root/main/trunk/greenstone2/runtime-src/src/colservr/collectserver.cpp @ 23492

Revision 23492, 19.8 KB (checked in by davidb, 9 years ago)

Additionoal check added before printing out warning about a filter not being available/recognized.

  • Property svn:keywords set to Author Date Id Revision
Line 
1 
2/**********************************************************************
3 *
4 * collectserver.cpp --
5 * Copyright (C) 1999  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 "collectserver.h"
28#include "OIDtools.h"
29#include <assert.h>
30#include "display.h"
31
32void check_if_valid_buildtype(const text_t& buildtype)
33{
34  if (buildtype=="mg") {
35#ifndef ENABLE_MG
36    cerr << "Warning: Greenstone installation has not been compiled to support buildtype 'mg'." << endl;
37#endif
38  }
39
40  else if (buildtype=="mgpp") {
41#ifndef ENABLE_MGPP
42    cerr << "Warning: Greenstone installation has not been compiled to support buildtype 'mgpp'." << endl;
43#endif
44  }
45
46  else if (buildtype=="lucene") {
47#ifndef ENABLE_LUCENE
48    cerr << "Warning: Greenstone installation has not been compiled to support buildtype 'lucene'." << endl;
49#endif
50  }
51
52  else {
53    cerr << "Error: buildtype '" << buildtype << "' is not a recognized indexer for Greenstone." << endl;
54  }
55
56}
57
58
59void check_if_valid_infodbtype(const text_t& infodbtype)
60{
61  if (infodbtype=="gdbm") {
62#ifndef USE_GDBM
63    cerr << "Warning: Greenstone installation has not been compiled to support infodbtype 'gdbm'." << endl;
64#endif
65  }
66  else if (infodbtype=="gdbm-txtgz") {
67#ifndef USE_GDBM
68    cerr << "Warning: Greenstone installation has not been compiled to support infodbtype 'gdbm-txtgz'." << endl;
69#endif
70  }
71  else if (infodbtype=="jdbm") {
72#ifndef USE_JDBM
73    cerr << "Warning: Greenstone installation has not been compiled to support infodbtype 'jdbm'." << endl;
74#endif
75  }
76  else if (infodbtype=="sqlite") {
77#ifndef USE_SQLITE
78    cerr << "Warning: Greenstone installation has not been compiled to support infodbtype 'sqlite'." << endl;
79#endif
80  }
81  else if (infodbtype=="mssql") {
82#ifndef USE_MSSQL
83    cerr << "Warning: Greenstone installation has not been compiled to support infodbtype 'mssql'." << endl;
84#endif
85  }
86
87  else {
88    cerr << "Error: infodbtype '" << infodbtype << "' is not a recognized database type for Greenstone." << endl;
89  }
90
91}
92
93
94
95collectserver::collectserver ()
96  : collectinfo()
97{
98  configinfo.collection = "null";
99}
100
101collectserver::~collectserver () {
102
103  // clean up the sources
104  sourcelistclass::iterator source_here = sources.begin();
105  sourcelistclass::iterator source_end = sources.end();
106  while (source_here != source_end) {
107    if ((*source_here).s != NULL)
108      delete (*source_here).s;
109    ++source_here;
110  }
111  sources.clear();
112
113  // clean up the filters
114  filtermapclass::iterator filter_here = filters.begin();
115  filtermapclass::iterator filter_end = filters.end();
116  while (filter_here != filter_end) {
117    if ((*filter_here).second.f != NULL)
118      delete (*filter_here).second.f;
119    ++filter_here;
120  }
121  filters.clear();
122}
123
124// configure should be called for each line in the
125// configuration files to configure the collection server and everything
126// it contains. The configuration should take place just before initialisation
127void collectserver::configure (const text_t &key, const text_tarray &cfgline) {
128  if (cfgline.size() >= 1) {
129    const text_t &value = cfgline[0];
130    if (key == "plugin")
131    {
132        //get the plugin name
133    const text_t &name = cfgline[0];
134   
135    if (name == "HTMLPlugin")
136    {
137        for (int hI = 1; hI < cfgline.size(); hI++)
138        {
139            const text_t &plugOption = cfgline[hI];
140           
141            if (plugOption == "-use_realistic_book")
142            {
143                collectinfo.useBook = true;
144                break;
145            }
146        }
147    }
148    }
149    else if (key == "gsdlhome") configinfo.gsdlhome = value;
150    else if (key == "gdbmhome") configinfo.dbhome = value;
151    else if (key == "collecthome") configinfo.collecthome = value;
152    else if (key == "collection") {
153      configinfo.collection = value;
154      collectinfo.shortInfo.name = value;
