michael@0: /******* BEGIN LICENSE BLOCK *******
michael@0:  * Version: MPL 1.1/GPL 2.0/LGPL 2.1
michael@0:  * 
michael@0:  * The contents of this file are subject to the Mozilla Public License Version
michael@0:  * 1.1 (the "License"); you may not use this file except in compliance with
michael@0:  * the License. You may obtain a copy of the License at
michael@0:  * http://www.mozilla.org/MPL/
michael@0:  * 
michael@0:  * Software distributed under the License is distributed on an "AS IS" basis,
michael@0:  * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
michael@0:  * for the specific language governing rights and limitations under the
michael@0:  * License.
michael@0:  * 
michael@0:  * The Initial Developers of the Original Code are Kevin Hendricks (MySpell)
michael@0:  * and László Németh (Hunspell). Portions created by the Initial Developers
michael@0:  * are Copyright (C) 2002-2005 the Initial Developers. All Rights Reserved.
michael@0:  * 
michael@0:  * Contributor(s): Kevin Hendricks (kevin.hendricks@sympatico.ca)
michael@0:  *                 David Einstein (deinst@world.std.com)
michael@0:  *                 László Németh (nemethl@gyorsposta.hu)
michael@0:  *                 Caolan McNamara (caolanm@redhat.com)
michael@0:  *                 Davide Prina
michael@0:  *                 Giuseppe Modugno
michael@0:  *                 Gianluca Turconi
michael@0:  *                 Simon Brouwer
michael@0:  *                 Noll Janos
michael@0:  *                 Biro Arpad
michael@0:  *                 Goldman Eleonora
michael@0:  *                 Sarlos Tamas
michael@0:  *                 Bencsath Boldizsar
michael@0:  *                 Halacsy Peter
michael@0:  *                 Dvornik Laszlo
michael@0:  *                 Gefferth Andras
michael@0:  *                 Nagy Viktor
michael@0:  *                 Varga Daniel
michael@0:  *                 Chris Halls
michael@0:  *                 Rene Engelhard
michael@0:  *                 Bram Moolenaar
michael@0:  *                 Dafydd Jones
michael@0:  *                 Harri Pitkanen
michael@0:  *                 Andras Timar
michael@0:  *                 Tor Lillqvist
michael@0:  * 
michael@0:  * Alternatively, the contents of this file may be used under the terms of
michael@0:  * either the GNU General Public License Version 2 or later (the "GPL"), or
michael@0:  * the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
michael@0:  * in which case the provisions of the GPL or the LGPL are applicable instead
michael@0:  * of those above. If you wish to allow use of your version of this file only
michael@0:  * under the terms of either the GPL or the LGPL, and not to allow others to
michael@0:  * use your version of this file under the terms of the MPL, indicate your
michael@0:  * decision by deleting the provisions above and replace them with the notice
michael@0:  * and other provisions required by the GPL or the LGPL. If you do not delete
michael@0:  * the provisions above, a recipient may use your version of this file under
michael@0:  * the terms of any one of the MPL, the GPL or the LGPL.
