Index: src/test_spellfix.c ================================================================== --- src/test_spellfix.c +++ src/test_spellfix.c @@ -8,280 +8,13 @@ ** May you find forgiveness for yourself and forgive others. ** May you share freely, never taking more than you give. ** ************************************************************************* ** -** This module implements a VIRTUAL TABLE that can be used to search -** a large vocabulary for close matches. For example, this virtual -** table can be used to suggest corrections to misspelled words. Or, -** it could be used with FTS4 to do full-text search using potentially -** misspelled words. -** -** Create an instance of the virtual table this way: -** -** CREATE VIRTUAL TABLE demo USING spellfix1; -** -** The "spellfix1" term is the name of this module. The "demo" is the -** name of the virtual table you will be creating. The table is initially -** empty. You have to populate it with your vocabulary. Suppose you -** have a list of words in a table named "big_vocabulary". Then do this: -** -** INSERT INTO demo(word) SELECT word FROM big_vocabulary; -** -** If you intend to use this virtual table in cooperation with an FTS4 -** table (for spelling correctly of search terms) then you can extract -** the vocabulary using an fts3aux table: -** -** INSERT INTO demo(word) SELECT term FROM search_aux WHERE col='*'; -** -** You can also provide the virtual table with a "rank" for each word. -** The "rank" is an estimate of how common the word is. Larger numbers -** mean the word is more common. If you omit the rank when populating -** the table, then a rank of 1 is assumed. But if you have rank -** information, you can supply it and the virtual table will show a -** slight preference for selecting more commonly used terms. To -** populate the rank from an fts4aux table "search_aux" do something -** like this: -** -** INSERT INTO demo(word,rank) -** SELECT term, documents FROM search_aux WHERE col='*'; -** -** To query the virtual table, include a MATCH operator in the WHERE -** clause. For example: -** -** SELECT word FROM demo WHERE word MATCH 'kennasaw'; -** -** Using a dataset of American place names (derived from -** http://geonames.usgs.gov/domestic/download_data.htm) the query above -** returns 20 results beginning with: -** -** kennesaw -** kenosha -** kenesaw -** kenaga -** keanak -** -** If you append the character '*' to the end of the pattern, then -** a prefix search is performed. For example: -** -** SELECT word FROM demo WHERE word MATCH 'kennes*'; -** -** Yields 20 results beginning with: -** -** kennesaw -** kennestone -** kenneson -** kenneys -** keanes -** keenes -** -** The virtual table actually has a unique rowid with five columns plus three -** extra hidden columns. The columns are as follows: -** -** rowid A unique integer number associated with each -** vocabulary item in the table. This can be used -** as a foreign key on other tables in the database. -** -** word The text of the word that matches the pattern. -** Both word and pattern can contains unicode characters -** and can be mixed case. -** -** rank This is the rank of the word, as specified in the -** original INSERT statement. -** -** distance This is an edit distance or Levensthein distance going -** from the pattern to the word. -** -** langid This is the language-id of the word. All queries are -** against a single language-id, which defaults to 0. -** For any given query this value is the same on all rows. -** -** score The score is a combination of rank and distance. The -** idea is that a lower score is better. The virtual table -** attempts to find words with the lowest score and -** by default (unless overridden by ORDER BY) returns -** results in order of increasing score. -** -** matchlen For prefix queries, the number of characters in the prefix -** of the returned value (word) that matched the query term. -** For non-prefix queries, the number of characters in the -** returned value. -** -** top (HIDDEN) For any query, this value is the same on all -** rows. It is an integer which is the maximum number of -** rows that will be output. The actually number of rows -** output might be less than this number, but it will never -** be greater. The default value for top is 20, but that -** can be changed for each query by including a term of -** the form "top=N" in the WHERE clause of the query. -** -** scope (HIDDEN) For any query, this value is the same on all -** rows. The scope is a measure