasprevious))
|
---|
| 340 | tmetaptr->values.push_back("1");
|
---|
| 341 | else
|
---|
| 342 | tmetaptr->values.push_back("0");
|
---|
| 343 | } else
|
---|
| 344 | tmetaptr->values.push_back("0");
|
---|
| 345 | }
|
---|
| 346 | }
|
---|
| 347 | //else if (pos_metadata != NULL && *fields_here != "contains") {
|
---|
| 348 | else if (pos_metadata != NULL) {
|
---|
| 349 | tmetaptr->values = *pos_metadata;
|
---|
| 350 | }
|
---|
| 351 | else
|
---|
| 352 | tmetaptr->values.push_back("");
|
---|
| 353 |
|
---|
| 354 | ++this_info;
|
---|
| 355 | if (this_info != end_info) {
|
---|
| 356 | tmetaptr->parent = new MetadataInfo_t();
|
---|
| 357 | tmetaptr = tmetaptr->parent;
|
---|
| 358 | }
|
---|
| 359 | }
|
---|
| 360 | metadata[*fields_here] = this_metadata;
|
---|
| 361 | ++fields_here;
|
---|
| 362 | }
|
---|
| 363 |
|
---|
| 364 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
| 365 | return true;
|
---|
[226] | 366 | }
|
---|
[219] | 367 |
|
---|
[15591] | 368 | bool sourceclass::get_document (const text_t &OID, text_t &doc,
|
---|
| 369 | comerror_t &err, ostream &logout)
|
---|
| 370 | {
|
---|
| 371 | outconvertclass text_t2ascii;
|
---|
| 372 |
|
---|
[259] | 373 | err = noError;
|
---|
[15591] | 374 | if (db_ptr == NULL) {
|
---|
| 375 | // most likely a configuration problem
|
---|
| 376 | logout << text_t2ascii
|
---|
| 377 | << "configuration error: " << classname << " contains a null dbclass\n\n";
|
---|
| 378 | err = configurationError;
|
---|
| 379 | return true;
|
---|
| 380 | }
|
---|
[219] | 381 |
|
---|
[15591] | 382 | // open the database
|
---|
| 383 | db_ptr->setlogout(&logout);
|
---|
| 384 | if (!db_ptr->opendatabase (db_filename, DB_READER, 100, false)) {
|
---|
| 385 | // most likely a system problem (we have already checked that the database exists)
|
---|
| 386 | logout << text_t2ascii
|
---|
| 387 | << "system problem: open on database \"" << db_filename << "\" failed\n\n";
|
---|
| 388 | err = systemProblem;
|
---|
| 389 | return true;
|
---|
| 390 | }
|
---|
| 391 |
|
---|
| 392 | text_t tOID = OID;
|
---|
| 393 | if (needs_translating (OID))
|
---|
| 394 | translate_OID (OID, tOID, err, logout);
|
---|
| 395 | infodbclass info;
|
---|
| 396 | if (!db_ptr->getinfo(tOID, info)) {
|
---|
| 397 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
| 398 | return false;
|
---|
| 399 | }
|
---|
| 400 |
|
---|
| 401 | if (info["hastxt"].getint() == 1) {
|
---|
| 402 | int docnum = info["docnum"].getint();
|
---|
| 403 |
|
---|
| 404 | // set the collection directory
|
---|
| 405 | textsearchptr->setcollectdir (collectdir);
|
---|
| 406 |
|
---|
| 407 | // get the text
|
---|
| 408 | textsearchptr->docTargetDocument(defaultindex, defaultsubcollection,
|
---|
[21772] | 409 | defaultlanguage, collection, docnum, doc);
|
---|
| 410 |
|
---|
| 411 | // remove the <Doc></Doc> and <Sec></Sec> tags
|
---|
| 412 | doc.replace("<Doc>", "");
|
---|
| 413 | doc.replace("</Doc>", "");
|
---|
| 414 | doc.replace("<Sec>", "");
|
---|
| 415 | doc.replace("</Sec>", "");
|
---|
[15591] | 416 | }
|
---|
| 417 |
|
---|
| 418 | db_ptr->closedatabase(); // Important that local library doesn't leave any files open
|
---|
| 419 | return true;
|
---|
[259] | 420 | }
|
---|
[226] | 421 |
|
---|
[15591] | 422 | bool sourceclass::is_searchable(bool &issearchable, comerror_t &err, ostream &logout)
|
---|
| 423 | {
|
---|
| 424 | err = noError;
|
---|
| 425 | issearchable = false;
|
---|
[259] | 426 |
|
---|
[15591] | 427 | text_tarray fromarray;
|
---|
| 428 | indexmap.getfromarray(fromarray);
|
---|
| 429 | if (fromarray.size() == 0) {
|
---|
| 430 | return true;
|
---|
| 431 | } else if (fromarray.size() == 1) {
|
---|
| 432 | if (fromarray[0] == "dummy:text") {
|
---|
| 433 | // always return true - issearchable is false here though
|
---|
| 434 | return true;
|
---|
| 435 | }
|
---|
| 436 | }
|
---|
[4774] | 437 | issearchable = true;
|
---|
[15591] | 438 | return true;
|
---|
[4774] | 439 | }
|
---|
| 440 |
|
---|
[15591] | 441 |
|
---|
[504] | 442 | bool operator==(const sourceptr &x, const sourceptr &y) {
|
---|
| 443 | return (x.s == y.s);
|
---|
| 444 | }
|
---|
| 445 |
|
---|
| 446 | bool operator<(const sourceptr &x, const sourceptr &y) {
|
---|
| 447 | return (x.s < y.s);
|
---|
| 448 | }
|
---|
| 449 |
|
---|
| 450 |
|
---|
[219] | 451 | // thesource remains the property of the calling code but
|
---|
| 452 | // should not be deleted until it is removed from this list.
|
---|
| 453 | void sourcelistclass::addsource (sourceclass *thesource) {
|
---|
[226] | 454 | // can't add a source that doesn't exist
|
---|
[219] | 455 | assert (thesource != NULL);
|
---|
| 456 | if (thesource == NULL) return;
|
---|
| 457 |
|
---|
| 458 | sourceptr sp;
|
---|
| 459 | sp.s = thesource;
|
---|
| 460 |
|
---|
| 461 | sourceptrs.push_back(sp);
|
---|
| 462 | }
|
---|