includes/search/SearchMySQL.php

   1 <?php
   2 /**
   3  * MySQL search engine
   4  *
   5  * Copyright (C) 2004 Brion Vibber <brion@pobox.com>
   6  * http://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 SearchEngine {
  32         var $strictMatching = true;
  33         static $mMinSearchLength;
  34
  35         /**
  36          * Creates an instance of this class
  37          * @param $db DatabaseMysql: database object
  38          */
  39         function __construct( $db ) {
  40                 parent::__construct( $db );
  41         }
  42
  43         /**
  44          * Parse the user's query and transform it into an SQL fragment which will
  45          * become part of a WHERE clause
  46          *
  47          * @return string
  48          */
  49         function parseQuery( $filteredText, $fulltext ) {
  50                 global $wgContLang;
  51                 $lc = SearchEngine::legalSearchChars(); // Minus format chars
  52                 $searchon = '';
  53                 $this->searchTerms = array();
  54
  55                 # @todo FIXME: This doesn't handle parenthetical expressions.
  56                 $m = array();
  57                 if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  58                           $filteredText, $m, PREG_SET_ORDER ) ) {
  59                         foreach( $m as $bits ) {
  60                                 @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  61
  62                                 if( $nonQuoted != '' ) {
  63                                         $term = $nonQuoted;
  64                                         $quote = '';
  65                                 } else {
  66                                         $term = str_replace( '"', '', $term );
  67                                         $quote = '"';
  68                                 }
  69
  70                                 if( $searchon !== '' ) $searchon .= ' ';
  71                                 if( $this->strictMatching && ($modifier == '') ) {
  72                                         // If we leave this out, boolean op defaults to OR which is rarely helpful.
  73                                         $modifier = '+';
  74                                 }
  75
  76                                 // Some languages such as Serbian store the input form in the search index,
  77                                 // so we may need to search for matches in multiple writing system variants.
  78                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  79                                 if( is_array( $convertedVariants ) ) {
  80                                         $variants = array_unique( array_values( $convertedVariants ) );
  81                                 } else {
  82                                         $variants = array( $term );
  83                                 }
  84
  85                                 // The low-level search index does some processing on input to work
  86                                 // around problems with minimum lengths and encoding in MySQL's
  87                                 // fulltext engine.
  88                                 // For Chinese this also inserts spaces between adjacent Han characters.
  89                                 $strippedVariants = array_map(
  90                                         array( $wgContLang, 'normalizeForSearch' ),
  91                                         $variants );
  92
  93                                 // Some languages such as Chinese force all variants to a canonical
  94                                 // form when stripping to the low-level search index, so to be sure
  95                                 // let's check our variants list for unique items after stripping.
  96                                 $strippedVariants = array_unique( $strippedVariants );
  97
  98                                 $searchon .= $modifier;
  99                                 if( count( $strippedVariants) > 1 )
 100                                         $searchon .= '(';
 101                                 foreach( $strippedVariants as $stripped ) {
 102                                         $stripped = $this->normalizeText( $stripped );
 103                                         if( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
 104                                                 // Hack for Chinese: we need to toss in quotes for
 105                                                 // multiple-character phrases since normalizeForSearch()
 106                                                 // added spaces between them to make word breaks.
 107                                                 $stripped = '"' . trim( $stripped ) . '"';
 108                                         }
 109                                         $searchon .= "$quote$stripped$quote$wildcard ";
 110                                 }
 111                                 if( count( $strippedVariants) > 1 )
 112                                         $searchon .= ')';
 113
 114                                 // Match individual terms or quoted phrase in result highlighting...
 115                                 // Note that variants will be introduced in a later stage for highlighting!
 116                                 $regexp = $this->regexTerm( $term, $wildcard );
 117                                 $this->searchTerms[] = $regexp;
 118                         }
 119                         wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
 120                         wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
 121                 } else {
 122                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 123                 }
 124
 125                 $searchon = $this->db->strencode( $searchon );
 126                 $field = $this->getIndexField( $fulltext );
 127                 return " MATCH($field) AGAINST('$searchon' IN BOOLEAN MODE) ";
 128         }
 129
 130         function regexTerm( $string, $wildcard ) {
 131                 global $wgContLang;
 132
 133                 $regex = preg_quote( $string, '/' );
 134                 if( $wgContLang->hasWordBreaks() ) {
 135                         if( $wildcard ) {
 136                                 // Don't cut off the final bit!