of how widely the virtual -** table looks for matching words. Smaller values of -** scope cause a broader search. The scope is normally -** choosen automatically and is capped at 4. Applications -** can change the scope by including a term of the form -** "scope=N" in the WHERE clause of the query. Increasing -** the scope will make the query run faster, but will reduce -** the possible corrections. -** -** srchcnt (HIDDEN) For any query, this value is the same on all -** rows. This value is an integer which is the number of -** of words examined using the edit-distance algorithm to -** find the top matches that are ultimately displayed. This -** value is for diagnostic use only. -** -** soundslike (HIDDEN) When inserting vocabulary entries, this field -** can be set to an spelling that matches what the word -** sounds like. See the DEALING WITH UNUSUAL AND DIFFICULT -** SPELLINGS section below for details. -** -** When inserting into or updating the virtual table, only the rowid, word, -** rank, and langid may be changes. Any attempt to set or modify the values -** of distance, score, top, scope, or srchcnt is silently ignored. -** -** ALGORITHM -** -** A shadow table named "%_vocab" (where the % is replaced by the name of -** the virtual table; Ex: "demo_vocab" for the "demo" virtual table) is -** constructed with these columns: -** -** id The unique id (INTEGER PRIMARY KEY) -** -** rank The rank of word. -** -** langid The language id for this entry. -** -** word The original UTF8 text of the vocabulary word -** -** k1 The word transliterated into lower-case ASCII. -** There is a standard table of mappings from non-ASCII -** characters into ASCII. Examples: "æ" -> "ae", -** "þ" -> "th", "ß" -> "ss", "á" -> "a", ... The -** accessory function spellfix1_translit(X) will do -** the non-ASCII to ASCII mapping. The built-in lower(X) -** function will convert to lower-case. Thus: -** k1 = lower(spellfix1_translit(word)). -** -** k2 This field holds a phonetic code derived from k1. Letters -** that have similar sounds are mapped into the same symbol. -** For example, all vowels and vowel clusters become the -** single symbol "A". And the letters "p", "b", "f", and -** "v" all become "B". All nasal sounds are represented -** as "N". And so forth. The mapping is base on -** ideas found in Soundex, Metaphone, and other -** long-standing phonetic matching systems. This key can -** be generated by the function spellfix1_phonehash(X). -** Hence: k2 = spellfix1_phonehash(k1) -** -** There is also a function for computing the Wagner edit distance or the -** Levenshtein distance between a pattern and a word. This function -** is exposed as spellfix1_editdist(X,Y). The edit distance function -** returns the "cost" of converting X into Y. Some transformations -** cost more than others. Changing one vowel into a different vowel, -** for example is relatively cheap, as is doubling a constant, or -** omitting the second character of a double-constant. Other transformations -** or more expensive. The idea is that the edit distance function returns -** a low cost of words that are similar and a higher cost for words -** that are futher apart. In this implementation, the maximum cost -** of any single-character edit (delete, insert, or substitute) is 100, -** with lower costs for some edits (such as transforming vowels). -** -** The "score" for a comparison is the edit distance between the pattern -** and the word, adjusted down by the base-2 logorithm of the word rank. -** For example, a match with distance 100 but rank 1000 would have a -** score of 122 (= 100 - log2(1000) + 32) where as a match with distance -** 100 with a rank of 1 would have a score of 131 (100 - log2(1) + 32). -** (NB: The constant 32 is added to each score to keep it from going -** negative in case the edit distance is zero.) In this way, frequently -** used words get a slightly lower cost which tends to move them toward -** the top of the list of alternative spellings. -** -** A straightforward implementation of a spelling corrector would be -** to compare the search term against every word in the vocabulary -** and select the 20 with the lowest scores. However, there will -** typically be hundreds of thousands or millions of words in the -** vocabulary, and so this approach is not fast enough. -** -** Suppose the term that is being spell-corrected is X. To limit -** the search space, X is converted to a k2-like key using the -** equivalent of: -** -** key = spellfix1_phonehash(lower(spellfix1_translit(X))) -** -** This key is then limited to "scope" characters. The default scope -** value is 4, but an alternative scope can be specified using the -** "scope=N" term in the WHERE clause. After the key has been truncated, -** the edit distance is run against every term in the vocabulary that -** has a k2 value that begins with the abbreviated key. -** -** For example, suppose the input word is "Paskagula". The phonetic -** key is "BACACALA" which is then truncated to 4 characters "BACA". -** The edit distance is then run on the 4980 entries (out of -** 272,597 entries total) of the vocabulary whose k2 values begin with -** BACA, yielding "Pascagoula" as the best match. -** -** Only terms of the vocabulary with a matching langid are searched. -** Hence, the same table can contain entries from multiple languages -** and only the requested language will be used. The default langid -** is 0. -** -** DEALING WITH UNUSUAL AND DIFFICULT SPELLINGS -** -** The algorithm above works quite well for most cases, but there are -** exceptions. These exceptions can be dealt with by making additional -** entries in the virtual table using the "soundslike" column. -** -** For example, many words of Greek origin begin with letters "ps" where -** the "p" is silent. Ex: psalm, pseudonym, psoriasis, psyche. In -** another example, many Scottish surnames can be spelled with an -** initial "Mac" or "Mc". Thus, "MacKay" and "McKay" are both pronounced -** the same. -** -** Accommodation can be made for words that are not spelled as they -** sound by making additional entries into the virtual table for the -** same word, but adding an alternative spelling in the "soundslike" -** column. For example, the canonical entry for "psalm" would be this: -** -** INSERT INTO demo(word) VALUES('psalm'); -** -** To enhance the ability to correct the spelling of "salm" into -** "psalm", make an addition entry like this: -** -** INSERT INTO demo(word,soundslike) VALUES('psalm','salm'); -** -** It is ok to make multiple entries for the same word as long as -** each entry has a different soundslike value. Note that if no -** soundslike value is specified, the soundslike defaults to the word -** itself. -** -** Listed below are some cases where it might make sense to add additional -** soundslike entries. The specific entries will depend on the application -** and the target language. -** -** * Silent "p" in words beginning with "ps": psalm, psyche -** -** * Silent "p" in words beginning with "pn": pneumonia, pneumatic -** -** * Silent "p" in words beginning with "pt": pterodactyl, ptolemaic -** -** * Silent "d" in words beginning with "dj": djinn, Djikarta -** -** * Silent "k" in words beginning with "kn": knight, Knuthson -** -** * Silent "g" in words beginning with "gn": gnarly, gnome, gnat -** -** * "Mac" versus "Mc" beginning Scottish surnames -** -** * "Tch" sounds in Slavic words: Tchaikovsky vs. Chaykovsky -** -** * The letter "j" pronounced like "h" in Spanish: LaJolla -** -** * Words beginning with "wr" versus "r": write vs. rite -** -** * Miscellanous problem words such as "debt", "tsetse", -** "Nguyen", "Van Nuyes". +** This module implements the spellfix1 VIRTUAL TABLE that can be used +** to search a large vocabulary for close matches. See separate +** documentation files (spellfix1.wiki and editdist3.wiki) for details. */ #if SQLITE_CORE # include "sqliteInt.h" #else # include @@ -304,11 +37,11 @@ ** 6 'L' Glide: L ** 7 'R' Semivowel: R ** 8 'M' Nasals: M N ** 9 'W' Letter W at the beginning of a word ** 10 'Y' Letter Y at the beginning of a word. -** 11 '9' A digit: 0 1 2 3 4 5 6 7 8 9 +** 11 '9' Digits: 0 1 2 3 4 5 6 7 8 9 ** 12 ' ' White space ** 13 '?' Other. */ #define CCLASS_SILENT 0 #define CCLASS_VOWEL 1 @@ -464,11 +197,10 @@ if( zIn[1]=='n' ){ zIn++; nIn--; } break; } } } - if( zIn[0]=='k' && zIn[1]=='n' ){ zIn++, nIn--; } for(i=0; i=CCLASS_B && classFrom<=CCLASS_Y && classTo>=CCLASS_B && classTo<=CCLASS_Y ){ /* Convert from one consonant to another, but in a different class */ return 75; @@ -614,11 +346,11 @@ ** of zB that matched the pattern in zA. If zA does not end with a '*', ** then this value is always the number of bytes in zB (i.e. strlen(zB)). ** If zA does end in a '*', then it is the number of bytes in the prefix ** of zB that was deemed to match zA. */ -static int editdist1(const char *zA, const char *zB, int iLangId, int *pnMatch){ +static int editdist1(const char *zA, const char *zB, int *pnMatch){ int nA, nB; /* Number