From 06eaf7232e9a920468c0f8d74dcf2fe8b555501c Mon Sep 17 00:00:00 2001 From: Daniel Baumann Date: Sat, 13 Apr 2024 14:24:36 +0200 Subject: Adding upstream version 1:10.11.6. Signed-off-by: Daniel Baumann --- storage/maria/ma_ft_parser.c | 398 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 398 insertions(+) create mode 100644 storage/maria/ma_ft_parser.c (limited to 'storage/maria/ma_ft_parser.c') diff --git a/storage/maria/ma_ft_parser.c b/storage/maria/ma_ft_parser.c new file mode 100644 index 00000000..f600873d --- /dev/null +++ b/storage/maria/ma_ft_parser.c @@ -0,0 +1,398 @@ +/* Copyright (C) 2006 MySQL AB & MySQL Finland AB & TCX DataKonsult AB + Copyright (c) 2020, MariaDB Corporation. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; version 2 of the License. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335 USA */ + +/* Written by Sergei A. Golubchik, who has a shared copyright to this code */ + +#include "ma_ftdefs.h" + +typedef struct st_maria_ft_docstat { + FT_WORD *list; + uint uniq; + double sum; +} FT_DOCSTAT; + + +typedef struct st_my_maria_ft_parser_param +{ + TREE *wtree; + MEM_ROOT *mem_root; +} MY_FT_PARSER_PARAM; + + +static int FT_WORD_cmp(CHARSET_INFO* cs, FT_WORD *w1, FT_WORD *w2) +{ + return ha_compare_word(cs, (uchar*) w1->pos, w1->len, + (uchar*) w2->pos, w2->len); +} + +static int walk_and_copy(FT_WORD *word,uint32 count,FT_DOCSTAT *docstat) +{ + word->weight=LWS_IN_USE; + docstat->sum+=word->weight; + memcpy((docstat->list)++, word, sizeof(FT_WORD)); + return 0; +} + +/* transforms tree of words into the array, applying normalization */ + +FT_WORD * maria_ft_linearize(TREE *wtree, MEM_ROOT *mem_root) +{ + FT_WORD *wlist,*p; + FT_DOCSTAT docstat; + DBUG_ENTER("maria_ft_linearize"); + + if ((wlist=(FT_WORD *) alloc_root(mem_root, sizeof(FT_WORD)* + (1+wtree->elements_in_tree)))) + { + docstat.list=wlist; + docstat.uniq=wtree->elements_in_tree; + docstat.sum=0; + tree_walk(wtree,(tree_walk_action)&walk_and_copy,&docstat,left_root_right); + } + delete_tree(wtree, 0); + if (!wlist) + DBUG_RETURN(NULL); + + docstat.list->pos=NULL; + + for (p=wlist;p->pos;p++) + { + p->weight=PRENORM_IN_USE; + } + + for (p=wlist;p->pos;p++) + { + p->weight/=NORM_IN_USE; + } + + DBUG_RETURN(wlist); +} + +/* + RETURN VALUE + 0 - eof + 1 - word found + 2 - left bracket + 3 - right bracket + 4 - stopword found +*/ +uchar maria_ft_get_word(CHARSET_INFO *cs, const uchar **start, + const uchar *end, + FT_WORD *word, MYSQL_FTPARSER_BOOLEAN_INFO *param) +{ + const uchar *doc= *start; + int ctype; + uint mwc, length; + int mbl; + + param->yesno=(FTB_YES==' ') ? 1 : (param->quot != 0); + param->weight_adjust= param->wasign= 0; + param->type= FT_TOKEN_EOF; + + while (doc 0 ? mbl : (mbl < 0 ? -mbl : 1))) + { + mbl= my_ci_ctype(cs, &ctype, doc, end); + if (true_word_char(ctype, *doc)) + break; + if (*doc == FTB_RQUOT && param->quot) + { + param->quot= (char *) doc; + *start=doc+1; + param->type= FT_TOKEN_RIGHT_PAREN; + goto ret; + } + if (!param->quot) + { + if (*doc == FTB_LBR || *doc == FTB_RBR || *doc == FTB_LQUOT) + { + /* param->prev=' '; */ + *start=doc+1; + if (*doc == FTB_LQUOT) + param->quot= (char *) *start; + param->type= (*doc == FTB_RBR ? FT_TOKEN_RIGHT_PAREN : FT_TOKEN_LEFT_PAREN); + goto ret; + } + if (param->prev == ' ') + { + if (*doc == FTB_YES ) { param->yesno=+1; continue; } else + if (*doc == FTB_EGAL) { param->yesno= 0; continue; } else + if (*doc == FTB_NO ) { param->yesno=-1; continue; } else + if (*doc == FTB_INC ) { param->weight_adjust++; continue; } else + if (*doc == FTB_DEC ) { param->weight_adjust--; continue; } else + if (*doc == FTB_NEG ) { param->wasign= !param->wasign; continue; } + } + } + param->prev=*doc; + param->yesno=(FTB_YES==' ') ? 