includes/search/SearchSqlite.php

   1 <?php
   2 # SQLite search backend, based upon SearchMysql
   3 #
   4 # This program is free software; you can redistribute it and/or modify
   5 # it under the terms of the GNU General Public License as published by
   6 # the Free Software Foundation; either version 2 of the License, or
   7 # (at your option) any later version.
   8 #
   9 # This program is distributed in the hope that it will be useful,
  10 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  12 # GNU General Public License for more details.
  13 #
  14 # You should have received a copy of the GNU General Public License along
  15 # with this program; if not, write to the Free Software Foundation, Inc.,
  16 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  17 # http://www.gnu.org/copyleft/gpl.html
  18
  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         var $strictMatching = true;
  30
  31         // Cached because SearchUpdate keeps recreating our class
  32         private static $fulltextSupported = NULL;
  33
  34         /**
  35          * Creates an instance of this class
  36          * @param $db DatabaseSqlite: database object
  37          */
  38         function __construct( $db ) {
  39                 $this->db = $db;
  40         }
  41
  42         /**
  43          * Whether fulltext search is supported by current schema
  44          * @return Boolean
  45          */
  46         function fulltextSearchSupported() {
  47                 if ( self::$fulltextSupported === NULL ) {
  48                         $res = $this->db->selectField( 'updatelog', 'ul_key', array( 'ul_key' => 'fts3' ), __METHOD__ );
  49                         self::$fulltextSupported = $res && $this->db->numRows( $res ) > 0;
  50                 }
  51                 return self::$fulltextSupported;
  52         }
  53
  54         /**
  55          * Parse the user's query and transform it into an SQL fragment which will
  56          * become part of a WHERE clause
  57          */
  58         function parseQuery( $filteredText, $fulltext ) {
  59                 global $wgContLang;
  60                 $lc = SearchEngine::legalSearchChars(); // Minus format chars
  61                 $searchon = '';
  62                 $this->searchTerms = array();
  63
  64                 # FIXME: This doesn't handle parenthetical expressions.
  65                 $m = array();
  66                 if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  67                           $filteredText, $m, PREG_SET_ORDER ) ) {
  68                         foreach( $m as $bits ) {
  69                                 @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  70
  71                                 if( $nonQuoted != '' ) {
  72                                         $term = $nonQuoted;
  73                                         $quote = '';
  74                                 } else {
  75                                         $term = str_replace( '"', '', $term );
  76                                         $quote = '"';
  77                                 }
  78
  79                                 if( $searchon !== '' ) $searchon .= ' ';
  80                                 if( $this->strictMatching && ($modifier == '') ) {
  81                                         // If we leave this out, boolean op defaults to OR which is rarely helpful.
  82                                         $modifier = '+';
  83                                 }
  84
  85                                 // Some languages such as Serbian store the input form in the search index,
  86                                 // so we may need to search for matches in multiple writing system variants.
  87                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  88                                 if( is_array( $convertedVariants ) ) {
  89                                         $variants = array_unique( array_values( $convertedVariants ) );
  90                                 } else {
  91                                         $variants = array( $term );
  92                                 }
  93
  94                                 // The low-level search index does some processing on input to work
  95                                 // around problems with minimum lengths and encoding in MySQL's
  96                                 // fulltext engine.
  97                                 // For Chinese this also inserts spaces between adjacent Han characters.
  98                                 $strippedVariants = array_map(
  99                                         array( $wgContLang, 'stripForSearch' ),
 100                                         $variants );
 101
 102                                 // Some languages such as Chinese force all variants to a canonical
 103                                 // form when stripping to the low-level search index, so to be sure
 104                                 // let's check our variants list for unique items after stripping.
 105                                 $strippedVariants = array_unique( $strippedVariants );
 106
 107                                 $searchon .= $modifier;
 108                                 if( count( $strippedVariants) > 1 )
 109                                         $searchon .= '(';
 110                                 foreach( $strippedVariants as $stripped ) {
 111                                         if( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
 112                                                 // Hack for Chinese: we need to toss in quotes for
 113                                                 // multiple-character phrases since stripForSearch()
 114                                                 // added spaces between them to make word breaks.
