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         function parseQuery( $filteredText, $fulltext ) {
  48                 global $wgContLang;
  49                 $lc = SearchEngine::legalSearchChars(); // Minus format chars
  50                 $searchon = '';
  51                 $this->searchTerms = array();
  52
  53                 # FIXME: This doesn't handle parenthetical expressions.
  54                 $m = array();
  55                 if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  56                           $filteredText, $m, PREG_SET_ORDER ) ) {
  57                         foreach( $m as $bits ) {
  58                                 @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  59
  60                                 if( $nonQuoted != '' ) {
  61                                         $term = $nonQuoted;
  62                                         $quote = '';
  63                                 } else {
  64                                         $term = str_replace( '"', '', $term );
  65                                         $quote = '"';
  66                                 }
  67
  68                                 if( $searchon !== '' ) $searchon .= ' ';
  69                                 if( $this->strictMatching && ($modifier == '') ) {
  70                                         // If we leave this out, boolean op defaults to OR which is rarely helpful.
  71                                         $modifier = '+';
  72                                 }
  73
  74                                 // Some languages such as Serbian store the input form in the search index,
  75                                 // so we may need to search for matches in multiple writing system variants.
  76                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  77                                 if( is_array( $convertedVariants ) ) {
  78                                         $variants = array_unique( array_values( $convertedVariants ) );
  79                                 } else {
  80                                         $variants = array( $term );
  81                                 }
  82
  83                                 // The low-level search index does some processing on input to work
  84                                 // around problems with minimum lengths and encoding in MySQL's
  85                                 // fulltext engine.
  86                                 // For Chinese this also inserts spaces between adjacent Han characters.
  87                                 $strippedVariants = array_map(
  88                                         array( $wgContLang, 'normalizeForSearch' ),
  89                                         $variants );
  90
  91                                 // Some languages such as Chinese force all variants to a canonical
  92                                 // form when stripping to the low-level search index, so to be sure
  93                                 // let's check our variants list for unique items after stripping.
  94                                 $strippedVariants = array_unique( $strippedVariants );
  95
  96                                 $searchon .= $modifier;
  97                                 if( count( $strippedVariants) > 1 )
  98                                         $searchon .= '(';
  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                                 // Match individual terms or quoted phrase in result highlighting...
 113                                 // Note that variants will be introduced in a later stage for highlighting!
 114                                 $regexp = $this->regexTerm( $term, $wildcard );
 115                                 $this->searchTerms[] = $regexp;
 116                         }
 117                         wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
 118                         wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
 119                 } else {
 120                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 121                 }
 122
 123                 $searchon = $this->db->strencode( $searchon );
 124                 $field = $this->getIndexField( $fulltext );
 125                 return " MATCH($field) AGAINST('$searchon' IN BOOLEAN MODE) ";
 126         }
 127
 128         function regexTerm( $string, $wildcard ) {
 129                 global $wgContLang;
 130
 131                 $regex = preg_quote( $string, '/' );
 132                 if( $wgContLang->hasWordBreaks() ) {
 133                         if( $wildcard ) {
 134                                 // Don't cut off the final bit!
 135                                 $regex = "\b$regex";
 136                         } else {
 137                                 $regex = "\b$regex\b";
 138                         }
 139                 } else {
 140                         // For Chinese, words may legitimately abut other words in the text literal.
 141                         // Don't add \b boundary checks... note this could cause false positives
 142                         // for latin chars.
 143                 }
 144                 return $regex;
 145         }
 146
 147         public static function legalSearchChars() {
 148                 return "\"*" . parent::legalSearchChars();
 149         }
 150
 151         /**
 152          * Perform a full text search query and return a result set.
 153          *
 154          * @param $term String: raw search term
 155          * @return MySQLSearchResultSet
 156          */
 157         function searchText( $term ) {
 158                 return $this->searchInternal( $term, true );
 159         }
 160
 161         /**
 162          * Perform a title-only search query and return a result set.
 163          *
 164          * @param $term String: raw search term
 165          * @return MySQLSearchResultSet
 166          */
 167         function searchTitle( $term ) {
 168                 return $this->searchInternal( $term, false );
 169         }
 170
 171         protected function searchInternal( $term, $fulltext ) {
 172                 global $wgCountTotalSearchHits;
 173
 174                 // This seems of place, why is this called with empty term?