michael@0:  *
michael@0:  ******* END LICENSE BLOCK *******/
michael@0: 
michael@0: #ifndef _AFFIXMGR_HXX_
michael@0: #define _AFFIXMGR_HXX_
michael@0: 
michael@0: #include "hunvisapi.h"
michael@0: 
michael@0: #include <stdio.h>
michael@0: 
michael@0: #include "atypes.hxx"
michael@0: #include "baseaffix.hxx"
michael@0: #include "hashmgr.hxx"
michael@0: #include "phonet.hxx"
michael@0: #include "replist.hxx"
michael@0: 
michael@0: // check flag duplication
michael@0: #define dupSFX        (1 << 0)
michael@0: #define dupPFX        (1 << 1)
michael@0: 
michael@0: class PfxEntry;
michael@0: class SfxEntry;
michael@0: 
michael@0: class LIBHUNSPELL_DLL_EXPORTED AffixMgr
michael@0: {
michael@0: 
michael@0:   PfxEntry *          pStart[SETSIZE];
michael@0:   SfxEntry *          sStart[SETSIZE];
michael@0:   PfxEntry *          pFlag[SETSIZE];
michael@0:   SfxEntry *          sFlag[SETSIZE];
michael@0:   HashMgr *           pHMgr;
michael@0:   HashMgr **          alldic;
michael@0:   int *               maxdic;
michael@0:   char *              keystring;
michael@0:   char *              trystring;
michael@0:   char *              encoding;
michael@0:   struct cs_info *    csconv;
michael@0:   int                 utf8;
michael@0:   int                 complexprefixes;
michael@0:   FLAG                compoundflag;
michael@0:   FLAG                compoundbegin;
michael@0:   FLAG                compoundmiddle;
michael@0:   FLAG                compoundend;
michael@0:   FLAG                compoundroot;
michael@0:   FLAG                compoundforbidflag;
michael@0:   FLAG                compoundpermitflag;
michael@0:   int                 checkcompounddup;
michael@0:   int                 checkcompoundrep;
michael@0:   int                 checkcompoundcase;
michael@0:   int                 checkcompoundtriple;
michael@0:   int                 simplifiedtriple;
michael@0:   FLAG                forbiddenword;
michael@0:   FLAG                nosuggest;
michael@0:   FLAG                nongramsuggest;
michael@0:   FLAG                needaffix;
michael@0:   int                 cpdmin;
michael@0:   int                 numrep;
michael@0:   replentry *         reptable;
michael@0:   RepList *           iconvtable;
michael@0:   RepList *           oconvtable;
michael@0:   int                 nummap;
michael@0:   mapentry *          maptable;
michael@0:   int                 numbreak;
michael@0:   char **             breaktable;
michael@0:   int                 numcheckcpd;
michael@0:   patentry *          checkcpdtable;
michael@0:   int                 simplifiedcpd;
michael@0:   int                 numdefcpd;
michael@0:   flagentry *         defcpdtable;
michael@0:   phonetable *        phone;
michael@0:   int                 maxngramsugs;
michael@0:   int                 maxcpdsugs;
michael@0:   int                 maxdiff;
michael@0:   int                 onlymaxdiff;
michael@0:   int                 nosplitsugs;
michael@0:   int                 sugswithdots;
michael@0:   int                 cpdwordmax;
michael@0:   int                 cpdmaxsyllable;
michael@0:   char *              cpdvowels;
michael@0:   w_char *            cpdvowels_utf16;
michael@0:   int                 cpdvowels_utf16_len;
michael@0:   char *              cpdsyllablenum;
michael@0:   const char *        pfxappnd; // BUG: not stateless
michael@0:   const char *        sfxappnd; // BUG: not stateless
michael@0:   FLAG                sfxflag;  // BUG: not stateless
michael@0:   char *              derived;  // BUG: not stateless
michael@0:   SfxEntry *          sfx;      // BUG: not stateless
michael@0:   PfxEntry *          pfx;      // BUG: not stateless
michael@0:   int                 checknum;
michael@0:   char *              wordchars;
michael@0:   unsigned short *    wordchars_utf16;
michael@0:   int                 wordchars_utf16_len;
michael@0:   char *              ignorechars;
michael@0:   unsigned short *    ignorechars_utf16;
michael@0:   int                 ignorechars_utf16_len;
michael@0:   char *              version;
michael@0:   char *              lang;
michael@0:   int                 langnum;
michael@0:   FLAG                lemma_present;
michael@0:   FLAG                circumfix;
michael@0:   FLAG                onlyincompound;
michael@0:   FLAG                keepcase;
michael@0:   FLAG                forceucase;
