void LA_Recognizer::LoadTxt_Misspelling( lem::Iridium::Macro_Parser &txtfile, Dictionary &dict )
{
txtfile.read_it(B_LANGUAGE);
// дальше идет наименование языка, в рамках которого действует правило.
txtfile.read_it( B_EQUAL );
lem::Iridium::BethToken t1 = txtfile.read();
int id_language = dict.GetSynGram().Find_Language(t1.string());
if( id_language==UNKNOWN )
{
lem::Iridium::Print_Error(t1,txtfile);
dict.GetIO().merr().printf( "Unknown language name %us\n", t1.c_str() );
throw lem::E_BaseException();
}
txtfile.read_it( B_OFIGPAREN );
txtfile.read_it( B_IF );
Solarix::Lexem old_word = txtfile.read().string();
old_word.strip(L'"');
dict.GetLexAuto().TranslateLexem( old_word, true, id_language );
txtfile.read_it( B_THEN );
Solarix::Lexem new_word = txtfile.read().string();
new_word.strip(L'"');
dict.GetLexAuto().TranslateLexem( new_word, true, id_language );
txtfile.read_it( B_CFIGPAREN );
storage->AddMisspelling( id_language, old_word, new_word );
return;
}
void TreeScorerResult::LoadTxt(
Dictionary & dict,
lem::Iridium::Macro_Parser & txtfile,
const TreeScorerMarkers & markers
)
{
// ќценка может быть отрицательной.
if( txtfile.probe( B_SUB ) )
{
type=NumberScoreType;
score = -txtfile.read_int();
}
else
{
if( lem::is_int( txtfile.pick().string() ) )
{
type=NumberScoreType;
score = txtfile.read_int();
}
else
{
const lem::Iridium::BethToken & t = txtfile.read();
id_fact = dict.GetLexAuto().GetKnowledgeBase().FindFacts( t.string() );
if( id_fact==UNKNOWN )
{
type = FuncScoreType;
Solarix::TrKnownVars vars;
Solarix::TrTreeType tree_type;
for( lem::Container::size_type k=0; k<markers.CountMarkers(); ++k )
{
vars.RegisterVar( tree_type, markers.GetMarker(k) );
args.push_back( markers.GetMarker(k) );
}
txtfile.seekp(t);
score_fun = dict.GetLexAuto().GetFunctions().Get().CompileCall( dict.GetLexAuto(), txtfile, vars );
}
else
{
LoadBoundVars( dict, txtfile, markers );
type=NGramScoreType;
}
}
}
return;
}
void SynPatterns::LoadTxt( Dictionary &dict, lem::Iridium::Macro_Parser & txtfile )
{
lem::Iridium::BSourceState beg = txtfile.tellp();
SynPatternOptions *x = new SynPatternOptions();
x->LoadTxt( dict, txtfile );
if( IsPatternName(x->GetName()) )
{
dict.GetIO().merr().printf( "Patterns group [%us] is already declared\n", x->GetName().c_str() );
lem::Iridium::Print_Error(beg,txtfile);
throw lem::E_BaseException();
}
if( dict.GetLexAuto().GetWordEntrySet().IsSetName(x->GetName()) )
{
dict.GetIO().merr().printf( "%vfC%us%vn is a name of word entry set, word set or collocation set\n", x->GetName().c_str() );
lem::Iridium::Print_Error(beg,txtfile);
throw lem::E_BaseException();
}
const int id = GetNextTreeID();
options.push_back(x);
lem::UCString uname( lem::to_upper(x->GetName() ) );
patterns.insert( std::make_pair( uname, x ) );
name2id.insert( std::make_pair( uname, id ) );
return;
}
int TreeScorerResult::Calculate( Dictionary & dict, const TreeScorerBoundVariables & bound_variables ) const
{
if( type==0 )
return score;
else
{
lem::MCollect<const Solarix::Word_Form*> vars;
for( lem::Container::size_type i=0; i<args.size(); ++i )
{
const Solarix::Word_Form * wf = bound_variables.GetVariable( args[i] );
if( wf==NULL )
{
lem::MemFormatter mem;
mem.printf( "tree_scorer: can not find bound variable %us", args[i].c_str() );
throw lem::E_BaseException( mem.string() );
}
vars.push_back(wf);
}
KB_CheckingResult res = dict.GetLexAuto().GetKnowledgeBase().Prove( id_fact, vars );
if( res.IsMatched() )
return res.GetInt();
else
return 0;
}
}
void TreeScorerResult::LoadTxt( Dictionary & dict, lem::Iridium::Macro_Parser & txtfile, const TreeScorerMarkers & markers )
{
// ќценка может быть отрицательной.