 137                                 $regex = "\b$regex";
 138                         } else {
 139                                 $regex = "\b$regex\b";
 140                         }
 141                 } else {
 142                         // For Chinese, words may legitimately abut other words in the text literal.
 143                         // Don't add \b boundary checks... note this could cause false positives
 144                         // for latin chars.
 145                 }
 146                 return $regex;
 147         }
 148
 149         public static function legalSearchChars() {
 150                 return "\"*" . parent::legalSearchChars();
 151         }
 152
 153         /**
 154          * Perform a full text search query and return a result set.
 155          *
 156          * @param $term String: raw search term
 157          * @return MySQLSearchResultSet
 158          */
 159         function searchText( $term ) {
 160                 return $this->searchInternal( $term, true );
 161         }
 162
 163         /**
 164          * Perform a title-only search query and return a result set.
 165          *
 166          * @param $term String: raw search term
 167          * @return MySQLSearchResultSet
 168          */
 169         function searchTitle( $term ) {
 170                 return $this->searchInternal( $term, false );
 171         }
 172
 173         protected function searchInternal( $term, $fulltext ) {
 174                 global $wgCountTotalSearchHits;
 175
 176                 // This seems out of place, why is this called with empty term?
 177                 if ( trim( $term ) === '' ) return null;
 178
 179                 $filteredTerm = $this->filter( $term );
 180                 $query = $this->getQuery( $filteredTerm, $fulltext );
 181                 $resultSet = $this->db->select(
 182                         $query['tables'], $query['fields'], $query['conds'],
 183                         __METHOD__, $query['options'], $query['joins']
 184                 );
 185
 186                 $total = null;
 187                 if( $wgCountTotalSearchHits ) {
 188                         $query = $this->getCountQuery( $filteredTerm, $fulltext );
 189                         $totalResult = $this->db->select(
 190                                 $query['tables'], $query['fields'], $query['conds'],
 191                                 __METHOD__, $query['options'], $query['joins']
 192                         );
 193
 194                         $row = $totalResult->fetchObject();
 195                         if( $row ) {
 196                                 $total = intval( $row->c );
 197                         }
 198                         $totalResult->free();
 199                 }
 200
 201                 return new MySQLSearchResultSet( $resultSet, $this->searchTerms, $total );
 202         }
 203
 204         public function supports( $feature ) {
 205                 switch( $feature ) {
 206                 case 'list-redirects':
 207                 case 'title-suffix-filter':
 208                         return true;
 209                 default:
 210                         return false;
 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 ( $feature ===  'list-redirects' && !$value ) {
 222                                 $query['conds']['page_is_redirect'] = 0;
 223                         } elseif( $feature === 'title-suffix-filter' && $value ) {
 224                                 $query['conds'][] = 'page_title' . $this->db->buildLike( $this->db->anyString(), $value );
 225                         }
 226                 }
 227         }
 228
 229         /**
 230          * Add namespace conditions
 231          * @param $query Array
 232          * @since 1.18 (changed)
 233          */
 234         function queryNamespaces( &$query ) {
 235                 if ( is_array( $this->namespaces ) ) {
 236                         if ( count( $this->namespaces ) === 0 ) {
 237                                 $this->namespaces[] = '0';
 238                         }
 239                         $query['conds']['page_namespace'] = $this->namespaces;
 240                 }
 241         }
 242
 243         /**
 244          * Add limit options
 245          * @param $query Array
 246          * @since 1.18
 247          */
 248         protected function limitResult( &$query ) {
 249                 $query['options']['LIMIT'] = $this->limit;
 250                 $query['options']['OFFSET'] = $this->offset;
 251         }
 252
 253         /**
 254          * Construct the SQL query to do the search.
 255          * The guts shoulds be constructed in queryMain()
 256          * @param $filteredTerm String
 257          * @param $fulltext Boolean
 258          * @return Array
 259          * @since 1.18 (changed)
 260          */
 261         function getQuery( $filteredTerm, $fulltext ) {
 262                 $query = array(
 263                         'tables' => array(),
 264                         'fields' => array(),
 265                         'conds' => array(),
 266                         'options' => array(),
 267                         'joins' => array(),
 268                 );
 269
 270                 $this->queryMain( $query, $filteredTerm, $fulltext );
 271                 $this->queryFeatures( $query );
 272                 $this->queryNamespaces( $query );
 273                 $this->limitResult( $query );
 274
 275                 return $query;
 276         }
 277
 278         /**
 279          * Picks which field to index on, depending on what type of query.
