1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/extensions/spellcheck/hunspell/src/affixmgr.hxx Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,307 @@
1.4 +/******* BEGIN LICENSE BLOCK *******
1.5 + * Version: MPL 1.1/GPL 2.0/LGPL 2.1
1.6 + *
1.7 + * The contents of this file are subject to the Mozilla Public License Version
1.8 + * 1.1 (the "License"); you may not use this file except in compliance with
1.9 + * the License. You may obtain a copy of the License at
1.10 + * http://www.mozilla.org/MPL/
1.11 + *
1.12 + * Software distributed under the License is distributed on an "AS IS" basis,
1.13 + * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
1.14 + * for the specific language governing rights and limitations under the
1.15 + * License.
1.16 + *
1.17 + * The Initial Developers of the Original Code are Kevin Hendricks (MySpell)
1.18 + * and László Németh (Hunspell). Portions created by the Initial Developers
1.19 + * are Copyright (C) 2002-2005 the Initial Developers. All Rights Reserved.
1.20 + *
1.21 + * Contributor(s): Kevin Hendricks (kevin.hendricks@sympatico.ca)
1.22 + * David Einstein (deinst@world.std.com)
1.23 + * László Németh (nemethl@gyorsposta.hu)
1.24 + * Caolan McNamara (caolanm@redhat.com)
1.25 + * Davide Prina
1.26 + * Giuseppe Modugno
1.27 + * Gianluca Turconi
1.28 + * Simon Brouwer
1.29 + * Noll Janos
1.30 + * Biro Arpad
1.31 + * Goldman Eleonora
1.32 + * Sarlos Tamas
1.33 + * Bencsath Boldizsar
1.34 + * Halacsy Peter
1.35 + * Dvornik Laszlo
1.36 + * Gefferth Andras
1.37 + * Nagy Viktor
1.38 + * Varga Daniel
1.39 + * Chris Halls
1.40 + * Rene Engelhard
1.41 + * Bram Moolenaar
1.42 + * Dafydd Jones
1.43 + * Harri Pitkanen
1.44 + * Andras Timar
1.45 + * Tor Lillqvist
1.46 + *
1.47 + * Alternatively, the contents of this file may be used under the terms of
1.48 + * either the GNU General Public License Version 2 or later (the "GPL"), or
1.49 + * the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
1.50 + * in which case the provisions of the GPL or the LGPL are applicable instead
1.51 + * of those above. If you wish to allow use of your version of this file only
1.52 + * under the terms of either the GPL or the LGPL, and not to allow others to
1.53 + * use your version of this file under the terms of the MPL, indicate your
1.54 + * decision by deleting the provisions above and replace them with the notice
1.55 + * and other provisions required by the GPL or the LGPL. If you do not delete
1.56 + * the provisions above, a recipient may use your version of this file under
1.57 + * the terms of any one of the MPL, the GPL or the LGPL.
1.58 + *
1.59 + ******* END LICENSE BLOCK *******/
1.60 +
1.61 +#ifndef _AFFIXMGR_HXX_
1.62 +#define _AFFIXMGR_HXX_
1.63 +
1.64 +#include "hunvisapi.h"
1.65 +
1.66 +#include <stdio.h>
1.67 +
1.68 +#include "atypes.hxx"
1.69 +#include "baseaffix.hxx"
1.70 +#include "hashmgr.hxx"
1.71 +#include "phonet.hxx"
1.72 +#include "replist.hxx"
1.73 +
1.74 +// check flag duplication
1.75 +#define dupSFX (1 << 0)
1.76 +#define dupPFX (1 << 1)
1.77 +
1.78 +class PfxEntry;
1.79 +class SfxEntry;
1.80 +
1.81 +class LIBHUNSPELL_DLL_EXPORTED AffixMgr
1.82 +{
1.83 +
1.84 + PfxEntry * pStart[SETSIZE];
1.85 + SfxEntry * sStart[SETSIZE];
1.86 + PfxEntry * pFlag[SETSIZE];
1.87 + SfxEntry * sFlag[SETSIZE];
1.88 + HashMgr * pHMgr;
1.89 + HashMgr ** alldic;
1.90 + int * maxdic;
1.91 + char * keystring;
1.92 + char * trystring;
1.93 + char * encoding;
1.94 + struct cs_info * csconv;
1.95 + int utf8;
1.96 + int complexprefixes;
1.97 + FLAG compoundflag;
1.98 + FLAG compoundbegin;
1.99 + FLAG compoundmiddle;
1.100 + FLAG compoundend;
1.101 + FLAG compoundroot;
1.102 + FLAG compoundforbidflag;
1.103 + FLAG compoundpermitflag;
1.104 + int checkcompounddup;
1.105 + int checkcompoundrep;
1.106 + int checkcompoundcase;
1.107 + int checkcompoundtriple;
1.108 + int simplifiedtriple;
1.109 + FLAG forbiddenword;
1.110 + FLAG nosuggest;
1.111 + FLAG nongramsuggest;
1.112 + FLAG needaffix;
1.113 + int cpdmin;
1.114 + int numrep;
1.115 + replentry * reptable;
1.116 + RepList * iconvtable;
1.117 + RepList * oconvtable;