155    }
156    else if (key == "collectdir") configinfo.collectdir = value;
157    else if (key == "host") collectinfo.shortInfo.host = value;
158    else if (key == "port") collectinfo.shortInfo.port = value.getint();
159    else if (key == "public") {
160      if (value == "true") collectinfo.isPublic = true;
161      else collectinfo.isPublic = false;
162    } else if (key == "beta") {
163      if (value == "true") collectinfo.isBeta = true;
164      else collectinfo.isBeta = false;
165    } else if (key == "collectgroup") {
166      if (value == "true") collectinfo.isCollectGroup = true;
167      else collectinfo.isCollectGroup = false;
168    } else if ((key == "ccscols") || (key == "supercollection")) collectinfo.ccsCols = cfgline;
169    else if (key == "supercollectionoptions") {
170      text_tarray::const_iterator begin = cfgline.begin();
171      text_tarray::const_iterator end = cfgline.end();
172      while(begin != end) {
173   
174    if (*begin == "uniform_search_results_formatting") {
175      collectinfo.ccsOptions |= CCSUniformSearchResultsFormatting;
176    }
177    begin++;
178      }
179    }
180    else if (key == "builddate") collectinfo.buildDate = value.getint();
181    else if (key == "languages") collectinfo.languages = cfgline;
182    else if (key == "numdocs") collectinfo.numDocs = value.getint();
183    else if (key == "numsections") collectinfo.numSections = value.getint();
184    else if (key == "numwords") collectinfo.numWords = value.getint();
185    else if (key == "numbytes") collectinfo.numBytes = value.getint();
186    else if (key == "stemindexes") collectinfo.stemIndexes = value.getint();
187    else if (key == "collectionmeta") {
188      // genuine collmeta get added as collectionmeta and collection_macros
189      // .collmeta just get added as collection_macros
190      text_t params;
191      if (cfgline.size() == 3) {
192    // get the params for later
193    text_t::const_iterator first=cfgline[1].begin()+1;
194    text_t::const_iterator last=cfgline[1].end()-1;
195    params=substr(first, last);
196      }
197     
198      text_t meta_name = cfgline[0];
199      if (*(meta_name.begin())=='.') {
200    // a .xxx collectionmeta. strip off the . and
201    // look it up in the indexmap to get the actual value
202   
203    text_t name = substr(cfgline[0].begin()+1,cfgline[0].end());
204    text_t new_name;
205    if (indexmap.from2to(name, new_name)) {
206      meta_name = new_name;
207    }
208      } else {
209    // add them to collectionmeta
210    text_tmap lang_map = collectinfo.collectionmeta[cfgline[0]];
211    if (cfgline.size() == 2) {
212      lang_map[g_EmptyText] = cfgline[1];
213    } else if (cfgline.size() == 3 ) {
214      // get the lang out of params
215      paramhashtype params_hash;
216      splitparams(params, params_hash);
217     
218      text_t lang = params_hash["l"];
219      lang_map[lang] = cfgline[2];
220      if (lang_map[g_EmptyText].empty()) {
221        // want the first one as the default if no default specified
222        lang_map[g_EmptyText] = cfgline[2];
223      }
224    }
225    collectinfo.collectionmeta[cfgline[0]] = lang_map;
226   
227      }
228     
229      // add all collectionmeta to macro list
230      text_tmap params_map = collectinfo.collection_macros[meta_name];
231     
232      if (cfgline.size() == 2) {// no params for this macro
233    params_map[g_EmptyText] = cfgline[1];
234      }
235      else if (cfgline.size() == 3) {// has params
236    params_map[params] = cfgline[2];
237    if (params_map[g_EmptyText].empty()) {
238      params_map[g_EmptyText] = cfgline[2];
239    }
240      }
241      collectinfo.collection_macros[meta_name] = params_map;
242    }
243    else if (key == "collectionmacro") {
244      text_t nobrackets;
245      text_tmap params_map = collectinfo.collection_macros[cfgline[0]];
246      // add all to macro list
247      if (cfgline.size() == 2) { // no params for this macro
248    params_map[g_EmptyText] = cfgline[1];
249      }
250      else if (cfgline.size() == 3) {// has params
251    // strip [ ] brackets from params
252    text_t::const_iterator first=cfgline[1].begin()+1;
253    text_t::const_iterator last=cfgline[1].end()-1;
254    nobrackets=substr(first, last);
255    params_map[nobrackets] = cfgline[2];
256      }
257      collectinfo.collection_macros[cfgline[0]] = params_map;