of characters in zA[] and zB[] */ int xA, xB; /* Loop counters for zA[] and zB[] */ char cA, cB; /* Current character of zA and zB */ char cAprev, cBprev; /* Previous character of zA and zB */ char cAnext, cBnext; /* Next character in zA and zB */ @@ -643,14 +375,14 @@ printf("A=\"%s\" B=\"%s\" dc=%c\n", zA, zB, dc?dc:' '); #endif /* Verify input strings and measure their lengths */ for(nA=0; zA[nA]; nA++){ - if( zA[nA]>127 ) return -2; + if( zA[nA]&0x80 ) return -2; } for(nB=0; zB[nB]; nB++){ - if( zB[nB]>127 ) return -2; + if( zB[nB]&0x80 ) return -2; } /* Special processing if either string is empty */ if( nA==0 ){ cBprev = dc; @@ -754,19 +486,20 @@ if( pnMatch ) *pnMatch = xB+nMatch; } } }else{ res = m[nB]; - if( pnMatch ) *pnMatch = -1; + /* In the current implementation, pnMatch is always NULL if zA does + ** not end in "*" */ + assert( pnMatch==0 ); } sqlite3_free(toFree); return res; } /* ** Function: editdist(A,B) -** editdist(A,B,langid) ** ** Return the cost of transforming string A into string B. Both strings ** must be pure ASCII text. If A ends with '*' then it is assumed to be ** a prefix of B and extra characters on the end of B have minimal additional ** cost. @@ -774,15 +507,14 @@ static void editdistSqlFunc( sqlite3_context *context, int argc, sqlite3_value **argv ){ - int langid = argc==2 ? 0 : sqlite3_value_int(argv[2]); int res = editdist1( (const char*)sqlite3_value_text(argv[0]), (const char*)sqlite3_value_text(argv[1]), - langid, 0); + 0); if( res<0 ){ if( res==(-3) ){ sqlite3_result_error_nomem(context); }else if( res==(-2) ){ sqlite3_result_error(context, "non-ASCII input to editdist()", -1); @@ -922,11 +654,11 @@ EditDist3Config *p, /* The edit distance configuration to load */ sqlite3 *db, /* Load from this database */ const char *zTable /* Name of the table from which to load */ ){ sqlite3_stmt *pStmt; - int rc; + int rc, rc2; char *zSql; int iLangPrev = -9999; EditDist3Lang *pLang; zSql = sqlite3_mprintf("SELECT iLang, cFrom, cTo, iCost" @@ -937,28 +669,30 @@ if( rc ) return rc; editDist3ConfigClear(p); while( sqlite3_step(pStmt)==SQLITE_ROW ){ int iLang = sqlite3_column_int(pStmt, 0); const char *zFrom = (const char*)sqlite3_column_text(pStmt, 1); - int nFrom = sqlite3_column_bytes(pStmt, 1); + int nFrom = zFrom ? sqlite3_column_bytes(pStmt, 1) : 0; const char *zTo = (const char*)sqlite3_column_text(pStmt, 2); - int nTo = sqlite3_column_bytes(pStmt, 2); + int nTo = zTo ? sqlite3_column_bytes(pStmt, 2) : 0; int iCost = sqlite3_column_int(pStmt, 3); - if( nFrom>100 || nFrom<0 || nTo>100 || nTo<0 ) continue; + assert( zFrom!=0 || nFrom==0 ); + assert( zTo!=0 || nTo==0 ); + if( nFrom>100 || nTo>100 ) continue; if( iCost<0 ) continue; if( iLang!=iLangPrev ){ EditDist3Lang *pNew; - p->nLang++; - pNew = sqlite3_realloc(p->a, p->nLang*sizeof(p->a[0])); + pNew = sqlite3_realloc(p->a, (p->nLang+1)*sizeof(p->a[0])); if( pNew==0 ){ rc = SQLITE_NOMEM; break; } p->a = pNew; - pLang = &p->a[p->nLang-1]; + pLang = &p->a[p->nLang]; + p->nLang++; pLang->iLang = iLang; pLang->iInsCost = 100; pLang->iDelCost = 100; - pLang->iSubCost = 200; + pLang->iSubCost = 150; pLang->pCost = 0; iLangPrev = iLang; } if( nFrom==1 && zFrom[0]=='?' && nTo==0 ){ pLang->iDelCost = iCost; @@ -979,11 +713,12 @@ memcpy(pCost->a + nFrom, zTo, nTo); pCost->pNext = pLang->pCost; pLang->pCost = pCost; } } - sqlite3_finalize(pStmt); + rc2 = sqlite3_finalize(pStmt); + if( rc==SQLITE_OK ) rc = rc2; return rc; } /* ** Return the length (in bytes) of a utf-8 character. Or return a maximum @@ -1017,11 +752,11 @@ /* ** Return TRUE (non-zero) of the To side of the given cost matches ** the given string. */ static int matchFrom(EditDist3Cost *p, const char *z, int n){ - if( p->nFrom>n ) return 0; + assert( p->nFrom<=n ); if( memcmp(p->a, z, p->nFrom)!=0 ) return 0; return 1; } /* @@ -1064,14 +799,16 @@ ){ EditDist3FromString *pStr; EditDist3Cost *p; int i; + if( z==0 ) return 0; if( n<0 ) n = (int)strlen(z); pStr = sqlite3_malloc( sizeof(*pStr) + sizeof(pStr->a[0])*n + n + 1 ); if( pStr==0 ) return 0; pStr->a = (EditDist3From*)&pStr[1]; + memset(pStr->a, 0, sizeof(pStr->a[0])*n); pStr->n = n; pStr->z = (char*)&pStr->a[n]; memcpy(pStr->z, z, n+1); if( n && z[n-1]=='*' ){ pStr->isPrefix = 1; @@ -1111,34 +848,10 @@ } } return pStr; } -#if 0 /* No longer used */ -/* -** Return the number of bytes in the common prefix of two UTF8 strings. -** Only complete characters are considered. -*/ -static int editDist3PrefixLen(const char *z1, const char *z2){ - int n = 