<!doctype linuxdoc system>
<!-- @ ʵ -->

<article>

<!-- Title information -->
<title>libtabe -- TaBE 禡w

<author>ʵ <it><htmlurl url="mailto:shawn@iis.sinica.edu.tw"
                     name="&lt;shawn@iis.sinica.edu.tw&gt;"></it></author>
<date>$Id: libtabe.sgml,v 1.2 2001/08/20 03:53:02 thhsieh Exp $

<abstract>
bqWBzA즳 C 禡w@QXC
H Big5 XӻA@ӦrF byteAެOprסA
Ϊ̬OnҼ{rrǡAƦܭnBzΥylɭԡAݭnSOվA
즳 C 禡wڥŨϥΡC
t@譱AѩCӵ{}oHbBz媺ɭԡAۤvkA
ҥHP{ɨ禡wA󤣥δЧQΤFC
TaBE peƱѤ@ӧnϥҡA]A
Nܻݭn@ӯBzrBByΤ@P禡wA
ӰҦε{¦C
libtabeANOƱ榨o˪@M禡wAѲΤ@P\A
ε{jj¦iHoC
</abstract>

<!-- Table of contents -->
<toc>

<!-- Begin of document -->
<sect>
<p>
libtabe N媺BzXӶqAѤpjOOG
<itemize>
<item>`Ÿ(ZuYin Symbols)G
      C@ӤrѤ@өΦhӪ`Ÿ[W@nղŸҲզC
      `Ÿ[WnղŸ`ƬO|QG
      (TQCӪ`Ÿ[WnղŸ)
<item>Ū(Yin)G
      @ŪOѤ@өΦhӪ`Ÿ[W@nղŸզC
      CŪO㦳NqAiHN@ӤrCMA
      Pr]\|ۦPŪC@Ӧr]i঳ƺŪC
<item>r(Zhi)G
      ]NOrC
<item>(TsiYin)G
      @ӵŪC@ӵi঳ƺŪC
<item>(Tsi)G
      Ѥ@ƭӦrҲզCbyWAO̰򥻤S㦳NqC
<item>y(Chu)G
      @ӥyliHF@ӧ㪺AѤ@ƭӵҲզC
</itemize>
</p>

<p>
libtabe wHWUӳBzqѤ@Ǩ禡C]b{BzqA
ഫ禡C᪺`AN@@ШCӳBzqҴѪ禡A
λP۾FBzqഫ禡C
</p>

<p>
ϥΤPsXBzɡAi|PʽA
ҥHoӨ禡wĥ Big5 XXAݭnഫLXɡA
iHϥഫơC
</p>
</sect>

<sect>Ƶc
<sect1>e
<p>
b`Aڭ̱NЦb libtabe ҷ|ϥΨ쪺XӰ򥻸ƵcC
ǸƵcMuOLƧκAswqӤwALA
FP C yƧκA۲VcAڭ٬Oϥγo˪覡C
</p>

<sect1>wq
<p>
<code>
typedef unsigned char    *ZuYinSymbol;
typedef unsigned char    *ZuYinSymbolSequence;
typedef unsigned int      ZuYinIndex;

typedef unsigned int      ZhiCode;
typedef unsigned char    *Zhi;
typedef unsigned char    *ZhiStr;

typedef unsigned long int Yin;

struct ZhiInfo {
  ZhiCode           code;
  Zhi               chct;
  Yin               yin[4];
  unsigned long int refcount; /* should be obsoleted soon */
};

struct TsiInfo {
  ZhiStr             tsi;
  unsigned long int  refcount;
  unsigned long int  yinnum;
  Yin               *yindata;
};

struct TsiYinInfo {
  Yin               *yin ;
  unsigned long int  yinlen;
  unsigned long int  tsinum;
  ZhiStr             tsidata;
};

struct ChunkInfo {
  ZhiStr          chunk;
  int             num_tsi;
  struct TsiInfo *tsi;
};

struct ChuInfo {
  ZhiStr            chu;
  int               num_chunk;
  struct ChunkInfo *chunk;
};

struct TsiDB {
  int type;
  int flags;
  char *db_name;
  void *dbp;
  void *dbcp;
  .
  .
  .
};

struct TsiYinDB {
  int type;
  int flags;
  char *db_name;
  void *dbp;
  void *dbcp;
  .
  .
  .
};
</code>
</p>
</sect1>

