void WrittenTextAnalysisSession::AnalyzeIt(const lem::UFString & str, bool ApplyPatterns, bool DoSyntacticLinks, const ElapsedTimeConstraint & constraints)
{
processed_str = str;
if (params.LanguageUnknown())
{
// Язык предложения не задан явно, определим его на основе статистических критериев.
const int id_language = dict->GetLexAuto().GuessLanguage(str);
if (id_language == UNKNOWN)
{
lem::MemFormatter msg;
msg.printf("Can not guess the language of the phrase [%us]", str.c_str());
throw E_BaseException(msg.string());
}
params.SetLanguageID(id_language);
#if defined SOL_DEBUGGING
if (trace != nullptr)
{
trace->LanguageGuessed(str, id_language);
}
#endif
}
delete lexer;
lexer = new WrittenTextLexer(str, params, dict, trace);
Analyze(ApplyPatterns, DoSyntacticLinks, constraints);
return;
}
void TreeScorerResult::DeserializeExpression( int expr_type, const lem::UFString & serialized )
{
type=expr_type;
if( type==NumberScoreType )
{
score = lem::to_int( serialized);
}
else if( type==NGramScoreType )
{
lem::StrParser<lem::UFString> txt(serialized);
id_fact = lem::to_int( txt.read() );
txt.read_it( L"(" );
while( !txt.eof() )
{
if( txt.probe( L")" ) )
break;
if( !args.empty() )
txt.read_it(L",");
args.push_back( lem::UCString(txt.read().c_str()) );
}
}
else if( type==FuncScoreType )
{
lem::MemReadHexStream bin2( serialized.c_str() );
args.LoadBin(bin2);
score_fun = TrFunCall::load_bin(bin2);
}
else
{
LEM_STOPIT;
}
return;
}
SG_DeclensionForm::SG_DeclensionForm(
const lem::UFString & lexem_generator,
const lem::UFString & rx_condition,
const lem::UFString & rx_flexer_flags,
const CP_Array & dims
)
: dim(dims), form(lexem_generator.c_str())
{
valid_condition = !rx_condition.empty();
valid_flexer_flags = !rx_flexer_flags.empty();
condition_str = rx_condition;
flexer_flags_str = rx_flexer_flags;
if( !condition_str.empty() )
condition = boost::wregex( condition_str.c_str(), boost::basic_regex<wchar_t>::icase );
if( !flexer_flags.empty() )
flexer_flags = boost::wregex( flexer_flags_str.c_str(), boost::basic_regex<wchar_t>::icase );
return;
}
bool SyntaxShell::TryCommand( const lem::UFString &_str )
{
LEM_CHECKIT_Z( !_str.empty() );
if( _str==L"#help" || _str==L"?" )
{
ShowHelp();
return true;
}
if( _str.front()!=L'#' )
return false;
if( _str.eq_beg( L"# " ) )
return true; // комментарий
if( _str.eq_beg( L"#timeout" ) )
{
lem::MCollect<UCString> toks;
lem::parse( _str, toks, false );
MaxTimeout = lem::to_int( toks[1] );
return true;
}
if( _str.eq_beg( L"#maxalt" ) )
{
lem::MCollect<UCString> toks;
lem::parse( _str, toks, false );
MaxAlt = lem::to_int( toks[1] );
lem::mout->printf( "MaxAlt=%d\n", MaxAlt );
return true;
}
if( _str.eq_beg( L"#maxskiptoken" ) )
{
lem::MCollect<UCString> toks;
lem::parse( _str, toks, false );
MaxSkipToken = lem::to_int( toks[1] );
lem::mout->printf( "MaxSkipToken=%d\n", MaxSkipToken );
if( MaxSkipToken>0 )
CompleteAnalysisOnly = false;
if( MaxAlt==0 || MaxAlt==lem::int_max )
{
