/*************************************************************************************** * * WRITEPAD(r): Handwriting Recognition Engine (HWRE) and components. * Copyright (c) 2001-2016 PhatWare (r) Corp. All rights reserved. * * Licensing and other inquires: * Developer: Stan Miasnikov, et al. (c) PhatWare Corp. * * WRITEPAD HWRE is free software: you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation, either version 3 of the License, or * (at your option) any later version. * * THE MATERIAL EMBODIED ON THIS SOFTWARE IS PROVIDED TO YOU "AS-IS" * AND WITHOUT WARRANTY OF ANY KIND, EXPRESS, IMPLIED OR OTHERWISE, * INCLUDING WITHOUT LIMITATION, ANY WARRANTY OF MERCHANTABILITY OR * FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL PHATWARE CORP. * BE LIABLE TO YOU OR ANYONE ELSE FOR ANY DIRECT, SPECIAL, INCIDENTAL, * INDIRECT OR CONSEQUENTIAL DAMAGES OF ANY KIND, OR ANY DAMAGES WHATSOEVER, * INCLUDING WITHOUT LIMITATION, LOSS OF PROFIT, LOSS OF USE, SAVINGS * OR REVENUE, OR THE CLAIMS OF THIRD PARTIES, WHETHER OR NOT PHATWARE CORP. * HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH LOSS, HOWEVER CAUSED AND ON * ANY THEORY OF LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE * POSSESSION, USE OR PERFORMANCE OF THIS SOFTWARE. * See the GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with WritePad. If not, see . * **************************************************************************************/ #ifndef PYDICT_H_INCLUDED #define PYDICT_H_INCLUDED //typedef unsigned char uchar; #include "pzdict.h" //-------------- Defines ------------------------------------------------------- #define PYDICT_ID_LEN 16 #define PYDICT_VER_ID (('1' << 0) | ('.' << 8) | ('1' << 16) | ('0' << 24)) #define PYDICT_VER_ID_LEN 4 #define PYDICT_VER_ID_PREV (('1' << 0) | ('.' << 8) | ('0' << 16) | ('1' << 24)) #define PYDICT_TREE_ID_LEN 16 typedef struct { char stamp[5]; unsigned char bType; unsigned short usDataOffset; unsigned int dwDataSize; } SRecDictFileHeader; #define PYDICT_NOERR 0 #define PYDICT_ERR 1 #define PYDICT_OLDWORD 2 #define PYDICT_MAX_WORDLEN HW_MAXWORDLEN #define PZDICT_MAX_WORDLEN HW_MAXWORDLEN #define DICT_INIT 1 /* Empty vertex (the only vertex in empty voc)*/ #define LHDR_STEP_LOG 4 //6 #define LHDR_STEP_MASK 0x0F //0x3F //0011 1111 #define LHDR_STEP (1< 0 ) ? 1 : 0 ) #define PutTreeMerge(pV,b) ( *(int *)VBeg(pV) = (b) ? 1 : 0 ) #define IsVocChanged(pV) ( ( *(int *)VBeg(pV) < 0 ) ? 1 : 0 ) #define PutVocIsChanged(pV) { if (IsTreeMerged(pV)==0) *(int *)VBeg(pV)=-1; } #define PYDictHeaderSize(IsMerged) ( (IsMerged) ? \ PYDICT_VER_ID_LEN+sizeof(int)+sizeof(int)+sizeof(int)+sizeof(int) : \ PYDICT_VER_ID_LEN+sizeof(int)+sizeof(int) ) #define PYDictGetVocHeaderSize(pV) ( PYDictHeaderSize(IsTreeMerged(pV)) ) #define PYDictGetGraphSize(pV) (*(int *)( VBeg(pV)+sizeof(int) )) #define PYDictPutGraphSize(pV,s) ( PYDictGetGraphSize(pV) = s ) #define PYDictGetChsetTablSize(pV) (*(int *)( VBeg(pV)+sizeof(int)+sizeof(int) )) #define PYDictPutChsetTablSize(pV,s) ( PYDictGetChsetTablSize(pV) = s ) #define PYDictGetDvsetTablSize(pV) (*(int *)( VBeg(pV)+sizeof(int)+sizeof(int)+sizeof(int) )) #define PYDictPutDvsetTablSize(pV,s) ( PYDictGetDvsetTablSize(pV) = s ) #define PYDictGetGraph(pV) ( (_UCHAR *)pV + PYDictGetVocHeaderSize(pV) ) #define PYDictGetChsetTabl(pV)( (void *)((_UCHAR *)PYDictGetGraph(pV)+PYDictGetGraphSize(pV)) ) #define PYDictGetDvsetTabl(pV)( (void *)((_UCHAR *)PYDictGetGraph(pV)+PYDictGetGraphSize(pV)+PYDictGetChsetTablSize(pV)) ) //--------- Proto -------------------------------------------------------------- int PYDictGetNextSyms(void *cur_fw, void *fwb, void *pd); int PYDictAddWord(_UCHAR *word, _UCHAR attr, void **pd); int PYDictCreateDict(void **pd); int PYDictFreeDict(void **pd); int PYDictLoadDict(_UCHAR *name, void **pd); int PYDictSaveDict(_UCHAR *name, void *pd); int PYDictCheckWord(_UCHAR *word, _UCHAR *status, _UCHAR *attr,void *pd); int PYDictGetDictStatus(int *len, void *pd); int PYDictGetDictMemSize(void *pVoc); _INT PYDictSaveDictToFile( _UCHAR *name, p_VOID pVoc ); _INT PYDictLoadDictFromFile( _UCHAR *name, p_VOID _PTR pd ); #endif //PYDICT_H_INCLUDED /* *************************************************************************** */ /* * BRIEF DESCRIPTION * */ /* *************************************************************************** * There are 2 types of PYDICT dictionary: PLAIN TREE and MERGED TREE. PLAIN TREE is usual uncompressed dictionary tree; this type is used for User Voc, since new words can be easily added to PLAIN TREE. PYDictCreateDict creates empty PLAIN TREE with PYDICT_MAX_WORDLEN levels; PYDictAddWord adds words to it. Other PYDict functions work with both dict types. MERGED TREE represents a Deterministic Finite State Machine with minimum number of states generating list of words L, i.e. it is a Labeled (i.e. with a letter on each edge) Directed Acyclic Graph G, satisfying the following conditions: (1) Every full path of G represents a word from list L; (2) Every word from list L is represented by a full path of G; (3) Any 2 edges with common starting node are labeled by different symbols; (4) G has minimal (with respect to first 3 properties) number of nodes. Merged Tree is constructed from Plain Tree first by merging leaves (rank 0), then by merging appropriate nodes of rank 1, and so on, (here node rank is defined by max path length from node to a leaf). All edges of final graph G can be divided into 2 sets: 1) non-diagonal (or nd_childs): these are edges from initial tree, each of them lead to a first-in-a-set-of-merging-nodes. 2) diagonal (or d_childs), which appear in the process of merging. As graph G without diagonal edges form a tree structure, it can be represented in a similar to PZDICT format: All nodes are ordered with respect to this tree structure. Graph header contains relative pointers to each level and number of nodes in prev levels. Each level header contains rel. pointer and number of prev nd_childs for each LHDR_STEP-th node, thus # of first (and other) nd_child of a node can be easily calculated by scanning only prev nodes in corresponding segment of LHDR_STEP length. Thus every node should contain only (a) list of symbols for all childs (nd_childs symbols - first) [chset], (b) list of addresses (#-s in graph) for d_childs [dvset]. Those chsets and dvsets, which are frequently used, are coded: sets are extracted in ChsetTabl and DvsetTabl; corresponding nodes in Graph contain only # of a set in a table. (# of a coded set, length of an uncoded dvset and # of a vertex in a dvset can be written down in either long or short form, with corresponding one bit flag). Sets in Tabls are ordered according to their length; for each length there is an entry in Tabl header, which contains length and # and rel. pointer to the first set of this length. Spec. notes: 1. In Plain Tree length of (uncoded) chset is indicated in a node before the chset, either in short or long form. In Merged Tree length is not indicated, last sym in chset is marked by LAST_SYM_FLAG. Thus, chsets, containing sym>=128, should be coded. 2. END_WRD_FLAG is instead additional '\n'-child. 3. One byte node has one child, non-diag, with sym<128; no END_WRD. 4. PLAIN TREE always has PYDICT_MAX_WORDLEN levels; MERGED TREE has only necessary (non-empty) levels. *************************************************************************** */