 175                 if ( trim( $term ) === '' ) return null;
 176
 177                 $filteredTerm = $this->filter( $term );
 178                 $query = $this->getQuery( $filteredTerm, $fulltext );
 179                 $resultSet = $this->db->select(
 180                         $query['tables'], $query['fields'], $query['conds'],
 181                         __METHOD__, $query['options'], $query['joins']
 182                 );
 183
 184                 $total = null;
 185                 if( $wgCountTotalSearchHits ) {
 186                         $query = $this->getCountQuery( $filteredTerm, $fulltext );
 187                         $totalResult = $this->db->select(
 188                                 $query['tables'], $query['fields'], $query['conds'],
 189                                 __METHOD__, $query['options'], $query['joins']
 190                         );
 191
 192                         $row = $totalResult->fetchObject();
 193                         if( $row ) {
 194                                 $total = intval( $row->c );
 195                         }
 196                         $totalResult->free();
 197                 }
 198
 199                 return new MySQLSearchResultSet( $resultSet, $this->searchTerms, $total );
 200         }
 201
 202
 203         /**
 204          * Add redirect conditions
 205          * @param $query Array
 206          * @since 1.18 (changed)
 207          */
 208         function queryRedirect( &$query ) {
 209                 if( !$this->showRedirects ) {
 210                         $query['conds']['page_is_redirect'] = 0;
 211                 }
 212         }
 213
 214         /**
 215          * Add namespace conditions
 216          * @param $query Array
 217          * @since 1.18 (changed)
 218          */
 219         function queryNamespaces( &$query ) {
 220                 if ( is_array( $this->namespaces ) ) {
 221                         if ( count( $this->namespaces ) === 0 ) {
 222                                 $this->namespaces[] = '0';
 223                         }
 224                         $query['conds']['page_namespace'] = $this->namespaces;
 225                 }
 226         }
 227
 228         /**
 229          * Add limit options
 230          * @param $query Array
 231          * @since 1.18
 232          */
 233         protected function limitResult( &$query ) {
 234                 $query['options']['LIMIT'] = $this->limit;
 235                 $query['options']['OFFSET'] = $this->offset;
 236         }
 237
 238         /**
 239          * Construct the SQL query to do the search.
 240          * The guts shoulds be constructed in queryMain()
 241          * @param $filteredTerm String
 242          * @param $fulltext Boolean
 243          * @return Array
 244          * @since 1.18 (changed)
 245          */
 246         function getQuery( $filteredTerm, $fulltext ) {
 247                 $query = array(
 248                         'tables' => array(),
 249                         'fields' => array(),
 250                         'conds' => array(),
 251                         'options' => array(),
 252                         'joins' => array(),
 253                 );
 254
 255                 $this->queryMain( $query, $filteredTerm, $fulltext );
 256                 $this->queryRedirect( $query );
 257                 $this->queryNamespaces( $query );
 258                 $this->limitResult( $query );
 259
 260                 return $query;
 261         }
 262
 263         /**
 264          * Picks which field to index on, depending on what type of query.
 265          * @param $fulltext Boolean
 266          * @return String
 267          */
 268         function getIndexField( $fulltext ) {
 269                 return $fulltext ? 'si_text' : 'si_title';
 270         }
 271
 272         /**
 273          * Get the base part of the search query.
 274          *
 275          * @param $filteredTerm String
 276          * @param $fulltext Boolean
 277          * @since 1.18 (changed)
 278          */
 279         function queryMain( &$query, $filteredTerm, $fulltext ) {
 280                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 281                 $query['tables'][] = 'page';
 282                 $query['tables'][] = 'searchindex';
 283                 $query['fields'][] = 'page_id';
 284                 $query['fields'][] = 'page_namespace';
 285                 $query['fields'][] = 'page_title';
 286                 $query['conds'][] = 'page_id=si_page';
 287                 $query['conds'][] = $match;
 288         }
 289
 290         /**
 291          * @since 1.18 (changed)
 292          */
 293         function getCountQuery( $filteredTerm, $fulltext ) {
 294                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 295
 296                 $query = array(
 297                         'tables' => array( 'page', 'searchindex' ),
 298                         'fields' => array( 'COUNT(*) as c' ),
 299                         'conds' => array( 'page_id=si_page', $match ),
 300                         'options' => array(),
 301                         'joins' => array(),
 302                 );
 303
 304                 $this->queryRedirect( $query );
 305                 $this->queryNamespaces( $query );
 306
 307                 return $query;
 308         }
 309
 310         /**
 311          * Create or update the search index record for the given page.
 312          * Title and text should be pre-processed.
