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                 $filteredTerm = $this->filter( $term );
 175                 $resultSet = $this->db->query( $this->getQuery( $filteredTerm, $fulltext ) );
 176
 177                 $total = null;
 178                 if( $wgCountTotalSearchHits ) {
 179                         $totalResult = $this->db->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 180                         $row = $totalResult->fetchObject();
 181                         if( $row ) {
 182                                 $total = intval( $row->c );
 183                         }
 184                         $totalResult->free();
 185                 }
 186
 187                 return new MySQLSearchResultSet( $resultSet, $this->searchTerms, $total );
 188         }
 189
 190
 191         /**
 192          * Return a partial WHERE clause to exclude redirects, if so set
 193          * @return String
 194          */
 195         function queryRedirect() {
 196                 if( $this->showRedirects ) {
 197                         return '';
 198                 } else {
 199                         return 'AND page_is_redirect=0';
 200                 }
 201         }
 202
 203         /**
 204          * Return a partial WHERE clause to limit the search to the given namespaces
 205          * @return String
 206          */
 207         function queryNamespaces() {
 208                 if( is_null($this->namespaces) )
 209                         return '';  # search all
 210                 if ( !count( $this->namespaces ) ) {
 211                         $namespaces = '0';
 212                 } else {
 213                         $namespaces = $this->db->makeList( $this->namespaces );
 214                 }
 215                 return 'AND page_namespace IN (' . $namespaces . ')';
 216         }
 217
 218         /**
 219          * Return a LIMIT clause to limit results on the query.
 220          * @return String
 221          */
 222         function queryLimit() {
 223                 return $this->db->limitResult( '', $this->limit, $this->offset );
 224         }
 225
 226         /**
 227          * Does not do anything for generic search engine
 228          * subclasses may define this though
 229          * @return String
 230          */
 231         function queryRanking( $filteredTerm, $fulltext ) {
 232                 return '';
 233         }
 234
 235         /**
 236          * Construct the full SQL query to do the search.
 237          * The guts shoulds be constructed in queryMain()
 238          * @param $filteredTerm String
 239          * @param $fulltext Boolean
 240          */
 241         function getQuery( $filteredTerm, $fulltext ) {
 242                 return $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 243                         $this->queryRedirect() . ' ' .
 244                         $this->queryNamespaces() . ' ' .
 245                         $this->queryRanking( $filteredTerm, $fulltext ) . ' ' .
 246                         $this->queryLimit();
 247         }
 248
 249         /**
 250          * Picks which field to index on, depending on what type of query.
 251          * @param $fulltext Boolean
 252          * @return String
 253          */
 254         function getIndexField( $fulltext ) {
 255                 return $fulltext ? 'si_text' : 'si_title';
 256         }
 257
 258         /**
 259          * Get the base part of the search query.
 260          * The actual match syntax will depend on the server
 261          * version; MySQL 3 and MySQL 4 have different capabilities
 262          * in their fulltext search indexes.
 263          *
 264          * @param $filteredTerm String
 265          * @param $fulltext Boolean
 266          * @return String
 267          */
 268         function queryMain( $filteredTerm, $fulltext ) {
 269                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 270                 $page        = $this->db->tableName( 'page' );
 271                 $searchindex = $this->db->tableName( 'searchindex' );
 272                 return 'SELECT page_id, page_namespace, page_title ' .
 273                         "FROM $page,$searchindex " .
 274                         'WHERE page_id=si_page AND ' . $match;
 275         }
 276
 277         function getCountQuery( $filteredTerm, $fulltext ) {
 278                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 279                 $page        = $this->db->tableName( 'page' );
 280                 $searchindex = $this->db->tableName( 'searchindex' );
 281                 return "SELECT COUNT(*) AS c " .
 282                         "FROM $page,$searchindex " .
 283                         'WHERE page_id=si_page AND ' . $match .
 284                         $this->queryRedirect() . ' ' .
 285                         $this->queryNamespaces();
 286         }
 287
 288         /**
 289          * Create or update the search index record for the given page.
 290          * Title and text should be pre-processed.
