1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/os/persistentdata/persistentstorage/sqlite3api/SQLite/fts1_tokenizer1.c Fri Jun 15 03:10:57 2012 +0200
1.3 @@ -0,0 +1,221 @@
1.4 +/*
1.5 +** The author disclaims copyright to this source code.
1.6 +**
1.7 +*************************************************************************
1.8 +** Implementation of the "simple" full-text-search tokenizer.
1.9 +*/
1.10 +
1.11 +/*
1.12 +** The code in this file is only compiled if:
1.13 +**
1.14 +** * The FTS1 module is being built as an extension
1.15 +** (in which case SQLITE_CORE is not defined), or
1.16 +**
1.17 +** * The FTS1 module is being built into the core of
1.18 +** SQLite (in which case SQLITE_ENABLE_FTS1 is defined).
1.19 +*/
1.20 +#if !defined(SQLITE_CORE) || defined(SQLITE_ENABLE_FTS1)
1.21 +
1.22 +
1.23 +#include <assert.h>
1.24 +#include <stdlib.h>
1.25 +#include <stdio.h>
1.26 +#include <string.h>
1.27 +#include <ctype.h>
1.28 +
1.29 +#include "fts1_tokenizer.h"
1.30 +
1.31 +typedef struct simple_tokenizer {
1.32 + sqlite3_tokenizer base;
1.33 + char delim[128]; /* flag ASCII delimiters */
1.34 +} simple_tokenizer;
1.35 +
1.36 +typedef struct simple_tokenizer_cursor {
1.37 + sqlite3_tokenizer_cursor base;
1.38 + const char *pInput; /* input we are tokenizing */
1.39 + int nBytes; /* size of the input */
1.40 + int iOffset; /* current position in pInput */
1.41 + int iToken; /* index of next token to be returned */
1.42 + char *pToken; /* storage for current token */
1.43 + int nTokenAllocated; /* space allocated to zToken buffer */
1.44 +} simple_tokenizer_cursor;
1.45 +
1.46 +
1.47 +/* Forward declaration */
1.48 +static const sqlite3_tokenizer_module simpleTokenizerModule;
1.49 +
1.50 +static int isDelim(simple_tokenizer *t, unsigned char c){
1.51 + return c<0x80 && t->delim[c];
1.52 +}
1.53 +
1.54 +/*
1.55 +** Create a new tokenizer instance.
1.56 +*/
1.57 +static int simpleCreate(
1.58 + int argc, const char * const *argv,
1.59 + sqlite3_tokenizer **ppTokenizer
1.60 +){
1.61 + simple_tokenizer *t;
1.62 +
1.63 + t = (simple_tokenizer *) calloc(sizeof(*t), 1);
1.64 + if( t==NULL ) return SQLITE_NOMEM;
1.65 +
1.66 + /* TODO(shess) Delimiters need to remain the same from run to run,
1.67 + ** else we need to reindex. One solution would be a meta-table to
1.68 + ** track such information in the database, then we'd only want this
1.69 + ** information on the initial create.
1.70 + */
1.71 + if( argc>1 ){
1.72 + int i, n = strlen(argv[1]);
1.73 + for(i=0; i<n; i++){
1.74 + unsigned char ch = argv[1][i];
1.75 + /* We explicitly don't support UTF-8 delimiters for now. */
1.76 + if( ch>=0x80 ){
1.77 + free(t);
1.78 + return SQLITE_ERROR;
1.79 + }
1.80 + t->delim[ch] = 1;
1.81 + }
1.82 + } else {
1.83 + /* Mark non-alphanumeric ASCII characters as delimiters */
1.84 + int i;
1.85 + for(i=1; i<0x80; i++){
1.86 + t->delim[i] = !isalnum(i);
1.87 + }
1.88 + }
1.89 +
1.90 + *ppTokenizer = &t->base;
1.91 + return SQLITE_OK;
1.92 +}
1.93 +
1.94 +/*
1.95 +** Destroy a tokenizer
1.96 +*/
1.97 +static int simpleDestroy(sqlite3_tokenizer *pTokenizer){
1.98 + free(pTokenizer);
1.99 + return SQLITE_OK;
1.100 +}
1.101 +
1.102 +/*
1.103 +** Prepare to begin tokenizing a particular string. The input
1.104 +** string to be tokenized is pInput[0..nBytes-1]. A cursor
1.105 +** used to incrementally tokenize this string is returned in
1.106 +** *ppCursor.