258     
259    } else if (key == "format" && cfgline.size() == 2)
260      collectinfo.format[cfgline[0]] = cfgline[1];
261    else if (key == "building" && cfgline.size() == 2)
262      collectinfo.building[cfgline[0]] = cfgline[1];
263    else if (key == "httpdomain") collectinfo.httpdomain = value;
264    else if (key == "httpprefix") collectinfo.httpprefix = value;
265    else if (key == "receptionist") collectinfo.receptionist = value;
266    else if (key == "buildtype") {
267      check_if_valid_buildtype(value); // prints warning if value (indexer) is invalid
268      collectinfo.buildType = value;
269    }
270    // backwards compatibility - searchytpes is now a format statement
271    else if (key == "searchtype") { // means buildtype is mgpp
272      if (collectinfo.buildType.empty()) {
273    check_if_valid_buildtype("mgpp"); // prints warning if value (indexer) is invalid
274    collectinfo.buildType = "mgpp";
275      }
276      joinchar(cfgline, ',', collectinfo.format["SearchTypes"]);
277      //collectinfo.searchTypes = cfgline;
278    }
279    else if (key == "infodbtype") {
280      check_if_valid_infodbtype(value); // prints warning if value (database type) is invalid
281      collectinfo.infodbType = value;
282    }
283    else if (key == "separate_cjk") {
284      if (value == "true") collectinfo.isSegmented = true;
285      else collectinfo.isSegmented = false;
286    }
287    // What have we set in our collect.cfg file :  document or collection ?
288    else if (key == "authenticate") collectinfo.authenticate = value;
289
290    // What have we set for our group list
291    else if ((key == "auth_group") || (key == "auth_groups")) joinchar(cfgline,',',collectinfo.auth_group);
292
293    // store all the mappings for use when collection meta is read later
294    // (build.cfg read before collect.cfg)
295    else if (key == "indexmap" || key == "indexfieldmap" || key == "subcollectionmap" || key == "languagemap" || key == "levelmap") {
296      indexmap.importmap (cfgline, true);
297     
298    }
299    // In the map the key-value pair contain the same
300    // data i.e key == data, if key is 2 then data is 2
301   
302    // What have we set for our public_documents ACL
303    else if (key == "public_documents")
304       {
305      text_tarray::const_iterator begin = cfgline.begin();
306      text_tarray::const_iterator end = cfgline.end();
307      while(begin != end)
308         {
309        // key = data i.e if key is 2 then data is 2
310        // collectinfo.public_documents[*begin] is the key
311        // *begin is the data value
312
313        collectinfo.public_documents[*begin] = *begin;
314        ++begin;
315         }
316       }
317   
318    // What have we set for our private_documents ACL
319    else if (key == "private_documents")
320       {
321      text_tarray::const_iterator begin = cfgline.begin();
322      text_tarray::const_iterator end = cfgline.end();
323      while(begin != end)
324         {
325        // key = data i.e if key is 2 then data is 2
326        // collectinfo.public_documents[*begin] is the key
327        // *begin is the data value
328       
329        collectinfo.private_documents[*begin] = *begin;
330        ++begin;
331         }
332       }
333
334    // dynamic_classifier <UniqueID> "<Options>"
335    else if (key == "dynamic_classifier")
336    {
337      collectinfo.dynamic_classifiers[cfgline[0]] = cfgline[1];
338    }
339  }
340 
341  // configure the filters
342  filtermapclass::iterator filter_here = filters.begin();
343  filtermapclass::iterator filter_end = filters.end();
344  while (filter_here != filter_end) {
345    assert ((*filter_here).second.f != NULL);
346    if ((*filter_here).second.f != NULL)
347      (*filter_here).second.f->configure(key, cfgline);
348
349    ++filter_here;
350  }
351
352  // configure the sources
353  sourcelistclass::iterator source_here = sources.begin();
354  sourcelistclass::iterator source_end = sources.end();
355  while (source_here != source_end) {
356    assert ((*source_here).s != NULL);
357    if ((*source_here).s != NULL)
358      (*source_here).s->configure(key, cfgline);
359   
360    ++source_here;
361  }
362}
363
364
365void collectserver::configure (const text_t &key, const text_t &value) {
366  text_tarray cfgline;
367  cfgline.push_back (value);
368  configure(key, cfgline);