michael@0:   FLAG                warn;
michael@0:   int                 forbidwarn;
michael@0:   FLAG                substandard;
michael@0:   int                 checksharps;
michael@0:   int                 fullstrip;
michael@0: 
michael@0:   int                 havecontclass; // boolean variable
michael@0:   char                contclasses[CONTSIZE]; // flags of possible continuing classes (twofold affix)
michael@0: 
michael@0: public:
michael@0: 
michael@0:   AffixMgr(const char * affpath, HashMgr** ptr, int * md,
michael@0:     const char * key = NULL);
michael@0:   ~AffixMgr();
michael@0:   struct hentry *     affix_check(const char * word, int len,
michael@0:             const unsigned short needflag = (unsigned short) 0,
michael@0:             char in_compound = IN_CPD_NOT);
michael@0:   struct hentry *     prefix_check(const char * word, int len,
michael@0:             char in_compound, const FLAG needflag = FLAG_NULL);
michael@0:   inline int isSubset(const char * s1, const char * s2);
michael@0:   struct hentry *     prefix_check_twosfx(const char * word, int len,
michael@0:             char in_compound, const FLAG needflag = FLAG_NULL);
michael@0:   inline int isRevSubset(const char * s1, const char * end_of_s2, int len);
michael@0:   struct hentry *     suffix_check(const char * word, int len, int sfxopts,
michael@0:             PfxEntry* ppfx, char ** wlst, int maxSug, int * ns,
michael@0:             const FLAG cclass = FLAG_NULL, const FLAG needflag = FLAG_NULL,
michael@0:             char in_compound = IN_CPD_NOT);
michael@0:   struct hentry *     suffix_check_twosfx(const char * word, int len,
michael@0:             int sfxopts, PfxEntry* ppfx, const FLAG needflag = FLAG_NULL);
michael@0: 
michael@0:   char * affix_check_morph(const char * word, int len,
michael@0:             const FLAG needflag = FLAG_NULL, char in_compound = IN_CPD_NOT);
michael@0:   char * prefix_check_morph(const char * word, int len,
michael@0:             char in_compound, const FLAG needflag = FLAG_NULL);
michael@0:   char * suffix_check_morph (const char * word, int len, int sfxopts,
michael@0:             PfxEntry * ppfx, const FLAG cclass = FLAG_NULL,
michael@0:             const FLAG needflag = FLAG_NULL, char in_compound = IN_CPD_NOT);
michael@0: 
michael@0:   char * prefix_check_twosfx_morph(const char * word, int len,
michael@0:             char in_compound, const FLAG needflag = FLAG_NULL);
michael@0:   char * suffix_check_twosfx_morph(const char * word, int len,
michael@0:             int sfxopts, PfxEntry * ppfx, const FLAG needflag = FLAG_NULL);
michael@0: 
michael@0:   char * morphgen(char * ts, int wl, const unsigned short * ap,
michael@0:             unsigned short al, char * morph, char * targetmorph, int level);
michael@0: 
michael@0:   int    expand_rootword(struct guessword * wlst, int maxn, const char * ts,
michael@0:             int wl, const unsigned short * ap, unsigned short al, char * bad,
michael@0:             int, char *);
michael@0: 
michael@0:   short       get_syllable (const char * word, int wlen);
michael@0:   int         cpdrep_check(const char * word, int len);
michael@0:   int         cpdpat_check(const char * word, int len, hentry * r1, hentry * r2,
michael@0:                     const char affixed);
michael@0:   int         defcpd_check(hentry *** words, short wnum, hentry * rv,
michael@0:                     hentry ** rwords, char all);
michael@0:   int         cpdcase_check(const char * word, int len);
michael@0:   inline int  candidate_check(const char * word, int len);
michael@0:   void        setcminmax(int * cmin, int * cmax, const char * word, int len);
michael@0:   struct hentry * compound_check(const char * word, int len, short wordnum,
michael@0:             short numsyllable, short maxwordnum, short wnum, hentry ** words,
michael@0:             char hu_mov_rule, char is_sug, int * info);
michael@0: 
michael@0:   int compound_check_morph(const char * word, int len, short wordnum,
michael@0:             short numsyllable, short maxwordnum, short wnum, hentry ** words,
michael@0:             char hu_mov_rule, char ** result, char * partresult);
michael@0: 
michael@0:   struct hentry * lookup(const char * word);
michael@0:   int                 get_numrep() const;
michael@0:   struct replentry *  get_reptable() const;
