includes/search/SearchSqlite.php

   1 <?php
   2 /**
   3  * SQLite search backend, based upon SearchMysql
   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  * @file
  21  * @ingroup Search
  22  */
  23
  24 /**
  25  * Search engine hook for SQLite
  26  * @ingroup Search
  27  */
  28 class SearchSqlite extends SearchEngine {
  29
  30         /**
  31          * @var DatabaseSqlite
  32          */
  33         protected $db;
  34
  35         /**
  36          * Creates an instance of this class
  37          * @param $db DatabaseSqlite: database object
  38          */
  39         function __construct( $db ) {
  40                 parent::__construct( $db );
  41         }
  42
  43         /**
  44          * Whether fulltext search is supported by current schema
  45          * @return Boolean
  46          */
  47         function fulltextSearchSupported() {
  48                 return $this->db->checkForEnabledSearch();
  49         }
  50
  51         /**
  52          * Parse the user's query and transform it into an SQL fragment which will
  53          * become part of a WHERE clause
  54          *
  55          * @return string
  56          */
  57         function parseQuery( $filteredText, $fulltext ) {
  58                 global $wgContLang;
  59                 $lc = SearchEngine::legalSearchChars(); // Minus format chars
  60                 $searchon = '';
  61                 $this->searchTerms = array();
  62
  63                 $m = array();
  64                 if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  65                                 $filteredText, $m, PREG_SET_ORDER ) ) {
  66                         foreach( $m as $bits ) {
  67                                 @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  68
  69                                 if( $nonQuoted != '' ) {
  70                                         $term = $nonQuoted;
  71                                         $quote = '';
  72                                 } else {
  73                                         $term = str_replace( '"', '', $term );
  74                                         $quote = '"';
  75                                 }
  76
  77                                 if( $searchon !== '' ) {
  78                                         $searchon .= ' ';
  79                                 }
  80
  81                                 // Some languages such as Serbian store the input form in the search index,
  82                                 // so we may need to search for matches in multiple writing system variants.
  83                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  84                                 if( is_array( $convertedVariants ) ) {
  85                                         $variants = array_unique( array_values( $convertedVariants ) );
  86                                 } else {
  87                                         $variants = array( $term );
  88                                 }
  89
  90                                 // The low-level search index does some processing on input to work
  91                                 // around problems with minimum lengths and encoding in MySQL's
  92                                 // fulltext engine.
  93                                 // For Chinese this also inserts spaces between adjacent Han characters.
  94                                 $strippedVariants = array_map(
  95                                         array( $wgContLang, 'normalizeForSearch' ),
  96                                         $variants );
  97
  98                                 // Some languages such as Chinese force all variants to a canonical
  99                                 // form when stripping to the low-level search index, so to be sure
 100                                 // let's check our variants list for unique items after stripping.
 101                                 $strippedVariants = array_unique( $strippedVariants );
 102
 103                                 $searchon .= $modifier;
 104                                 if( count( $strippedVariants) > 1 )
 105                                         $searchon .= '(';
 106                                 foreach( $strippedVariants as $stripped ) {
 107                                         if( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
 108                                                 // Hack for Chinese: we need to toss in quotes for
 109                                                 // multiple-character phrases since normalizeForSearch()
 110                                                 // added spaces between them to make word breaks.
 111                                                 $stripped = '"' . trim( $stripped ) . '"';
 112                                         }
 113                                         $searchon .= "$quote$stripped$quote$wildcard ";
 114                                 }
 115                                 if( count( $strippedVariants) > 1 )
 116                                         $searchon .= ')';
 117
 118                                 // Match individual terms or quoted phrase in result highlighting...
 119                                 // Note that variants will be introduced in a later stage for highlighting!
 120                                 $regexp = $this->regexTerm( $term, $wildcard );
 121                                 $this->searchTerms[] = $regexp;
 122                         }
 123
 124                 } else {
 125                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 126                 }
 127
 128                 $searchon = $this->db->strencode( $searchon );
 129                 $field = $this->getIndexField( $fulltext );
 130                 return " $field MATCH '$searchon' ";
 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 $term String: raw search term
 160          * @return SqliteSearchResultSet
 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 $term String: raw search term
 170          * @return SqliteSearchResultSet
 171          */
 172         function searchTitle( $term ) {
 173                 return $this->searchInternal( $term, false );
 174         }
 175
 176         protected function searchInternal( $term, $fulltext ) {
 177                 global $wgCountTotalSearchHits, $wgContLang;
 178
 179                 if ( !$this->fulltextSearchSupported() ) {
 180                         return null;
 181                 }
 182
 183                 $filteredTerm = $this->filter( $wgContLang->lc( $term ) );
 184                 $resultSet = $this->db->query( $this->getQuery( $filteredTerm, $fulltext ) );
 185
 186                 $total = null;
 187                 if( $wgCountTotalSearchHits ) {
 188                         $totalResult = $this->db->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 189                         $row = $totalResult->fetchObject();
 190                         if( $row ) {
 191                                 $total = intval( $row->c );
 192                         }
 193                         $totalResult->free();
 194                 }
 195
 196                 return new SqliteSearchResultSet( $resultSet, $this->searchTerms, $total );
 197         }
 198
 199         /**
 200          * Return a partial WHERE clause to exclude redirects, if so set
 201          * @return String
 202          */
 203         function queryRedirect() {
 204                 if( $this->showRedirects ) {
 205                         return '';
 206                 } else {
 207                         return 'AND page_is_redirect=0';
 208                 }
 209         }
 210
 211         /**
 212          * Return a partial WHERE clause to limit the search to the given namespaces
 213          * @return String
 214          */
 215         function queryNamespaces() {
 216                 if( is_null( $this->namespaces ) )
 217                         return '';  # search all
 218                 if ( !count( $this->namespaces ) ) {
 219                         $namespaces = '0';
 220                 } else {
 221                         $namespaces = $this->db->makeList( $this->namespaces );
 222                 }
 223                 return 'AND page_namespace IN (' . $namespaces . ')';
 224         }
 225
 226         /**
 227          * Returns a query with limit for number of results set.
 228          * @param $sql String:
 229          * @return String
 230          */
 231         function limitResult( $sql ) {
 232                 return $this->db->limitResult( $sql, $this->limit, $this->offset );
 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          * @return String
 241          */
 242         function getQuery( $filteredTerm, $fulltext ) {
 243                 return $this->limitResult(
 244                         $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 245                         $this->queryRedirect() . ' ' .
 246                         $this->queryNamespaces()
 247                 );
 248         }
 249
 250         /**
 251          * Picks which field to index on, depending on what type of query.
 252          * @param $fulltext Boolean
 253          * @return String
 254          */
 255         function getIndexField( $fulltext ) {
 256                 return $fulltext ? 'si_text' : 'si_title';
 257         }
 258
 259         /**
 260          * Get the base part of the search query.
 261          *
 262          * @param $filteredTerm String
 263          * @param $fulltext Boolean
 264          * @return String
 265          */
 266         function queryMain( $filteredTerm, $fulltext ) {
 267                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 268                 $page = $this->db->tableName( 'page' );
 269                 $searchindex = $this->db->tableName( 'searchindex' );
 270                 return "SELECT $searchindex.rowid, page_namespace, page_title " .
 271                         "FROM $page,$searchindex " .
 272                         "WHERE page_id=$searchindex.rowid AND $match";
 273         }
 274
 275         function getCountQuery( $filteredTerm, $fulltext ) {
 276                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 277                 $page = $this->db->tableName( 'page' );
 278                 $searchindex = $this->db->tableName( 'searchindex' );
 279                 return "SELECT COUNT(*) AS c " .
 280                         "FROM $page,$searchindex " .
 281                         "WHERE page_id=$searchindex.rowid AND $match" .
 282                         $this->queryRedirect() . ' ' .
 283                         $this->queryNamespaces();
 284         }
 285
 286         /**
 287          * Create or update the search index record for the given page.
 288          * Title and text should be pre-processed.
 289          *
 290          * @param $id Integer
 291          * @param $title String
 292          * @param $text String
 293          */
 294         function update( $id, $title, $text ) {
 295                 if ( !$this->fulltextSearchSupported() ) {
 296                         return;
 297                 }
 298                 // @todo: find a method to do it in a single request,
 299                 // couldn't do it so far due to typelessness of FTS3 tables.
 300                 $dbw = wfGetDB( DB_MASTER );
 301
 302                 $dbw->delete( 'searchindex', array( 'rowid' => $id ), __METHOD__ );
 303
 304                 $dbw->insert( 'searchindex',
 305                         array(
 306                                 'rowid' => $id,
 307                                 'si_title' => $title,
 308                                 'si_text' => $text
 309                         ), __METHOD__ );
 310         }
 311
 312         /**
 313          * Update a search index record's title only.
 314          * Title should be pre-processed.
 315          *
 316          * @param $id Integer
 317          * @param $title String
 318          */
 319         function updateTitle( $id, $title ) {
 320                 if ( !$this->fulltextSearchSupported() ) {
 321                         return;
 322                 }
 323                 $dbw = wfGetDB( DB_MASTER );
 324
 325                 $dbw->update( 'searchindex',
 326                         array( 'si_title' => $title ),
 327                         array( 'rowid'  => $id ),
 328                         __METHOD__ );
 329         }
 330 }
 331
 332 /**
 333  * @ingroup Search
 334  */
 335 class SqliteSearchResultSet extends SqlSearchResultSet {
 336         function __construct( $resultSet, $terms, $totalHits=null ) {
 337                 parent::__construct( $resultSet, $terms );
 338                 $this->mTotalHits = $totalHits;
 339         }
 340
 341         function getTotalHits() {
 342                 return $this->mTotalHits;
 343         }
 344 }