/*============================================================================ Microsoft Simplified Chinese Proofreading Engine Microsoft Confidential. Copyright 1997-1999 Microsoft Corporation. All Rights Reserved. Component: LexDef Purpose: Declare the file structure of lexicon. This is only a header file w/o any CPP, this header will be included by both Lexicon and LexMan module. Notes: We drop this file in Engine sub project only because we want to make Engine code self-contained Owner: donghz@microsoft.com Platform: Win32 Revise: First created by: donghz 12/5/97 ============================================================================*/ #ifndef _LEXDEF_H_ #define _LEXDEF_H_ // define Unicode character blocks #define LEX_LATIN_FIRST 0x0020 #define LEX_LATIN_LAST 0x00bf #define LEX_GENPUNC_FIRST 0x2010 #define LEX_GENPUNC_LAST 0x2046 #define LEX_NUMFORMS_FIRST 0x2153 #define LEX_NUMFORMS_LAST 0x2182 #define LEX_ENCLOSED_FIRST 0x2460 #define LEX_ENCLOSED_LAST 0x24ea #define LEX_CJKPUNC_FIRST 0x3000 #define LEX_CJKPUNC_LAST 0x33ff #define LEX_CJK_FIRST 0x4e00 #define LEX_CJK_LAST 0x9fff #define LEX_FORMS_FIRST 0xff01 #define LEX_FORMS_LAST 0xff64 // basic latin [0x20,0x7e] #define LEX_IDX_OFST_LATIN 0 #define LEX_LATIN_TOTAL (LEX_LATIN_LAST - LEX_LATIN_FIRST + 1) // General punctuation [0x2010,0x2046] #define LEX_IDX_OFST_GENPUNC (LEX_IDX_OFST_LATIN + LEX_LATIN_TOTAL) #define LEX_GENPUNC_TOTAL (LEX_GENPUNC_LAST - LEX_GENPUNC_FIRST + 1) // Number Forms : ¢ñ¢ò¢ó¢ô¢õ¢ö¢÷¢ø¢ù¢ú¢û¢ü ... #define LEX_IDX_OFST_NUMFORMS (LEX_IDX_OFST_GENPUNC + LEX_GENPUNC_TOTAL) #define LEX_NUMFORMS_TOTAL (LEX_NUMFORMS_LAST - LEX_NUMFORMS_FIRST + 1) // Enclosed Alphanumerics; ¢±¢²¢³... ¢Å¢Æ¢Ç... ¢Ù¢Ú¢Û... #define LEX_IDX_OFST_ENCLOSED (LEX_IDX_OFST_NUMFORMS + LEX_NUMFORMS_TOTAL) #define LEX_ENCLOSED_TOTAL (LEX_ENCLOSED_LAST - LEX_ENCLOSED_FIRST + 1) // CJK symbols and punctuation [0x3000,0x301f] #define LEX_IDX_OFST_CJKPUNC (LEX_IDX_OFST_ENCLOSED + LEX_ENCLOSED_TOTAL) #define LEX_CJKPUNC_TOTAL (LEX_CJKPUNC_LAST - LEX_CJKPUNC_FIRST + 1) // CJK unified idographs [0x4e00,0x9fff] #define LEX_IDX_OFST_CJK (LEX_IDX_OFST_CJKPUNC + LEX_CJKPUNC_TOTAL) #define LEX_CJK_TOTAL (LEX_CJK_LAST - LEX_CJK_FIRST + 1) // halfwidth and fullwidth forms [0xff01,0xff64] #define LEX_IDX_OFST_FORMS (LEX_IDX_OFST_CJK + LEX_CJK_TOTAL) #define LEX_FORMS_TOTAL ((LEX_FORMS_LAST - LEX_FORMS_FIRST + 1) + 1) #define LEX_IDX_OFST_OTHER (LEX_IDX_OFST_FORMS + LEX_FORMS_TOTAL) #define LEX_INDEX_COUNT (LEX_IDX_OFST_OTHER + 1) // define encoding/decodeing magic number #define LEX_CJK_MAGIC 0x5000 #define LEX_LATIN_MAGIC 0x8000 #define LEX_GENPUNC_MAGIC (LEX_LATIN_MAGIC + LEX_LATIN_TOTAL) #define LEX_NUMFORMS_MAGIC (LEX_GENPUNC_MAGIC + LEX_GENPUNC_TOTAL) #define LEX_ENCLOSED_MAGIC (LEX_NUMFORMS_MAGIC + LEX_NUMFORMS_TOTAL) #define LEX_CJKPUNC_MAGIC (LEX_ENCLOSED_MAGIC + LEX_ENCLOSED_TOTAL) #define LEX_FORMS_MAGIC 0xff00 // all encoded word in lex would has MostSignificant set to 1 #define LEX_MSBIT 0x8000 // no lex flag and offset mask for lex index item #define LEX_INDEX_NOLEX 0x80000000 #define LEX_OFFSET_MASK 0x7fffffff /*============================================================================ Struct CRTLexHeader Desc: File header of lexicon. The offset of both Index and Lex data section will be defined, and some version info are also defined in this structure ============================================================================*/ #pragma pack(1) struct CRTLexHeader { DWORD m_dwVersion; DWORD m_ofbIndex; // Offset of Index section, it's the length of CLexHeader DWORD m_ofbText; // Offset of Lex section DWORD m_ofbProp; // property, attributes and index of feature set DWORD m_ofbFeature; DWORD m_cbLexSize; // size of the whole lexicon }; /*============================================================================ Struct CRTLexIndex Desc: Index node data structure (m_dwIndex & 0x80000000), if no multi-char word in lexicon. Use (m_dwIndex & 0x7FFFFFFF) to keep tracking the position in lex section The offset is count in 2 bytes, bytes offset = (m_ofwLex * 2) (same as WIC) For words with no property, just set m_ofbProp = 0 ============================================================================*/ struct CRTLexIndex { DWORD m_ofwLex; // offset in the lex area count by WORD ! DWORD m_ofbProp; // offset in the property area }; /*============================================================================ Struct CRTLexRec Desc: Structure of word mark of multi-char word in the lexicon For words with no property, just set m_ofbProp = 0 ============================================================================*/ struct CRTLexRec { DWORD m_ofbProp; // Both HiWord and LoWord's high bit == 0 // follows by the lex text // WCHAR m_rgchLex[]; }; /*============================================================================ Struct CRTLexProp Desc: Structure of lex property ============================================================================*/ struct CRTLexProp { WORD m_iWordID; // WORD m_wFlags; // more property of lex can be add here USHORT m_ciAttri; USHORT m_ciFeature; // follows the attributes and feature index // USHORT m_rgAttri[]; // CRTLexFeature m_rgFertureIndex[]; }; /*============================================================================ Struct CRTLexFeature Desc: Structure of lex feature index ============================================================================*/ struct CRTLexFeature { USHORT m_wFID; USHORT m_cwchLen; DWORD m_ofbFSet; // offset point to the feature area }; /*============================================================================ The feature text of lex stored as "abcd\0efgh\0ijk\0...." There is no separators between features, also no separators between lex feature sets ============================================================================*/ #pragma pack() #endif // _LEXDEF_H_