lem::mout->printf( "Attention: it is highly recommended to use %vfE#maxalt%vn NNN in order to limit the search tree depth\n" );
}
return true;
}
if( _str.eq_beg( L"#sem" ) )
{
lem::MCollect<UCString> toks;
lem::parse( _str, toks, false );
FindFacts = lem::to_bool( toks[1] );
return true;
}
if( _str.eqi( L"#info" ) )
{
ShowDictionaryInfo();
return true;
}
if( _str.eqi( L"#disconnect" ) )
{
sol_id.Delete();
lem::mout->printf( "Dictionary database is disconnected.\n" );
return true;
}
if( _str.eqi( L"#connect" ) )
{
LoadDictionary();
return true;
}
if( _str.eq_begi( L"#tag" ) )
{
if( _str==L"#tag-" )
{
// Сбрасываем установленный фильтр
tags_ptr.Delete();
tags.clear();
return true;
}
lem::Collect<lem::UFString> toks;
lem::parse( UFString(_str.c_str()+4), toks, L"=" );
UCString tag_name, tag_value;
if( toks.size()>0 )
tag_name = toks[0].c_str();
if( toks.size()>1 )
tag_value = toks[1].c_str();
tag_name.trim();
tag_value.trim();
const int itag = sol_id->GetSynGram().Get_Net().FindTag(tag_name);
if( itag==UNKNOWN )
{
lem::mout->printf( "Tag [%vfE%us%vn] not found\n", tag_name.c_str() );
return true;
}
const ThesaurusTag &tt = sol_id->GetSynGram().Get_Net().GetTagDefs()[itag];
if( tt.CountValues()>0 )
{
int ivalue = tt[tag_value];
if( ivalue==UNKNOWN )
{
lem::mout->printf( "Tag value [%vfE%us%vn] not found\n", tag_value.c_str() );
return true;
}
}
tags_ptr = new TF_TagOrNullFilter( *sol_id, tag_name, tag_value );
return true;
}
if( _str.eq_begi( L"#param" ) )
{
if( _str==L"#param-" )
{
// Очищаем список параметров.
params.clear();
return true;
}
lem::Collect<lem::UFString> toks;
lem::parse( UFString(_str.c_str()+7), toks, L"=" );
UCString param_name, param_value;
if( toks.size()>0 )
param_name = toks[0].c_str();
if( toks.size()>1 )
param_value = toks[1].c_str();
param_name.trim();
param_value.trim();
params.push_back( std::make_pair( param_name, param_value ) );
return true;
}
lem::UFString str = lem::right( _str, _str.length()-1 );
lem::zbool ret;
if( str==L"debug" )
{
SetDebug(true);
ret=true;
}
else if( str==L"nodebug" )
{
SetDebug(false);
ret=true;
}
else if( str==L"traceon" )
{
SetDebug(true);
traceon=true;
debugger->Trace(true);
ret=true;
}
else if( str==L"traceoff" )
{
traceon=false;
if( debugger.NotNull() )
debugger->Trace(true);
ret=true;
}
else if( str==L"fuzzyon" )
{
allow_fuzzy = true;
mout->printf( "Fuzzy projection is now %vfAON%vn\n" );
ret=true;
}
else if( str==L"fuzzyoff" )
{
allow_fuzzy = false;
mout->printf( "Fuzzy projection is now %vfDOFF%vn\n" );
ret=true;
}
else if( str=="disable_filters" )
{
EnableFilters=false;
ret = true;
}
else if( str=="enable_filters" )
{
EnableFilters=true;
ret = true;
}
else if( str=="schedule1" )
{
CompleteAnalysisOnly=true;
UseTopDownThenSparse=true;
mout->printf( "Workflow=%vfATOP-DOWN, TOP-DOWN INCOMPLETE%vn\n" );
ret=true;
}
else if( str==L"topdown" )
{
UseTopDownThenSparse=false;
CompleteAnalysisOnly=true;
mout->printf( "%vfAtop-down%vn analyzer is activated\n" );
ret=true;
}
else if( str==L"allow_incomplete" )
{
CompleteAnalysisOnly = false;
mout->printf( "Incomplete analysis is %vfAALLOWED%vn\n" );
ret=true;
}
else if( str==L"disallow_incomplete" )
{
CompleteAnalysisOnly = true;
mout->printf( "Incomplete analysis is %vfDDISALLOWED%vn\n" );
ret=true;
}
else if( str==L"allow_reco" )
{
UseReconstructor = true;
mout->printf( "Token reconstructor is %vfAALLOWED%vn\n" );
ret=true;
}
else if( str==L"disallow_reco" )
{
UseReconstructor = false;
mout->printf( "Token reconstructor is %vfDDISALLOWED%vn\n" );
ret=true;
}
else if( str==L"allow_model" )
{
if( sol_id->GetLexAuto().GetModel().GetSequenceLabeler().IsAvailable() || sol_id->GetLexAuto().GetModel().GetClassifier().IsAvailable() )
{
ApplyModel = true;
mout->printf( "Morphology model is enabled\n" );
}
else
{
mout->printf( "Morphology model is not available\n" );
}
ret=true;
}
else if( str==L"disallow_model" )
{
ApplyModel = false;
mout->printf( "Morphology model is disabled\n" );
ret=true;
}
else if( str==L"show" )
{
if( current_analysis.NotNull() )
{
const Res_Pack &pack = current_analysis->GetPack();
mout->printf( "\nResult pack contains %vfE%d%vn variators:\n", pack.vars().size() );
if( run_mode==MorphologyMode )
{
for( lem::Container::size_type i=0; i<pack.vars().size(); i++ )
{
const Variator * var = pack.vars()[i];
for( lem::Container::size_type k=0; k<var->size(); ++k )
{
const Tree_Node & root = var->get(k);
mout->printf( "%d: ", CastSizeToInt(k) );
root.Print( *lem::mout, sol_id->GetSynGram(), -1, true );
mout->eol();
}
mout->eol();
mout->eol();
}
}
else
{
for( lem::Container::size_type i=0; i<pack.vars().size(); i++ )
{
pack.vars()[i]->PrintV( *mout, sol_id->GetSynGram(), true );
mout->eol();
mout->eol();
}
}
}
ret=true;
}
else if( str==L"tree" )
{
if( current_analysis.NotNull() )
{
const Res_Pack &pack = current_analysis->GetPack();
Solarix::print_syntax_tree( current_analysis->GetString(), current_analysis->GetPack(), *sol_id, *lem::mout, false, true );
}
ret=true;
}
else if( str.eq_beg("recog" ) )
{
if( current_analysis.NotNull() )
{
lem::mout->eol();
current_analysis->GetLexer().PrintRecognitions( *lem::mout );
}
return true;
}
else if( str==L"tokenize" )
{
SetMode(TokenizerMode);
ret=true;
}
else if( str==L"lemmatize" )
{
SetMode(LemmatizerMode);
ret=true;
}
else if( str==L"speak" )
{
SetMode(SpeakerMode);
ret=true;
}
else if( str==L"syntax" )
{
SetMode(SyntaxMode);
ret=true;
}
else if( str==L"morphology" )
{
SetMode(MorphologyMode);
ret=true;
}
else if( str==L"debugger" )
{
if( debugger.NotNull() )
debugger->ManageBreakpoints();
ret=true;
}
else
{
lem::mout->printf( "Invalid command %vfC%us%vn\n", str.c_str() );
ret=true;
}
return ret;
}
void LexicalAutomat::TranslateLexem(
lem::UFString &str,
bool substitute_entries,
int LanguageID
)
{
WideStringUcs4 cenum(str.c_str());
lem::uint32_t c;
lem::uint32_t prev_c=0;
lem::Ucs4ToUFString yield;
while( (c=cenum.Fetch())!=0 )
{
if( c==L' ' )
{
// Оставляем всегда единственный пробел (точнее - символ разделения лексем
// в мультилексеме).
