SQLite

#include <tcl.h>
#include <string.h>

/*
** Implementation of a special SQL scalar function for testing tokenizers 
** designed to be used in concert with the Tcl testing framework. This
** function must be called with two or more arguments:
**
**   SELECT <function-name>(<key-name>, ..., <input-string>);

**
** where <function-name> is the name passed as the second argument
** to the sqlite3Fts3InitHashTable() function (e.g. 'fts3_tokenizer')
** concatenated with the string '_test' (e.g. 'fts3_tokenizer_test').
**
** The return value is a string that may be interpreted as a Tcl
** list. For each token in the <input-string>, three elements are

  const char *zErr = 0;

  const char *zName;
  int nName;
  const char *zInput;
  int nInput;

  const char *azArg[64];

  const char *zToken;
  int nToken;
  int iStart;
  int iEnd;
  int iPos;
  int i;

  Tcl_Obj *pRet;

  if( argc<2 ){
    sqlite3_result_error(context, "insufficient arguments", -1);
    return;
  }

  nName = sqlite3_value_bytes(argv[0]);
  zName = (const char *)sqlite3_value_text(argv[0]);
  nInput = sqlite3_value_bytes(argv[argc-1]);
  zInput = (const char *)sqlite3_value_text(argv[argc-1]);





  pHash = (Fts3Hash *)sqlite3_user_data(context);
  p = (sqlite3_tokenizer_module *)sqlite3Fts3HashFind(pHash, zName, nName+1);

  if( !p ){
    char *zErr = sqlite3_mprintf("unknown tokenizer: %s", zName);
    sqlite3_result_error(context, zErr, -1);
    sqlite3_free(zErr);
    return;
  }

  pRet = Tcl_NewObj();
  Tcl_IncrRefCount(pRet);

  for(i=1; i<argc-1; i++){
    azArg[i-1] = (const char *)sqlite3_value_text(argv[i]);
  }

  if( SQLITE_OK!=p->xCreate(argc-2, azArg, &pTokenizer) ){
    zErr = "error in xCreate()";
    goto finish;
  }
  pTokenizer->pModule = p;
  if( sqlite3Fts3OpenTokenizer(pTokenizer, 0, zInput, nInput, &pCsr) ){
    zErr = "error in xOpen()";
    goto finish;

    rc = sqlite3_create_function(db, zName, 1, any, p, scalarFunc, 0, 0);
  }
  if( SQLITE_OK==rc ){
    rc = sqlite3_create_function(db, zName, 2, any, p, scalarFunc, 0, 0);
  }
#ifdef SQLITE_TEST
  if( SQLITE_OK==rc ){
    rc = sqlite3_create_function(db, zTest, -1, any, p, testFunc, 0, 0);



  }
  if( SQLITE_OK==rc ){
    rc = sqlite3_create_function(db, zTest2, 0, any, pdb, intTestFunc, 0, 0);
  }
#endif

#ifdef SQLITE_TEST

typedef struct unicode_tokenizer unicode_tokenizer;
typedef struct unicode_cursor unicode_cursor;

struct unicode_tokenizer {
  sqlite3_tokenizer base;
  int bRemoveDiacritic;
  int nException;
  int *aiException;
};

struct unicode_cursor {
  sqlite3_tokenizer_cursor base;
  const unsigned char *aInput;    /* Input text being tokenized */
  int nInput;                     /* Size of aInput[] in bytes */
  int iOff;                       /* Current offset within aInput[] */
  int iToken;                     /* Index of next token to be returned */
  char *zToken;                   /* storage for current token */
  int nAlloc;                     /* space allocated at zToken */
};


/*
** Destroy a tokenizer allocated by unicodeCreate().
*/
static int unicodeDestroy(sqlite3_tokenizer *pTokenizer){
  if( pTokenizer ){
    unicode_tokenizer *p = (unicode_tokenizer *)pTokenizer;
    sqlite3_free(p->aiException);
    sqlite3_free(p);
  }
  return SQLITE_OK;
}