369}
370
371void collectserver::ping (bool &wasSuccess, comerror_t &error, ostream &logout) {
372  // if we've not been properly configured, then it is a foregone
373  // conclusion that we cannot be active
374  if (this->configinfo.collection == "null")
375    {
376      wasSuccess = false;
377    }
378  // if no build date exists, then the collection was probably not built;
379  // ditto if the number of documents is zero, then something is pretty
380  // wrong
381  else if (this->collectinfo.buildDate == 0 ||
382      this->collectinfo.numDocs == 0)
383    {
384      wasSuccess =  false;
385    }
386  // it is probably okay
387  else
388    wasSuccess = true;
389}
390
391
392bool collectserver::init (ostream &logout) {
393  // delete the indexmap
394  indexmap.clear();
395
396  // init the filters
397  filtermapclass::iterator filter_here = filters.begin();
398  filtermapclass::iterator filter_end = filters.end();
399  while (filter_here != filter_end) {
400    assert ((*filter_here).second.f != NULL);
401    if (((*filter_here).second.f != NULL) &&
402    !(*filter_here).second.f->init(logout)) return false;
403   
404    ++filter_here;
405  }
406
407  // init the sources
408  sourcelistclass::iterator source_here = sources.begin();
409  sourcelistclass::iterator source_end = sources.end();
410  while (source_here != source_end) {
411    assert ((*source_here).s != NULL);
412    if (((*source_here).s != NULL) &&
413    !(*source_here).s->init(logout)) return false;
414   
415    ++source_here;
416  }
417
418  return true;
419}
420
421
422void collectserver::get_collectinfo (ColInfoResponse_t &reponse,
423              comerror_t &err, ostream &/*logout*/) {
424  reponse = collectinfo;
425  err = noError;
426}
427
428void collectserver::get_filterinfo (InfoFiltersResponse_t &response,
429                    comerror_t &err, ostream &/*logout*/) {
430  response.clear ();
431
432  // get a list of filter names
433  filtermapclass::iterator filter_here = filters.begin();
434  filtermapclass::iterator filter_end = filters.end();
435  while (filter_here != filter_end) {
436    response.filterNames.insert ((*filter_here).first);
437    ++filter_here;
438  }
439
440  err = noError;
441}
442
443void collectserver::get_filteroptions (const InfoFilterOptionsRequest_t &request,
444                       InfoFilterOptionsResponse_t &response,
445                       comerror_t &err, ostream &logout) {
446  outconvertclass text_t2ascii;
447
448  filterclass *thisfilter = filters.getfilter(request.filterName);
449  if (thisfilter != NULL) {
450    thisfilter->get_filteroptions (response, err, logout);
451  } else {
452    response.clear ();
453    err = protocolError;
454    text_t& infodbtype = collectinfo.infodbType;
455
456    // Don't print out the warning if were's asking about SQLQueryFilter
457    // when we know the infodbtype is something other than .*sql.*
458
459    if ((request.filterName != "SQLQueryFilter")
460    || (findword(infodbtype.begin(),infodbtype.end(),"sql") != infodbtype.end())) {
461      logout << text_t2ascii << "Protocol Error: filter options requested for non-existent\n"
462         << "filter \"" << request.filterName << "\".\n\n";
463    }
464  }
465}
466
467void collectserver::filter (FilterRequest_t &request,
468                FilterResponse_t &response,
469                comerror_t &err, ostream &logout) {
470  outconvertclass text_t2ascii;
471
472  // translate any ".fc", ".pr" etc. stuff in the docSet
473  text_t translatedOID;
474  text_tarray translatedOIDs;
475  text_tarray::iterator doc_here = request.docSet.begin();
476  text_tarray::iterator doc_end = request.docSet.end();
477  while (doc_here != doc_end) {
478    if (needs_translating (*doc_here)) {
479      sourcelistclass::iterator source_here = sources.begin();
480      sourcelistclass::iterator source_end = sources.end();
481      while (source_here != source_end) {
482    assert ((*source_here).s != NULL);
483    if (((*source_here).s != NULL) &&
484        ((*source_here).s->translate_OID (*doc_here, translatedOID, err, logout))) {
485      if (err != noError) return;
486      break;
487    }
488    ++source_here;
489      }
490      translatedOIDs.push_back (translatedOID);
491    } else {
492      translatedOIDs.push_back (*doc_here);
493    }
494    ++doc_here;
495  }
496  request.docSet = translatedOIDs;
497
498  response.clear();
499
500  filterclass *thisfilter = filters.getfilter(request.filterName);
501  if (thisfilter != NULL) {
502    // filter the data
503    thisfilter->filter (request, response, err, logout);
504    if (err != noError) return;
505    // fill in the metadata for each of the OIDs (if it is requested)
506    if (request.filterResultOptions & FRmetadata) {
507      bool processed = false;
508      ResultDocInfo_tarray::iterator resultdoc_here = response.docInfo.begin();
509      ResultDocInfo_tarray::iterator resultdoc_end = response.docInfo.end();
510      while (resultdoc_here != resultdoc_end) {
511    // try each of the sources in turn
512    sourcelistclass::iterator source_here = sources.begin();
513    sourcelistclass::iterator source_end = sources.end();
514    while (source_here != source_end) {
515      assert ((*source_here).s != NULL);
516      if (((*source_here).s != NULL) &&
517          ((*source_here).s->get_metadata(request.requestParams, request.refParams,
518                          request.getParents, request.fields,
519                          (*resultdoc_here).OID, (*resultdoc_here).metadata,
520                          err, logout))) {
521        if (err != noError) return;
522        processed = true;
523        break;
524      }
525      ++source_here;
526    }
527    if (!processed) {
528
529      logout << text_t2ascii << "Protocol Error: nothing processed for "
530       << "filter \"" << request.filterName << "\".\n\n";
531
532      err = protocolError;
533      return;
534    }
535    ++resultdoc_here;
536      }
537    } 
538 
539    err = noError;
540  }
541  else
542  {
543    response.clear ();
544    err = protocolError;
545    logout << text_t2ascii << "Protocol Error: filter options requested for non-existent\n"
546       << "filter \"" << request.filterName << "\".\n\n";
547  }
548}
549
550void collectserver::get_document (const DocumentRequest_t &request,
551                  DocumentResponse_t &response,
552                  comerror_t &err, ostream &logout) {
553
554  sourcelistclass::iterator source_here = sources.begin();
555  sourcelistclass::iterator source_end = sources.end();
556  while (source_here != source_end) {
557    assert ((*source_here).s != NULL);
558    if (((*source_here).s != NULL) &&
559    ((*source_here).s->get_document (request.OID, response.doc, err, logout))) {
560      if (err != noError) return;
561      break;
562    }
563    ++source_here;
564  }
565}
566
567void collectserver::is_searchable (bool &issearchable, comerror_t &err,
568                   ostream &logout) {
569
570  sourcelistclass::iterator source_here = sources.begin();
571  sourcelistclass::iterator source_end = sources.end();
572  while (source_here != source_end) {
573    assert ((*source_here).s != NULL);
574    if (((*source_here).s != NULL) &&
575    ((*source_here).s->is_searchable (issearchable, err, logout))) {
576      if (err != noError) return;
577      break;
578    }
579    ++source_here;
580  }
581}
582
583
584bool operator==(const collectserverptr &x, const collectserverptr &y) {
585  return (x.c == y.c);
586}
587
588bool operator<(const collectserverptr &x, const collectserverptr &y) {
589  return (x.c < y.c);
590}
591
592
593// thecollectserver remains the property of the calling code but
594// should not be deleted until it is removed from this list.
595void collectservermapclass::addcollectserver (collectserver *thecollectserver) {
596  // can't add a null collection server
597  assert (thecollectserver != NULL);
598  if (thecollectserver == NULL) return;
599 
600  // can't add an collection server with no collection name
601  assert (!(thecollectserver->get_collection_name()).empty());
602  if ((thecollectserver->get_collection_name()).empty()) return;
603
604  collectserverptr cptr;
605  cptr.c = thecollectserver;
606  collectserverptrs[thecollectserver->get_collection_name()] = cptr;
607}
608
609// getcollectserver will return NULL if the collectserver could not be found
610collectserver *collectservermapclass::getcollectserver (const text_t &collection) {
611  // can't find a collection with no name
612  if (collection.empty()) return NULL;
613
614  iterator here = collectserverptrs.find (collection);
615  if (here == collectserverptrs.end()) return NULL;
616 
617  return (*here).second.c;
618}
Note: See TracBrowser for help on using the browser.