includes/search/SearchMySQL.php

   1 <?php
   2 /**
   3  * MySQL search engine
   4  *
   5  * Copyright (C) 2004 Brion Vibber <brion@pobox.com>
   6  * https://www.mediawiki.org/
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Search
  25  */
  26
  27 /**
  28  * Search engine hook for MySQL 4+
  29  * @ingroup Search
  30  */
  31 class SearchMySQL extends SearchDatabase {
  32         var $strictMatching = true;
  33         static $mMinSearchLength;
  34
  35         /**
  36          * Parse the user's query and transform it into an SQL fragment which will
  37          * become part of a WHERE clause
  38          *
  39          * @param $filteredText string
  40          * @param $fulltext string
  41          *
  42          * @return string
  43          */
  44         function parseQuery( $filteredText, $fulltext ) {
  45                 global $wgContLang;
  46                 $lc = SearchEngine::legalSearchChars(); // Minus format chars
  47                 $searchon = '';
  48                 $this->searchTerms = array();
  49
  50                 # @todo FIXME: This doesn't handle parenthetical expressions.
  51                 $m = array();
  52                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  53                                 $filteredText, $m, PREG_SET_ORDER ) ) {
  54                         foreach ( $m as $bits ) {
  55                                 @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  56
  57                                 if ( $nonQuoted != '' ) {
  58                                         $term = $nonQuoted;
  59                                         $quote = '';
  60                                 } else {
  61                                         $term = str_replace( '"', '', $term );
  62                                         $quote = '"';
  63                                 }
  64
  65                                 if ( $searchon !== '' ) {
  66                                         $searchon .= ' ';
  67                                 }
  68                                 if ( $this->strictMatching && ( $modifier == '' ) ) {
  69                                         // If we leave this out, boolean op defaults to OR which is rarely helpful.
  70                                         $modifier = '+';
  71                                 }
  72
  73                                 // Some languages such as Serbian store the input form in the search index,
  74                                 // so we may need to search for matches in multiple writing system variants.
  75                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  76                                 if ( is_array( $convertedVariants ) ) {
  77                                         $variants = array_unique( array_values( $convertedVariants ) );
  78                                 } else {
  79                                         $variants = array( $term );
  80                                 }
  81
  82                                 // The low-level search index does some processing on input to work
  83                                 // around problems with minimum lengths and encoding in MySQL's
  84                                 // fulltext engine.
  85                                 // For Chinese this also inserts spaces between adjacent Han characters.
  86                                 $strippedVariants = array_map(
  87                                         array( $wgContLang, 'normalizeForSearch' ),
  88                                         $variants );
  89
  90                                 // Some languages such as Chinese force all variants to a canonical
  91                                 // form when stripping to the low-level search index, so to be sure
  92                                 // let's check our variants list for unique items after stripping.
  93                                 $strippedVariants = array_unique( $strippedVariants );
  94
  95                                 $searchon .= $modifier;
  96                                 if ( count( $strippedVariants ) > 1 ) {
  97                                         $searchon .= '(';
  98                                 }
  99                                 foreach ( $strippedVariants as $stripped ) {
 100                                         $stripped = $this->normalizeText( $stripped );
 101                                         if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
 102                                                 // Hack for Chinese: we need to toss in quotes for
 103                                                 // multiple-character phrases since normalizeForSearch()
 104                                                 // added spaces between them to make word breaks.
 105                                                 $stripped = '"' . trim( $stripped ) . '"';
 106                                         }
 107                                         $searchon .= "$quote$stripped$quote$wildcard ";
 108                                 }
 109                                 if ( count( $strippedVariants ) > 1 ) {
 110                                         $searchon .= ')';
 111                                 }
 112
 113                                 // Match individual terms or quoted phrase in result highlighting...
 114                                 // Note that variants will be introduced in a later stage for highlighting!