/*
** As part of a tokenchars= or separators= option, the CREATE VIRTUAL TABLE
** statement has specified that the tokenizer for this table shall consider
** all characters in string zIn/nIn to be separators (if bAlnum==0) or
** token characters (if bAlnum==1).
**
** For each codepoint in the zIn/nIn string, this function checks if the
** sqlite3FtsUnicodeIsalnum() function already returns the desired result.
** If so, no action is taken. Otherwise, the codepoint is added to the 
** unicode_tokenizer.aiException[] array. For the purposes of tokenization,
** the return value of sqlite3FtsUnicodeIsalnum() is inverted for all
** codepoints in the aiException[] array.
**
** If a standalone diacritic mark (one that sqlite3FtsUnicodeIsdiacritic()
** identifies as a diacritic) occurs in the zIn/nIn string it is ignored.
** It is not possible to change the behaviour of the tokenizer with respect
** to these codepoints.
*/
static int unicodeAddExceptions(
  unicode_tokenizer *p,           /* Tokenizer to add exceptions to */
  int bAlnum,                     /* Replace Isalnum() return value with this */
  const char *zIn,                /* Array of characters to make exceptions */
  int nIn                         /* Length of z in bytes */
){
  const unsigned char *z = (const unsigned char *)zIn;
  const unsigned char *zTerm = &z[nIn];
  int iCode;
  int nEntry = 0;

  assert( bAlnum==0 || bAlnum==1 );

  while( z<zTerm ){
    READ_UTF8(z, zTerm, iCode);
    assert( (sqlite3FtsUnicodeIsalnum(iCode) & 0xFFFFFFFE)==0 );
    if( sqlite3FtsUnicodeIsalnum(iCode)!=bAlnum 
     && sqlite3FtsUnicodeIsdiacritic(iCode)==0 
    ){
      nEntry++;
    }
  }

  if( nEntry ){
    int *aNew;                    /* New aiException[] array */
    int nNew;                     /* Number of valid entries in array aNew[] */

    aNew = sqlite3_realloc(p->aiException, (p->nException+nEntry)*sizeof(int));
    if( aNew==0 ) return SQLITE_NOMEM;
    nNew = p->nException;

    z = (const unsigned char *)zIn;
    while( z<zTerm ){
      READ_UTF8(z, zTerm, iCode);
      if( sqlite3FtsUnicodeIsalnum(iCode)!=bAlnum 
       && sqlite3FtsUnicodeIsdiacritic(iCode)==0
      ){
        int i, j;
        for(i=0; i<nNew && aNew[i]<iCode; i++);
        for(j=nNew; j>i; j--) aNew[j] = aNew[j-1];
        aNew[i] = iCode;
        nNew++;
      }
    }
    p->aiException = aNew;
    p->nException = nNew;
  }

  return SQLITE_OK;
}

/*
** Return true if the p->aiException[] array contains the value iCode.
*/
static int unicodeIsException(unicode_tokenizer *p, int iCode){
  if( p->nException>0 ){
    int *a = p->aiException;
    int iLo = 0;
    int iHi = p->nException-1;

    while( iHi>=iLo ){
      int iTest = (iHi + iLo) / 2;
      if( iCode==a[iTest] ){
        return 1;
      }else if( iCode>a[iTest] ){
        iLo = iTest+1;
      }else{
        iHi = iTest-1;
      }
    }
  }

  return 0;
}

/*
** Return true if, for the purposes of tokenization, codepoint iCode is
** considered a token character (not a separator).
*/
static int unicodeIsAlnum(unicode_tokenizer *p, int iCode){
  assert( (sqlite3FtsUnicodeIsalnum(iCode) & 0xFFFFFFFE)==0 );
  return sqlite3FtsUnicodeIsalnum(iCode) ^ unicodeIsException(p, iCode);
}

