int CSphConfigSection::GetSize ( const char * sKey, int iDefault ) const
{
CSphVariant * pEntry = (*this)( sKey );
if ( !pEntry )
return iDefault;
char sMemLimit[256];
strncpy ( sMemLimit, pEntry->cstr(), sizeof(sMemLimit) );
sMemLimit [ sizeof(sMemLimit)-1 ] = '\0';
int iLen = strlen ( sMemLimit );
if ( !iLen )
return iDefault;
iLen--;
int iScale = 1;
if ( toupper(sMemLimit[iLen])=='K' )
{
iScale = 1024;
sMemLimit[iLen] = '\0';
} else if ( toupper(sMemLimit[iLen])=='M' )
{
iScale = 1048576;
sMemLimit[iLen] = '\0';
}
char * sErr;
int iRes = strtol ( sMemLimit, &sErr, 10 );
if ( !*sErr )
return iScale*iRes;
// FIXME! report syntax error here
return iDefault;
}
ISphTokenizer * sphConfTokenizer ( const CSphConfigSection & hIndex, CSphString & sError )
{
// charset_type
CSphScopedPtr<ISphTokenizer> pTokenizer ( NULL );
if ( !hIndex("charset_type") || hIndex["charset_type"]=="sbcs" )
{
pTokenizer = sphCreateSBCSTokenizer ();
} else if ( hIndex["charset_type"]=="utf-8" )
{
pTokenizer = hIndex("ngram_chars")
? sphCreateUTF8NgramTokenizer ()
: sphCreateUTF8Tokenizer ();
} else
{
sError.SetSprintf ( "unknown charset type '%s'", hIndex["charset_type"].cstr() );
return NULL;
}
assert ( pTokenizer.Ptr() );
// charset_table
if ( hIndex("charset_table") )
if ( !pTokenizer->SetCaseFolding ( hIndex["charset_table"].cstr(), sError ) )
{
sError.SetSprintf ( "'charset_table': %s", sError.cstr() );
return NULL;
}
// min_word_len
int iMinWordLen = hIndex("min_word_len") ? Max ( hIndex["min_word_len"].intval(), 0 ) : 0;
if ( iMinWordLen )
pTokenizer->SetMinWordLen ( iMinWordLen );
// ngram_chars
if ( hIndex("ngram_chars") )
if ( !pTokenizer->SetNgramChars ( hIndex["ngram_chars"].cstr(), sError ) )
{
sError.SetSprintf ( "'ngram_chars': %s", sError.cstr() );
return NULL;
}
// ngram_len
int iNgramLen = hIndex("ngram_len") ? Max ( hIndex["ngram_len"].intval(), 0 ) : 0;
if ( iNgramLen )
pTokenizer->SetNgramLen ( iNgramLen );
// synonyms
CSphVariant * pExceptions = hIndex("exceptions"); // new option name
if ( !pExceptions )
pExceptions = hIndex("synonyms"); // deprecated option name
if ( pExceptions )
if ( !pTokenizer->LoadSynonyms ( pExceptions->cstr(), sError ) )
{
sError.SetSprintf ( "'exceptions': %s", sError.cstr() );
return NULL;
}
// phrase boundaries
int iBoundaryStep = hIndex("phrase_boundary_step") ? Max ( hIndex["phrase_boundary_step"].intval(), 0 ) : 0;
if ( iBoundaryStep>0 && hIndex("phrase_boundary") )
if ( !pTokenizer->SetBoundary ( hIndex["phrase_boundary"].cstr(), sError ) )
{
sError.SetSprintf ( "'phrase_boundary': %s", sError.cstr() );
return NULL;
}
// ignore_chars
if ( hIndex("ignore_chars") )
if ( !pTokenizer->SetIgnoreChars ( hIndex["ignore_chars"].cstr(), sError ) )
{
sError.SetSprintf ( "'ignore_chars': %s", sError.cstr() );
return NULL;
}
return pTokenizer.LeakPtr();
}
