tanl: tanl: parse/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 
00036 struct MorphExtractor
00037 {
00038   MorphExtractor() { }
00039 
00040   struct Features {
00041     Features() {
00042       Case[0] = '\0';
00043       extra[0] = '\0';
00044       gender[0] = '\0';
00045       mode[0] = '\0';
00046       subcat[0] = '\0';
00047       negative[0] = '\0';
00048       number[0] = '\0';
00049       person[0] = '\0';
00050       tense[0] = '\0';
00051       trans[0] = '\0';
00052       full[0] = '\0';
00053     }
00054 
00055     char Case[20];              
00056     char extra[20];             
00057     char gender[20];            
00058     char mode[20];              
00059     char subcat[20];            
00060     char negative[20];          
00061     char number[20];            
00062     char person[20];            
00063     char tense[20];             
00064     char trans[20];             
00065     char full[256];             
00066   };
00067 
00068   virtual void operator() (char const* start, char const* end,
00069                            Features& mf) const;
00070 };
00071 
00075 struct ArabicMorphExtractor : public MorphExtractor
00076 {
00077   static Tanl::Text::RegExp::Pattern reCase;
00078   static Tanl::Text::RegExp::Pattern reGender;
00079   static Tanl::Text::RegExp::Pattern reNumber;
00080   static Tanl::Text::RegExp::Pattern rePerson;
00081 
00082   void operator() (char const* start, char const* end, Features& mf) const;
00083 };
00084 
00088 struct BasqueMorphExtractor : public MorphExtractor
00089 {
00090   static Tanl::Text::RegExp::Pattern reNumber;
00091 
00092   BasqueMorphExtractor() {}
00093 
00094   void operator() (char const* start, char const* end, Features& mf) const;
00095 };
00096 
00100 struct BulgarianMorphExtractor : public MorphExtractor
00101 {
00102   static Tanl::Text::RegExp::Pattern reGender;
00103   static Tanl::Text::RegExp::Pattern reNumber;
00104   static Tanl::Text::RegExp::Pattern rePerson;
00105 
00106   void operator() (char const* start, char const* end, Features& mf) const;
00107 };
00108 
00112 struct CzechMorphExtractor : public MorphExtractor
00113 {
00114   static Tanl::Text::RegExp::Pattern reCase;
00115   static Tanl::Text::RegExp::Pattern reGender;
00116   static Tanl::Text::RegExp::Pattern reNumber;
00117   static Tanl::Text::RegExp::Pattern rePerson;
00118   static Tanl::Text::RegExp::Pattern reNegative;
00119   static Tanl::Text::RegExp::Pattern reGra;
00120 
00121   void operator() (char const* start, char const* end, Features& mf) const;
00122 };
00123 
00127 struct DanishMorphExtractor : public MorphExtractor
00128 {
00129   static Tanl::Text::RegExp::Pattern reGender;
00130   static Tanl::Text::RegExp::Pattern reNumber;
00131   static Tanl::Text::RegExp::Pattern rePerson;
00132   static Tanl::Text::RegExp::Pattern reCase;
00133 
00134   void operator() (char const* start, char const* end, Features& mf) const;
00135 };
00136 
00140 struct DutchMorphExtractor : public MorphExtractor
00141 {
00142   static Tanl::Text::RegExp::Pattern reGender;
00143   static Tanl::Text::RegExp::Pattern reNumber;
00144   static Tanl::Text::RegExp::Pattern rePerson;
00145   static Tanl::Text::RegExp::Pattern reCase;
00146 
00147   void operator() (char const* start, char const* end, Features& mf) const;
00148 };
00149 
00153 struct FrenchMorphExtractor : public MorphExtractor
00154 {
00155   static Tanl::Text::RegExp::Pattern reGender;
00156   static Tanl::Text::RegExp::Pattern reNumber;
00157   static Tanl::Text::RegExp::Pattern rePerson;
00216   static Tanl::Text::RegExp::Pattern reSubcat;
00217 
00218   void operator() (char const* start, char const* end, Features& mf) const;
00219 };
00220 
00221 // ======================================================================