 280          * @param $fulltext Boolean
 281          * @return String
 282          */
 283         function getIndexField( $fulltext ) {
 284                 return $fulltext ? 'si_text' : 'si_title';
 285         }
 286
 287         /**
 288          * Get the base part of the search query.
 289          *
 290          * @param $filteredTerm String
 291          * @param $fulltext Boolean
 292          * @since 1.18 (changed)
 293          */
 294         function queryMain( &$query, $filteredTerm, $fulltext ) {
 295                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 296                 $query['tables'][] = 'page';
 297                 $query['tables'][] = 'searchindex';
 298                 $query['fields'][] = 'page_id';
 299                 $query['fields'][] = 'page_namespace';
 300                 $query['fields'][] = 'page_title';
 301                 $query['conds'][] = 'page_id=si_page';
 302                 $query['conds'][] = $match;
 303         }
 304
 305         /**
 306          * @since 1.18 (changed)
 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          * Converts some characters for MySQL's indexing to grok it correctly,
 363          * and pads short words to overcome limitations.
 364          */
 365         function normalizeText( $string ) {
 366                 global $wgContLang;
 367
 368                 wfProfileIn( __METHOD__ );
 369
 370                 $out = parent::normalizeText( $string );
 371
 372                 // MySQL fulltext index doesn't grok utf-8, so we
 373                 // need to fold cases and convert to hex
 374                 $out = preg_replace_callback(
 375                         "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
 376                         array( $this, 'stripForSearchCallback' ),
 377                         $wgContLang->lc( $out ) );
 378
 379                 // And to add insult to injury, the default indexing
 380                 // ignores short words... Pad them so we can pass them
 381                 // through without reconfiguring the server...
 382                 $minLength = $this->minSearchLength();
 383                 if( $minLength > 1 ) {
 384                         $n = $minLength - 1;
 385                         $out = preg_replace(
 386                                 "/\b(\w{1,$n})\b/",
 387                                 "$1u800",
 388                                 $out );
 389                 }
 390
 391                 // Periods within things like hostnames and IP addresses
 392                 // are also important -- we want a search for "example.com"
 393                 // or "192.168.1.1" to work sanely.
 394                 //
 395                 // MySQL's search seems to ignore them, so you'd match on
 396                 // "example.wikipedia.com" and "192.168.83.1" as well.
 397                 $out = preg_replace(
 398                         "/(\w)\.(\w|\*)/u",
 399                         "$1u82e$2",
 400                         $out );
 401
 402                 wfProfileOut( __METHOD__ );
 403
 404                 return $out;
 405         }
 406
 407         /**
 408          * Armor a case-folded UTF-8 string to get through MySQL's
 409          * fulltext search without being mucked up by funny charset
 410          * settings or anything else of the sort.
 411          */
 412         protected function stripForSearchCallback( $matches ) {
 413                 return 'u8' . bin2hex( $matches[1] );
 414         }
 415
 416         /**
 417          * Check MySQL server's ft_min_word_len setting so we know
 418          * if we need to pad short words...
 419          *
 420          * @return int
 421          */
 422         protected function minSearchLength() {
 423                 if( is_null( self::$mMinSearchLength ) ) {
 424                         $sql = "SHOW GLOBAL VARIABLES LIKE 'ft\\_min\\_word\\_len'";
 425
 426                         $dbr = wfGetDB( DB_SLAVE );
 427                         $result = $dbr->query( $sql );
 428                         $row = $result->fetchObject();
 429                         $result->free();
 430
 431                         if( $row && $row->Variable_name == 'ft_min_word_len' ) {
 432                                 self::$mMinSearchLength = intval( $row->Value );
 433                         } else {
 434                                 self::$mMinSearchLength = 0;
 435                         }
 436                 }
 437                 return self::$mMinSearchLength;
 438         }
 439 }
 440
 441 /**
 442  * @ingroup Search
 443  */
 444 class MySQLSearchResultSet extends SqlSearchResultSet {
 445         function __construct( $resultSet, $terms, $totalHits=null ) {
 446                 parent::__construct( $resultSet, $terms );
 447                 $this->mTotalHits = $totalHits;
 448         }
 449
 450         function getTotalHits() {
 451                 return $this->mTotalHits;
 452         }
 453 }