 115                                                 $stripped = '"' . trim( $stripped ) . '"';
 116                                         }
 117                                         $searchon .= "$quote$stripped$quote$wildcard ";
 118                                 }
 119                                 if( count( $strippedVariants) > 1 )
 120                                         $searchon .= ')';
 121
 122                                 // Match individual terms or quoted phrase in result highlighting...
 123                                 // Note that variants will be introduced in a later stage for highlighting!
 124                                 $regexp = $this->regexTerm( $term, $wildcard );
 125                                 $this->searchTerms[] = $regexp;
 126                         }
 127                         wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
 128                         wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
 129                 } else {
 130                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 131                 }
 132
 133                 $searchon = $this->db->strencode( $searchon );
 134                 $field = $this->getIndexField( $fulltext );
 135                 return " $field MATCH '$searchon' ";
 136         }
 137
 138         function regexTerm( $string, $wildcard ) {
 139                 global $wgContLang;
 140
 141                 $regex = preg_quote( $string, '/' );
 142                 if( $wgContLang->hasWordBreaks() ) {
 143                         if( $wildcard ) {
 144                                 // Don't cut off the final bit!
 145                                 $regex = "\b$regex";
 146                         } else {
 147                                 $regex = "\b$regex\b";
 148                         }
 149                 } else {
 150                         // For Chinese, words may legitimately abut other words in the text literal.
 151                         // Don't add \b boundary checks... note this could cause false positives
 152                         // for latin chars.
 153                 }
 154                 return $regex;
 155         }
 156
 157         public static function legalSearchChars() {
 158                 return "\"*" . parent::legalSearchChars();
 159         }
 160
 161         /**
 162          * Perform a full text search query and return a result set.
 163          *
 164          * @param $term String: raw search term
 165          * @return SqliteSearchResultSet
 166          */
 167         function searchText( $term ) {
 168                 return $this->searchInternal( $term, true );
 169         }
 170
 171         /**
 172          * Perform a title-only search query and return a result set.
 173          *
 174          * @param $term String: raw search term
 175          * @return SqliteSearchResultSet
 176          */
 177         function searchTitle( $term ) {
 178                 return $this->searchInternal( $term, false );
 179         }
 180
 181         protected function searchInternal( $term, $fulltext ) {
 182                 global $wgSearchMySQLTotalHits, $wgContLang;
 183
 184                 if ( !$this->fulltextSearchSupported() ) {
 185                         return null;
 186                 }
 187
 188                 $filteredTerm = $this->filter( $wgContLang->lc( $term ) );
 189                 $resultSet = $this->db->query( $this->getQuery( $filteredTerm, $fulltext ) );
 190
 191                 $total = null;
 192                 if( $wgSearchMySQLTotalHits ) {
 193                         $totalResult = $this->db->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 194                         $row = $totalResult->fetchObject();
 195                         if( $row ) {
 196                                 $total = intval( $row->c );
 197                         }
 198                         $totalResult->free();
 199                 }
 200
 201                 return new SqliteSearchResultSet( $resultSet, $this->searchTerms, $total );
 202         }
 203
 204
 205         /**
 206          * Return a partial WHERE clause to exclude redirects, if so set
 207          * @return String
 208          */
 209         function queryRedirect() {
 210                 if( $this->showRedirects ) {
 211                         return '';
 212                 } else {
 213                         return 'AND page_is_redirect=0';
 214                 }
 215         }
 216
 217         /**
 218          * Return a partial WHERE clause to limit the search to the given namespaces
 219          * @return String
 220          */
 221         function queryNamespaces() {
 222                 if( is_null($this->namespaces) )
 223                         return '';  # search all
 224                 if ( !count( $this->namespaces ) ) {
 225                         $namespaces = '0';
 226                 } else {
 227                         $namespaces = $this->db->makeList( $this->namespaces );
 228                 }
 229                 return 'AND page_namespace IN (' . $namespaces . ')';
 230         }
 231
 232         /**
 233          * Returns a query with limit for number of results set.
 234          * @param $sql String:
 235          * @return String
 236          */
 237         function limitResult( $sql ) {
 238                 return $this->db->limitResult( $sql, $this->limit, $this->offset );
 239         }
 240
 241         /**
 242          * Does not do anything for generic search engine
 243          * subclasses may define this though
 244          * @return String
 245          */
 246         function queryRanking( $filteredTerm, $fulltext ) {
 247                 return '';
 248         }
 249
 250         /**
 251          * Construct the full SQL query to do the search.
