fts3_porter.c

最新的sqlite3.6.2源代码

/*
** 2006 September 30
**
** The author disclaims copyright to this source code.  In place of
** a legal notice, here is a blessing:
**
**    May you do good and not evil.
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** Implementation of the full-text-search tokenizer that implements
** a Porter stemmer.
*/

/*
** The code in this file is only compiled if:
**
**     * The FTS3 module is being built as an extension
**       (in which case SQLITE_CORE is not defined), or
**
**     * The FTS3 module is being built into the core of
**       SQLite (in which case SQLITE_ENABLE_FTS3 is defined).
*/
#if !defined(SQLITE_CORE) || defined(SQLITE_ENABLE_FTS3)


#include <assert.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <ctype.h>

#include "fts3_tokenizer.h"

/*
** Class derived from sqlite3_tokenizer
*/
typedef struct porter_tokenizer {
  sqlite3_tokenizer base;      /* Base class */
} porter_tokenizer;

/*
** Class derived from sqlit3_tokenizer_cursor
*/
typedef struct porter_tokenizer_cursor {
  sqlite3_tokenizer_cursor base;
  const char *zInput;          /* input we are tokenizing */
  int nInput;                  /* size of the input */
  int iOffset;                 /* current position in zInput */
  int iToken;                  /* index of next token to be returned */
  char *zToken;                /* storage for current token */
  int nAllocated;              /* space allocated to zToken buffer */
} porter_tokenizer_cursor;


/* Forward declaration */
static const sqlite3_tokenizer_module porterTokenizerModule;


/*
** Create a new tokenizer instance.
*/
static int porterCreate(
  int argc, const char * const *argv,
  sqlite3_tokenizer **ppTokenizer
){
  porter_tokenizer *t;
  t = (porter_tokenizer *) sqlite3_malloc(sizeof(*t));
  if( t==NULL ) return SQLITE_NOMEM;
  memset(t, 0, sizeof(*t));
  *ppTokenizer = &t->base;
  return SQLITE_OK;
}

/*
** Destroy a tokenizer
*/
static int porterDestroy(sqlite3_tokenizer *pTokenizer){
  sqlite3_free(pTokenizer);
  return SQLITE_OK;
}

/*
** Prepare to begin tokenizing a particular string.  The input
** string to be tokenized is zInput[0..nInput-1].  A cursor
** used to incrementally tokenize this string is returned in 
** *ppCursor.
*/
static int porterOpen(
  sqlite3_tokenizer *pTokenizer,         /* The tokenizer */
  const char *zInput, int nInput,        /* String to be tokenized */
  sqlite3_tokenizer_cursor **ppCursor    /* OUT: Tokenization cursor */
){
  porter_tokenizer_cursor *c;

  c = (porter_tokenizer_cursor *) sqlite3_malloc(sizeof(*c));
  if( c==NULL ) return SQLITE_NOMEM;

  c->zInput = zInput;
  if( zInput==0 ){
    c->nInput = 0;
  }else if( nInput<0 ){
    c->nInput = (int)strlen(zInput);
  }else{
    c->nInput = nInput;
  }
  c->iOffset = 0;                 /* start tokenizing at the beginning */
  c->iToken = 0;
  c->zToken = NULL;               /* no space allocated, yet. */
  c->nAllocated = 0;

  *ppCursor = &c->base;
  return SQLITE_OK;
}

/*
** Close a tokenization cursor previously opened by a call to
** porterOpen() above.
*/
static int porterClose(sqlite3_tokenizer_cursor *pCursor){
  porter_tokenizer_cursor *c = (porter_tokenizer_cursor *) pCursor;
  sqlite3_free(c->zToken);
  sqlite3_free(c);
  return SQLITE_OK;
}
/*
** Vowel or consonant
*/
static const char cType[] = {
   0, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0,
   1, 1, 1, 2, 1
};

/*
** isConsonant() and isVowel() determine if their first character in
** the string they point to is a consonant or a vowel, according
** to Porter ruls.  
**
** A consonate is any letter other than 'a', 'e', 'i', 'o', or 'u'.
** 'Y' is a consonant unless it follows another consonant,
** in which case it is a vowel.
**
** In these routine, the letters are in reverse order.  So the 'y' rule
** is that 'y' is a consonant unless it is followed by another
** consonent.
*/
static int isVowel(const char*);
static int isConsonant(const char *z){
  int j;
  char x = *z;
  if( x==0 ) return 0;
  assert( x>='a' && x<='z' );
  j = cType[x-'a'];
  if( j<2 ) return j;
  return z[1]==0 || isVowel(z + 1);
}
static int isVowel(const char *z){
  int j;
  char x = *z;
  if( x==0 ) return 0;
  assert( x>='a' && x<='z' );
  j = cType[x-'a'];
  if( j<2 ) return 1-j;
  return isConsonant(z + 1);
}

/*
** Let any sequence of one or more vowels be represented by V and let
** C be sequence of one or more consonants.  Then every word can be
** represented as:
**
**           [C] (VC){m} [V]
**
** In prose:  A word is an optional consonant followed by zero or
** vowel-consonant pairs followed by an optional vowel.  "m" is the
** number of vowel consonant pairs.  This routine computes the value
** of m for the first i bytes of a word.
**
** Return true if the m-value for z is 1 or more.  In other words,
** return true if z contains at least one vowel that is followed
** by a consonant.
**
** In this routine z[] is in reverse order.  So we are really looking
** for an instance of of a consonant followed by a vowel.
*/
static int m_gt_0(const char *z){
  while( isVowel(z) ){ z++; }
  if( *z==0 ) return 0;
  while( isConsonant(z) ){ z++; }
  return *z!=0;
}

/* Like mgt0 above except we are looking for a value of m which is
** exactly 1
*/
static int m_eq_1(const char *z){
  while( isVowel(z) ){ z++; }
  if( *z==0 ) return 0;
  while( isConsonant(z) ){ z++; }
  if( *z==0 ) return 0;
  while( isVowel(z) ){ z++; }
  if( *z==0 ) return 1;
  while( isConsonant(z) ){ z++; }
  return *z==0;
}

/* Like mgt0 above except we are looking for a value of m>1 instead
** or m>0
*/
static int m_gt_1(const char *z){
  while( isVowel(z) ){ z++; }
  if( *z==0 ) return 0;
  while( isConsonant(z) ){ z++; }
  if( *z==0 ) return 0;
  while( isVowel(z) ){ z++; }
  if( *z==0 ) return 0;
  while( isConsonant(z) ){ z++; }
  return *z!=0;
}

/*
** Return TRUE if there is a vowel anywhere within z[0..n-1]
*/
static int hasVowel(const char *z){
  while( isConsonant(z) ){ z++; }
  return *z!=0;
}

/*
** Return TRUE if the word ends in a double consonant.
**
** The text is reversed here. So we are really looking at
** the first two characters of z[].
*/
static int doubleConsonant(const char *z){
  return isConsonant(z) && z[0]==z[1] && isConsonant(z+1);
}

/*
** Return TRUE if the word ends with three letters which
** are consonant-vowel-consonent and where the final consonant
** is not 'w', 'x', or 'y'.
**
** The word is reversed here.  So we are really checking the
** first three letters and the first one cannot be in [wxy].
*/
static int star_oh(const char *z){
  return
    z[0]!=0 && isConsonant(z) &&
    z[0]!='w' && z[0]!='x' && z[0]!='y' &&
    z[1]!=0 && isVowel(z+1) &&
    z[2]!=0 && isConsonant(z+2);
}

/*
** If the word ends with zFrom and xCond() is true for the stem
** of the word that preceeds the zFrom ending, then change the 
** ending to zTo.
**
** The input word *pz and zFrom are both in reverse order.  zTo
** is in normal order. 
**
** Return TRUE if zFrom matches.  Return FALSE if zFrom does not
** match.  Not that TRUE is returned even if xCond() fails and
** no substitution occurs.
*/
static int stem(
  char **pz,             /* The word being stemmed (Reversed) */
  const char *zFrom,     /* If the ending matches this... (Reversed) */
  const char *zTo,       /* ... change the ending to this (not reversed) */
  int (*xCond)(const char*)   /* Condition that must be true */
){
  char *z = *pz;
  while( *zFrom && *zFrom==*z ){ z++; zFrom++; }
  if( *zFrom!=0 ) return 0;
  if( xCond && !xCond(z) ) return 1;
  while( *zTo ){
    *(--z) = *(zTo++);
  }
  *pz = z;
  return 1;
}

/*
** This is the fallback stemmer used when the porter stemmer is
** inappropriate.  The input word is copied into the output with
** US-ASCII case folding.  If the input word is too long (more
** than 20 bytes if it contains no digits or more than 6 bytes if
** it contains digits) then word is truncated to 20 or 6 bytes
** by taking 10 or 3 bytes from the beginning and end.
*/
static void copy_stemmer(const char *zIn, int nIn, char *zOut, int *pnOut){
  int i, mx, j;
  int hasDigit = 0;
  for(i=0; i<nIn; i++){
    int c = zIn[i];
    if( c>='A' && c<='Z' ){
      zOut[i] = c - 'A' + 'a';
    }else{
      if( c>='0' && c<='9' ) hasDigit = 1;
      zOut[i] = c;
    }
  }
  mx = hasDigit ? 3 : 10;
  if( nIn>mx*2 ){
    for(j=mx, i=nIn-mx; i<nIn; i++, j++){
      zOut[j] = zOut[i];
    }
    i = j;
  }
  zOut[i] = 0;
  *pnOut = i;
}


/*
** Stem the input word zIn[0..nIn-1].  Store the output in zOut.
** zOut is at least big enough to hold nIn bytes.  Write the actual
** size of the output word (exclusive of the '\0' terminator) into *pnOut.
**
** Any upper-case characters in the US-ASCII character set ([A-Z])