tris/recoll
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Allow stem expansion for several (all) stemming languages at a time

Browse Source
This commit is contained in:
dockes
2007-08-01 10:04:53 +00:00
parent 4cffc83104
commit e883b24ffc
8 changed files with 77 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/rcldb/stemdb.cpp
+31 -6
View File
@@ -1,5 +1,5 @@
#ifndef lint
static char rcsid[] = "@(#$Id: stemdb.cpp,v 1.9 2007-05-24 09:35:02 dockes Exp $ (C) 2005 J.F.Dockes";
static char rcsid[] = "@(#$Id: stemdb.cpp,v 1.10 2007-08-01 10:04:53 dockes Exp $ (C) 2005 J.F.Dockes";
#endif
/**
@@ -218,12 +218,13 @@ static string stringlistdisp(const list<string>& sl)
}
/**
* Expand term to list of all terms which stem to the same term.
* Expand term to list of all terms which stem to the same term, for one
* expansion language
*/
bool stemExpand(const std::string& dbdir,
const std::string& lang,
const std::string& term,
list<string>& result)
static bool stemExpandOne(const std::string& dbdir,
const std::string& lang,
const std::string& term,
list<string>& result)
{
try {
Xapian::Stem stemmer(lang);
@@ -281,6 +282,30 @@ bool stemExpand(const std::string& dbdir,
return true;
}
/**
* Expand term to list of all terms which stem to the same term, add the
* expansion sets for possibly multiple expansion languages
*/
bool stemExpand(const std::string& dbdir,
const std::string& langs,
const std::string& term,
list<string>& result)
{
list<string> llangs;
stringToStrings(langs, llangs);
for (list<string>::const_iterator it = llangs.begin();
it != llangs.end(); it++) {
list<string> oneexp;
stemExpandOne(dbdir, *it, term, oneexp);
result.insert(result.end(), oneexp.begin(), oneexp.end());
}
result.sort();
result.unique();
return true;
}
}
}
src/rcldb/stemdb.h
+2 -2
View File
@@ -1,6 +1,6 @@
#ifndef _STEMDB_H_INCLUDED_
#define _STEMDB_H_INCLUDED_
/* @(#$Id: stemdb.h,v 1.3 2006-12-19 12:11:21 dockes Exp $ (C) 2004 J.F.Dockes */
/* @(#$Id: stemdb.h,v 1.4 2007-08-01 10:04:53 dockes Exp $ (C) 2004 J.F.Dockes */
/// Stem database code
///
/// Stem databases list stems and the set of index terms they expand to. They
@@ -30,7 +30,7 @@ extern bool createDb(Xapian::Database& xdb,
const std::string& dbdir, const std::string& lang);
/// Expand term to stem siblings
extern bool stemExpand(const std::string& dbdir,
const std::string& lang,
const std::string& langs,
const std::string& term,
list<string>& result);
#ifndef NO_NAMESPACES