<sect1>
<p>
<itemize>
<item>ZhiCodeGs Big5 XC
<item>ZhiGs@ Big5 XܪrCo@wO@ӦrA
           iu byte ӤwC
<item>ZhiStrGs@ Big5 XܪrCo@wO@ӦrA
                NULL ׵C
<item>YinGs@ŪC
<item>ZhiInfoG@ӤrƵcCzLoӸƵcA
               iHo\hoӦrTCyin soӦrŪA
               ̦h|ӡC
<item>TsiInfoG@ӵƵcCzLoӸƵcA
               iHo\hoӦrTC
               yinnum P yindata s񪺬OoӵTŪA
               ]\u@ӡC
<item>TsiYinInfoG@ӵƵcCyinlen OoӵסA
                  tsinum OoŪ@XӵA
                  tsidata Ooǵ۱sC
<item>TsiDBGwƵcCzLoӸƵcA
             iHswAPLTC
<item>TsiYinDBGwƵcCzLoӸƵcA
                iHswPAPLTC
<item>ChunkInfoGΨ_ƵcCn_qsbo̡A
                 _XӪhsb tsi C
<item>ChuInfoGyz@ӥylƵcC@ӥylSiAƭӤqC
</itemize>
</sect1>
</sect>

<sect>`Ÿ
<sect1>e
<p>
`ŸOxWaϩҦHbDz߰yɡAΨӾDzߵokuC
o˪kAiHTܤ@ӦrŪA|yVcCҥHA
libtabe ϥΪ`ŸŪpBzC
</p>
<p>
`Ÿ@TQCӡAA[WnղŸA`pO|QGӡC
FKbqBz`ŸAlibtabe q "t"  ""A
̧ǵP 1  37 `s (Index)C38 On()A
39  42 OO@n|nC0 OdϥΡAHKPŪBzqഫC
</p>
</sect1>

<sect1>tabeZuYinIndexToZuYinSymbol()
<p>
<tscreen>
const Zhi tabeZuYinIndexToZuyinSymbol(ZuYinIndex idx);
</tscreen>
</p>

<p>
beҴL`sX (Index) P`Ÿ (ZuYinSymbol) ഫC
</p>
</sect1>

<sect1>tabeZuYinSymbolToZuYinIndex()
<p>
<tscreen>
int tabeZuYinSymbolToZuYinIndex(ZuYinSymbol sym);
</tscreen>
</p>

<p>
P tabeZuYinIndexToZuYinSymbol() ۦAuLOۤϪഫC
</p>
</sect1>

<sect1>tabeZozyKeyToZuYinIndex()
<p>
<tscreen>
int tabeZozyKeyToZuYinIndex(int key);
</tscreen>
</p>
<p>
禡ѹs@`JkP`sഫC
</p>
</sect1>
</sect>

<sect>Ū
<sect1>e
<p>
qLqs| (CMEX) ҴѪݩʪέpoA
Big5 Xҥ]tr@ 1302 ӤPŪC
ӨCŪOѪ`Ÿ[nղŸҲզA
ҥHڭ̴NΪ`sXӲզŪsXAKഫC
SAFPDzΪϥκDҵXAŪOĤ@nɡAڭ̨äϥθnղŸC
</p>

<p>
軡A"" oŪNO
((((<bf>15</bf>*43)+<bf>23</bf>)*43+<bf>36</bf>)*43)+<bf>0</bf>)*43C
15B23B36 OO ""B""B"" `sC
ѩ@Ū̦hTӪ`Ÿ[@nղŸA
ĥγoؤ覡N|VcA]ݭn`ŸX{mSC
</p>

<p>
CMEX ݩɤA@Ӥr̦h|ŪC
b Big5 XDΪ 13060 ӤrA12098 Ӧru@ŪA
888 ӦrŪA62 ӦrTŪA10 Ӧr|ŪCt~A
٦ӦrSŪAOOJ (0xC94A) P (0xDDFC) oӦ`ΦrC
(]L̬OЪrCӥB CMEX rɤ]SoӦr)
ѩڭ̨ϥ CMEX ݩXӪӪAҥHoӦrNŪC
</p>
</sect1>

<sect1>tabeYinToZuYinSymbolSequence()
<p>
<tscreen>
ZuYinSymbolSequence tabeYinToZuyinSymbolSequence(Yin yin);
</tscreen>
</p>
<p>
NŪରҲզ`ŸPnղŸC
</p>
</sect1>

<sect1>tabeZuYinSymbolSequenceToYin()
<p>
<tscreen>
Yin tabeZuYinSymbolSequenceToYin(ZuYinSymbolSequence str);
</tscreen>
</p>
<p>
N`ŸPnղŸզXনŪC
</p>
</sect1>

<sect1>tabeYinLookupZhiList()
<p>
<tscreen>
ZhiStr tabeYinLookupZhiList(Yin yin);
</tscreen>
</p>

<p>
dߩҦ㦳oŪrCoǦrsb@ӦrꤤC
</p>
</sect>

<sect>r
<sect1>e
<p>
bxWҺDΪ Big5 XA@ 13060 ӤrA
OO 5401 ӱ`ΦrA7652 Ӧ`ΦrAA[Wʤѩr 7 ӤrC
</p>
</sect1>

