1/* Copyright (C) 2006 MySQL AB & MySQL Finland AB & TCX DataKonsult AB
2
3 This program is free software; you can redistribute it and/or modify
4 it under the terms of the GNU General Public License as published by
5 the Free Software Foundation; version 2 of the License.
6
7 This program is distributed in the hope that it will be useful,
8 but WITHOUT ANY WARRANTY; without even the implied warranty of
9 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
10 GNU General Public License for more details.
11
12 You should have received a copy of the GNU General Public License
13 along with this program; if not, write to the Free Software
14 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111-1301 USA */
15
16/* Written by Sergei A. Golubchik, who has a shared copyright to this code */
17
18/* some definitions for full-text indices */
19
20#include "ma_fulltext.h"
21#include <m_ctype.h>
22#include <my_tree.h>
23#include <queues.h>
24#include <mysql/plugin.h>
25
26#define true_word_char(ctype, character) \
27 ((ctype) & (_MY_U | _MY_L | _MY_NMR) || \
28 (character) == '_')
29#define misc_word_char(X) 0
30
31#define FT_MAX_WORD_LEN_FOR_SORT 31
32
33#define FTPARSER_MEMROOT_ALLOC_SIZE 65536
34
35#define COMPILE_STOPWORDS_IN
36
37/* Interested readers may consult SMART
38 (ftp://ftp.cs.cornell.edu/pub/smart/smart.11.0.tar.Z)
39 for an excellent implementation of vector space model we use.
40 It also demonstrate the usage of different weghting techniques.
41 This code, though, is completely original and is not based on the
42 SMART code but was in some cases inspired by it.
43
44 NORM_PIVOT was taken from the article
45 A.Singhal, C.Buckley, M.Mitra, "Pivoted Document Length Normalization",
46 ACM SIGIR'96, 21-29, 1996
47 */
48
49#define LWS_FOR_QUERY LWS_TF
50#define LWS_IN_USE LWS_LOG
51#define PRENORM_IN_USE PRENORM_AVG
52#define NORM_IN_USE NORM_PIVOT
53#define GWS_IN_USE GWS_PROB
54/*==============================================================*/
55#define LWS_TF (count)
56#define LWS_BINARY (count>0)
57#define LWS_SQUARE (count*count)
58#define LWS_LOG (count?(log( (double) count)+1):0)
59/*--------------------------------------------------------------*/
60#define PRENORM_NONE (p->weight)
61#define PRENORM_MAX (p->weight/docstat.max)
62#define PRENORM_AUG (0.4+0.6*p->weight/docstat.max)
63#define PRENORM_AVG (p->weight/docstat.sum*docstat.uniq)
64#define PRENORM_AVGLOG ((1+log(p->weight))/(1+log(docstat.sum/docstat.uniq)))
65/*--------------------------------------------------------------*/
66#define NORM_NONE (1)
67#define NORM_SUM (docstat.nsum)
68#define NORM_COS (sqrt(docstat.nsum2))
69
70#define PIVOT_VAL (0.0115)
71#define NORM_PIVOT (1+PIVOT_VAL*docstat.uniq)
72/*---------------------------------------------------------------*/
73#define GWS_NORM (1/sqrt(sum2))
74#define GWS_GFIDF (sum/doc_cnt)
75/* Mysterious, but w/o (double) GWS_IDF performs better :-o */
76#define GWS_IDF log(aio->info->state->records/doc_cnt)
77#define GWS_IDF1 log((double)aio->info->state->records/doc_cnt)
78#define GWS_PROB ((aio->info->state->records > doc_cnt) ? log(((double)(aio->info->state->records-doc_cnt))/doc_cnt) : 0 )
79#define GWS_FREQ (1.0/doc_cnt)