 115                                 $regexp = $this->regexTerm( $term, $wildcard );
 116                                 $this->searchTerms[] = $regexp;
 117                         }
 118                         wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
 119                         wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
 120                 } else {
 121                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 122                 }
 123
 124                 $searchon = $this->db->strencode( $searchon );
 125                 $field = $this->getIndexField( $fulltext );
 126                 return " MATCH($field) AGAINST('$searchon' IN BOOLEAN MODE) ";
 127         }
 128
 129         function regexTerm( $string, $wildcard ) {
 130                 global $wgContLang;
 131
 132                 $regex = preg_quote( $string, '/' );
 133                 if ( $wgContLang->hasWordBreaks() ) {
 134                         if ( $wildcard ) {
 135                                 // Don't cut off the final bit!
 136                                 $regex = "\b$regex";
 137                         } else {
 138                                 $regex = "\b$regex\b";
 139                         }
 140                 } else {
 141                         // For Chinese, words may legitimately abut other words in the text literal.
 142                         // Don't add \b boundary checks... note this could cause false positives
 143                         // for latin chars.
 144                 }
 145                 return $regex;
 146         }
 147
 148         public static function legalSearchChars() {
 149                 return "\"*" . parent::legalSearchChars();
 150         }
 151
 152         /**
 153          * Perform a full text search query and return a result set.
 154          *
 155          * @param string $term raw search term
 156          * @return MySQLSearchResultSet
 157          */
 158         function searchText( $term ) {
 159                 return $this->searchInternal( $term, true );
 160         }
 161
 162         /**
 163          * Perform a title-only search query and return a result set.
 164          *
 165          * @param string $term raw search term
 166          * @return MySQLSearchResultSet
 167          */
 168         function searchTitle( $term ) {
 169                 return $this->searchInternal( $term, false );
 170         }
 171
 172         protected function searchInternal( $term, $fulltext ) {
 173                 global $wgCountTotalSearchHits;
 174
 175                 // This seems out of place, why is this called with empty term?
 176                 if ( trim( $term ) === '' ) {
 177                         return null;
 178                 }
 179
 180                 $filteredTerm = $this->filter( $term );
 181                 $query = $this->getQuery( $filteredTerm, $fulltext );
 182                 $resultSet = $this->db->select(
 183                         $query['tables'], $query['fields'], $query['conds'],
 184                         __METHOD__, $query['options'], $query['joins']
 185                 );
 186
 187                 $total = null;
 188                 if ( $wgCountTotalSearchHits ) {
 189                         $query = $this->getCountQuery( $filteredTerm, $fulltext );
 190                         $totalResult = $this->db->select(
 191                                 $query['tables'], $query['fields'], $query['conds'],
 192                                 __METHOD__, $query['options'], $query['joins']
 193                         );
 194
 195                         $row = $totalResult->fetchObject();
 196                         if ( $row ) {
 197                                 $total = intval( $row->c );
 198                         }
 199                         $totalResult->free();
 200                 }
 201
 202                 return new MySQLSearchResultSet( $resultSet, $this->searchTerms, $total );
 203         }
 204
 205         public function supports( $feature ) {
 206                 switch ( $feature ) {
 207                 case 'title-suffix-filter':
 208                         return true;
 209                 default:
 210                         return parent::supports( $feature );
 211                 }
 212         }
 213
 214         /**
 215          * Add special conditions
 216          * @param $query Array
 217          * @since 1.18
 218          */
 219         protected function queryFeatures( &$query ) {
 220                 foreach ( $this->features as $feature => $value ) {
 221                         if ( $feature === 'title-suffix-filter' && $value ) {
 222                                 $query['conds'][] = 'page_title' . $this->db->buildLike( $this->db->anyString(), $value );
 223                         }
 224                 }
 225         }
 226
 227         /**
 228          * Add namespace conditions
 229          * @param $query Array
 230          * @since 1.18 (changed)
 231          */
 232         function queryNamespaces( &$query ) {
 233                 if ( is_array( $this->namespaces ) ) {
 234                         if ( count( $this->namespaces ) === 0 ) {
 235                                 $this->namespaces[] = '0';
 236                         }
 237                         $query['conds']['page_namespace'] = $this->namespaces;
 238                 }
 239         }
 240
 241         /**
 242          * Add limit options
 243          * @param $query Array
 244          * @since 1.18
 245          */
 246         protected function limitResult( &$query ) {
 247                 $query['options']['LIMIT'] = $this->limit;
 248                 $query['options']['OFFSET'] = $this->offset;
 249         }
 250
 251         /**
 252          * Construct the SQL query to do the search.