1.118 + int nummap;
1.119 + mapentry * maptable;
1.120 + int numbreak;
1.121 + char ** breaktable;
1.122 + int numcheckcpd;
1.123 + patentry * checkcpdtable;
1.124 + int simplifiedcpd;
1.125 + int numdefcpd;
1.126 + flagentry * defcpdtable;
1.127 + phonetable * phone;
1.128 + int maxngramsugs;
1.129 + int maxcpdsugs;
1.130 + int maxdiff;
1.131 + int onlymaxdiff;
1.132 + int nosplitsugs;
1.133 + int sugswithdots;
1.134 + int cpdwordmax;
1.135 + int cpdmaxsyllable;
1.136 + char * cpdvowels;
1.137 + w_char * cpdvowels_utf16;
1.138 + int cpdvowels_utf16_len;
1.139 + char * cpdsyllablenum;
1.140 + const char * pfxappnd; // BUG: not stateless
1.141 + const char * sfxappnd; // BUG: not stateless
1.142 + FLAG sfxflag; // BUG: not stateless
1.143 + char * derived; // BUG: not stateless
1.144 + SfxEntry * sfx; // BUG: not stateless
1.145 + PfxEntry * pfx; // BUG: not stateless
1.146 + int checknum;
1.147 + char * wordchars;
1.148 + unsigned short * wordchars_utf16;
1.149 + int wordchars_utf16_len;
1.150 + char * ignorechars;
1.151 + unsigned short * ignorechars_utf16;
1.152 + int ignorechars_utf16_len;
1.153 + char * version;
1.154 + char * lang;
1.155 + int langnum;
1.156 + FLAG lemma_present;
1.157 + FLAG circumfix;
1.158 + FLAG onlyincompound;
1.159 + FLAG keepcase;
1.160 + FLAG forceucase;
1.161 + FLAG warn;
1.162 + int forbidwarn;
1.163 + FLAG substandard;
1.164 + int checksharps;
1.165 + int fullstrip;
1.166 +
1.167 + int havecontclass; // boolean variable
1.168 + char contclasses[CONTSIZE]; // flags of possible continuing classes (twofold affix)
1.169 +
1.170 +public:
1.171 +
1.172 + AffixMgr(const char * affpath, HashMgr** ptr, int * md,
1.173 + const char * key = NULL);
1.174 + ~AffixMgr();
1.175 + struct hentry * affix_check(const char * word, int len,
1.176 + const unsigned short needflag = (unsigned short) 0,
1.177 + char in_compound = IN_CPD_NOT);
1.178 + struct hentry * prefix_check(const char * word, int len,
1.179 + char in_compound, const FLAG needflag = FLAG_NULL);
1.180 + inline int isSubset(const char * s1, const char * s2);
1.181 + struct hentry * prefix_check_twosfx(const char * word, int len,
1.182 + char in_compound, const FLAG needflag = FLAG_NULL);
1.183 + inline int isRevSubset(const char * s1, const char * end_of_s2, int len);
1.184 + struct hentry * suffix_check(const char * word, int len, int sfxopts,
1.185 + PfxEntry* ppfx, char ** wlst, int maxSug, int * ns,
1.186 + const FLAG cclass = FLAG_NULL, const FLAG needflag = FLAG_NULL,
1.187 + char in_compound = IN_CPD_NOT);
1.188 + struct hentry * suffix_check_twosfx(const char * word, int len,
1.189 + int sfxopts, PfxEntry* ppfx, const FLAG needflag = FLAG_NULL);
1.190 +
1.191 + char * affix_check_morph(const char * word, int len,
1.192 + const FLAG needflag = FLAG_NULL, char in_compound = IN_CPD_NOT);
1.193 + char * prefix_check_morph(const char * word, int len,
1.194 + char in_compound, const FLAG needflag = FLAG_NULL);
1.195 + char * suffix_check_morph (const char * word, int len, int sfxopts,
1.196 + PfxEntry * ppfx, const FLAG cclass = FLAG_NULL,
1.197 + const FLAG needflag = FLAG_NULL, char in_compound = IN_CPD_NOT);
1.198 +
1.199 + char * prefix_check_twosfx_morph(const char * word, int len,
1.200 + char in_compound, const FLAG needflag = FLAG_NULL);
1.201 + char * suffix_check_twosfx_morph(const char * word, int len,
1.202 + int sfxopts, PfxEntry * ppfx, const FLAG needflag = FLAG_NULL);
1.203 +
1.204 + char * morphgen(char * ts, int wl, const unsigned short * ap,
1.205 + unsigned short al, char * morph, char * targetmorph, int level);
1.206 +
1.207 + int expand_rootword(struct guessword * wlst, int maxn, const char * ts,
1.208 + int wl, const unsigned short * ap, unsigned short al, char * bad,
1.209 + int, char *);
1.210 +
1.211 + short get_syllable (const char * word, int wlen);
1.212 + int cpdrep_check(const char * word, int len);
1.213 + int cpdpat_check(const char * word, int len, hentry * r1, hentry * r2,
1.214 + const char affixed);
1.215 + int defcpd_check(hentry *** words, short wnum, hentry * rv,
1.216 + hentry ** rwords, char all);
