sl@0: /*
sl@0: ** 2007 June 22
sl@0: **
sl@0: ** The author disclaims copyright to this source code.  In place of
sl@0: ** a legal notice, here is a blessing:
sl@0: **
sl@0: **    May you do good and not evil.
sl@0: **    May you find forgiveness for yourself and forgive others.
sl@0: **    May you share freely, never taking more than you give.
sl@0: **
sl@0: ******************************************************************************
sl@0: **
sl@0: ** This is part of an SQLite module implementing full-text search.
sl@0: ** This particular file implements the generic tokenizer interface.
sl@0: */
sl@0: 
sl@0: /*
sl@0: ** The code in this file is only compiled if:
sl@0: **
sl@0: **     * The FTS3 module is being built as an extension
sl@0: **       (in which case SQLITE_CORE is not defined), or
sl@0: **
sl@0: **     * The FTS3 module is being built into the core of
sl@0: **       SQLite (in which case SQLITE_ENABLE_FTS3 is defined).
sl@0: */
sl@0: #if !defined(SQLITE_CORE) || defined(SQLITE_ENABLE_FTS3)
sl@0: 
sl@0: #include "sqlite3ext.h"
sl@0: #ifndef SQLITE_CORE
sl@0:   SQLITE_EXTENSION_INIT1
sl@0: #endif
sl@0: 
sl@0: #include "fts3_hash.h"
sl@0: #include "fts3_tokenizer.h"
sl@0: #include <assert.h>
sl@0: 
sl@0: /*
sl@0: ** Implementation of the SQL scalar function for accessing the underlying 
sl@0: ** hash table. This function may be called as follows:
sl@0: **
sl@0: **   SELECT <function-name>(<key-name>);
sl@0: **   SELECT <function-name>(<key-name>, <pointer>);
sl@0: **
sl@0: ** where <function-name> is the name passed as the second argument
sl@0: ** to the sqlite3Fts3InitHashTable() function (e.g. 'fts3_tokenizer').
sl@0: **
sl@0: ** If the <pointer> argument is specified, it must be a blob value
sl@0: ** containing a pointer to be stored as the hash data corresponding
sl@0: ** to the string <key-name>. If <pointer> is not specified, then
sl@0: ** the string <key-name> must already exist in the has table. Otherwise,
sl@0: ** an error is returned.
sl@0: **
sl@0: ** Whether or not the <pointer> argument is specified, the value returned
sl@0: ** is a blob containing the pointer stored as the hash data corresponding
sl@0: ** to string <key-name> (after the hash-table is updated, if applicable).
sl@0: */
sl@0: static void scalarFunc(
sl@0:   sqlite3_context *context,
sl@0:   int argc,
sl@0:   sqlite3_value **argv
sl@0: ){
sl@0:   fts3Hash *pHash;
sl@0:   void *pPtr = 0;
sl@0:   const unsigned char *zName;
sl@0:   int nName;
sl@0: 
sl@0:   assert( argc==1 || argc==2 );
sl@0: 
sl@0:   pHash = (fts3Hash *)sqlite3_user_data(context);
sl@0: 
sl@0:   zName = sqlite3_value_text(argv[0]);
sl@0:   nName = sqlite3_value_bytes(argv[0])+1;
sl@0: 
sl@0:   if( argc==2 ){
sl@0:     void *pOld;
sl@0:     int n = sqlite3_value_bytes(argv[1]);
sl@0:     if( n!=sizeof(pPtr) ){
sl@0:       sqlite3_result_error(context, "argument type mismatch", -1);
sl@0:       return;
sl@0:     }
sl@0:     pPtr = *(void **)sqlite3_value_blob(argv[1]);
sl@0:     pOld = sqlite3Fts3HashInsert(pHash, (void *)zName, nName, pPtr);
sl@0:     if( pOld==pPtr ){
sl@0:       sqlite3_result_error(context, "out of memory", -1);
sl@0:       return;
sl@0:     }
sl@0:   }else{
sl@0:     pPtr = sqlite3Fts3HashFind(pHash, zName, nName);
sl@0:     if( !pPtr ){
sl@0:       char *zErr = sqlite3_mprintf("unknown tokenizer: %s", zName);
sl@0:       sqlite3_result_error(context, zErr, -1);
sl@0:       sqlite3_free(zErr);
sl@0:       return;
sl@0:     }
sl@0:   }
sl@0: 
sl@0:   sqlite3_result_blob(context, (void *)&pPtr, sizeof(pPtr), SQLITE_TRANSIENT);
sl@0: }
sl@0: 
sl@0: #ifdef SQLITE_TEST
sl@0: 
sl@0: #include "tcl.h"
sl@0: #include <string.h>
sl@0: 
sl@0: /*
sl@0: ** Implementation of a special SQL scalar function for testing tokenizers 
sl@0: ** designed to be used in concert with the Tcl testing framework. This
sl@0: ** function must be called with two arguments:
sl@0: **
sl@0: **   SELECT <function-name>(<key-name>, <input-string>);
sl@0: **   SELECT <function-name>(<key-name>, <pointer>);
sl@0: **
sl@0: ** where <function-name> is the name passed as the second argument
sl@0: ** to the sqlite3Fts3InitHashTable() function (e.g. 'fts3_tokenizer')
sl@0: ** concatenated with the string '_test' (e.g. 'fts3_tokenizer_test').
sl@0: **
sl@0: ** The return value is a string that may be interpreted as a Tcl
sl@0: ** list. For each token in the <input-string>, three elements are
sl@0: ** added to the returned list. The first is the token position, the 
sl@0: ** second is the token text (folded, stemmed, etc.) and the third is the
sl@0: ** substring of <input-string> associated with the token. For example, 
sl@0: ** using the built-in "simple" tokenizer:
sl@0: **
sl@0: **   SELECT fts_tokenizer_test('simple', 'I don't see how');
sl@0: **
sl@0: ** will return the string:
sl@0: **
sl@0: **   "{0 i I 1 dont don't 2 see see 3 how how}"
sl@0: **   
sl@0: */
sl@0: static void testFunc(
sl@0:   sqlite3_context *context,
sl@0:   int argc,
sl@0:   sqlite3_value **argv
sl@0: ){
sl@0:   fts3Hash *pHash;
sl@0:   sqlite3_tokenizer_module *p;
sl@0:   sqlite3_tokenizer *pTokenizer = 0;
sl@0:   sqlite3_tokenizer_cursor *pCsr = 0;
sl@0: 
sl@0:   const char *zErr = 0;
sl@0: 
sl@0:   const char *zName;
sl@0:   int nName;
sl@0:   const char *zInput;
sl@0:   int nInput;
sl@0: 
sl@0:   const char *zArg = 0;
sl@0: 
sl@0:   const char *zToken;
sl@0:   int nToken;
sl@0:   int iStart;
sl@0:   int iEnd;
sl@0:   int iPos;
sl@0: 
sl@0:   Tcl_Obj *pRet;
sl@0: 
sl@0:   assert( argc==2 || argc==3 );
sl@0: 
sl@0:   nName = sqlite3_value_bytes(argv[0]);
sl@0:   zName = (const char *)sqlite3_value_text(argv[0]);
sl@0:   nInput = sqlite3_value_bytes(argv[argc-1]);
sl@0:   zInput = (const char *)sqlite3_value_text(argv[argc-1]);
sl@0: 
sl@0:   if( argc==3 ){
sl@0:     zArg = (const char *)sqlite3_value_text(argv[1]);
sl@0:   }
sl@0: 
sl@0:   pHash = (fts3Hash *)sqlite3_user_data(context);
sl@0:   p = (sqlite3_tokenizer_module *)sqlite3Fts3HashFind(pHash, zName, nName+1);
sl@0: 
sl@0:   if( !p ){
sl@0:     char *zErr = sqlite3_mprintf("unknown tokenizer: %s", zName);
sl@0:     sqlite3_result_error(context, zErr, -1);
sl@0:     sqlite3_free(zErr);
sl@0:     return;
sl@0:   }
sl@0: 
sl@0:   pRet = Tcl_NewObj();
sl@0:   Tcl_IncrRefCount(pRet);
sl@0: 
sl@0:   if( SQLITE_OK!=p->xCreate(zArg ? 1 : 0, &zArg, &pTokenizer) ){
sl@0:     zErr = "error in xCreate()";
sl@0:     goto finish;
sl@0:   }
sl@0:   pTokenizer->pModule = p;
sl@0:   if( SQLITE_OK!=p->xOpen(pTokenizer, zInput, nInput, &pCsr) ){
sl@0:     zErr = "error in xOpen()";
sl@0:     goto finish;
sl@0:   }
sl@0:   pCsr->pTokenizer = pTokenizer;
sl@0: 
sl@0:   while( SQLITE_OK==p->xNext(pCsr, &zToken, &nToken, &iStart, &iEnd, &iPos) ){
sl@0:     Tcl_ListObjAppendElement(0, pRet, Tcl_NewIntObj(iPos));
sl@0:     Tcl_ListObjAppendElement(0, pRet, Tcl_NewStringObj(zToken, nToken));
sl@0:     zToken = &zInput[iStart];
sl@0:     nToken = iEnd-iStart;
sl@0:     Tcl_ListObjAppendElement(0, pRet, Tcl_NewStringObj(zToken, nToken));
sl@0:   }
sl@0: 
sl@0:   if( SQLITE_OK!=p->xClose(pCsr) ){
sl@0:     zErr = "error in xClose()";
sl@0:     goto finish;
sl@0:   }
sl@0:   if( SQLITE_OK!=p->xDestroy(pTokenizer) ){
sl@0:     zErr = "error in xDestroy()";
sl@0:     goto finish;
sl@0:   }
sl@0: 
sl@0: finish:
sl@0:   if( zErr ){
sl@0:     sqlite3_result_error(context, zErr, -1);
sl@0:   }else{
sl@0:     sqlite3_result_text(context, Tcl_GetString(pRet), -1, SQLITE_TRANSIENT);
sl@0:   }
sl@0:   Tcl_DecrRefCount(pRet);
sl@0: }
sl@0: 
sl@0: static
sl@0: int registerTokenizer(
sl@0:   sqlite3 *db, 
sl@0:   char *zName, 
sl@0:   const sqlite3_tokenizer_module *p
sl@0: ){
sl@0:   int rc;
sl@0:   sqlite3_stmt *pStmt;
sl@0:   const char zSql[] = "SELECT fts3_tokenizer(?, ?)";
sl@0: 
sl@0:   rc = sqlite3_prepare_v2(db, zSql, -1, &pStmt, 0);
sl@0:   if( rc!=SQLITE_OK ){
sl@0:     return rc;
sl@0:   }
sl@0: 
sl@0:   sqlite3_bind_text(pStmt, 1, zName, -1, SQLITE_STATIC);
sl@0:   sqlite3_bind_blob(pStmt, 2, &p, sizeof(p), SQLITE_STATIC);
sl@0:   sqlite3_step(pStmt);
sl@0: 
sl@0:   return sqlite3_finalize(pStmt);
sl@0: }
sl@0: 
sl@0: static
sl@0: int queryTokenizer(
sl@0:   sqlite3 *db, 
sl@0:   char *zName,  
sl@0:   const sqlite3_tokenizer_module **pp
sl@0: ){
sl@0:   int rc;
sl@0:   sqlite3_stmt *pStmt;
sl@0:   const char zSql[] = "SELECT fts3_tokenizer(?)";
sl@0: 
sl@0:   *pp = 0;
sl@0:   rc = sqlite3_prepare_v2(db, zSql, -1, &pStmt, 0);
sl@0:   if( rc!=SQLITE_OK ){
sl@0:     return rc;
sl@0:   }
sl@0: 
sl@0:   sqlite3_bind_text(pStmt, 1, zName, -1, SQLITE_STATIC);
sl@0:   if( SQLITE_ROW==sqlite3_step(pStmt) ){
sl@0:     if( sqlite3_column_type(pStmt, 0)==SQLITE_BLOB ){
sl@0:       memcpy(pp, sqlite3_column_blob(pStmt, 0), sizeof(*pp));
sl@0:     }
sl@0:   }
sl@0: 
sl@0:   return sqlite3_finalize(pStmt);
sl@0: }
sl@0: 
sl@0: void sqlite3Fts3SimpleTokenizerModule(sqlite3_tokenizer_module const**ppModule);
sl@0: 
sl@0: /*
sl@0: ** Implementation of the scalar function fts3_tokenizer_internal_test().
sl@0: ** This function is used for testing only, it is not included in the
sl@0: ** build unless SQLITE_TEST is defined.
sl@0: **
sl@0: ** The purpose of this is to test that the fts3_tokenizer() function
sl@0: ** can be used as designed by the C-code in the queryTokenizer and
sl@0: ** registerTokenizer() functions above. These two functions are repeated
sl@0: ** in the README.tokenizer file as an example, so it is important to
sl@0: ** test them.
sl@0: **
sl@0: ** To run the tests, evaluate the fts3_tokenizer_internal_test() scalar
sl@0: ** function with no arguments. An assert() will fail if a problem is
sl@0: ** detected. i.e.:
sl@0: **
sl@0: **     SELECT fts3_tokenizer_internal_test();
sl@0: **
sl@0: */
sl@0: static void intTestFunc(
sl@0:   sqlite3_context *context,
sl@0:   int argc,
sl@0:   sqlite3_value **argv
sl@0: ){
sl@0:   int rc;
sl@0:   const sqlite3_tokenizer_module *p1;
sl@0:   const sqlite3_tokenizer_module *p2;
sl@0:   sqlite3 *db = (sqlite3 *)sqlite3_user_data(context);
sl@0: 
sl@0:   /* Test the query function */
sl@0:   sqlite3Fts3SimpleTokenizerModule(&p1);
sl@0:   rc = queryTokenizer(db, "simple", &p2);
sl@0:   assert( rc==SQLITE_OK );
sl@0:   assert( p1==p2 );
sl@0:   rc = queryTokenizer(db, "nosuchtokenizer", &p2);
sl@0:   assert( rc==SQLITE_ERROR );
sl@0:   assert( p2==0 );
sl@0:   assert( 0==strcmp(sqlite3_errmsg(db), "unknown tokenizer: nosuchtokenizer") );
sl@0: 
sl@0:   /* Test the storage function */
sl@0:   rc = registerTokenizer(db, "nosuchtokenizer", p1);
sl@0:   assert( rc==SQLITE_OK );
sl@0:   rc = queryTokenizer(db, "nosuchtokenizer", &p2);
sl@0:   assert( rc==SQLITE_OK );
sl@0:   assert( p2==p1 );
sl@0: 
sl@0:   sqlite3_result_text(context, "ok", -1, SQLITE_STATIC);
sl@0: }
sl@0: 
sl@0: #endif
sl@0: 
sl@0: /*
sl@0: ** Set up SQL objects in database db used to access the contents of
sl@0: ** the hash table pointed to by argument pHash. The hash table must
sl@0: ** been initialised to use string keys, and to take a private copy 
sl@0: ** of the key when a value is inserted. i.e. by a call similar to:
sl@0: **
sl@0: **    sqlite3Fts3HashInit(pHash, FTS3_HASH_STRING, 1);
sl@0: **
sl@0: ** This function adds a scalar function (see header comment above
sl@0: ** scalarFunc() in this file for details) and, if ENABLE_TABLE is
sl@0: ** defined at compilation time, a temporary virtual table (see header 
sl@0: ** comment above struct HashTableVtab) to the database schema. Both 
sl@0: ** provide read/write access to the contents of *pHash.
sl@0: **
sl@0: ** The third argument to this function, zName, is used as the name
sl@0: ** of both the scalar and, if created, the virtual table.
sl@0: */
sl@0: int sqlite3Fts3InitHashTable(
sl@0:   sqlite3 *db, 
sl@0:   fts3Hash *pHash, 
sl@0:   const char *zName
sl@0: ){
sl@0:   int rc = SQLITE_OK;
sl@0:   void *p = (void *)pHash;
sl@0:   const int any = SQLITE_ANY;
sl@0:   char *zTest = 0;
sl@0:   char *zTest2 = 0;
sl@0: 
sl@0: #ifdef SQLITE_TEST
sl@0:   void *pdb = (void *)db;
sl@0:   zTest = sqlite3_mprintf("%s_test", zName);
sl@0:   zTest2 = sqlite3_mprintf("%s_internal_test", zName);
sl@0:   if( !zTest || !zTest2 ){
sl@0:     rc = SQLITE_NOMEM;
sl@0:   }
sl@0: #endif
sl@0: 
sl@0:   if( rc!=SQLITE_OK
sl@0:    || (rc = sqlite3_create_function(db, zName, 1, any, p, scalarFunc, 0, 0))
sl@0:    || (rc = sqlite3_create_function(db, zName, 2, any, p, scalarFunc, 0, 0))
sl@0: #ifdef SQLITE_TEST
sl@0:    || (rc = sqlite3_create_function(db, zTest, 2, any, p, testFunc, 0, 0))
sl@0:    || (rc = sqlite3_create_function(db, zTest, 3, any, p, testFunc, 0, 0))
sl@0:    || (rc = sqlite3_create_function(db, zTest2, 0, any, pdb, intTestFunc, 0, 0))
sl@0: #endif
sl@0:   );
sl@0: 
sl@0:   sqlite3_free(zTest);
sl@0:   sqlite3_free(zTest2);
sl@0:   return rc;
sl@0: }
sl@0: 
sl@0: #endif /* !defined(SQLITE_CORE) || defined(SQLITE_ENABLE_FTS3) */