 253          * The guts shoulds be constructed in queryMain()
 254          * @param $filteredTerm String
 255          * @param $fulltext Boolean
 256          * @return Array
 257          * @since 1.18 (changed)
 258          */
 259         function getQuery( $filteredTerm, $fulltext ) {
 260                 $query = array(
 261                         'tables' => array(),
 262                         'fields' => array(),
 263                         'conds' => array(),
 264                         'options' => array(),
 265                         'joins' => array(),
 266                 );
 267
 268                 $this->queryMain( $query, $filteredTerm, $fulltext );
 269                 $this->queryFeatures( $query );
 270                 $this->queryNamespaces( $query );
 271                 $this->limitResult( $query );
 272
 273                 return $query;
 274         }
 275
 276         /**
 277          * Picks which field to index on, depending on what type of query.
 278          * @param $fulltext Boolean
 279          * @return String
 280          */
 281         function getIndexField( $fulltext ) {
 282                 return $fulltext ? 'si_text' : 'si_title';
 283         }
 284
 285         /**
 286          * Get the base part of the search query.
 287          *
 288          * @param &$query array Search query array
 289          * @param $filteredTerm String
 290          * @param $fulltext Boolean
 291          * @since 1.18 (changed)
 292          */
 293         function queryMain( &$query, $filteredTerm, $fulltext ) {
 294                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 295                 $query['tables'][] = 'page';
 296                 $query['tables'][] = 'searchindex';
 297                 $query['fields'][] = 'page_id';
 298                 $query['fields'][] = 'page_namespace';
 299                 $query['fields'][] = 'page_title';
 300                 $query['conds'][] = 'page_id=si_page';
 301                 $query['conds'][] = $match;
 302         }
 303
 304         /**
 305          * @since 1.18 (changed)
 306          * @return array
 307          */
 308         function getCountQuery( $filteredTerm, $fulltext ) {
 309                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 310
 311                 $query = array(
 312                         'tables' => array( 'page', 'searchindex' ),
 313                         'fields' => array( 'COUNT(*) as c' ),
 314                         'conds' => array( 'page_id=si_page', $match ),
 315                         'options' => array(),
 316                         'joins' => array(),
 317                 );
 318
 319                 $this->queryFeatures( $query );
 320                 $this->queryNamespaces( $query );
 321
 322                 return $query;
 323         }
 324
 325         /**
 326          * Create or update the search index record for the given page.
 327          * Title and text should be pre-processed.
 328          *
 329          * @param $id Integer
 330          * @param $title String
 331          * @param $text String
 332          */
 333         function update( $id, $title, $text ) {
 334                 $dbw = wfGetDB( DB_MASTER );
 335                 $dbw->replace( 'searchindex',
 336                         array( 'si_page' ),
 337                         array(
 338                                 'si_page' => $id,
 339                                 'si_title' => $this->normalizeText( $title ),
 340                                 'si_text' => $this->normalizeText( $text )
 341                         ), __METHOD__ );
 342         }
 343
 344         /**
 345          * Update a search index record's title only.
 346          * Title should be pre-processed.
 347          *
 348          * @param $id Integer
 349          * @param $title String
 350          */
 351         function updateTitle( $id, $title ) {
 352                 $dbw = wfGetDB( DB_MASTER );
 353
 354                 $dbw->update( 'searchindex',
 355                         array( 'si_title' => $this->normalizeText( $title ) ),
 356                         array( 'si_page' => $id ),
 357                         __METHOD__,
 358                         array( $dbw->lowPriorityOption() ) );
 359         }
 360
 361         /**
 362          * Delete an indexed page
 363          * Title should be pre-processed.