0; - while( z1[n] && z1[n]==z2[n] ){ n++; } - while( n && (z1[n]&0xc0)==0x80 ){ n--; } - return n; -} - -/* -** Return the number of bytes in the common suffix of two UTF8 strings. -** Only complete characters are considered. -*/ -static int editDist3SuffixLen(const char *z1, int n1, const char *z2, int n2){ - int origN1 = n1; - while( n1>0 && n2>0 && z1[n1-1]==z2[n2-1] ){ n1--; n2--; } - while( n10. */ + if( NEVER(n==0) ){ c = i = 0; }else{ c = z[0]; i = 1; if( c>=0xc0 ){ @@ -1878,14 +1581,14 @@ ** Space to hold the returned string comes from sqlite3_malloc() and ** should be freed by the caller. */ static unsigned char *transliterate(const unsigned char *zIn, int nIn){ unsigned char *zOut = sqlite3_malloc( nIn*4 + 1 ); - int i, c, sz, nOut; + int c, sz, nOut; if( zOut==0 ) return 0; - i = nOut = 0; - while( i0 ){ c = utf8Read(zIn, nIn, &sz); zIn += sz; nIn -= sz; if( c<=127 ){ zOut[nOut++] = c; @@ -2033,140 +1736,18 @@ } /* End transliterate ****************************************************************************** ****************************************************************************** -** Begin Polloc & Zamora SPEEDCOP style keying functions. -*/ -/* -** The Pollock & Zamora skeleton function. Move all consonants to the -** front and all vowels to the end, removing duplicates. Except if the -** first letter is a vowel then it remains as the first letter. -*/ -static void pollockSkeletonKey(const char *zIn, char *zOut){ - int i, j; - unsigned char c; - char seen[26]; - static const unsigned char isVowel[] = { 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, - 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 }; - memset(seen, 0, sizeof(seen)); - for(i=j=0; (c = (unsigned char)zIn[i])!=0; i++){ - if( c<'a' || c>'z' ) continue; - if( j>0 || isVowel[c-'a'] ) continue; - if( seen[c-'a'] ) continue; - seen[c-'a'] = 1; - zOut[j++] = c; - } - for(i=0; (c = (unsigned char)zIn[i])!=0; i++){ - if( c<'a' || c>'z' ) continue; - if( seen[c-'a'] ) continue; - if( !isVowel[c-'a'] ) continue; - seen[c-'a'] = 1; - zOut[j++] = c; - } - zOut[j] = 0; -} - -/* -** Function: pollock_skeleton(X) -** -** Return the Pollock and Zamora skeleton key for a string X of all -** lower-case letters. -*/ -static void pollockSkeletonSqlFunc( - sqlite3_context *context, - int argc, - sqlite3_value **argv -){ - const char *zIn = (const char*)sqlite3_value_text(argv[0]); - int nIn = sqlite3_value_bytes(argv[0]); - char *zOut; - if( zIn ){ - zOut = sqlite3_malloc( nIn + 1 ); - if( zOut==0 ){ - sqlite3_result_error_nomem(context); - }else{ - pollockSkeletonKey(zIn, zOut); - sqlite3_result_text(context, (char*)zOut, -1, sqlite3_free); - } - } -} - -/* -** The Pollock & Zamora omission key. -** -** The key consists of unique consonants in the following order: -** -** jkqxzvwybfmgpdhclntsr -** -** These are followed by unique vowels in input order. -*/ -static void pollockOmissionKey(const char *zIn, char *zOut){ - int i, j; - unsigned char c; - char seen[26]; - static const unsigned char isVowel[] = { 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, - 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 }; - static const unsigned char constOrder[] = "jkqxzvwybfmgpdhclntsr"; - - memset(seen, 0, sizeof(seen)); - for(i=j=0; (c = (unsigned char)zIn[i])!=0; i++){ - if( c<'a' || c>'z' ) continue; - if( isVowel[c-'a'] ) continue; - if( seen[c-'a'] ) continue; - seen[c-'a'] = 1; - } - for(i=0; (c = constOrder[i])!=0; i++){ - if( seen[c-'a'] ) zOut[j++] = c; - } - for(i=0; (c = (unsigned char)zIn[i])!=0; i++){ - if( c<'a' || c>'z' ) continue; - if( seen[c-'a'] ) continue; - if( !isVowel[c-'a'] ) continue; - seen[c-'a'] = 1; - zOut[j++] = c; - } - zOut[j] = 0; -} - -/* -** Function: pollock_omission(X) -** -** Return the Pollock and Zamora omission key for a string X of all -** lower-case letters. -*/ -static void pollockOmissionSqlFunc( - sqlite3_context *context, - int argc, - sqlite3_value **argv -){ - const char *zIn = (const char*)sqlite3_value_text(argv[0]); - int nIn = sqlite3_value_bytes(argv[0]); - char *zOut; - if( zIn ){ - zOut = sqlite3_malloc( nIn + 1 ); - if( zOut==0 ){ - sqlite3_result_error_nomem(context); - }else{ - pollockOmissionKey(zIn, zOut); - sqlite3_result_text(context, (char*)zOut, -1, sqlite3_free); - } - } -} - - -/* End SPEEDCOP keying functions -****************************************************************************** -****************************************************************************** ** Begin spellfix1 virtual