 252          * The guts shoulds be constructed in queryMain()
 253          * @param $filteredTerm String
 254          * @param $fulltext Boolean
 255          */
 256         function getQuery( $filteredTerm, $fulltext ) {
 257                 return $this->limitResult(
 258                         $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 259                         $this->queryRedirect() . ' ' .
 260                         $this->queryNamespaces() . ' ' .
 261                         $this->queryRanking( $filteredTerm, $fulltext )
 262                 );
 263         }
 264
 265         /**
 266          * Picks which field to index on, depending on what type of query.
 267          * @param $fulltext Boolean
 268          * @return String
 269          */
 270         function getIndexField( $fulltext ) {
 271                 return $fulltext ? 'si_text' : 'si_title';
 272         }
 273
 274         /**
 275          * Get the base part of the search query.
 276          *
 277          * @param $filteredTerm String
 278          * @param $fulltext Boolean
 279          * @return String
 280          */
 281         function queryMain( $filteredTerm, $fulltext ) {
 282                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 283                 $page        = $this->db->tableName( 'page' );
 284                 $searchindex = $this->db->tableName( 'searchindex' );
 285                 return "SELECT $searchindex.rowid, page_namespace, page_title " .
 286                         "FROM $page,$searchindex " .
 287                         "WHERE page_id=$searchindex.rowid AND $match";
 288         }
 289
 290         function getCountQuery( $filteredTerm, $fulltext ) {
 291                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 292                 $page        = $this->db->tableName( 'page' );
 293                 $searchindex = $this->db->tableName( 'searchindex' );
 294                 return "SELECT COUNT(*) AS c " .
 295                         "FROM $page,$searchindex " .
 296                         "WHERE page_id=$searchindex.rowid AND $match" .
 297                         $this->queryRedirect() . ' ' .
 298                         $this->queryNamespaces();
 299         }
 300
 301         /**
 302          * Create or update the search index record for the given page.
 303          * Title and text should be pre-processed.
 304          *
 305          * @param $id Integer
 306          * @param $title String
 307          * @param $text String
 308          */
 309         function update( $id, $title, $text ) {
 310                 if ( !$this->fulltextSearchSupported() ) {
 311                         return;
 312                 }
 313                 // @todo: find a method to do it in a single request,
 314                 // couldn't do it so far due to typelessness of FTS3 tables.
 315                 $dbw = wfGetDB( DB_MASTER );
 316
 317                 $dbw->delete( 'searchindex', array( 'rowid' => $id ), __METHOD__ );
 318
 319                 $dbw->insert( 'searchindex',
 320                         array(
 321                                 'rowid' => $id,
 322                                 'si_title' => $title,
 323                                 'si_text' => $text
 324                         ), __METHOD__ );
 325         }
 326
 327         /**
 328          * Update a search index record's title only.
 329          * Title should be pre-processed.
 330          *
 331          * @param $id Integer
 332          * @param $title String
 333          */
 334     function updateTitle( $id, $title ) {
 335                 if ( !$this->fulltextSearchSupported() ) {
 336                         return;
 337                 }
 338                 $dbw = wfGetDB( DB_MASTER );
 339
 340                 $dbw->update( 'searchindex',
 341                         array( 'rowid'  => $id ),
 342                         array( 'si_title' => $title ),
 343                         __METHOD__ );
 344         }
 345 }
 346
 347 /**
 348  * @ingroup Search
 349  */
 350 class SqliteSearchResultSet extends SearchResultSet {
 351         function SqliteSearchResultSet( $resultSet, $terms, $totalHits=null ) {
 352                 $this->mResultSet = $resultSet;
 353                 $this->mTerms = $terms;
 354                 $this->mTotalHits = $totalHits;
 355         }
 356
 357         function termMatches() {
 358                 return $this->mTerms;
 359         }
 360
 361         function numRows() {
 362                 return $this->mResultSet->numRows();
 363         }
 364
 365         function next() {
 366                 $row = $this->mResultSet->fetchObject();
 367                 if( $row === false ) {
 368                         return false;
 369                 } else {
 370                         return new SearchResult( $row );
 371                 }
 372         }
 373
 374         function free() {
 375                 $this->mResultSet->free();
 376         }
 377
 378
 379         function getTotalHits() {
 380                 return $this->mTotalHits;
 381         }
 382 }