/*
** Create a new tokenizer instance.
*/
static int unicodeCreate(
  int nArg,                       /* Size of array argv[] */
  const char * const *azArg,      /* Tokenizer creation arguments */
  sqlite3_tokenizer **pp          /* OUT: New tokenizer handle */
){
  unicode_tokenizer *pNew;        /* New tokenizer object */
  int i;
  int rc = SQLITE_OK;

  pNew = (unicode_tokenizer *) sqlite3_malloc(sizeof(unicode_tokenizer));
  if( pNew==NULL ) return SQLITE_NOMEM;


  memset(pNew, 0, sizeof(unicode_tokenizer));
  pNew->bRemoveDiacritic = 1;

  for(i=0; rc==SQLITE_OK && i<nArg; i++){
    const char *z = azArg[i];
    int n = strlen(z);

    if( n==19 && memcmp("remove_diacritics=1", z, 19)==0 ){
      pNew->bRemoveDiacritic = 1;
    }
    else if( n==19 && memcmp("remove_diacritics=0", z, 19)==0 ){
      pNew->bRemoveDiacritic = 0;
    }
    else if( n>=11 && memcmp("tokenchars=", z, 11)==0 ){
      rc = unicodeAddExceptions(pNew, 1, &z[11], n-11);
    }
    else if( n>=11 && memcmp("separators=", z, 11)==0 ){
      rc = unicodeAddExceptions(pNew, 0, &z[11], n-11);
    }
    else{
      /* Unrecognized argument */
      rc  = SQLITE_ERROR;
    }
  }


  if( rc!=SQLITE_OK ){
    unicodeDestroy((sqlite3_tokenizer *)pNew);
    pNew = 0;
  }




  *pp = (sqlite3_tokenizer *)pNew;

  return rc;
}

/*
** Prepare to begin tokenizing a particular string.  The input
** string to be tokenized is pInput[0..nBytes-1].  A cursor
** used to incrementally tokenize this string is returned in 
** *ppCursor.

}

/*
** Extract the next token from a tokenization cursor.  The cursor must
** have been opened by a prior call to simpleOpen().
*/
static int unicodeNext(
  sqlite3_tokenizer_cursor *pC,   /* Cursor returned by simpleOpen */
  const char **paToken,           /* OUT: Token text */
  int *pnToken,                   /* OUT: Number of bytes at *paToken */
  int *piStart,                   /* OUT: Starting offset of token */
  int *piEnd,                     /* OUT: Ending offset of token */
  int *piPos                      /* OUT: Position integer of token */
){
  unicode_cursor *pCsr = (unicode_cursor *)pC;
  unicode_tokenizer *p = ((unicode_tokenizer *)pCsr->base.pTokenizer);
  int iCode;
  char *zOut;
  const unsigned char *z = &pCsr->aInput[pCsr->iOff];
  const unsigned char *zStart = z;
  const unsigned char *zEnd;
  const unsigned char *zTerm = &pCsr->aInput[pCsr->nInput];

  /* Scan past any delimiter characters before the start of the next token.
  ** Return SQLITE_DONE early if this takes us all the way to the end of 
  ** the input.  */
  while( z<zTerm ){
    READ_UTF8(z, zTerm, iCode);
    if( unicodeIsAlnum(p, iCode) ) break;
    zStart = z;
  }
  if( zStart>=zTerm ) return SQLITE_DONE;

  zOut = pCsr->zToken;
  do {
    int iOut;

    /* Grow the output buffer if required. */
    if( (zOut-pCsr->zToken)>=(pCsr->nAlloc-4) ){
      char *zNew = sqlite3_realloc(pCsr->zToken, pCsr->nAlloc+64);
      if( !zNew ) return SQLITE_NOMEM;
      zOut = &zNew[zOut - pCsr->zToken];
      pCsr->zToken = zNew;
      pCsr->nAlloc += 64;
    }

    /* Write the folded case of the last character read to the output */
    zEnd = z;
    iOut = sqlite3FtsUnicodeFold(iCode, p->bRemoveDiacritic);


    if( iOut ){
      WRITE_UTF8(zOut, iOut);
    }

    /* If the cursor is not at EOF, read the next character */
    if( z>=zTerm ) break;
    READ_UTF8(z, zTerm, iCode);
  }while( unicodeIsAlnum(p, iCode) 
       || sqlite3FtsUnicodeIsdiacritic(iCode)
  );