if( prev_c!=Lexem::DelimiterChar )
{
yield.Add(Lexem::DelimiterChar);
prev_c = c;
}
continue;
}
if( c==L'-' || c==L',' || c==L'\'' )
{
if( prev_c!=Lexem::DelimiterChar && prev_c!=0 )
yield.Add(Lexem::DelimiterChar);
yield.Add(c);
if( cenum.Fetch()!=0 )
{
yield.Add( Lexem::DelimiterChar );
cenum.Back();
}
prev_c = Lexem::DelimiterChar;
continue;
}
Word_Coord wc;
if( LanguageID==ANY_STATE )
wc = GG->FindSymbol(c);
else if( LanguageID==UNKNOWN )
wc = GG->FindSymbol(c,GetDefaultAlphabets());
else
wc = GG->FindSymbol(c,LanguageID);
if( wc.GetEntry()==UNKNOWN )
{
// Символ не найден в графической грамматике.
yield.Add(c);
}
else
{
const GG_Entry &entry = GG->entries()[ wc.GetEntry() ];
if(
!substitute_entries ||
wc.GetForm()==UNKNOWN ||
wc.GetForm()==ANY_STATE
)
yield.Add( entry.GetForm(wc.GetForm()).GetName() );
else
yield.Add( entry.GetName() );
}
prev_c = c;
}
str = yield.ToString();
return;
}
void CasingCoder::RestoreCasing( int external_casing_state, lem::UFString &res, int ekey )
{
if( (ekey!=UNKNOWN && ekey==UnknownEntries_ekey ) )
return;
const XLAT *xlat = GetXLAT(ekey);
switch( x(external_casing_state) )
{
case 1:
{
if( xlat->use_unicode )
{
res.to_Aa();
}
else
{
WideStringUcs4 src_enum( res.c_str() );
lem::UFString out;
out.reserve(res.length()+1);
for( int i=0; ; ++i )
{
const lem::uint32_t src_ucs4 = src_enum.Fetch();
if(!src_ucs4)
break;
if(i==0)
AddUpper( xlat, src_ucs4, out );
else
AddLower( xlat, src_ucs4, out );
}
res = out;
}
break;
}
case 2:
{
if( xlat->use_unicode )
{
res.to_upper();
}
else
{
WideStringUcs4 src_enum( res.c_str() );
lem::UFString out;
out.reserve(res.length()+1);
for(;;)
{
const lem::uint32_t src_ucs4 = src_enum.Fetch();
if(!src_ucs4)
break;
AddUpper( xlat, src_ucs4, out );
}
res = out;
}
break;
}
case 3:
{
if( xlat->use_unicode )
{
Solarix::MakeEachLexemAa(res);
}
else
{
bool capitalize=true;
WideStringUcs4 src_enum( res.c_str() );
lem::UFString out;
out.reserve(res.length()+1);
for(;;)
{
const lem::uint32_t src_ucs4 = src_enum.Fetch();
if(!src_ucs4)
break;
if( capitalize )
{
AddUpper( xlat, src_ucs4, out );
capitalize=false;
}
else
{
AddLower( xlat, src_ucs4, out );
if( src_ucs4==L' ' || src_ucs4==L'-' )
capitalize=true;
}
}
res = out;
}
break;
}
case 0:
default:
{
if( xlat->use_unicode )
{
res.to_lower();
}
else
{
WideStringUcs4 src_enum( res.c_str() );
lem::UFString out;
out.reserve(res.length()+1);
for(;;)
{
const lem::uint32_t src_ucs4 = src_enum.Fetch();
if(!src_ucs4)
break;
AddLower( xlat, src_ucs4, out );
}
res = out;
}
break;
}
}
// res.subst_all( L" - ", L"-" );
// res.subst_all( L" ' ", L"'" );
return;
}
int SegmentingSentenceTokenizer::MatchLen(const lem::UFString & s, int i0) const
{
return lookup->match_len(s.c_str() + i0);
}