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
  32         /** @todo document */
  33         function __construct( $db ) {
  34                 $this->db = $db;
  35         }
  36
  37         /**
  38          * Parse the user's query and transform it into an SQL fragment which will
  39          * become part of a WHERE clause
  40          */
  41         function parseQuery( $filteredText, $fulltext ) {
  42                 global $wgContLang;
  43                 $lc = SearchEngine::legalSearchChars(); // Minus format chars
  44                 $searchon = '';
  45                 $this->searchTerms = array();
  46
  47                 # FIXME: This doesn't handle parenthetical expressions.
  48                 $m = array();
  49                 if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  50                           $filteredText, $m, PREG_SET_ORDER ) ) {
  51                         foreach( $m as $bits ) {
  52                                 @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  53
  54                                 if( $nonQuoted != '' ) {
  55                                         $term = $nonQuoted;
  56                                         $quote = '';
  57                                 } else {
  58                                         $term = str_replace( '"', '', $term );
  59                                         $quote = '"';
  60                                 }
  61
  62                                 if( $searchon !== '' ) $searchon .= ' ';
  63                                 if( $this->strictMatching && ($modifier == '') ) {
  64                                         // If we leave this out, boolean op defaults to OR which is rarely helpful.
  65                                         $modifier = '+';
  66                                 }
  67
  68                                 // Some languages such as Serbian store the input form in the search index,
  69                                 // so we may need to search for matches in multiple writing system variants.
  70                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  71                                 if( is_array( $convertedVariants ) ) {
  72                                         $variants = array_unique( array_values( $convertedVariants ) );
  73                                 } else {
  74                                         $variants = array( $term );
  75                                 }
  76
  77                                 // The low-level search index does some processing on input to work
  78                                 // around problems with minimum lengths and encoding in MySQL's
  79                                 // fulltext engine.
  80                                 // For Chinese this also inserts spaces between adjacent Han characters.
  81                                 $strippedVariants = array_map(
  82                                         array( $wgContLang, 'stripForSearch' ),
  83                                         $variants );
  84
  85                                 // Some languages such as Chinese force all variants to a canonical
  86                                 // form when stripping to the low-level search index, so to be sure
  87                                 // let's check our variants list for unique items after stripping.
  88                                 $strippedVariants = array_unique( $strippedVariants );
  89
  90                                 $searchon .= $modifier;
  91                                 if( count( $strippedVariants) > 1 )
  92                                         $searchon .= '(';
  93                                 foreach( $strippedVariants as $stripped ) {
  94                                         if( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
  95                                                 // Hack for Chinese: we need to toss in quotes for
  96                                                 // multiple-character phrases since stripForSearch()
  97                                                 // added spaces between them to make word breaks.
  98                                                 $stripped = '"' . trim( $stripped ) . '"';
  99                                         }
 100                                         $searchon .= "$quote$stripped$quote$wildcard ";
 101                                 }
 102                                 if( count( $strippedVariants) > 1 )
 103                                         $searchon .= ')';
 104
 105                                 // Match individual terms or quoted phrase in result highlighting...
 106                                 // Note that variants will be introduced in a later stage for highlighting!
 107                                 $regexp = $this->regexTerm( $term, $wildcard );
 108                                 $this->searchTerms[] = $regexp;
 109                         }
 110                         wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
 111                         wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
 112                 } else {
 113                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 114                 }
 115
 116                 $searchon = $this->db->strencode( $searchon );
 117                 $field = $this->getIndexField( $fulltext );
 118                 return " MATCH($field) AGAINST('$searchon' IN BOOLEAN MODE) ";
 119         }
 120
 121         function regexTerm( $string, $wildcard ) {
 122                 global $wgContLang;
 123
 124                 $regex = preg_quote( $string, '/' );
 125                 if( $wgContLang->hasWordBreaks() ) {
 126                         if( $wildcard ) {
 127                                 // Don't cut off the final bit!
 128                                 $regex = "\b$regex";
 129                         } else {
 130                                 $regex = "\b$regex\b";
 131                         }
 132                 } else {
 133                         // For Chinese, words may legitimately abut other words in the text literal.
 134                         // Don't add \b boundary checks... note this could cause false positives
 135                         // for latin chars.
 136                 }
 137                 return $regex;
 138         }
 139
 140         public static function legalSearchChars() {
 141                 return "\"*" . parent::legalSearchChars();
 142         }
 143
 144         /**
 145          * Perform a full text search query and return a result set.
 146          *
 147          * @param $term String: raw search term
 148          * @return MySQLSearchResultSet
 149          */
 150         function searchText( $term ) {
 151                 return $this->searchInternal( $term, true );
 152         }
 153
 154         /**
 155          * Perform a title-only search query and return a result set.
 156          *
 157          * @param $term String: raw search term
 158          * @return MySQLSearchResultSet
 159          */
 160         function searchTitle( $term ) {
 161                 return $this->searchInternal( $term, false );
 162         }
 163
 164         protected function searchInternal( $term, $fulltext ) {
 165                 global $wgSearchMySQLTotalHits;
 166
 167                 $filteredTerm = $this->filter( $term );
 168                 $resultSet = $this->db->query( $this->getQuery( $filteredTerm, $fulltext ) );
 169
 170                 $total = null;
 171                 if( $wgSearchMySQLTotalHits ) {
 172                         $totalResult = $this->db->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 173                         $row = $totalResult->fetchObject();
 174                         if( $row ) {
 175                                 $total = intval( $row->c );
 176                         }
 177                         $totalResult->free();
 178                 }
 179
 180                 return new MySQLSearchResultSet( $resultSet, $this->searchTerms, $total );
 181         }
 182
 183
 184         /**
 185          * Return a partial WHERE clause to exclude redirects, if so set
 186          * @return String
 187          */
 188         function queryRedirect() {
 189                 if( $this->showRedirects ) {
 190                         return '';
 191                 } else {
 192                         return 'AND page_is_redirect=0';
 193                 }
 194         }
 195
 196         /**
 197          * Return a partial WHERE clause to limit the search to the given namespaces
 198          * @return String
 199          */
 200         function queryNamespaces() {
 201                 if( is_null($this->namespaces) )
 202                         return '';  # search all
 203                 if ( !count( $this->namespaces ) ) {
 204                         $namespaces = '0';
 205                 } else {
 206                         $namespaces = $this->db->makeList( $this->namespaces );
 207                 }
 208                 return 'AND page_namespace IN (' . $namespaces . ')';
 209         }
 210
 211         /**
 212          * Return a LIMIT clause to limit results on the query.
 213          * @return String
 214          */
 215         function queryLimit() {
 216                 return $this->db->limitResult( '', $this->limit, $this->offset );
 217         }
 218
 219         /**
 220          * Does not do anything for generic search engine
 221          * subclasses may define this though
 222          * @return String
 223          */
 224         function queryRanking( $filteredTerm, $fulltext ) {
 225                 return '';
 226         }
 227
 228         /**
 229          * Construct the full SQL query to do the search.
 230          * The guts shoulds be constructed in queryMain()
 231          * @param $filteredTerm String
 232          * @param $fulltext Boolean
 233          */
 234         function getQuery( $filteredTerm, $fulltext ) {
 235                 return $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 236                         $this->queryRedirect() . ' ' .
 237                         $this->queryNamespaces() . ' ' .
 238                         $this->queryRanking( $filteredTerm, $fulltext ) . ' ' .
 239                         $this->queryLimit();
 240         }
 241
 242         /**
 243          * Picks which field to index on, depending on what type of query.
 244          * @param $fulltext Boolean
 245          * @return String
 246          */
 247         function getIndexField( $fulltext ) {
 248                 return $fulltext ? 'si_text' : 'si_title';
 249         }
 250
 251         /**
 252          * Get the base part of the search query.
 253          * The actual match syntax will depend on the server
 254          * version; MySQL 3 and MySQL 4 have different capabilities
 255          * in their fulltext search indexes.
 256          *
 257          * @param $filteredTerm String
 258          * @param $fulltext Boolean
 259          * @return String
 260          */
 261         function queryMain( $filteredTerm, $fulltext ) {
 262                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 263                 $page        = $this->db->tableName( 'page' );
 264                 $searchindex = $this->db->tableName( 'searchindex' );
 265                 return 'SELECT page_id, page_namespace, page_title ' .
 266                         "FROM $page,$searchindex " .
 267                         'WHERE page_id=si_page AND ' . $match;
 268         }
 269
 270         function getCountQuery( $filteredTerm, $fulltext ) {
 271                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 272                 $page        = $this->db->tableName( 'page' );
 273                 $searchindex = $this->db->tableName( 'searchindex' );
 274                 return "SELECT COUNT(*) AS c " .
 275                         "FROM $page,$searchindex " .
 276                         'WHERE page_id=si_page AND ' . $match .
 277                         $this->queryRedirect() . ' ' .
 278                         $this->queryNamespaces();
 279         }
 280
 281         /**
 282          * Create or update the search index record for the given page.
 283          * Title and text should be pre-processed.
 284          *
 285          * @param $id Integer
 286          * @param $title String
 287          * @param $text String
 288          */
 289         function update( $id, $title, $text ) {
 290                 $dbw = wfGetDB( DB_MASTER );
 291                 $dbw->replace( 'searchindex',
 292                         array( 'si_page' ),
 293                         array(
 294                                 'si_page' => $id,
 295                                 'si_title' => $title,
 296                                 'si_text' => $text
 297                         ), __METHOD__ );
 298         }
 299
 300         /**
 301          * Update a search index record's title only.
 302          * Title should be pre-processed.
 303          *
 304          * @param $id Integer
 305          * @param $title String
 306          */
 307     function updateTitle( $id, $title ) {
 308                 $dbw = wfGetDB( DB_MASTER );
 309
 310                 $dbw->update( 'searchindex',
 311                         array( 'si_title' => $title ),
 312                         array( 'si_page'  => $id ),
 313                         __METHOD__,
 314                         array( $dbw->lowPriorityOption() ) );
 315         }
 316 }
 317
 318 /**
 319  * @ingroup Search
 320  */
 321 class MySQLSearchResultSet extends SqlSearchResultSet {
 322         function MySQLSearchResultSet( $resultSet, $terms, $totalHits=null ) {
 323                 parent::__construct( $resultSet, $terms );
 324                 $this->mTotalHits = $totalHits;
 325         }
 326
 327         function getTotalHits() {
 328                 return $this->mTotalHits;
 329         }
 330 }