  /* Set the output variables and return. */
  pCsr->iOff = (z - pCsr->aInput);
  *paToken = pCsr->zToken;
  *pnToken = zOut - pCsr->zToken;

proc do_unicode_token_test2 {tn input res} {
  set input [string map {' ''} $input]
  uplevel [list do_execsql_test $tn "
    SELECT fts3_tokenizer_test('unicode61', '$input');
  " [list [list {*}$res]]]
}

proc do_unicode_token_test3 {tn args} {
  set res   [lindex $args end]
  set sql "SELECT fts3_tokenizer_test('unicode61'"
  foreach a [lrange $args 0 end-1] {
    append sql ", '"
    append sql [string map {' ''} $a]
    append sql "'"
  }
  append sql ")"
  uplevel [list do_execsql_test $tn $sql [list [list {*}$res]]]
}

do_unicode_token_test 1.0 {a B c D} {0 a a 1 b B 2 c c 3 d D}
do_unicode_token_test 1.1 {Ä Ö Ü} {0 ä Ä 1 ö Ö 2 ü Ü}
do_unicode_token_test 1.2 {xÄx xÖx xÜx} {0 xäx xÄx 1 xöx xÖx 2 xüx xÜx}

# 0x00DF is a small "sharp s". 0x1E9E is a capital sharp s.
do_unicode_token_test 1.3 "\uDF" "0 \uDF \uDF"

    INSERT INTO t1 VALUES($a);
    INSERT INTO t1 VALUES($b);
    INSERT INTO t1 VALUES($c);
    INSERT INTO t1 VALUES($d);
  }
} {}

#-------------------------------------------------------------------------

do_unicode_token_test3 5.1 {tokenchars=} {
  sqlite3_reset sqlite3_column_int
} {
  0 sqlite3 sqlite3 
  1 reset reset 
  2 sqlite3 sqlite3 
  3 column column 
  4 int int
}

do_unicode_token_test3 5.2 {tokenchars=_} {
  sqlite3_reset sqlite3_column_int
} {
  0 sqlite3_reset sqlite3_reset 
  1 sqlite3_column_int sqlite3_column_int
}

do_unicode_token_test3 5.3 {separators=xyz} {
  Laotianxhorseyrunszfast
} {
  0 laotian Laotian
  1 horse horse
  2 runs runs
  3 fast fast
}

do_unicode_token_test3 5.4 {tokenchars=xyz} {
  Laotianxhorseyrunszfast
} {
  0 laotianxhorseyrunszfast Laotianxhorseyrunszfast
}

do_unicode_token_test3 5.5 {tokenchars=_} {separators=zyx} {
  sqlite3_resetxsqlite3_column_intyhonda_phantom
} {
  0 sqlite3_reset sqlite3_reset 
  1 sqlite3_column_int sqlite3_column_int
  2 honda_phantom honda_phantom
}

do_unicode_token_test3 5.6 "separators=\u05D1" "abc\u05D1def" {
  0 abc abc 1 def def
}

do_unicode_token_test3 5.7                             \
  "tokenchars=\u2444\u2445"                            \
  "separators=\u05D0\u05D1\u05D2"                      \
  "\u2444fre\u2445sh\u05D0water\u05D2fish.\u2445timer" \
  [list                                                \
    0 \u2444fre\u2445sh \u2444fre\u2445sh              \
    1 water water                                      \
    2 fish fish                                        \
    3 \u2445timer \u2445timer                          \
  ]

# Check that it is not possible to add a standalone diacritic codepoint 
# to either separators or tokenchars.
do_unicode_token_test3 5.8 "separators=\u0301" \
  "hello\u0301world \u0301helloworld"          \
  "0 helloworld hello\u0301world 1 helloworld helloworld"

do_unicode_token_test3 5.9 "tokenchars=\u0301" \
  "hello\u0301world \u0301helloworld"          \
  "0 helloworld hello\u0301world 1 helloworld helloworld"

do_unicode_token_test3 5.10 "separators=\u0301" \
  "remove_diacritics=0"                        \
  "hello\u0301world \u0301helloworld"          \
  "0 hello\u0301world hello\u0301world 1 helloworld helloworld"

do_unicode_token_test3 5.11 "tokenchars=\u0301" \
  "remove_diacritics=0"                         \
  "hello\u0301world \u0301helloworld"           \
  "0 hello\u0301world hello\u0301world 1 helloworld helloworld"


finish_test