 364          *
 365          * @param Integer $id Page id that was deleted
 366          * @param String $title Title of page that was deleted
 367          */
 368         function delete( $id, $title ) {
 369                 $dbw = wfGetDB( DB_MASTER );
 370
 371                 $dbw->delete( 'searchindex', array( 'si_page' => $id ), __METHOD__ );
 372         }
 373
 374         /**
 375          * Converts some characters for MySQL's indexing to grok it correctly,
 376          * and pads short words to overcome limitations.
 377          * @return mixed|string
 378          */
 379         function normalizeText( $string ) {
 380                 global $wgContLang;
 381
 382                 wfProfileIn( __METHOD__ );
 383
 384                 $out = parent::normalizeText( $string );
 385
 386                 // MySQL fulltext index doesn't grok utf-8, so we
 387                 // need to fold cases and convert to hex
 388                 $out = preg_replace_callback(
 389                         "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
 390                         array( $this, 'stripForSearchCallback' ),
 391                         $wgContLang->lc( $out ) );
 392
 393                 // And to add insult to injury, the default indexing
 394                 // ignores short words... Pad them so we can pass them
 395                 // through without reconfiguring the server...
 396                 $minLength = $this->minSearchLength();
 397                 if ( $minLength > 1 ) {
 398                         $n = $minLength - 1;
 399                         $out = preg_replace(
 400                                 "/\b(\w{1,$n})\b/",
 401                                 "$1u800",
 402                                 $out );
 403                 }
 404
 405                 // Periods within things like hostnames and IP addresses
 406                 // are also important -- we want a search for "example.com"
 407                 // or "192.168.1.1" to work sanely.
 408                 //
 409                 // MySQL's search seems to ignore them, so you'd match on
 410                 // "example.wikipedia.com" and "192.168.83.1" as well.
 411                 $out = preg_replace(
 412                         "/(\w)\.(\w|\*)/u",
 413                         "$1u82e$2",
 414                         $out );
 415
 416                 wfProfileOut( __METHOD__ );
 417
 418                 return $out;
 419         }
 420
 421         /**
 422          * Armor a case-folded UTF-8 string to get through MySQL's
 423          * fulltext search without being mucked up by funny charset
 424          * settings or anything else of the sort.
 425          * @return string
 426          */
 427         protected function stripForSearchCallback( $matches ) {
 428                 return 'u8' . bin2hex( $matches[1] );
 429         }
 430
 431         /**
 432          * Check MySQL server's ft_min_word_len setting so we know
 433          * if we need to pad short words...
 434          *
 435          * @return int
 436          */
 437         protected function minSearchLength() {
 438                 if ( is_null( self::$mMinSearchLength ) ) {
 439                         $sql = "SHOW GLOBAL VARIABLES LIKE 'ft\\_min\\_word\\_len'";
 440
 441                         $dbr = wfGetDB( DB_SLAVE );
 442                         $result = $dbr->query( $sql );
 443                         $row = $result->fetchObject();
 444                         $result->free();
 445
 446                         if ( $row && $row->Variable_name == 'ft_min_word_len' ) {
 447                                 self::$mMinSearchLength = intval( $row->Value );
 448                         } else {
 449                                 self::$mMinSearchLength = 0;
 450                         }
 451                 }
 452                 return self::$mMinSearchLength;
 453         }
 454 }
 455
 456 /**
 457  * @ingroup Search
 458  */
 459 class MySQLSearchResultSet extends SqlSearchResultSet {
 460         function __construct( $resultSet, $terms, $totalHits = null ) {
 461                 parent::__construct( $resultSet, $terms );
 462                 $this->mTotalHits = $totalHits;
 463         }
 464
 465         function getTotalHits() {
 466                 return $this->mTotalHits;
 467         }
 468 }