 291          *
 292          * @param $id Integer
 293          * @param $title String
 294          * @param $text String
 295          */
 296         function update( $id, $title, $text ) {
 297                 $dbw = wfGetDB( DB_MASTER );
 298                 $dbw->replace( 'searchindex',
 299                         array( 'si_page' ),
 300                         array(
 301                                 'si_page' => $id,
 302                                 'si_title' => $this->normalizeText( $title ),
 303                                 'si_text' => $this->normalizeText( $text )
 304                         ), __METHOD__ );
 305         }
 306
 307         /**
 308          * Update a search index record's title only.
 309          * Title should be pre-processed.
 310          *
 311          * @param $id Integer
 312          * @param $title String
 313          */
 314     function updateTitle( $id, $title ) {
 315                 $dbw = wfGetDB( DB_MASTER );
 316
 317                 $dbw->update( 'searchindex',
 318                         array( 'si_title' => $this->normalizeText( $title ) ),
 319                         array( 'si_page'  => $id ),
 320                         __METHOD__,
 321                         array( $dbw->lowPriorityOption() ) );
 322         }
 323
 324         /**
 325          * Converts some characters for MySQL's indexing to grok it correctly,
 326          * and pads short words to overcome limitations.
 327          */
 328         function normalizeText( $string ) {
 329                 global $wgContLang;
 330
 331                 wfProfileIn( __METHOD__ );
 332
 333                 $out = parent::normalizeText( $string );
 334
 335                 // MySQL fulltext index doesn't grok utf-8, so we
 336                 // need to fold cases and convert to hex
 337                 $out = preg_replace_callback(
 338                         "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
 339                         array( $this, 'stripForSearchCallback' ),
 340                         $wgContLang->lc( $out ) );
 341
 342                 // And to add insult to injury, the default indexing
 343                 // ignores short words... Pad them so we can pass them
 344                 // through without reconfiguring the server...
 345                 $minLength = $this->minSearchLength();
 346                 if( $minLength > 1 ) {
 347                         $n = $minLength - 1;
 348                         $out = preg_replace(
 349                                 "/\b(\w{1,$n})\b/",
 350                                 "$1u800",
 351                                 $out );
 352                 }
 353
 354                 // Periods within things like hostnames and IP addresses
 355                 // are also important -- we want a search for "example.com"
 356                 // or "192.168.1.1" to work sanely.
 357                 //
 358                 // MySQL's search seems to ignore them, so you'd match on
 359                 // "example.wikipedia.com" and "192.168.83.1" as well.
 360                 $out = preg_replace(
 361                         "/(\w)\.(\w|\*)/u",
 362                         "$1u82e$2",
 363                         $out );
 364
 365                 wfProfileOut( __METHOD__ );
 366
 367                 return $out;
 368         }
 369
 370         /**
 371          * Armor a case-folded UTF-8 string to get through MySQL's
 372          * fulltext search without being mucked up by funny charset
 373          * settings or anything else of the sort.
 374          */
 375         protected function stripForSearchCallback( $matches ) {
 376                 return 'u8' . bin2hex( $matches[1] );
 377         }
 378
 379         /**
 380          * Check MySQL server's ft_min_word_len setting so we know
 381          * if we need to pad short words...
 382          *
 383          * @return int
 384          */
 385         protected function minSearchLength() {
 386                 if( is_null( self::$mMinSearchLength ) ) {
 387                         $sql = "SHOW GLOBAL VARIABLES LIKE 'ft\\_min\\_word\\_len'";
 388
 389                         $dbr = wfGetDB( DB_SLAVE );
 390                         $result = $dbr->query( $sql );
 391                         $row = $result->fetchObject();
 392                         $result->free();
 393
 394                         if( $row && $row->Variable_name == 'ft_min_word_len' ) {
 395                                 self::$mMinSearchLength = intval( $row->Value );
 396                         } else {
 397                                 self::$mMinSearchLength = 0;
 398                         }
 399                 }
 400                 return self::$mMinSearchLength;
 401         }
 402 }
 403
 404 /**
 405  * @ingroup Search
 406  */
 407 class MySQLSearchResultSet extends SqlSearchResultSet {
 408         function __construct( $resultSet, $terms, $totalHits=null ) {
 409                 parent::__construct( $resultSet, $terms );
 410                 $this->mTotalHits = $totalHits;
 411         }
 412
 413         function getTotalHits() {
 414                 return $this->mTotalHits;
 415         }
 416 }