00222 // German
00223 
00227 struct GreekMorphExtractor : public MorphExtractor
00228 {
00229   static Tanl::Text::RegExp::Pattern reGender;
00230   static Tanl::Text::RegExp::Pattern reNumber;
00231   static Tanl::Text::RegExp::Pattern rePerson;
00232   static Tanl::Text::RegExp::Pattern reCase;
00233 
00234   void operator() (char const* start, char const* end, Features& mf) const;
00235 };
00236 
00240 struct HungarianMorphExtractor : public MorphExtractor
00241 {
00242   static Tanl::Text::RegExp::Pattern reNumber;
00243   static Tanl::Text::RegExp::Pattern rePerson;
00244   static Tanl::Text::RegExp::Pattern reCase;
00245 
00246   void operator() (char const* start, char const* end, Features& mf) const;
00247 };
00248 
00252 struct IndianMorphExtractor : public MorphExtractor
00253 {
00254   static Tanl::Text::RegExp::Pattern reCase;
00255   static Tanl::Text::RegExp::Pattern reGender;
00256   static Tanl::Text::RegExp::Pattern reNumber;
00257   static Tanl::Text::RegExp::Pattern rePerson;
00258   static Tanl::Text::RegExp::Pattern reVibhakti;
00259   static Tanl::Text::RegExp::Pattern reTam;
00260 
00261   void operator() (char const* start, char const* end, Features& mf) const;
00262 };
00263 
00267 struct ItalianMorphExtractor : public MorphExtractor
00268 {
00269   static Tanl::Text::RegExp::Pattern reGender;
00270   static Tanl::Text::RegExp::Pattern reMode;
00271   static Tanl::Text::RegExp::Pattern reNumber;
00272   static Tanl::Text::RegExp::Pattern rePerson;
00273   static Tanl::Text::RegExp::Pattern reTense;
00274 
00275   void operator() (char const* start, char const* end, Features& mf) const;
00276 };
00277 
00281 struct ItalianTutMorphExtractor : public MorphExtractor
00282 {
00283   static Tanl::Text::RegExp::Pattern reCase;
00284   static Tanl::Text::RegExp::Pattern reGender;
00285   static Tanl::Text::RegExp::Pattern reMode;
00286   static Tanl::Text::RegExp::Pattern reNumber;
00287   static Tanl::Text::RegExp::Pattern rePerson;
00288   static Tanl::Text::RegExp::Pattern reSem;
00289   static Tanl::Text::RegExp::Pattern reTense;
00290   static Tanl::Text::RegExp::Pattern reTrans;
00291   static Tanl::Text::RegExp::Pattern reVTrans;
00292 
00293   void operator() (char const* start, char const* end, Features& mf) const;
00294 };
00295 
00299 struct PortugueseMorphExtractor : public MorphExtractor
00300 {
00301   static Tanl::Text::RegExp::Pattern reGender;
00302   static Tanl::Text::RegExp::Pattern reNumber;
00303   static Tanl::Text::RegExp::Pattern rePerson;
00304 
00305   PortugueseMorphExtractor() {}
00306 
00307   void operator() (char const* start, char const* end, Features& mf) const;
00308 };
00309 
00313 struct SloveneMorphExtractor : public MorphExtractor
00314 {
00315   static Tanl::Text::RegExp::Pattern reCase;
00316   static Tanl::Text::RegExp::Pattern reGender;
00317   static Tanl::Text::RegExp::Pattern reNegative;
00318   static Tanl::Text::RegExp::Pattern reNumber;
00319   static Tanl::Text::RegExp::Pattern rePerson;
00320 
00321   void operator() (char const* start, char const* end, Features& mf) const;
00322 };
00323 
00327 struct SpanishMorphExtractor : public MorphExtractor
00328 {
00329   static Tanl::Text::RegExp::Pattern reGender;
00330   static Tanl::Text::RegExp::Pattern reNumber;
00331   static Tanl::Text::RegExp::Pattern rePerson;
00332 
00333   void operator() (char const* start, char const* end, Features& mf) const;
00334 };
00335 
00339 struct TurkishMorphExtractor : public MorphExtractor
00340 {
00341   static Tanl::Text::RegExp::Pattern reAFeats;
00342   static Tanl::Text::RegExp::Pattern rePFeats;
00343 
00344   void operator() (char const* start, char const* end, Features& mf) const;
00345 };
00346 
00347 } // namespace Tanl
00348 
00349 #endif // DeSR_MorphExtractor_H