| 1 | /***************************************************************************** |
| 2 | |
| 3 | Copyright (c) 2013, 2015, Oracle and/or its affiliates. All Rights Reserved. |
| 4 | |
| 5 | This program is free software; you can redistribute it and/or modify it under |
| 6 | the terms of the GNU General Public License as published by the Free Software |
| 7 | Foundation; version 2 of the License. |
| 8 | |
| 9 | This program is distributed in the hope that it will be useful, but WITHOUT |
| 10 | ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS |
| 11 | FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. |
| 12 | |
| 13 | You should have received a copy of the GNU General Public License along with |
| 14 | this program; if not, write to the Free Software Foundation, Inc., |
| 15 | 51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA |
| 16 | |
| 17 | *****************************************************************************/ |
| 18 | |
| 19 | /******************************************************************//** |
| 20 | @file fts/fts0plugin.cc |
| 21 | Full Text Search plugin support. |
| 22 | |
| 23 | Created 2013/06/04 Shaohua Wang |
| 24 | ***********************************************************************/ |
| 25 | |
| 26 | #include "fts0ast.h" |
| 27 | #include "fts0plugin.h" |
| 28 | #include "fts0tokenize.h" |
| 29 | |
| 30 | #include "ft_global.h" |
| 31 | |
| 32 | /******************************************************************//** |
| 33 | FTS default parser init |
| 34 | @return 0 */ |
| 35 | static int fts_default_parser_init(MYSQL_FTPARSER_PARAM*) { return 0; } |
| 36 | |
| 37 | /******************************************************************//** |
| 38 | FTS default parser deinit |
| 39 | @return 0 */ |
| 40 | static int fts_default_parser_deinit(MYSQL_FTPARSER_PARAM*) { return 0; } |
| 41 | |
| 42 | /******************************************************************//** |
| 43 | FTS default parser parse from ft_static.c in MYISAM. |
| 44 | @return 0 if parse successfully, or return non-zero */ |
| 45 | static |
| 46 | int |
| 47 | fts_default_parser_parse( |
| 48 | /*=====================*/ |
| 49 | MYSQL_FTPARSER_PARAM *param) /*!< in: plugin parser param */ |
| 50 | { |
| 51 | return(param->mysql_parse(param, param->doc, param->length)); |
| 52 | } |
| 53 | |
| 54 | /* FTS default parser from ft_static.c in MYISAM. */ |
| 55 | struct st_mysql_ftparser fts_default_parser = |
| 56 | { |
| 57 | MYSQL_FTPARSER_INTERFACE_VERSION, |
| 58 | fts_default_parser_parse, |
| 59 | fts_default_parser_init, |
| 60 | fts_default_parser_deinit |
| 61 | }; |
| 62 | |
| 63 | /******************************************************************//** |
| 64 | Get a operator node from token boolean info |
| 65 | @return node */ |
| 66 | static |
| 67 | fts_ast_node_t* |
| 68 | fts_query_get_oper_node( |
| 69 | /*====================*/ |
| 70 | MYSQL_FTPARSER_BOOLEAN_INFO* info, /*!< in: token info */ |
| 71 | fts_ast_state_t* state) /*!< in/out: query parse state*/ |
| 72 | { |
| 73 | fts_ast_node_t* oper_node = NULL; |
| 74 | |
| 75 | if (info->yesno > 0) { |
| 76 | oper_node = fts_ast_create_node_oper(state, FTS_EXIST); |
| 77 | } else if (info->yesno < 0) { |
| 78 | oper_node = fts_ast_create_node_oper(state, FTS_IGNORE); |
| 79 | } else if (info->weight_adjust > 0) { |
| 80 | oper_node = fts_ast_create_node_oper(state, FTS_INCR_RATING); |