1.107 +*/
1.108 +static int simpleOpen(
1.109 + sqlite3_tokenizer *pTokenizer, /* The tokenizer */
1.110 + const char *pInput, int nBytes, /* String to be tokenized */
1.111 + sqlite3_tokenizer_cursor **ppCursor /* OUT: Tokenization cursor */
1.112 +){
1.113 + simple_tokenizer_cursor *c;
1.114 +
1.115 + c = (simple_tokenizer_cursor *) malloc(sizeof(*c));
1.116 + if( c==NULL ) return SQLITE_NOMEM;
1.117 +
1.118 + c->pInput = pInput;
1.119 + if( pInput==0 ){
1.120 + c->nBytes = 0;
1.121 + }else if( nBytes<0 ){
1.122 + c->nBytes = (int)strlen(pInput);
1.123 + }else{
1.124 + c->nBytes = nBytes;
1.125 + }
1.126 + c->iOffset = 0; /* start tokenizing at the beginning */
1.127 + c->iToken = 0;
1.128 + c->pToken = NULL; /* no space allocated, yet. */
1.129 + c->nTokenAllocated = 0;
1.130 +
1.131 + *ppCursor = &c->base;
1.132 + return SQLITE_OK;
1.133 +}
1.134 +
1.135 +/*
1.136 +** Close a tokenization cursor previously opened by a call to
1.137 +** simpleOpen() above.
1.138 +*/
1.139 +static int simpleClose(sqlite3_tokenizer_cursor *pCursor){
1.140 + simple_tokenizer_cursor *c = (simple_tokenizer_cursor *) pCursor;
1.141 + free(c->pToken);
1.142 + free(c);
1.143 + return SQLITE_OK;
1.144 +}
1.145 +
1.146 +/*
1.147 +** Extract the next token from a tokenization cursor. The cursor must
1.148 +** have been opened by a prior call to simpleOpen().
1.149 +*/
1.150 +static int simpleNext(
1.151 + sqlite3_tokenizer_cursor *pCursor, /* Cursor returned by simpleOpen */
1.152 + const char **ppToken, /* OUT: *ppToken is the token text */
1.153 + int *pnBytes, /* OUT: Number of bytes in token */
1.154 + int *piStartOffset, /* OUT: Starting offset of token */
1.155 + int *piEndOffset, /* OUT: Ending offset of token */
1.156 + int *piPosition /* OUT: Position integer of token */
1.157 +){
1.158 + simple_tokenizer_cursor *c = (simple_tokenizer_cursor *) pCursor;
1.159 + simple_tokenizer *t = (simple_tokenizer *) pCursor->pTokenizer;
1.160 + unsigned char *p = (unsigned char *)c->pInput;
1.161 +
1.162 + while( c->iOffset<c->nBytes ){
1.163 + int iStartOffset;
1.164 +
1.165 + /* Scan past delimiter characters */
1.166 + while( c->iOffset<c->nBytes && isDelim(t, p[c->iOffset]) ){
1.167 + c->iOffset++;
1.168 + }
1.169 +
1.170 + /* Count non-delimiter characters. */
1.171 + iStartOffset = c->iOffset;
1.172 + while( c->iOffset<c->nBytes && !isDelim(t, p[c->iOffset]) ){
1.173 + c->iOffset++;
1.174 + }
1.175 +
1.176 + if( c->iOffset>iStartOffset ){
1.177 + int i, n = c->iOffset-iStartOffset;
1.178 + if( n>c->nTokenAllocated ){
1.179 + c->nTokenAllocated = n+20;
1.180 + c->pToken = realloc(c->pToken, c->nTokenAllocated);
1.181 + if( c->pToken==NULL ) return SQLITE_NOMEM;
1.182 + }
1.183 + for(i=0; i<n; i++){
1.184 + /* TODO(shess) This needs expansion to handle UTF-8
1.185 + ** case-insensitivity.
1.186 + */
1.187 + unsigned char ch = p[iStartOffset+i];
1.188 + c->pToken[i] = ch<0x80 ? tolower(ch) : ch;
1.189 + }
1.190 + *ppToken = c->pToken;
1.191 + *pnBytes = n;
1.192 + *piStartOffset = iStartOffset;
1.193 + *piEndOffset = c->iOffset;
1.194 + *piPosition = c->iToken++;
1.195 +
1.196 + return SQLITE_OK;
1.197 + }
1.198 + }
1.199 + return SQLITE_DONE;
1.200 +}
1.201 +
1.202 +/*
1.203 +** The set of routines that implement the simple tokenizer
1.204 +*/
1.205 +static const sqlite3_tokenizer_module simpleTokenizerModule = {
1.206 + 0,
1.207 + simpleCreate,
1.208 + simpleDestroy,
1.209 + simpleOpen,
1.210 + simpleClose,
1.211 + simpleNext,
1.212 +};
1.213 +
1.214 +/*
1.215 +** Allocate a new simple tokenizer. Return a pointer to the new
1.216 +** tokenizer in *ppModule
1.217 +*/
1.218 +void sqlite3Fts1SimpleTokenizerModule(
1.219 + sqlite3_tokenizer_module const**ppModule
1.220 +){
1.221 + *ppModule = &simpleTokenizerModule;
1.222 +}
1.223 +
1.224 +#endif /* !defined(SQLITE_CORE) || defined(SQLITE_ENABLE_FTS1) */