 313          *
 314          * @param $id Integer
 315          * @param $title String
 316          * @param $text String
 317          */
 318         function update( $id, $title, $text ) {
 319                 $dbw = wfGetDB( DB_MASTER );
 320                 $dbw->replace( 'searchindex',
 321                         array( 'si_page' ),
 322                         array(
 323                                 'si_page' => $id,
 324                                 'si_title' => $this->normalizeText( $title ),
 325                                 'si_text' => $this->normalizeText( $text )
 326                         ), __METHOD__ );
 327         }
 328
 329         /**
 330          * Update a search index record's title only.
 331          * Title should be pre-processed.
 332          *
 333          * @param $id Integer
 334          * @param $title String
 335          */
 336         function updateTitle( $id, $title ) {
 337                 $dbw = wfGetDB( DB_MASTER );
 338
 339                 $dbw->update( 'searchindex',
 340                         array( 'si_title' => $this->normalizeText( $title ) ),
 341                         array( 'si_page'  => $id ),
 342                         __METHOD__,
 343                         array( $dbw->lowPriorityOption() ) );
 344         }
 345
 346         /**
 347          * Converts some characters for MySQL's indexing to grok it correctly,
 348          * and pads short words to overcome limitations.
 349          */
 350         function normalizeText( $string ) {
 351                 global $wgContLang;
 352
 353                 wfProfileIn( __METHOD__ );
 354
 355                 $out = parent::normalizeText( $string );
 356
 357                 // MySQL fulltext index doesn't grok utf-8, so we
 358                 // need to fold cases and convert to hex
 359                 $out = preg_replace_callback(
 360                         "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
 361                         array( $this, 'stripForSearchCallback' ),
 362                         $wgContLang->lc( $out ) );
 363
 364                 // And to add insult to injury, the default indexing
 365                 // ignores short words... Pad them so we can pass them
 366                 // through without reconfiguring the server...
 367                 $minLength = $this->minSearchLength();
 368                 if( $minLength > 1 ) {
 369                         $n = $minLength - 1;
 370                         $out = preg_replace(
 371                                 "/\b(\w{1,$n})\b/",
 372                                 "$1u800",
 373                                 $out );
 374                 }
 375
 376                 // Periods within things like hostnames and IP addresses
 377                 // are also important -- we want a search for "example.com"
 378                 // or "192.168.1.1" to work sanely.
 379                 //
 380                 // MySQL's search seems to ignore them, so you'd match on
 381                 // "example.wikipedia.com" and "192.168.83.1" as well.
 382                 $out = preg_replace(
 383                         "/(\w)\.(\w|\*)/u",
 384                         "$1u82e$2",
 385                         $out );
 386
 387                 wfProfileOut( __METHOD__ );
 388
 389                 return $out;
 390         }
 391
 392         /**
 393          * Armor a case-folded UTF-8 string to get through MySQL's
 394          * fulltext search without being mucked up by funny charset
 395          * settings or anything else of the sort.
 396          */
 397         protected function stripForSearchCallback( $matches ) {
 398                 return 'u8' . bin2hex( $matches[1] );
 399         }
 400
 401         /**
 402          * Check MySQL server's ft_min_word_len setting so we know
 403          * if we need to pad short words...
 404          *
 405          * @return int
 406          */
 407         protected function minSearchLength() {
 408                 if( is_null( self::$mMinSearchLength ) ) {
 409                         $sql = "SHOW GLOBAL VARIABLES LIKE 'ft\\_min\\_word\\_len'";
 410
 411                         $dbr = wfGetDB( DB_SLAVE );
 412                         $result = $dbr->query( $sql );
 413                         $row = $result->fetchObject();
 414                         $result->free();
 415
 416                         if( $row && $row->Variable_name == 'ft_min_word_len' ) {
 417                                 self::$mMinSearchLength = intval( $row->Value );
 418                         } else {
 419                                 self::$mMinSearchLength = 0;
 420                         }
 421                 }
 422                 return self::$mMinSearchLength;
 423         }
 424 }
 425
 426 /**
 427  * @ingroup Search
 428  */
 429 class MySQLSearchResultSet extends SqlSearchResultSet {
 430         function __construct( $resultSet, $terms, $totalHits=null ) {
 431                 parent::__construct( $resultSet, $terms );
 432                 $this->mTotalHits = $totalHits;
 433         }
 434
 435         function getTotalHits() {
 436                 return $this->mTotalHits;
 437         }
 438 }