<sect1>tabeZhiInfoLookupYin()
<p>
<tscreen>int tabeZhiInfoLookupYin(struct ZhiInfo *h);
</tscreen>
</p>

<p>
d ZhiInfo oӸƵc code ҥNrҦŪA
sb ZhiInfo  yin CǦ^ȦpGp 0ANӦrSŪC
</p>
</sect1>
</sect>

<sect>
<sect1>e
<p>
boM禡wAO̭nBzqC襤yӻA
^夤r@ˡAO㦳Nq̤pCO^夣POA
媺PS㪺jAάO^夤ťդ@ˡA
iHϹjӳrjŸC
</p>
<p>
]Aڭ̻ݭn_C_N@ꤤrjXӡC
</p>
<p>
媺_ثeOĥΤw覡ӧ@A
ҥH@ӨjBIJvnwOݭnC
ҥHAlibtabe  Berkeley DB Ӻ޲zwAp@ӡAJϵwܤjɡA
]ٯܦnįC
</p>
</sect1>
<sect1>tabeTsiInfoLookupPossibleYin()
<p>
<tscreen>
int tabeBig5TsiLookupPossibleTsiYin(struct TsiInfo *tsi);
</tscreen>
dX TsiInfo  tsi oӦrҦC
`Ʀsb TsiInfo  yinnum Ayindata hOoǵAssC
<p>

<sect1>tabeTsiDBOpen()
<p>
<tscreen>
struct TsiDB *tabeTsiDBOpen(int type, const char *db_name, int flags);
</tscreen>
</p>
<p>
}Ҥ@ӵwCtype OwAثeu DB_TYPE_DB @ءA
]NO Berkeley DBCflags OoӸƮw@ǯSʡApUG
<itemize>
<item>DB_FLAG_OVERWRITEGpGbgJwɤwЪAмgC
<item>DB_FLAG_CREATEDBGpG}ҵwɧ䤣즹wAsWɡC
<item>DB_FLAG_READONLYG}ҵwŪACPe flag VΡC
</itemize>
</p>
<p>
Ǧ^ TsiDB NOIsLwBzƮɲĤ@ӰѼơC
</p>
</sect1>

<sect1>tsidb->Close()
<p>
<tscreen>
void tsidb->Close(struct TsiDB *tsidb);
</tscreen>
</p>
<p>
wC
</p>
</sect1>

<sect1>tsidb->RecordNumber()
<p>
<tscreen>
int tsidb->RecordNumber(struct TsiDB *tsidb);
</tscreen>
</p>
<p>
d߸ӵw@hֵC
</p>
</sect1>

<sect1>tsidb->Put()
<p>
<tscreen>
int tsidb->Put(struct TsiDB *tsidb, struct TsiInfo *tsi);
</tscreen>
</p>
<p>
N TsiInfo  tsiBrefcountByinnumByindata sJwC
</p>
</sect1>

<sect1>tsidb->Get()
<p>
<tscreen>
int tsidb->Get(struct TsiDB *tsidb, struct TsiInfo *tsi);
</tscreen>
</p>
<p>
dߵwO_ TsiInfo  tsi oӵC
ܱNLƥѵwJ TsiInfo C
</p>
</sect1>

<sect1>tsidb->CursorSet()
<p>
<tscreen>
int tsidb->CursorSet(struct TsiDB *tsidb, struct TsiInfo *tsi);
</tscreen>
</p>
<p>
Nwг] TsiInfo  tsi ҫVaCpG tsi sbA
hVĤ@OCPɤ]NӵJ TsiInfo C
</p>
</sect1>

<sect1>tsidb->CursorNext()
<p>
<tscreen>
int tsidb->CursorNext(struct TsiDB *tsidb, struct TsiInfo *tsi);
</tscreen>
</p>
<p>
ŪU@C
</p>
</sect1>

<sect1>tsidb->CursorPrev()
<p>
<tscreen>
int tsidb->CursorPrev(struct TsiDB *tsidb, struct TsiInfo *tsi);
</tscreen>
</p>
<p>
ŪW@C
</p>
</sect1>
</sect>

<sect>
<sect1>e
<p>
ŪC
</p>
</sect1>
<sect1>tabeTsiYinDBOpen()
<p>
<tscreen>
struct TsiYinDB *tabeTsiYinDBOpen(int type, const char *db_name, int flags);
</tscreen>
</p>
<p>
}Ҥ@ӵwCtype OwAثeu DB_TYPE_DB @ءA
]NO Berkeley DBCflags OoӸƮw@ǯSʡApUG
<itemize>
<item>DB_FLAG_OVERWRITEGpGbgJwɤwЪAмgC
<item>DB_FLAG_CREATEDBGpG}ҵwɧ䤣즹wAsWɡC
<item>DB_FLAG_READONLYG}ҵwŪACPe flag VΡC
</itemize>
</p>
<p>
Ǧ^ TsiYinDB NOIsLwBzƮɲĤ@ӰѼơC
</p>
</sect1>