michael@0:   RepList *           get_iconvtable() const;
michael@0:   RepList *           get_oconvtable() const;
michael@0:   struct phonetable * get_phonetable() const;
michael@0:   int                 get_nummap() const;
michael@0:   struct mapentry *   get_maptable() const;
michael@0:   int                 get_numbreak() const;
michael@0:   char **             get_breaktable() const;
michael@0:   char *              get_encoding();
michael@0:   int                 get_langnum() const;
michael@0:   char *              get_key_string();
michael@0:   char *              get_try_string() const;
michael@0:   const char *        get_wordchars() const;
michael@0:   unsigned short *    get_wordchars_utf16(int * len) const;
michael@0:   char *              get_ignore() const;
michael@0:   unsigned short *    get_ignore_utf16(int * len) const;
michael@0:   int                 get_compound() const;
michael@0:   FLAG                get_compoundflag() const;
michael@0:   FLAG                get_compoundbegin() const;
michael@0:   FLAG                get_forbiddenword() const;
michael@0:   FLAG                get_nosuggest() const;
michael@0:   FLAG                get_nongramsuggest() const;
michael@0:   FLAG                get_needaffix() const;
michael@0:   FLAG                get_onlyincompound() const;
michael@0:   FLAG                get_compoundroot() const;
michael@0:   FLAG                get_lemma_present() const;
michael@0:   int                 get_checknum() const;
michael@0:   const char *        get_prefix() const;
michael@0:   const char *        get_suffix() const;
michael@0:   const char *        get_derived() const;
michael@0:   const char *        get_version() const;
michael@0:   int                 have_contclass() const;
michael@0:   int                 get_utf8() const;
michael@0:   int                 get_complexprefixes() const;
michael@0:   char *              get_suffixed(char ) const;
michael@0:   int                 get_maxngramsugs() const;
michael@0:   int                 get_maxcpdsugs() const;
michael@0:   int                 get_maxdiff() const;
michael@0:   int                 get_onlymaxdiff() const;
michael@0:   int                 get_nosplitsugs() const;
michael@0:   int                 get_sugswithdots(void) const;
michael@0:   FLAG                get_keepcase(void) const;
michael@0:   FLAG                get_forceucase(void) const;
michael@0:   FLAG                get_warn(void) const;
michael@0:   int                 get_forbidwarn(void) const;
michael@0:   int                 get_checksharps(void) const;
michael@0:   char *              encode_flag(unsigned short aflag) const;
michael@0:   int                 get_fullstrip() const;
michael@0: 
michael@0: private:
michael@0:   int  parse_file(const char * affpath, const char * key);
michael@0:   int  parse_flag(char * line, unsigned short * out, FileMgr * af);
michael@0:   int  parse_num(char * line, int * out, FileMgr * af);
michael@0:   int  parse_cpdsyllable(char * line, FileMgr * af);
michael@0:   int  parse_reptable(char * line, FileMgr * af);
michael@0:   int  parse_convtable(char * line, FileMgr * af, RepList ** rl, const char * keyword);
michael@0:   int  parse_phonetable(char * line, FileMgr * af);
michael@0:   int  parse_maptable(char * line, FileMgr * af);
michael@0:   int  parse_breaktable(char * line, FileMgr * af);
michael@0:   int  parse_checkcpdtable(char * line, FileMgr * af);
michael@0:   int  parse_defcpdtable(char * line, FileMgr * af);
michael@0:   int  parse_affix(char * line, const char at, FileMgr * af, char * dupflags);
michael@0: 
michael@0:   void reverse_condition(char *);
michael@0:   void debugflag(char * result, unsigned short flag);
michael@0:   int condlen(char *);
michael@0:   int encodeit(affentry &entry, char * cs);
michael@0:   int build_pfxtree(PfxEntry* pfxptr);
michael@0:   int build_sfxtree(SfxEntry* sfxptr);
michael@0:   int process_pfx_order();
michael@0:   int process_sfx_order();
michael@0:   PfxEntry * process_pfx_in_order(PfxEntry * ptr, PfxEntry * nptr);
michael@0:   SfxEntry * process_sfx_in_order(SfxEntry * ptr, SfxEntry * nptr);
michael@0:   int process_pfx_tree_to_list();
michael@0:   int process_sfx_tree_to_list();
michael@0:   int redundant_condition(char, char * strip, int stripl,
michael@0:       const char * cond, int);
michael@0: };
michael@0: 
michael@0: #endif
michael@0: