includes/search/SearchMySQL.php

   1 <?php
   2 /**
   3  * MySQL search engine
   4  *
   5  * Copyright (C) 2004 Brion Vibber <brion@pobox.com>
   6  * https://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 SearchDatabase {
  32         protected $strictMatching = true;
  33
  34         private static $mMinSearchLength;
  35
  36         /**
  37          * Parse the user's query and transform it into an SQL fragment which will
  38          * become part of a WHERE clause
  39          *
  40          * @param string $filteredText
  41          * @param string $fulltext
  42          *
  43          * @return string
  44          */
  45         function parseQuery( $filteredText, $fulltext ) {
  46                 global $wgContLang;
  47
  48                 $lc = $this->legalSearchChars(); // Minus format chars
  49                 $searchon = '';
  50                 $this->searchTerms = array();
  51
  52                 # @todo FIXME: This doesn't handle parenthetical expressions.
  53                 $m = array();
  54                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  55                                 $filteredText, $m, PREG_SET_ORDER ) ) {
  56                         foreach ( $m as $bits ) {
  57                                 wfSuppressWarnings();
  58                                 list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  59                                 wfRestoreWarnings();
  60
  61                                 if ( $nonQuoted != '' ) {
  62                                         $term = $nonQuoted;
  63                                         $quote = '';
  64                                 } else {
  65                                         $term = str_replace( '"', '', $term );
  66                                         $quote = '"';
  67                                 }
  68
  69                                 if ( $searchon !== '' ) {
  70                                         $searchon .= ' ';
  71                                 }
  72                                 if ( $this->strictMatching && ( $modifier == '' ) ) {
  73                                         // If we leave this out, boolean op defaults to OR which is rarely helpful.
  74                                         $modifier = '+';
  75                                 }
  76
  77                                 // Some languages such as Serbian store the input form in the search index,
  78                                 // so we may need to search for matches in multiple writing system variants.
  79                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  80                                 if ( is_array( $convertedVariants ) ) {
  81                                         $variants = array_unique( array_values( $convertedVariants ) );
  82                                 } else {
  83                                         $variants = array( $term );
  84                                 }
  85
  86                                 // The low-level search index does some processing on input to work
  87                                 // around problems with minimum lengths and encoding in MySQL's
  88                                 // fulltext engine.
  89                                 // For Chinese this also inserts spaces between adjacent Han characters.
  90                                 $strippedVariants = array_map(
  91                                         array( $wgContLang, 'normalizeForSearch' ),
  92                                         $variants );
  93
  94                                 // Some languages such as Chinese force all variants to a canonical
  95                                 // form when stripping to the low-level search index, so to be sure
  96                                 // let's check our variants list for unique items after stripping.
  97                                 $strippedVariants = array_unique( $strippedVariants );
  98
  99                                 $searchon .= $modifier;
 100                                 if ( count( $strippedVariants ) > 1 ) {
 101                                         $searchon .= '(';
 102                                 }
 103                                 foreach ( $strippedVariants as $stripped ) {
 104                                         $stripped = $this->normalizeText( $stripped );
 105                                         if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
 106                                                 // Hack for Chinese: we need to toss in quotes for
 107                                                 // multiple-character phrases since normalizeForSearch()
 108                                                 // added spaces between them to make word breaks.
 109                                                 $stripped = '"' . trim( $stripped ) . '"';
 110                                         }
 111                                         $searchon .= "$quote$stripped$quote$wildcard ";
 112                                 }
 113                                 if ( count( $strippedVariants ) > 1 ) {
 114                                         $searchon .= ')';
 115                                 }
 116
 117                                 // Match individual terms or quoted phrase in result highlighting...
 118                                 // Note that variants will be introduced in a later stage for highlighting!
 119                                 $regexp = $this->regexTerm( $term, $wildcard );
 120                                 $this->searchTerms[] = $regexp;
 121                         }
 122                         wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
 123                         wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
 124                 } else {
 125                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 126                 }
 127
 128                 $searchon = $this->db->strencode( $searchon );
 129                 $field = $this->getIndexField( $fulltext );
 130                 return " MATCH($field) AGAINST('$searchon' IN BOOLEAN MODE) ";
 131         }
 132
 133         function regexTerm( $string, $wildcard ) {
 134                 global $wgContLang;
 135
 136                 $regex = preg_quote( $string, '/' );
 137                 if ( $wgContLang->hasWordBreaks() ) {
 138                         if ( $wildcard ) {
 139                                 // Don't cut off the final bit!
 140                                 $regex = "\b$regex";
 141                         } else {
 142                                 $regex = "\b$regex\b";
 143                         }
 144                 } else {
 145                         // For Chinese, words may legitimately abut other words in the text literal.
 146                         // Don't add \b boundary checks... note this could cause false positives
 147                         // for latin chars.
 148                 }
 149                 return $regex;
 150         }
 151
 152         public static function legalSearchChars() {
 153                 return "\"*" . parent::legalSearchChars();
 154         }
 155
 156         /**
 157          * Perform a full text search query and return a result set.
 158          *
 159          * @param string $term Raw search term
 160          * @return SqlSearchResultSet
 161          */
 162         function searchText( $term ) {
 163                 return $this->searchInternal( $term, true );
 164         }
 165
 166         /**
 167          * Perform a title-only search query and return a result set.
 168          *
 169          * @param string $term Raw search term
 170          * @return SqlSearchResultSet
 171          */
 172         function searchTitle( $term ) {
 173                 return $this->searchInternal( $term, false );
 174         }
 175
 176         protected function searchInternal( $term, $fulltext ) {
 177                 global $wgCountTotalSearchHits;
 178
 179                 // This seems out of place, why is this called with empty term?
 180                 if ( trim( $term ) === '' ) {
 181                         return null;
 182                 }
 183
 184                 $filteredTerm = $this->filter( $term );
 185                 $query = $this->getQuery( $filteredTerm, $fulltext );
 186                 $resultSet = $this->db->select(
 187                         $query['tables'], $query['fields'], $query['conds'],
 188                         __METHOD__, $query['options'], $query['joins']
 189                 );
 190
 191                 $total = null;
 192                 if ( $wgCountTotalSearchHits ) {
 193                         $query = $this->getCountQuery( $filteredTerm, $fulltext );
 194                         $totalResult = $this->db->select(
 195                                 $query['tables'], $query['fields'], $query['conds'],
 196                                 __METHOD__, $query['options'], $query['joins']
 197                         );
 198
 199                         $row = $totalResult->fetchObject();
 200                         if ( $row ) {
 201                                 $total = intval( $row->c );
 202                         }
 203                         $totalResult->free();
 204                 }
 205
 206                 return new SqlSearchResultSet( $resultSet, $this->searchTerms, $total );
 207         }
 208
 209         public function supports( $feature ) {
 210                 switch ( $feature ) {
 211                 case 'title-suffix-filter':
 212                         return true;
 213                 default:
 214                         return parent::supports( $feature );
 215                 }
 216         }
 217
 218         /**
 219          * Add special conditions
 220          * @param array $query
 221          * @since 1.18
 222          */
 223         protected function queryFeatures( &$query ) {
 224                 foreach ( $this->features as $feature => $value ) {
 225                         if ( $feature === 'title-suffix-filter' && $value ) {
 226                                 $query['conds'][] = 'page_title' . $this->db->buildLike( $this->db->anyString(), $value );
 227                         }
 228                 }
 229         }
 230
 231         /**
 232          * Add namespace conditions
 233          * @param array $query
 234          * @since 1.18 (changed)
 235          */
 236         function queryNamespaces( &$query ) {
 237                 if ( is_array( $this->namespaces ) ) {
 238                         if ( count( $this->namespaces ) === 0 ) {
 239                                 $this->namespaces[] = '0';
 240                         }
 241                         $query['conds']['page_namespace'] = $this->namespaces;
 242                 }
 243         }
 244
 245         /**
 246          * Add limit options
 247          * @param array $query
 248          * @since 1.18
 249          */
 250         protected function limitResult( &$query ) {
 251                 $query['options']['LIMIT'] = $this->limit;
 252                 $query['options']['OFFSET'] = $this->offset;
 253         }
 254
 255         /**
 256          * Construct the SQL query to do the search.
 257          * The guts shoulds be constructed in queryMain()
 258          * @param string $filteredTerm
 259          * @param bool $fulltext
 260          * @return array
 261          * @since 1.18 (changed)
 262          */
 263         function getQuery( $filteredTerm, $fulltext ) {
 264                 $query = array(
 265                         'tables' => array(),
 266                         'fields' => array(),
 267                         'conds' => array(),
 268                         'options' => array(),
 269                         'joins' => array(),
 270                 );
 271
 272                 $this->queryMain( $query, $filteredTerm, $fulltext );
 273                 $this->queryFeatures( $query );
 274                 $this->queryNamespaces( $query );
 275                 $this->limitResult( $query );
 276
 277                 return $query;
 278         }
 279
 280         /**
 281          * Picks which field to index on, depending on what type of query.
 282          * @param bool $fulltext
 283          * @return string
 284          */
 285         function getIndexField( $fulltext ) {
 286                 return $fulltext ? 'si_text' : 'si_title';
 287         }
 288
 289         /**
 290          * Get the base part of the search query.
 291          *
 292          * @param array &$query Search query array
 293          * @param string $filteredTerm
 294          * @param bool $fulltext
 295          * @since 1.18 (changed)
 296          */
 297         function queryMain( &$query, $filteredTerm, $fulltext ) {
 298                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 299                 $query['tables'][] = 'page';
 300                 $query['tables'][] = 'searchindex';
 301                 $query['fields'][] = 'page_id';
 302                 $query['fields'][] = 'page_namespace';
 303                 $query['fields'][] = 'page_title';
 304                 $query['conds'][] = 'page_id=si_page';
 305                 $query['conds'][] = $match;
 306         }
 307
 308         /**
 309          * @since 1.18 (changed)
 310          * @return array
 311          */
 312         function getCountQuery( $filteredTerm, $fulltext ) {
 313                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 314
 315                 $query = array(
 316                         'tables' => array( 'page', 'searchindex' ),
 317                         'fields' => array( 'COUNT(*) as c' ),
 318                         'conds' => array( 'page_id=si_page', $match ),
 319                         'options' => array(),
 320                         'joins' => array(),
 321                 );
 322
 323                 $this->queryFeatures( $query );
 324                 $this->queryNamespaces( $query );
 325
 326                 return $query;
 327         }
 328
 329         /**
 330          * Create or update the search index record for the given page.
 331          * Title and text should be pre-processed.
 332          *
 333          * @param int $id
 334          * @param string $title
 335          * @param string $text
 336          */
 337         function update( $id, $title, $text ) {
 338                 $dbw = wfGetDB( DB_MASTER );
 339                 $dbw->replace( 'searchindex',
 340                         array( 'si_page' ),
 341                         array(
 342                                 'si_page' => $id,
 343                                 'si_title' => $this->normalizeText( $title ),
 344                                 'si_text' => $this->normalizeText( $text )
 345                         ), __METHOD__ );
 346         }
 347
 348         /**
 349          * Update a search index record's title only.
 350          * Title should be pre-processed.
 351          *
 352          * @param int $id
 353          * @param string $title
 354          */
 355         function updateTitle( $id, $title ) {
 356                 $dbw = wfGetDB( DB_MASTER );
 357
 358                 $dbw->update( 'searchindex',
 359                         array( 'si_title' => $this->normalizeText( $title ) ),
 360                         array( 'si_page' => $id ),
 361                         __METHOD__,
 362                         array( $dbw->lowPriorityOption() ) );
 363         }
 364
 365         /**
 366          * Delete an indexed page
 367          * Title should be pre-processed.
 368          *
 369          * @param int $id Page id that was deleted
 370          * @param string $title Title of page that was deleted
 371          */
 372         function delete( $id, $title ) {
 373                 $dbw = wfGetDB( DB_MASTER );
 374
 375                 $dbw->delete( 'searchindex', array( 'si_page' => $id ), __METHOD__ );
 376         }
 377
 378         /**
 379          * Converts some characters for MySQL's indexing to grok it correctly,
 380          * and pads short words to overcome limitations.
 381          * @return mixed|string
 382          */
 383         function normalizeText( $string ) {
 384                 global $wgContLang;
 385
 386                 wfProfileIn( __METHOD__ );
 387
 388                 $out = parent::normalizeText( $string );
 389
 390                 // MySQL fulltext index doesn't grok utf-8, so we
 391                 // need to fold cases and convert to hex
 392                 $out = preg_replace_callback(
 393                         "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
 394                         array( $this, 'stripForSearchCallback' ),
 395                         $wgContLang->lc( $out ) );
 396
 397                 // And to add insult to injury, the default indexing
 398                 // ignores short words... Pad them so we can pass them
 399                 // through without reconfiguring the server...
 400                 $minLength = $this->minSearchLength();
 401                 if ( $minLength > 1 ) {
 402                         $n = $minLength - 1;
 403                         $out = preg_replace(
 404                                 "/\b(\w{1,$n})\b/",
 405                                 "$1u800",
 406                                 $out );
 407                 }
 408
 409                 // Periods within things like hostnames and IP addresses
 410                 // are also important -- we want a search for "example.com"
 411                 // or "192.168.1.1" to work sanely.
 412                 //
 413                 // MySQL's search seems to ignore them, so you'd match on
 414                 // "example.wikipedia.com" and "192.168.83.1" as well.
 415                 $out = preg_replace(
 416                         "/(\w)\.(\w|\*)/u",
 417                         "$1u82e$2",
 418                         $out );
 419
 420                 wfProfileOut( __METHOD__ );
 421
 422                 return $out;
 423         }
 424
 425         /**
 426          * Armor a case-folded UTF-8 string to get through MySQL's
 427          * fulltext search without being mucked up by funny charset
 428          * settings or anything else of the sort.
 429          * @return string
 430          */
 431         protected function stripForSearchCallback( $matches ) {
 432                 return 'u8' . bin2hex( $matches[1] );
 433         }
 434
 435         /**
 436          * Check MySQL server's ft_min_word_len setting so we know
 437          * if we need to pad short words...
 438          *
 439          * @return int
 440          */
 441         protected function minSearchLength() {
 442                 if ( is_null( self::$mMinSearchLength ) ) {
 443                         $sql = "SHOW GLOBAL VARIABLES LIKE 'ft\\_min\\_word\\_len'";
 444
 445                         $dbr = wfGetDB( DB_SLAVE );
 446                         $result = $dbr->query( $sql );
 447                         $row = $result->fetchObject();
 448                         $result->free();
 449
 450                         if ( $row && $row->Variable_name == 'ft_min_word_len' ) {
 451                                 self::$mMinSearchLength = intval( $row->Value );
 452                         } else {
 453                                 self::$mMinSearchLength = 0;
 454                         }
 455                 }
 456                 return self::$mMinSearchLength;
 457         }
 458 }