80#define GWS_SQUARED pow(log((double)aio->info->state->records/doc_cnt),2)
81#define GWS_CUBIC pow(log((double)aio->info->state->records/doc_cnt),3)
82#define GWS_ENTROPY (1-(suml/sum-log(sum))/log(aio->info->state->records))
83/*=================================================================*/
84
85/* Boolean search operators */
86#define FTB_YES (ft_boolean_syntax[0])
87#define FTB_EGAL (ft_boolean_syntax[1])
88#define FTB_NO (ft_boolean_syntax[2])
89#define FTB_INC (ft_boolean_syntax[3])
90#define FTB_DEC (ft_boolean_syntax[4])
91#define FTB_LBR (ft_boolean_syntax[5])
92#define FTB_RBR (ft_boolean_syntax[6])
93#define FTB_NEG (ft_boolean_syntax[7])
94#define FTB_TRUNC (ft_boolean_syntax[8])
95#define FTB_LQUOT (ft_boolean_syntax[10])
96#define FTB_RQUOT (ft_boolean_syntax[11])
97
98typedef struct st_maria_ft_word {
99 const uchar * pos;
100 uint len;
101 double weight;
102} FT_WORD;
103
104int is_stopword(const char *word, size_t len);
105
106MARIA_KEY *_ma_ft_make_key(MARIA_HA *, MARIA_KEY *, uint , uchar *, FT_WORD *,
107 my_off_t);
108
109uchar maria_ft_get_word(CHARSET_INFO *, const uchar **, const uchar *,
110 FT_WORD *, MYSQL_FTPARSER_BOOLEAN_INFO *);
111uchar maria_ft_simple_get_word(CHARSET_INFO *, uchar **, const uchar *,
112 FT_WORD *, my_bool);
113
114typedef struct _st_maria_ft_seg_iterator {
115 uint num, len;
116 HA_KEYSEG *seg;
117 const uchar *rec, *pos;
118} FT_SEG_ITERATOR;
119
120void _ma_ft_segiterator_init(MARIA_HA *, uint, const uchar *, FT_SEG_ITERATOR *);
121void _ma_ft_segiterator_dummy_init(const uchar *, uint, FT_SEG_ITERATOR *);
122uint _ma_ft_segiterator(FT_SEG_ITERATOR *);
123
124void maria_ft_parse_init(TREE *, CHARSET_INFO *);
125int maria_ft_parse(TREE *, uchar *, int, struct st_mysql_ftparser *parser,
126 MYSQL_FTPARSER_PARAM *, MEM_ROOT *);
127FT_WORD * maria_ft_linearize(TREE *, MEM_ROOT *);
128FT_WORD * _ma_ft_parserecord(MARIA_HA *, uint, const uchar *, MEM_ROOT *);
129uint _ma_ft_parse(TREE *, MARIA_HA *, uint, const uchar *,
130 MYSQL_FTPARSER_PARAM *, MEM_ROOT *);
131
132FT_INFO *maria_ft_init_nlq_search(MARIA_HA *, uint, uchar *, uint, uint,
133 uchar *);
134FT_INFO *maria_ft_init_boolean_search(MARIA_HA *, uint, uchar *, uint,
135 CHARSET_INFO *);
136
137extern const struct _ft_vft _ma_ft_vft_nlq;
138int maria_ft_nlq_read_next(FT_INFO *, char *);
139float maria_ft_nlq_find_relevance(FT_INFO *, uchar *, uint);
140void maria_ft_nlq_close_search(FT_INFO *);
141float maria_ft_nlq_get_relevance(FT_INFO *);
142my_off_t maria_ft_nlq_get_docid(FT_INFO *);
143void maria_ft_nlq_reinit_search(FT_INFO *);
144
145extern const struct _ft_vft _ma_ft_vft_boolean;
146int maria_ft_boolean_read_next(FT_INFO *, char *);
147float maria_ft_boolean_find_relevance(FT_INFO *, uchar *, uint);
148void maria_ft_boolean_close_search(FT_INFO *);
149float maria_ft_boolean_get_relevance(FT_INFO *);
150my_off_t maria_ft_boolean_get_docid(FT_INFO *);
151void maria_ft_boolean_reinit_search(FT_INFO *);
152MYSQL_FTPARSER_PARAM* maria_ftparser_alloc_param(MARIA_HA *info);
153extern MYSQL_FTPARSER_PARAM *maria_ftparser_call_initializer(MARIA_HA *info,
154 uint keynr,
155 uint paramnr);
156extern void maria_ftparser_call_deinitializer(MARIA_HA *info);
157