includes/search/SearchMySQL.php

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