desr: desr: src/MorphExtractor.h Source File

00001 /*
00002 **  DeSR
00003 **  src/MorphExtractor.h
00004 **  ----------------------------------------------------------------------
00005 **  Copyright (c) 2005  Giuseppe Attardi (attardi@di.unipi.it).
00006 **  ----------------------------------------------------------------------
00007 **
00008 **  This file is part of DeSR.
00009 **
00010 **  DeSR is free software; you can redistribute it and/or modify it
00011 **  under the terms of the GNU General Public License, version 3,
00012 **  as published by the Free Software Foundation.
00013 **
00014 **  DeSR is distributed in the hope that it will be useful,
00015 **  but WITHOUT ANY WARRANTY; without even the implied warranty of
00016 **  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
00017 **  GNU General Public License for more details.
00018 **
00019 **  You should have received a copy of the GNU General Public License
00020 **  along with this program.  If not, see <http://www.gnu.org/licenses/>.
00021 **  ----------------------------------------------------------------------
00022 */
00023 
00024 #ifndef DeSR_MorphExtractor_H
00025 #define DeSR_MorphExtractor_H
00026 
00027 #include "text/RegExp.h"
00028 
00029 namespace Tanl {
00030 
00034 struct MorphExtractor
00035 {
00036   MorphExtractor() { }
00037 
00038   struct Features {
00039     Features() {
00040       Case[0] = '\0';
00041       extra[0] = '\0';
00042       gender[0] = '\0';
00043       mode[0] = '\0';
00044       negative[0] = '\0';
00045       number[0] = '\0';
00046       person[0] = '\0';
00047       tense[0] = '\0';
00048       trans[0] = '\0';
00049       full[0] = '\0';
00050     }
00051 
00052     char Case[20];              
00053     char extra[20];             
00054     char gender[20];            
00055     char mode[20];              
00056     char negative[20];          
00057     char number[20];            
00058     char person[20];            
00059     char tense[20];             
00060     char trans[20];             
00061     char full[256];             
00062   };
00063 
00064   virtual void operator() (char const* start, char const* end,
00065                            Features& mf) const;
00066 };
00067 
00068 struct ArabicMorphExtractor : public MorphExtractor
00069 {
00070   static Tanl::Text::RegExp::Pattern reCase;
00071   static Tanl::Text::RegExp::Pattern reGender;
00072   static Tanl::Text::RegExp::Pattern reNumber;
00073   static Tanl::Text::RegExp::Pattern rePerson;
00074 
00075   void operator() (char const* start, char const* end, Features& mf) const;
00076 };
00077 
00078 struct BasqueMorphExtractor : public MorphExtractor
00079 {
00080   static Tanl::Text::RegExp::Pattern reNumber;
00081 
00082   BasqueMorphExtractor() {}
00083 
00084   void operator() (char const* start, char const* end, Features& mf) const;
00085 };
00086 
00087 struct BulgarianMorphExtractor : public MorphExtractor
00088 {
00089   static Tanl::Text::RegExp::Pattern reGender;
00090   static Tanl::Text::RegExp::Pattern reNumber;
00091   static Tanl::Text::RegExp::Pattern rePerson;
00092 
00093   void operator() (char const* start, char const* end, Features& mf) const;
00094 };
00095 
00096 struct CzechMorphExtractor : public MorphExtractor
00097 {
00098   static Tanl::Text::RegExp::Pattern reCase;
00099   static Tanl::Text::RegExp::Pattern reGender;
00100   static Tanl::Text::RegExp::Pattern reNumber;
00101   static Tanl::Text::RegExp::Pattern rePerson;
00102   static Tanl::Text::RegExp::Pattern reNegative;
00103   static Tanl::Text::RegExp::Pattern reGra;
00104 
00105   void operator() (char const* start, char const* end, Features& mf) const;
00106 };
00107 
00108 struct DanishMorphExtractor : public MorphExtractor
00109 {
00110   static Tanl::Text::RegExp::Pattern reGender;
00111   static Tanl::Text::RegExp::Pattern reNumber;
00112   static Tanl::Text::RegExp::Pattern rePerson;
00113   static Tanl::Text::RegExp::Pattern reCase;
00114 
00115   void operator() (char const* start, char const* end, Features& mf) const;
00116 };
00117 
00118 struct DutchMorphExtractor : public MorphExtractor
00119 {
00120   static Tanl::Text::RegExp::Pattern reGender;
00121   static Tanl::Text::RegExp::Pattern reNumber;
00122   static Tanl::Text::RegExp::Pattern rePerson;
00123   static Tanl::Text::RegExp::Pattern reCase;
00124 
00125   void operator() (char const* start, char const* end, Features& mf) const;
00126 };
00127 
00128 struct GreekMorphExtractor : public MorphExtractor
00129 {
00130   static Tanl::Text::RegExp::Pattern reGender;
00131   static Tanl::Text::RegExp::Pattern reNumber;
00132   static Tanl::Text::RegExp::Pattern rePerson;
00133   static Tanl::Text::RegExp::Pattern reCase;
00134 
00135   void operator() (char const* start, char const* end, Features& mf) const;
00136 };
00137 
00138 struct HungarianMorphExtractor : public MorphExtractor
00139 {
00140   static Tanl::Text::RegExp::Pattern reNumber;
00141   static Tanl::Text::RegExp::Pattern rePerson;
00142   static Tanl::Text::RegExp::Pattern reCase;
00143 
00144   void operator() (char const* start, char const* end, Features& mf) const;
00145 };
00146 
00147 struct ItalianTutMorphExtractor : public MorphExtractor
00148 {
00149   static Tanl::Text::RegExp::Pattern reCase;
00150   static Tanl::Text::RegExp::Pattern reGender;
00151   static Tanl::Text::RegExp::Pattern reMode;
00152   static Tanl::Text::RegExp::Pattern reNumber;
00153   static Tanl::Text::RegExp::Pattern rePerson;
00154   static Tanl::Text::RegExp::Pattern reSem;
00155   static Tanl::Text::RegExp::Pattern reTense;
00156   static Tanl::Text::RegExp::Pattern reTrans;
00157   static Tanl::Text::RegExp::Pattern reVTrans;
00158 
00159   void operator() (char const* start, char const* end, Features& mf) const;
00160 };
00161 
00162 struct PortugueseMorphExtractor : public MorphExtractor
00163 {
00164   static Tanl::Text::RegExp::Pattern reGender;
00165   static Tanl::Text::RegExp::Pattern reNumber;
00166   static Tanl::Text::RegExp::Pattern rePerson;
00167 
00168   PortugueseMorphExtractor() {}
00169 
00170   void operator() (char const* start, char const* end, Features& mf) const;
00171 };
00172 
00173 struct SloveneMorphExtractor : public MorphExtractor
00174 {
00175   static Tanl::Text::RegExp::Pattern reCase;
00176   static Tanl::Text::RegExp::Pattern reGender;
00177   static Tanl::Text::RegExp::Pattern reNegative;
00178   static Tanl::Text::RegExp::Pattern reNumber;
00179   static Tanl::Text::RegExp::Pattern rePerson;
00180 
00181   void operator() (char const* start, char const* end, Features& mf) const;
00182 };
00183 
00184 struct SpanishMorphExtractor : public MorphExtractor
00185 {
00186   static Tanl::Text::RegExp::Pattern reGender;
00187   static Tanl::Text::RegExp::Pattern reNumber;
00188   static Tanl::Text::RegExp::Pattern rePerson;
00189 
00190   void operator() (char const* start, char const* end, Features& mf) const;
00191 };
00192 
00193 struct TurkishMorphExtractor : public MorphExtractor
00194 {
00195   static Tanl::Text::RegExp::Pattern reAFeats;
00196   static Tanl::Text::RegExp::Pattern rePFeats;
00197 
00198   void operator() (char const* start, char const* end, Features& mf) const;
00199 };
00200 
00201 } // namespace Tanl
00202 
00203 #endif // DeSR_MorphExtractor_H