📄 amiseventformat.h
字号:
//////////////////////////////////////////////////////////////////////////// Copyright (c) 2000, Yusuke Miyao/// You may distribute under the terms of the Artistic License.////// <id>$Id: AmisEventFormat.h,v 1.4 2003/05/16 05:11:05 yusuke Exp $</id>/// <collection>Maximum Entropy Estimator</collection>/// <name>AmisEventFormat.h</name>/// <overview>Parser for Amis-style event files</overview>/////////////////////////////////////////////////////////////////////////#ifndef Amis_AmisEventFormat_h_#define Amis_AmisEventFormat_h_#include <amis/configure.h>#include <amis/EventFormat.h>#include <amis/Model.h>#include <amis/EventSpace.h>#include <amis/Tokenizer.h>AMIS_NAMESPACE_BEGIN///////////////////////////////////////////////////////////////////////// <classdef>/// <name>AmisEventFormat</name>/// <overview>Parser for Amis-style event files</overview>/// <desc>/// The class imports event data from Amis-style data files./// The Amis format enables non-binary features and packed events./// For details, see README./// </desc>/// <see>Event, EventSpace, EventFormat</see>/// <body>template < class Feature, class Name = std::string >class AmisEventFormat : public EventFormat {public: typedef typename Feature::FeatureFreq FeatureFreq; static FeatureFreq extractFreq( std::string& feature ) { std::string::size_type pos = feature.find( ':' ); if ( pos == std::string::npos ) return 1; FeatureFreq freq; str2FeatureFreq( feature.substr( pos + 1 ), freq ); feature.erase( pos ); //string tmp_s( feature.substr(0, pos) ); //feature = tmp_s; //cerr << "feature=" << feature << std::endl; return freq; } /// Extract frequency from the string description of a featureprivate: std::string current_event;protected: static void str2FeatureFreq( const std::string& s, int& f ) { f = Tokenizer::str2int( s ); } /// Convert a string into feature frequency static void str2FeatureFreq( const std::string& s, Real& f ) { f = Tokenizer::str2Real( s ); } /// Convert a string into feature frequency void inputFeatureList( const Model< Name >& model, Tokenizer& t, std::vector< Feature >& fl ) const { std::string feature_str; while ( t.nextToken( feature_str ) ) { FeatureFreq freq = extractFreq( feature_str ); if ( freq <= static_cast< FeatureFreq >( 0 ) ) { throw IllegalEventFormatError( "Feature frequency must be positive in Event " + current_event, t.lineNumber() ); } IStringStream feature_is( feature_str ); Name feature; feature_is >> feature; fl.push_back( Feature( model.featureID( feature ), freq ) ); } } /// Input a feature listpublic: AmisEventFormat() : EventFormat(), current_event() { } /// Constructor virtual ~AmisEventFormat() {} /// Destructor std::string eventFormatName() const { return "AmisEvent"; } /// Name of the event format bool inputEvent( std::istream& s, const Model< Name >& model, Event< Feature >& event ) { Tokenizer t( s ); return inputEvent( t, model, event ); } /// Input an event from stream bool inputEvent( Tokenizer& t, const Model< Name >& model, Event< Feature >& event ) { if ( ! t.nextToken( current_event ) ) return false; // empty line std::string dummy; Real prob = 0.0; if ( t.nextToken( dummy ) ) { prob = t.str2Real( dummy ); if ( t.nextToken( dummy ) ) { throw IllegalEventFormatError( "Too many tokens found in Event " + current_event, t.lineNumber() ); } } else { prob = 1.0; }#ifdef AMIS_JOINT_PROB event.setEventProbability( prob );#endif // AMIS_JOINT_PROB while ( true ) { FeatureListFreq freq; if ( ! t.nextToken( freq ) ) break; if ( freq < 0 ) { throw IllegalEventFormatError( "Feature list frequency must be positive or zero in Event " + current_event, t.lineNumber() ); } //cerr << "freq = " << freq << std::endl; std::vector< Feature > fl; inputFeatureList( model, t, fl ); event.addFeatureList( fl, freq ); } return true; } void inputEventSpace( std::istream& s, const ModelBase& model_base, EventSpaceBase& event_space_base ) { const Model< Name >* model = dynamic_cast< const Model< Name >* >( &model_base ); if ( model == NULL ) { throw IllegalEventFormatError( "AmisEventFormat can be used only for Model< ... > class", 0 ); } EventSpace< Feature, Event< Feature > >* event_space = dynamic_cast< EventSpace< Feature, Event< Feature > >* >( &event_space_base ); if ( event_space == NULL ) { throw IllegalEventFormatError( "AmisEventFormat can be used only for EventSpace class", 0 ); } inputEventSpace( s, *model, *event_space ); } /// Interface defined by class EventFormat virtual void inputEventSpace( std::istream& s, const Model< Name >& model, EventSpace< Feature, Event< Feature > >& event_space ) { Tokenizer t( s ); AMIS_DEBUG_MESSAGE( 3, "\nInput events...\n" ); AMIS_DEBUG_MESSAGE( 5, "\t----------------------------------------\n" ); AMIS_DEBUG_MESSAGE( 5, "\tEvent\tFreq.\t# FLs\n" ); while ( ! t.endOfStream() ) { Event< Feature > event; if ( inputEvent( t, model, event ) ) { AMIS_DEBUG_MESSAGE( 5, '\t' << current_event << '\t' << event.eventFrequency() << '\t' << event.numFeatureList() + 1 << '\n' ); try { event_space.addEvent( event ); } catch ( IllegalEventError e ) { throw IllegalEventFormatError( e.message() + " in Event " + current_event, t.lineNumber() ); } } } //cerr << "Number of events = " << event_space->numEvents() << std::endl; //cerr << "Count of events = " << event_space->sumEventCount() << std::endl; AMIS_DEBUG_MESSAGE( 5, "\t----------------------------------------\n" ); } /// Input event data from an input stream};/// </body>/// </classdef>AMIS_NAMESPACE_END#endif // Amis_AmisEventFormat_h_// end of AmisEventFormat.h⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -