/* Copyright (C) 2005 J.F.Dockes
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the
* Free Software Foundation, Inc.,
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#include "autoconfig.h"
#include "debuglog.h"
#include "utf8iter.h"
#include "smallut.h"
#include "refcntr.h"
#include "textsplit.h"
#include "xmacros.h"
#include "rcldb.h"
#include "stemdb.h"
#include "expansiondbs.h"
using namespace std;
namespace Rcl {
/**
* Create all expansion dbs used to transform user input term to widen a query
* We use Xapian synonyms subsets to store the expansions.
*/
bool createExpansionDbs(Xapian::WritableDatabase& wdb,
const vector<string>& langs)
{
LOGDEB(("StemDb::createExpansionDbs: languages: %s\n",
stringsToString(langs).c_str()));
Chrono cron;
// Erase and recreate all the expansion groups
// If langs is empty and we don't need casediac expansion, then no need to
// walk the big list
if (langs.empty()) {
#ifndef RCL_INDEX_STRIPCHARS
if (o_index_stripchars)
#endif
return true;
}
// Stem dbs
vector<XapWritableComputableSynFamMember> stemdbs;
// Note: tried to make this to work with stack-allocated objects, couldn't.
// Looks like a bug in copy constructors somewhere, can't guess where
vector<RefCntr<SynTermTransStem> > stemmers;
for (unsigned int i = 0; i < langs.size(); i++) {
stemmers.push_back(RefCntr<SynTermTransStem>
(new SynTermTransStem(langs[i])));
stemdbs.push_back(
XapWritableComputableSynFamMember(wdb, synFamStem, langs[i],
stemmers.back().getptr()));
stemdbs.back().recreate();
}
#ifndef RCL_INDEX_STRIPCHARS
// Unaccented stem dbs
vector<XapWritableComputableSynFamMember> unacstemdbs;
// We can reuse the same stemmer pointers, the objects are stateless.
if (!o_index_stripchars) {
for (unsigned int i = 0; i < langs.size(); i++) {
unacstemdbs.push_back(
XapWritableComputableSynFamMember(wdb, synFamStemUnac, langs[i],
stemmers.back().getptr()));
unacstemdbs.back().recreate();
}
}
SynTermTransUnac transunac(UNACOP_UNACFOLD);
XapWritableComputableSynFamMember
diacasedb(wdb, synFamDiCa, "all", &transunac);
if (!o_index_stripchars)
diacasedb.recreate();
#endif
// Walk the list of all terms, and stem/unac each.
string ermsg;
try {
Xapian::TermIterator it = wdb.allterms_begin();
// We'd want to skip to the first non-prefixed term, but this is a bit
// complicated, so we just jump over most of the prefixed term and then
// skip the rest one by one.
it.skip_to(wrap_prefix("Z"));
for ( ;it != wdb.allterms_end(); it++) {
if (has_prefix(*it))
continue;
// Detect and skip CJK terms.
Utf8Iter utfit(*it);
if (TextSplit::isCJK(*utfit)) {
// LOGDEB(("stemskipped: Skipping CJK\n"));
continue;
}
string lower = *it;
#ifndef RCL_INDEX_STRIPCHARS
// If the index is raw, compute the case-folded term which
// is the input to the stem db, and add a synonym from the
// stripped term to the cased and accented one, for accent
// and case expansion at query time
if (!o_index_stripchars) {
unacmaybefold(*it, lower, "UTF-8", UNACOP_FOLD);
diacasedb.addSynonym(*it);
}
#endif
// Dont' apply stemming to terms which don't look like
// natural language words.
if (!Db::isSpellingCandidate(*it)) {
LOGDEB1(("createExpansionDbs: skipped: [%s]\n", (*it).c_str()));
continue;
}
// Create stemming synonym for every language. The input is the
// lowercase accented term
for (unsigned int i = 0; i < langs.size(); i++) {
stemdbs[i].addSynonym(lower);
}
#ifndef RCL_INDEX_STRIPCHARS
// For a raw index, also maybe create a stem expansion for
// the unaccented term. While this may be incorrect, it is
// also necessary for searching in a diacritic-unsensitive
// way on a raw index
if (!o_index_stripchars) {
string unac;
unacmaybefold(lower, unac, "UTF-8", UNACOP_UNAC);
if (unac != lower) {
for (unsigned int i = 0; i < langs.size(); i++) {
unacstemdbs[i].addSynonym(unac);
}
}
}
#endif
}
} XCATCHERROR(ermsg);
if (!ermsg.empty()) {
LOGERR(("Db::createStemDb: map build failed: %s\n", ermsg.c_str()));
return false;
}
LOGDEB(("StemDb::createExpansionDbs: done: %.2f S\n", cron.secs()));
return true;
}
}