1 : (param->quot != 0); + param->weight_adjust= param->wasign= 0; + } + + mwc=length=0; + for (word->pos= doc; doc < end; length++, + doc+= (mbl > 0 ? mbl : (mbl < 0 ? -mbl : 1))) + { + mbl= my_ci_ctype(cs, &ctype, doc, end); + if (true_word_char(ctype, *doc)) + mwc=0; + else if (!misc_word_char(*doc) || mwc) + break; + else + mwc++; + } + param->prev='A'; /* be sure *prev is true_word_char */ + word->len= (uint)(doc-word->pos) - mwc; + if ((param->trunc=(doc= ft_min_word_len && !is_stopword((char *) word->pos, + word->len)) + || param->trunc) && length < ft_max_word_len) + { + *start=doc; + param->type= FT_TOKEN_WORD; + goto ret; + } + else if (length) /* make sure length > 0 (if start contains spaces only) */ + { + *start= doc; + param->type= FT_TOKEN_STOPWORD; + goto ret; + } + } + if (param->quot) + { + param->quot= (char *)(*start= doc); + param->type= 3; /* FT_RBR */ + goto ret; + } +ret: + return param->type; +} + +uchar maria_ft_simple_get_word(CHARSET_INFO *cs, uchar **start, + const uchar *end, FT_WORD *word, + my_bool skip_stopwords) +{ + uchar *doc= *start; + uint mwc, length; + int ctype, mbl; + DBUG_ENTER("maria_ft_simple_get_word"); + + do + { + for (;; doc+= (mbl > 0 ? mbl : (mbl < 0 ? -mbl : 1))) + { + if (doc >= end) + DBUG_RETURN(0); + mbl= my_ci_ctype(cs, &ctype, doc, end); + if (true_word_char(ctype, *doc)) + break; + } + + mwc= length= 0; + for (word->pos= doc; doc < end; length++, + doc+= (mbl > 0 ? mbl : (mbl < 0 ? -mbl : 1))) + { + mbl= my_ci_ctype(cs, &ctype, doc, end); + if (true_word_char(ctype, *doc)) + mwc= 0; + else if (!misc_word_char(*doc) || mwc) + break; + else + mwc++; + } + + word->len= (uint)(doc-word->pos) - mwc; + + if (skip_stopwords == FALSE || + (length >= ft_min_word_len && length < ft_max_word_len && + !is_stopword((char *) word->pos, word->len))) + { + *start= doc; + DBUG_RETURN(1); + } + } while (doc < end); + DBUG_RETURN(0); +} + +void maria_ft_parse_init(TREE *wtree, CHARSET_INFO *cs) +{ + DBUG_ENTER("maria_ft_parse_init"); + if (!is_tree_inited(wtree)) + init_tree(wtree,0,0,sizeof(FT_WORD),(qsort_cmp2)&FT_WORD_cmp, NULL, + (void*) cs, MYF(0)); + DBUG_VOID_RETURN; +} + + +static int maria_ft_add_word(MYSQL_FTPARSER_PARAM *param, + const char *word, int word_len, + MYSQL_FTPARSER_BOOLEAN_INFO *boolean_info + __attribute__((unused))) +{ + TREE *wtree; + FT_WORD w; + MY_FT_PARSER_PARAM *ft_param=param->mysql_ftparam; + DBUG_ENTER("maria_ft_add_word"); + wtree= ft_param->wtree; + if (param->flags & MYSQL_FTFLAGS_NEED_COPY) + { + uchar *ptr; + DBUG_ASSERT(wtree->with_delete == 0); + ptr= (uchar *)alloc_root(ft_param->mem_root, word_len); + memcpy(ptr, word, word_len); + w.pos= ptr; + } + else + w.pos= (uchar*) word; + w.len= word_len; + if (!tree_insert(wtree, &w, 0, wtree->custom_arg)) + { + delete_tree(wtree, 0); + DBUG_RETURN(1); + } + DBUG_RETURN(0); +} + + +static int maria_ft_parse_internal(MYSQL_FTPARSER_PARAM *param, + const char *doc_arg, + int doc_len) +{ + uchar *doc= (uchar*) doc_arg; + uchar *end= doc + doc_len; + MY_FT_PARSER_PARAM *ft_param=param->mysql_ftparam; + TREE *wtree= ft_param->wtree; + FT_WORD w; + DBUG_ENTER("maria_ft_parse_internal"); + + while (maria_ft_simple_get_word(wtree->custom_arg, &doc, end, &w, TRUE)) + if (param->mysql_add_word(param, (char*)w.pos, w.len, 0)) + DBUG_RETURN(1); + DBUG_RETURN(0); +} + + +int maria_ft_parse(TREE *wtree, uchar *doc, int doclen, + struct st_mysql_ftparser *parser, + MYSQL_FTPARSER_PARAM *param, MEM_ROOT *mem_root) +{ + MY_FT_PARSER_PARAM my_param; + DBUG_ENTER("maria_ft_parse"); + DBUG_ASSERT(parser); + my_param.wtree= wtree; + my_param.mem_root= mem_root; + + param->mysql_parse= maria_ft_parse_internal; + param->mysql_add_word= maria_ft_add_word; + param->mysql_ftparam= &my_param; + param->cs= wtree->custom_arg; + param->doc= (char*)doc; + param->length= doclen; + param->mode= MYSQL_FTPARSER_SIMPLE_MODE; + DBUG_RETURN(parser->parse(param)); +} + + +#define MAX_PARAM_NR 2 + +MYSQL_FTPARSER_PARAM* maria_ftparser_alloc_param(MARIA_HA *info) +{ + if (!info->ftparser_param) + { + /* +. info->ftparser_param can not be zero after the initialization, + because it always includes built-in fulltext parser. And built-in + parser can be called even if the table has no fulltext indexes and + no varchar/text fields. + + ftb_find_relevance... parser (ftb_find_relevance_parse, + ftb_find_relevance_add_word) calls ftb_check_phrase... parser + (ftb_check_phrase_internal, ftb_phrase_add_word). Thus MAX_PARAM_NR=2. + */ + info->ftparser_param= (MYSQL_FTPARSER_PARAM *) + my_malloc(PSI_INSTRUMENT_ME, MAX_PARAM_NR * sizeof(MYSQL_FTPARSER_PARAM) * + info->s->ftkeys, MYF(MY_WME | MY_ZEROFILL)); + init_alloc_root(PSI_INSTRUMENT_ME, &info->ft_memroot, + FTPARSER_MEMROOT_ALLOC_SIZE, 0, MYF(0)); + } + return info->ftparser_param; +} + + +MYSQL_FTPARSER_PARAM *maria_ftparser_call_initializer(MARIA_HA *info, + uint keynr, uint paramnr) +{ + uint32 ftparser_nr; + struct st_mysql_ftparser *parser; + + if (!maria_ftparser_alloc_param(info)) + return 0; + + if (keynr == NO_SUCH_KEY) + { + ftparser_nr= 0; + parser= &ft_default_parser; + } + else + { + ftparser_nr= info->s->keyinfo[keynr].ftkey_nr; + parser= info->s->keyinfo[keynr].parser; + } + DBUG_ASSERT(paramnr < MAX_PARAM_NR); + ftparser_nr= ftparser_nr*MAX_PARAM_NR + paramnr; + if (! info->ftparser_param[ftparser_nr].mysql_add_word) + { + /* Note, that mysql_add_word is used here as a flag: + mysql_add_word == 0 - parser is not initialized + mysql_add_word != 0 - parser is initialized, or no + initialization needed. */ + info->ftparser_param[ftparser_nr].mysql_add_word= + (int (*)(struct st_mysql_ftparser_param *, const char *, + int, MYSQL_FTPARSER_BOOLEAN_INFO *)) 1; + if (parser->init && parser->init(&info->ftparser_param[ftparser_nr])) + return 0; + } + return &info->ftparser_param[ftparser_nr]; +} + + +void maria_ftparser_call_deinitializer(MARIA_HA *info) +{ + uint i, j, keys= info->s->state.header.keys; + free_root(&info->ft_memroot, MYF(0)); + if (! info->ftparser_param) + return; + for (i= 0; i < keys; i++) + { + MARIA_KEYDEF *keyinfo= &info->s->keyinfo[i]; + for (j=0; j < MAX_PARAM_NR; j++) + { + MYSQL_FTPARSER_PARAM *ftparser_param= + &info->ftparser_param[keyinfo->ftkey_nr*MAX_PARAM_NR + j]; + if (keyinfo->flag & HA_FULLTEXT && ftparser_param->mysql_add_word) + { + if (keyinfo->parser->deinit) + keyinfo->parser->deinit(ftparser_param); + ftparser_param->mysql_add_word= 0; + } + else + break; + } + } +} -- cgit v1.2.3