<sect1>yindb->Close()
<p>
<tscreen>
void yindb->Close(struct TsiYinDB *yindb);
</tscreen>
</p>
<p>
wC
</p>
</sect1>

<sect1>yindb->RecordNumber()
<p>
<tscreen>
int tabeTsiYinDBRecordNumber(struct Big5TsiYinDB *yindb);
</tscreen>
</p>
<p>
d߸ӵw@hֵC
</p>
</sect1>

<sect1>yindb->Put()
<p>
<tscreen>
int yindb->Put(struct TsiYinDB *yindb, struct TsiYinInfo *yin);
</tscreen>
</p>
<p>
N TsiYinInfo  yinByinlenBtsinumBtsidata sJwC
</p>
</sect1>

<sect1>yindb->Get()
<p>
<tscreen>
int yindb->Get(struct TsiYinDB *yindb, struct TsiYinInfo *yin);
</tscreen>
</p>
<p>
dߵwO_ TsiYinInfo  yin oӵC
ܱNLƥѵwJ TsiYinInfo C
</p>
</sect1>

<sect1>yindb->CursorSet()
<p>
<tscreen>
int yindb->CursorSet(struct TsiYinDB *yindb, struct TsiYinInfo *tsiyin);
</tscreen>
</p>
<p>
Nwг] TsiYinInfo  yin ҫVaCpG yin sbA
hVĤ@OCPɤ]NӵJ TsiYinInfo C
</p>
</sect1>

<sect1>yindb->CursorNext()
<p>
<tscreen>
int yindb->CursorNext(struct TsiYinDB *yindb, struct TsiYinInfo *yin);
</tscreen>
</p>
<p>
ŪU@C
</p>
</sect1>

<sect1>yindb->CursorPrev()
<p>
<tscreen>
int yindb->CursorPrev(struct TsiYinDB *yindb, struct TsiYinInfo *yin);
</tscreen>
</p>
<p>
ŪW@C
</p>
</sect1>
</sect>

<sect>y
<sect1>e
<p>
FHPqbrPBzWošA
ڭ̳]wHqq̤pOylCLylwqӹLҽkA
A[WӤHbqWϥΤ媺ߺDɬۦPAҥH󬰡yyzɭɩwC
yqzOڭ̦byP[J@ӤCC
@ӡyqzrҲզADrҰϹjC@ӥyliH\hqC
</p>
<p>
wylBzAF⥦ӤrA󭫭nOn[C
ҥH_OoӶqnBzơC
</p>
</sect1>

<sect1>tabeChuInfoToChunkInfo()
<p>
<tscreen>
int tabeChuInfoToChunkInfo(struct ChuInfo *chu);
</tscreen>
</p>
<p>
N ChuInfo  ChunkInfo XӡCChunkInfo `Ʀs num_chunk A
 chunk[i] O i+1  ChunkInfoC
</p>
</sect1>

<sect1>tabeChunkInfoSegmentationSimplex()
<p>
<tscreen>
int tabeChunkInfoSegmentationSimplex(struct TsiDB *tsidb,
                                     struct ChunkInfo *chunk);
</tscreen>
</p>
<p>
ϥγ̰򥻪yukhzN ChunkInfo XӡC
Cӵs@ TsiInfo ƵcAtsi[i] O i+1 ӵC
`Ʀs num_tsi C
</p>
</sect1>

<sect1>tabeChunkInfoSegmentationComplex()
<p>
<tscreen>
int tabeChunkInfoSegmentationComplex(struct TsiDB *tsidb,
                                     struct ChunkInfo *chunk);
</tscreen>
</p>
<p>
ϥνӯE_tkN ChunkInfo XӡC
Cӵs@ TsiInfo ƵcAtsi[i] O i+1 ӵC
`Ʀs num_tsi C
</p>
<p>
ӯE_tkiѦ
<htmlurl
 url="http://casper.beckman.uiuc.edu/~c-tsai4/chinese/wordseg/mmseg.html"
name="http://casper.beckman.uiuc.edu/~c-tsai4/chinese/wordseg/mmseg.html">
</p>
</sect1>

<sect1>tabeChunkInfoSegmentationBackward()
<p>
<tscreen>
int tabeChunkInfoSegmentationBackward(struct TsiDB *tsidb,
                                      struct ChunkInfo *chunk);
</tscreen>
</p>
<p>
ϥΪLŵ (shlin@iis.sinica.edu.tw) _tkN ChunkInfo
XӡCCӵs@ TsiInfo ƵcA
tsi[i] O i+1 ӵC`Ʀs num_tsi C
</p>
</sect1>
</sect>

</article>