if( txtfile.probe( B_SUB ) )
{
type=0;
score = -txtfile.read_int();
}
else
{
if( lem::is_int( txtfile.pick().string() ) )
{
type=0;
score = txtfile.read_int();
}
else
{
const lem::Iridium::BethToken & t = txtfile.read();
id_fact = dict.GetLexAuto().GetKnowledgeBase().FindFacts( t.string() );
if( id_fact==UNKNOWN )
{
// todo - тут могут быть другие варианты вызываемых вычислений.
lem::Iridium::Print_Error(t,txtfile);
dict.GetIO().merr().printf( "Unknown scoring expression starts with %us\n", t.string().c_str() );
throw lem::E_BaseException();
}
txtfile.read_it( B_OROUNDPAREN );
while( !txtfile.eof() )
{
if( txtfile.probe( B_CROUNDPAREN ) )
break;
if( !args.empty() )
txtfile.read_it( B_COMMA );
const lem::Iridium::BethToken & var = txtfile.read();
lem::UCString upper_var = lem::to_upper(var.string());
if( !markers.IsAlreadyBound(upper_var) )
{
lem::Iridium::Print_Error(var,txtfile);
dict.GetIO().merr().printf( "variable %us not bound\n", var.string().c_str() );
throw lem::E_BaseException();
}
args.push_back( upper_var );
}
type=1;
}
}
return;
}
int TreeScorerResult::Calculate(
Dictionary & dict,
const TreeScorerBoundVariables & bound_variables,
const ElapsedTimeConstraint & constraints,
TrTrace *trace_log
) const
{
if( type==NumberScoreType )
return score;
else if( type==NGramScoreType )
{
lem::MCollect<const Solarix::Word_Form*> vars;
for( lem::Container::size_type i=0; i<args.size(); ++i )
{
const Solarix::Word_Form * wf = bound_variables.GetVariable( args[i] );
if( wf==NULL )
{
lem::MemFormatter mem;
mem.printf( "tree_scorer: can not find bound variable %us", args[i].c_str() );
throw lem::E_BaseException( mem.string() );
}
vars.push_back(wf);
}
KB_CheckingResult res = dict.GetLexAuto().GetKnowledgeBase().Prove( id_fact, vars );
if( res.IsMatched() )
return res.GetInt();
else
return 0;
}
else if( type==FuncScoreType )
{
TrFunContext ctx0( (TrFunContext*)NULL );
TrContextInvokation ctx2( &ctx0 );
for( lem::Container::size_type i=0; i<args.size(); ++i )
{
const Solarix::Word_Form * wf = bound_variables.GetVariable( args[i] );
if( wf==NULL )
{
lem::MemFormatter mem;
mem.printf( "tree_scorer: can not find bound variable %us", args[i].c_str() );
throw lem::E_BaseException( mem.string() );
}
lem::Ptr<TrValue> this_wordform( new TrValue( new Tree_Node(*wf), true ) );
ctx2.AddVar( args[i], this_wordform );
}
lem::Ptr<TrValue> fun_res = score_fun->Run( constraints, dict.GetLexAuto(), ctx2, trace_log );
if( fun_res->GetType().IsInt() )
{
int res_int = fun_res->GetInt();
return res_int;
}
else
{
#if LEM_DEBUGGING==1
if( trace_log!=NULL )
trace_log->PrintStack(*lem::mout);
#endif
lem::UFString msg( lem::format_str( L"Score function must return int, not %s", fun_res->GetType().GetName().c_str() ) );
throw E_BaseException(msg.c_str());
}
}
else
{
LEM_STOPIT;
}
return 0;
}