| 81 | } else if (info->weight_adjust < 0) { |
| 82 | oper_node = fts_ast_create_node_oper(state, FTS_DECR_RATING); |
| 83 | } else if (info->wasign > 0) { |
| 84 | oper_node = fts_ast_create_node_oper(state, FTS_NEGATE); |
| 85 | } |
| 86 | |
| 87 | return(oper_node); |
| 88 | } |
| 89 | |
| 90 | /******************************************************************//** |
| 91 | FTS plugin parser 'myql_add_word' callback function for query parse. |
| 92 | Refer to 'st_mysql_ftparser_param' for more detail. |
| 93 | Note: |
| 94 | a. Parse logic refers to 'ftb_query_add_word' from ft_boolean_search.c in MYISAM; |
| 95 | b. Parse node or tree refers to fts0pars.y. |
| 96 | @return 0 if add successfully, or return non-zero. */ |
| 97 | static |
| 98 | int |
| 99 | fts_query_add_word_for_parser( |
| 100 | /*==========================*/ |
| 101 | MYSQL_FTPARSER_PARAM* param, /*!< in: parser param */ |
| 102 | const char* word, /*!< in: token */ |
| 103 | int word_len, /*!< in: token length */ |
| 104 | MYSQL_FTPARSER_BOOLEAN_INFO* info) /*!< in: token info */ |
| 105 | { |
| 106 | fts_ast_state_t* state = |
| 107 | static_cast<fts_ast_state_t*>(param->mysql_ftparam); |
| 108 | fts_ast_node_t* cur_node = state->cur_node; |
| 109 | fts_ast_node_t* oper_node = NULL; |
| 110 | fts_ast_node_t* term_node = NULL; |
| 111 | fts_ast_node_t* node = NULL; |
| 112 | |
| 113 | switch (info->type) { |
| 114 | case FT_TOKEN_STOPWORD: |
| 115 | /* We only handler stopword in phrase */ |
| 116 | if (cur_node->type != FTS_AST_PARSER_PHRASE_LIST) { |
| 117 | break; |
| 118 | } |
| 119 | /* fall through */ |
| 120 | |
| 121 | case FT_TOKEN_WORD: |
| 122 | term_node = fts_ast_create_node_term_for_parser( |
| 123 | state, word, ulint(word_len)); |
| 124 | |
| 125 | if (info->trunc) { |
| 126 | fts_ast_term_set_wildcard(term_node); |
| 127 | } |
| 128 | |
| 129 | if (cur_node->type == FTS_AST_PARSER_PHRASE_LIST) { |
| 130 | /* Ignore operator inside phrase */ |
| 131 | fts_ast_add_node(cur_node, term_node); |
| 132 | } else { |
| 133 | ut_ad(cur_node->type == FTS_AST_LIST |
| 134 | || cur_node->type == FTS_AST_SUBEXP_LIST); |
| 135 | oper_node = fts_query_get_oper_node(info, state); |
| 136 | |
| 137 | if (oper_node) { |
| 138 | node = fts_ast_create_node_list(state, oper_node); |
| 139 | fts_ast_add_node(node, term_node); |
| 140 | fts_ast_add_node(cur_node, node); |
| 141 | } else { |
| 142 | fts_ast_add_node(cur_node, term_node); |
| 143 | } |
| 144 | } |
| 145 | |
| 146 | break; |
| 147 | |
| 148 | case FT_TOKEN_LEFT_PAREN: |
| 149 | /* Check parse error */ |
| 150 | if (cur_node->type != FTS_AST_LIST |
| 151 | && cur_node->type != FTS_AST_SUBEXP_LIST) { |
| 152 | return(1); |
| 153 | } |
| 154 | |
| 155 | /* Set operator */ |
| 156 | oper_node = fts_query_get_oper_node(info, state); |
| 157 | if (oper_node != NULL) { |
| 158 | node = fts_ast_create_node_list(state, oper_node); |
| 159 | fts_ast_add_node(cur_node, node); |
| 160 | node->go_up = true; |
| 161 | node->up_node = cur_node; |
| 162 | cur_node = node; |
| 163 | } |
| 164 | |
| 165 | if (info->quot) { |
| 166 | /* Phrase node */ |
| 167 | node = fts_ast_create_node_phrase_list(state); |
| 168 | } else { |
| 169 | /* Subexp list node */ |
| 170 | node = fts_ast_create_node_subexp_list(state, NULL); |
| 171 | } |
| 172 | |
| 173 | fts_ast_add_node(cur_node, node); |
| 174 | |
| 175 | node->up_node = cur_node; |
| 176 | state->cur_node = node; |