1.217 + int cpdcase_check(const char * word, int len);
1.218 + inline int candidate_check(const char * word, int len);
1.219 + void setcminmax(int * cmin, int * cmax, const char * word, int len);
1.220 + struct hentry * compound_check(const char * word, int len, short wordnum,
1.221 + short numsyllable, short maxwordnum, short wnum, hentry ** words,
1.222 + char hu_mov_rule, char is_sug, int * info);
1.223 +
1.224 + int compound_check_morph(const char * word, int len, short wordnum,
1.225 + short numsyllable, short maxwordnum, short wnum, hentry ** words,
1.226 + char hu_mov_rule, char ** result, char * partresult);
1.227 +
1.228 + struct hentry * lookup(const char * word);
1.229 + int get_numrep() const;
1.230 + struct replentry * get_reptable() const;
1.231 + RepList * get_iconvtable() const;
1.232 + RepList * get_oconvtable() const;
1.233 + struct phonetable * get_phonetable() const;
1.234 + int get_nummap() const;
1.235 + struct mapentry * get_maptable() const;
1.236 + int get_numbreak() const;
1.237 + char ** get_breaktable() const;
1.238 + char * get_encoding();
1.239 + int get_langnum() const;
1.240 + char * get_key_string();
1.241 + char * get_try_string() const;
1.242 + const char * get_wordchars() const;
1.243 + unsigned short * get_wordchars_utf16(int * len) const;
1.244 + char * get_ignore() const;
1.245 + unsigned short * get_ignore_utf16(int * len) const;
1.246 + int get_compound() const;
1.247 + FLAG get_compoundflag() const;
1.248 + FLAG get_compoundbegin() const;
1.249 + FLAG get_forbiddenword() const;
1.250 + FLAG get_nosuggest() const;
1.251 + FLAG get_nongramsuggest() const;
1.252 + FLAG get_needaffix() const;
1.253 + FLAG get_onlyincompound() const;
1.254 + FLAG get_compoundroot() const;
1.255 + FLAG get_lemma_present() const;
1.256 + int get_checknum() const;
1.257 + const char * get_prefix() const;
1.258 + const char * get_suffix() const;
1.259 + const char * get_derived() const;
1.260 + const char * get_version() const;
1.261 + int have_contclass() const;
1.262 + int get_utf8() const;
1.263 + int get_complexprefixes() const;
1.264 + char * get_suffixed(char ) const;
1.265 + int get_maxngramsugs() const;
1.266 + int get_maxcpdsugs() const;
1.267 + int get_maxdiff() const;
1.268 + int get_onlymaxdiff() const;
1.269 + int get_nosplitsugs() const;
1.270 + int get_sugswithdots(void) const;
1.271 + FLAG get_keepcase(void) const;
1.272 + FLAG get_forceucase(void) const;
1.273 + FLAG get_warn(void) const;
1.274 + int get_forbidwarn(void) const;
1.275 + int get_checksharps(void) const;
1.276 + char * encode_flag(unsigned short aflag) const;
1.277 + int get_fullstrip() const;
1.278 +
1.279 +private:
1.280 + int parse_file(const char * affpath, const char * key);
1.281 + int parse_flag(char * line, unsigned short * out, FileMgr * af);
1.282 + int parse_num(char * line, int * out, FileMgr * af);
1.283 + int parse_cpdsyllable(char * line, FileMgr * af);
1.284 + int parse_reptable(char * line, FileMgr * af);
1.285 + int parse_convtable(char * line, FileMgr * af, RepList ** rl, const char * keyword);
1.286 + int parse_phonetable(char * line, FileMgr * af);
1.287 + int parse_maptable(char * line, FileMgr * af);
1.288 + int parse_breaktable(char * line, FileMgr * af);
1.289 + int parse_checkcpdtable(char * line, FileMgr * af);
1.290 + int parse_defcpdtable(char * line, FileMgr * af);
1.291 + int parse_affix(char * line, const char at, FileMgr * af, char * dupflags);
1.292 +
1.293 + void reverse_condition(char *);
1.294 + void debugflag(char * result, unsigned short flag);
1.295 + int condlen(char *);
1.296 + int encodeit(affentry &entry, char * cs);
1.297 + int build_pfxtree(PfxEntry* pfxptr);
1.298 + int build_sfxtree(SfxEntry* sfxptr);
1.299 + int process_pfx_order();
1.300 + int process_sfx_order();
1.301 + PfxEntry * process_pfx_in_order(PfxEntry * ptr, PfxEntry * nptr);
1.302 + SfxEntry * process_sfx_in_order(SfxEntry * ptr, SfxEntry * nptr);
1.303 + int process_pfx_tree_to_list();
1.304 + int process_sfx_tree_to_list();
1.305 + int redundant_condition(char, char * strip, int stripl,
1.306 + const char * cond, int);
1.307 +};
1.308 +
1.309 +#endif
1.310 +