table. */ /* Maximum length of a phonehash used for querying the shadow table */ #define SPELLFIX_MX_HASH 8 /* Maximum number of hash strings to examine per query */ -#define SPELLFIX_MX_RUN 8 +#define SPELLFIX_MX_RUN 1 typedef struct spellfix1_vtab spellfix1_vtab; typedef struct spellfix1_cursor spellfix1_cursor; /* Fuzzy-search virtual table object */ @@ -2185,14 +1766,15 @@ spellfix1_vtab *pVTab; /* The table to which this cursor belongs */ char *zPattern; /* rhs of MATCH clause */ int nRow; /* Number of rows of content */ int nAlloc; /* Number of allocated rows */ int iRow; /* Current row of content */ - int iLang; /* Value of the lang= constraint */ + int iLang; /* Value of the langid= constraint */ int iTop; /* Value of the top= constraint */ int iScope; /* Value of the scope= constraint */ int nSearch; /* Number of vocabulary items checked */ + sqlite3_stmt *pFullScan; /* Shadow query for a full table scan */ struct spellfix1_row { /* For each row of content */ sqlite3_int64 iRowid; /* Rowid for this row */ char *zWord; /* Text for this row */ int iRank; /* Rank for this row */ int iDistance; /* Distance from pattern for this row */ @@ -2265,15 +1847,17 @@ char c; while( isspace(zIn[0]) ) zIn++; zOut = sqlite3_mprintf("%s", zIn); if( zOut==0 ) return 0; i = (int)strlen(zOut); +#if 0 /* The parser will never leave spaces at the end */ while( i>0 && isspace(zOut[i-1]) ){ i--; } +#endif zOut[i] = 0; c = zOut[0]; if( c=='\'' || c=='"' ){ - for(i=1, j=0; zOut[i]; i++){ + for(i=1, j=0; ALWAYS(zOut[i]); i++){ zOut[j++] = zOut[i]; if( zOut[i]==c ){ if( zOut[i+1]==c ){ i++; }else{ @@ -2309,35 +1893,29 @@ const char *zTableName = argv[2]; int nDbName; int rc = SQLITE_OK; int i; - if( argc<3 ){ - *pzErr = sqlite3_mprintf( - "%s: wrong number of CREATE VIRTUAL TABLE arguments", argv[0] - ); - rc = SQLITE_ERROR; - }else{ - nDbName = strlen(zDbName); - pNew = sqlite3_malloc( sizeof(*pNew) + nDbName + 1); - if( pNew==0 ){ - rc = SQLITE_NOMEM; - }else{ - memset(pNew, 0, sizeof(*pNew)); - pNew->zDbName = (char*)&pNew[1]; - memcpy(pNew->zDbName, zDbName, nDbName+1); - pNew->zTableName = sqlite3_mprintf("%s", zTableName); - pNew->db = db; - if( pNew->zTableName==0 ){ - rc = SQLITE_NOMEM; - }else{ - rc = sqlite3_declare_vtab(db, - "CREATE TABLE x(word,rank,distance,langid, " - "score, matchlen, phonehash, " - "top HIDDEN, scope HIDDEN, srchcnt HIDDEN, " - "soundslike HIDDEN, command HIDDEN)" - ); + nDbName = strlen(zDbName); + pNew = sqlite3_malloc( sizeof(*pNew) + nDbName + 1); + if( pNew==0 ){ + rc = SQLITE_NOMEM; + }else{ + memset(pNew, 0, sizeof(*pNew)); + pNew->zDbName = (char*)&pNew[1]; + memcpy(pNew->zDbName, zDbName, nDbName+1); + pNew->zTableName = sqlite3_mprintf("%s", zTableName); + pNew->db = db; + if( pNew->zTableName==0 ){ + rc = SQLITE_NOMEM; + }else{ + rc = sqlite3_declare_vtab(db, + "CREATE TABLE x(word,rank,distance,langid, " + "score, matchlen, phonehash HIDDEN, " + "top HIDDEN, scope HIDDEN, srchcnt HIDDEN, " + "soundslike HIDDEN, command HIDDEN)" + ); #define SPELLFIX_COL_WORD 0 #define SPELLFIX_COL_RANK 1 #define SPELLFIX_COL_DISTANCE 2 #define SPELLFIX_COL_LANGID 3 #define SPELLFIX_COL_SCORE 4 @@ -2346,43 +1924,48 @@ #define SPELLFIX_COL_TOP 7 #define SPELLFIX_COL_SCOPE 8 #define SPELLFIX_COL_SRCHCNT 9 #define SPELLFIX_COL_SOUNDSLIKE 10 #define SPELLFIX_COL_COMMAND 11 - } - if( rc==SQLITE_OK && isCreate ){ - sqlite3_uint64 r; - spellfix1DbExec(&rc, db, - "CREATE TABLE IF NOT EXISTS \"%w\".\"%w_vocab\"(\n" - " id INTEGER PRIMARY KEY,\n" - " rank INT,\n" - " langid INT,\n" - " word TEXT,\n" - " k1 TEXT,\n" - " k2 TEXT\n" - ");\n", - zDbName, zTableName - ); - sqlite3_randomness(sizeof(r), &r); - spellfix1DbExec(&rc, db, - "CREATE INDEX IF NOT EXISTS \"%w\".\"%w_index_%llx\" " - "ON \"%w_vocab\"(langid,k2);", - zDbName, zModule, r, zTableName - ); - } - for(i=3; rc==SQLITE_OK && ibase); + }else{ + *ppVTab = (sqlite3_vtab *)pNew; + } return rc; } /* ** The xConnect and xCreate methods @@ -2415,10 +1998,14 @@ sqlite3_free(pCur->a[i].zWord); } pCur->nRow = 0; pCur->iRow = 0; pCur->nSearch = 0; + if( pCur->pFullScan ){ + sqlite3_finalize(pCur->pFullScan); + pCur->pFullScan = 0; + } } /* ** Resize the cursor to hold up to N rows of content */ @@ -2533,11 +2120,11 @@ } if( iPlan&1 ){ int idx = 2; pIdxInfo->idxNum = iPlan; if( pIdxInfo->nOrderBy==1 - && pIdxInfo->aOrderBy[0].iColumn==4 + && pIdxInfo->aOrderBy[0].iColumn==SPELLFIX_COL_SCORE && pIdxInfo->aOrderBy[0].desc==0 ){ pIdxInfo->orderByConsumed = 1; /* Default order by iScore */ } if( iPlan&2 ){ @@ -2638,13 +2225,13 @@ sqlite3_stmt *pStmt = p->pStmt; char zHash1[SPELLFIX_MX_HASH]; char zHash2[SPELLFIX_MX_HASH]; char *zClass; int nClass; + int rc; if( pCur->a==0 || p->rc ) return; /* Prior memory allocation failure */ - if( p->nRun>=SPELLFIX_MX_RUN ) return; zClass = (char*)phoneticHash((unsigned char*)zQuery, nQuery); if( zClass==0 ){ p->rc = SQLITE_NOMEM; return; } @@ -2664,22 +2251,31 @@ sqlite3_free(zClass); zHash1[iScope] = 0; memcpy(zHash2, zHash1, iScope); zHash2[iScope] = 'Z'; zHash2[iScope+1] = 0; +#if SPELLFIX_MX_RUN>1 for(i=0; inRun; i++){ if( strcmp(p->azPrior[i], zHash1)==0 ) return; } +#endif + assert( p->nRunazPrior[p->nRun++], zHash1, iScope+1); - sqlite3_bind_text(pStmt, 1, zHash1, -1, SQLITE_STATIC); - sqlite3_bind_text(pStmt, 2, zHash2, -1, SQLITE_STATIC); + if( sqlite3_bind_text(pStmt, 1, zHash1, -1, SQLITE_STATIC)==SQLITE_NOMEM + || sqlite3_bind_text(pStmt, 2, zHash2, -1, SQLITE_STATIC)==SQLITE_NOMEM + ){ + p->rc = SQLITE_NOMEM; + return; + } +#if SPELLFIX_MX_RUN>1 for(i=0; inRow; i++){ if( pCur->a[i].iScore>iWorst ){ iWorst = pCur->a[i].iScore; idxWorst = i; } } +#endif while( sqlite3_step(pStmt)==SQLITE_ROW ){ int iMatchlen = -1; iRank = sqlite3_column_int(pStmt, 2); if( p->pMatchStr3 ){ int nWord = sqlite3_column_bytes(pStmt, 1); @@ -2686,11 +2282,15 @@ zWord = (const char*)sqlite3_column_text(pStmt, 1); iDist = editDist3Core(p->pMatchStr3, zWord, nWord, p->pLang, &iMatchlen); }else{ zK1 = (const char*)sqlite3_column_text(pStmt, 3); if( zK1==0 ) continue; - iDist = editdist1(p->zPattern, zK1, pCur->iLang, 0); + iDist = editdist1(p->zPattern, zK1, 0); + } + if( iDist<0 ){ + p->rc = SQLITE_NOMEM; + break; } pCur->nSearch++; iScore = spellfix1Score(iDist,iRank); if( p->iMaxDist>=0 ){ if( iDist>p->iMaxDist ) continue; @@ -2706,10 +2306,14 @@ sqlite3_free(pCur->a[idx].zWord); }else{ continue; } pCur->a[idx].zWord = sqlite3_mprintf("%s", sqlite3_column_text(pStmt, 1)); + if( pCur->a[idx].zWord==0 ){ + p->rc = SQLITE_NOMEM; + break; + } pCur->a[idx].iRowid = sqlite3_column_int64(pStmt, 0); pCur->a[idx].iRank = iRank; pCur->a[idx].iDistance = iDist; pCur->a[idx].iScore = iScore; pCur->a[idx].iMatchlen = iMatchlen; @@ -2725,11 +2329,12 @@ idxWorst = i; } } } } - sqlite3_reset(pStmt); + rc = sqlite3_reset(pStmt); + if( rc ) p->rc = rc; } /* ** This version of the xFilter method work if the MATCH term is present ** and we are doing a scan. @@ -2746,11 +2351,11 @@ int nPattern; /* Length of zPattern */ int iLimit = 20; /* Max number of rows of output */ int iScope = 3; /* Use this many characters of zClass */ int iLang = 0; /* Language code */ char *zSql; /* SQL of shadow table query */ - sqlite3_stmt *pStmt; /* Shadow table query */ + sqlite3_stmt *pStmt = 0; /* Shadow table query */ int rc; /* Result code */ int idx = 1; /* Next available filter parameter */ spellfix1_vtab *p = pCur->pVTab; /* The virtual table that owns pCur */ MatchQuery x; /* For passing info to RunQuery() */ @@ -2788,25 +2393,37 @@ zMatchThis = sqlite3_value_text(argv[0]); if( zMatchThis==0 ) return SQLITE_OK; if( p->pConfig3 ){ x.pLang = editDist3FindLang(p->pConfig3, iLang); pMatchStr3 = editDist3FromStringNew(x.pLang, (const char*)zMatchThis, -1); + if( pMatchStr3==0 ){ + x.rc = SQLITE_NOMEM; + goto filter_exit; + } }else{ x.pLang = 0; } zPattern = (char*)transliterate(zMatchThis, sqlite3_value_bytes(argv[0])); sqlite3_free(pCur->zPattern); pCur->zPattern = zPattern; - if( zPattern==0 ) return SQLITE_NOMEM; + if( zPattern==0 ){ + x.rc = SQLITE_NOMEM; + goto filter_exit; + } nPattern = strlen(zPattern); if( zPattern[nPattern-1]=='*' ) nPattern--; zSql = sqlite3_mprintf( "SELECT id, word, rank, k1" " FROM \"%w\".\"%w_vocab\"" " WHERE langid=%d AND k2>=?1 AND k2zDbName, p->zTableName, iLang ); + if( zSql==0 ){ + x.rc = SQLITE_NOMEM; + pStmt = 0; + goto filter_exit; + } rc = sqlite3_prepare_v2(p->db, zSql, -1, &pStmt, 0); sqlite3_free(zSql); pCur->iLang = iLang; x.pCur = pCur; x.pStmt = pStmt; @@ -2818,38 +2435,22 @@ x.pConfig3 = p->pConfig3; if( x.rc==SQLITE_OK ){ spellfix1RunQuery(&x, zPattern, nPattern); } -#if 0 - /* Convert "ght" to "t" in the original pattern and try again */ - if( x.rc==SQLITE_OK ){ - int i, j; /* Loop counters */ - char zQuery[50]; /* Space for alternative query string */ - for(i=j=0; ia ){ qsort(pCur->a, pCur->nRow, sizeof(pCur->a[0]), spellfix1RowCompare); pCur->iTop = iLimit; pCur->iScope = iScope; + }else{ + x.rc = SQLITE_NOMEM; } + +filter_exit: sqlite3_finalize(pStmt); editDist3FromStringDelete(pMatchStr3); - return pCur->a ? x.rc : SQLITE_NOMEM; + return x.rc; } /* ** This version of xFilter handles a full-table scan case */ @@ -2857,13 +2458,29 @@ spellfix1_cursor *pCur, int idxNum, int argc, sqlite3_value **argv ){ + int rc; + char *zSql; + spellfix1_vtab *pVTab = pCur->pVTab; spellfix1ResetCursor(pCur); - spellfix1ResizeCursor(pCur, 0); - return SQLITE_OK; + zSql = sqlite3_mprintf( + "SELECT word, rank, NULL, langid, id FROM \"%w\".\"%w_vocab\"", + pVTab->zDbName, pVTab->zTableName); + if( zSql==0 ) return SQLITE_NOMEM; + rc = sqlite3_prepare_v2(pVTab->db, zSql, -1, &pCur->pFullScan, 0); + sqlite3_free(zSql); + pCur->nRow = pCur->iRow = 0; + if( rc==SQLITE_OK ){ + rc = sqlite3_step(pCur->pFullScan); + if( rc==SQLITE_ROW ){ pCur->iRow = -1; rc = SQLITE_OK; } + if( rc==SQLITE_DONE ){ rc = SQLITE_OK; } + }else{ + pCur->iRow = 0; + } + return rc; } /* ** Called to "rewind" a cursor back to the beginning so that @@ -2889,11 +2506,18 @@ /* ** Advance a cursor to its next row of output */ static int spellfix1Next(sqlite3_vtab_cursor *cur){ spellfix1_cursor *pCur = (spellfix1_cursor *)cur; - if( pCur->iRow < pCur->nRow ) pCur->iRow++; + if( pCur->iRow < pCur->nRow ){ + if( pCur->pFullScan ){ + int rc = sqlite3_step(pCur->pFullScan); + if( rc!=SQLITE_ROW ) pCur->iRow = pCur->nRow; + }else{ + pCur->iRow++; + } + } return SQLITE_OK; } /* ** Return TRUE if we are at the end-of-file @@ -2904,12 +2528,24 @@ } /* ** Return columns from the current row. */ -static int spellfix1Column(sqlite3_vtab_cursor *cur, sqlite3_context *ctx, int i){ +static int spellfix1Column( + sqlite3_vtab_cursor *cur, + sqlite3_context *ctx, + int i +){ spellfix1_cursor *pCur = (spellfix1_cursor*)cur; + if( pCur->pFullScan ){ + if( i<=SPELLFIX_COL_LANGID ){ + sqlite3_result_value(ctx, sqlite3_column_value(pCur->pFullScan, i)); + }else{ + sqlite3_result_null(ctx); + } + return SQLITE_OK; + } switch( i ){ case SPELLFIX_COL_WORD: { sqlite3_result_text(ctx, pCur->a[pCur->iRow].zWord, -1, SQLITE_STATIC); break; } @@ -2939,11 +2575,11 @@ if( nPattern>0 && pCur->zPattern[nPattern-1]=='*' ){ char *zTranslit; int res; zTranslit = (char *)transliterate((unsigned char *)zWord, nWord); if( !zTranslit ) return SQLITE_NOMEM; - res = editdist1(pCur->zPattern, zTranslit, pCur->iLang, &iMatchlen); + res = editdist1(pCur->zPattern, zTranslit, &iMatchlen); sqlite3_free(zTranslit); if( res<0 ) return SQLITE_NOMEM; iMatchlen = translen_to_charlen(zWord, nWord, iMatchlen); }else{ iMatchlen = utf8Charlen(zWord, nWord); @@ -2980,11 +2616,15 @@ /* ** The rowid. */ static int spellfix1Rowid(sqlite3_vtab_cursor *cur, sqlite_int64 *pRowid){ spellfix1_cursor *pCur = (spellfix1_cursor*)cur; - *pRowid = pCur->a[pCur->iRow].iRowid; + if( pCur->pFullScan ){ + *pRowid = sqlite3_column_int64(pCur->pFullScan, 4); + }else{ + *pRowid = pCur->a[pCur->iRow].iRowid; + } return SQLITE_OK; } /* ** The xUpdate() method. @@ -3064,12 +2704,12 @@ *pRowid = sqlite3_last_insert_rowid(db); }else{ rowid = sqlite3_value_int64(argv[0]); newRowid = *pRowid = sqlite3_value_int64(argv[1]); spellfix1DbExec(&rc, db, - "UPDATE \"%w\".\"%w_vocab\" SET id=%lld, rank=%d, lang=%d," - " word=%Q, rank=%d, k1=%Q, k2=%Q WHERE id=%lld", + "UPDATE \"%w\".\"%w_vocab\" SET id=%lld, rank=%d, langid=%d," + " word=%Q, k1=%Q, k2=%Q WHERE id=%lld", p->zDbName, p->zTableName, newRowid, iRank, iLang, zWord, zK1, zK2, rowid ); } sqlite3_free(zK1); @@ -3094,10 +2734,12 @@ p->zDbName, p->zTableName, zNewName ); if( rc==SQLITE_OK ){ sqlite3_free(p->zTableName); p->zTableName = zNewName; + }else{ + sqlite3_free(zNewName); } return rc; } @@ -3135,20 +2777,14 @@ int i; nErr += sqlite3_create_function(db, "spellfix1_translit", 1, SQLITE_UTF8, 0, transliterateSqlFunc, 0, 0); nErr += sqlite3_create_function(db, "spellfix1_editdist", 2, SQLITE_UTF8, 0, editdistSqlFunc, 0, 0); - nErr += sqlite3_create_function(db, "spellfix1_editdist", 3, SQLITE_UTF8, 0, - editdistSqlFunc, 0, 0); nErr += sqlite3_create_function(db, "spellfix1_phonehash", 1, SQLITE_UTF8, 0, phoneticHashSqlFunc, 0, 0); nErr += sqlite3_create_function(db, "spellfix1_scriptcode", 1, SQLITE_UTF8, 0, scriptCodeSqlFunc, 0, 0); - nErr += sqlite3_create_function(db, "pollock_skeleton", 1, SQLITE_UTF8, 0, - pollockSkeletonSqlFunc, 0, 0); - nErr += sqlite3_create_function(db, "pollock_omission", 1, SQLITE_UTF8, 0, - pollockOmissionSqlFunc, 0, 0); nErr += sqlite3_create_module(db, "spellfix1", &spellfix1Module, 0); nErr += editDist3Install(db); /* Verify sanity of the translit[] table */ for(i=0; i