| 177 | state->depth += 1; |
| 178 | |
| 179 | break; |
| 180 | |
| 181 | case FT_TOKEN_RIGHT_PAREN: |
| 182 | info->quot = 0; |
| 183 | |
| 184 | if (cur_node->up_node != NULL) { |
| 185 | cur_node = cur_node->up_node; |
| 186 | |
| 187 | if (cur_node->go_up) { |
| 188 | ut_a(cur_node->up_node |
| 189 | && !(cur_node->up_node->go_up)); |
| 190 | cur_node = cur_node->up_node; |
| 191 | } |
| 192 | } |
| 193 | |
| 194 | state->cur_node = cur_node; |
| 195 | |
| 196 | if (state->depth > 0) { |
| 197 | state->depth--; |
| 198 | } else { |
| 199 | /* Parentheses mismatch */ |
| 200 | return(1); |
| 201 | } |
| 202 | |
| 203 | break; |
| 204 | |
| 205 | case FT_TOKEN_EOF: |
| 206 | default: |
| 207 | break; |
| 208 | } |
| 209 | |
| 210 | return(0); |
| 211 | } |
| 212 | |
| 213 | /******************************************************************//** |
| 214 | FTS plugin parser 'myql_parser' callback function for query parse. |
| 215 | Refer to 'st_mysql_ftparser_param' for more detail. |
| 216 | @return 0 if parse successfully */ |
| 217 | static |
| 218 | int |
| 219 | fts_parse_query_internal( |
| 220 | /*=====================*/ |
| 221 | MYSQL_FTPARSER_PARAM* param, /*!< in: parser param */ |
| 222 | const char* query, /*!< in: query string */ |
| 223 | int len) /*!< in: query length */ |
| 224 | { |
| 225 | MYSQL_FTPARSER_BOOLEAN_INFO info; |
| 226 | const CHARSET_INFO* cs = param->cs; |
| 227 | uchar** start = (uchar**)(&query); |
| 228 | uchar* end = (uchar*)(query + len); |
| 229 | FT_WORD w = {NULL, 0, 0}; |
| 230 | |
| 231 | info.prev = ' '; |
| 232 | info.quot = 0; |
| 233 | memset(&w, 0, sizeof(w)); |
| 234 | /* Note: We don't handle simple parser mode here, |
| 235 | but user supplied plugin parser should handler it. */ |
| 236 | while (fts_get_word(cs, start, end, &w, &info)) { |
| 237 | int ret = param->mysql_add_word( |
| 238 | param, |
| 239 | reinterpret_cast<char*>(w.pos), |
| 240 | int(w.len), &info); |
| 241 | if (ret) { |
| 242 | return(ret); |
| 243 | } |
| 244 | } |
| 245 | |
| 246 | return(0); |
| 247 | } |
| 248 | |
| 249 | /******************************************************************//** |
| 250 | fts parse query by plugin parser. |
| 251 | @return 0 if parse successfully, or return non-zero. */ |
| 252 | int |
| 253 | fts_parse_by_parser( |
| 254 | /*================*/ |
| 255 | ibool mode, /*!< in: parse boolean mode */ |
| 256 | uchar* query_str, /*!< in: query string */ |
| 257 | ulint query_len, /*!< in: query string length */ |
| 258 | st_mysql_ftparser* parser, /*!< in: fts plugin parser */ |
| 259 | fts_ast_state_t* state) /*!< in/out: parser state */ |
| 260 | { |
| 261 | MYSQL_FTPARSER_PARAM param; |
| 262 | int ret; |
| 263 | |
| 264 | ut_ad(parser); |
| 265 | |
| 266 | /* Initial parser param */ |
| 267 | param.mysql_parse = fts_parse_query_internal; |
| 268 | param.mysql_add_word = fts_query_add_word_for_parser; |
| 269 | param.mysql_ftparam = static_cast<void*>(state); |
| 270 | param.cs = state->charset; |
| 271 | param.doc = reinterpret_cast<char*>(query_str); |
| 272 | param.length = static_cast<int>(query_len); |
| 273 | param.flags = 0; |
| 274 | param.mode = mode ? |
| 275 | MYSQL_FTPARSER_FULL_BOOLEAN_INFO : |
| 276 | MYSQL_FTPARSER_SIMPLE_MODE; |
| 277 | |
| 278 | PARSER_INIT(parser, ¶m); |
| 279 | ret = parser->parse(¶m); |
| 280 | PARSER_DEINIT(parser, ¶m); |
| 281 | |
| 282 | return(ret | state->depth); |
| 283 | } |
| 284 | |