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                         self::$fulltextSupported = $this->db->selectField(
  49                                 'updatelog',
  50                                 'ul_key',
  51                                 array( 'ul_key' => 'fts3' ),
  52                                 __METHOD__ ) !== false;
  53                 }
  54                 return self::$fulltextSupported;
  55         }
  56
  57         /**
  58          * Parse the user's query and transform it into an SQL fragment which will
  59          * become part of a WHERE clause
  60          */
  61         function parseQuery( $filteredText, $fulltext ) {
  62                 global $wgContLang;
  63                 $lc = SearchEngine::legalSearchChars(); // Minus format chars
  64                 $searchon = '';
  65                 $this->searchTerms = array();
  66
  67                 # FIXME: This doesn't handle parenthetical expressions.
  68                 $m = array();
  69                 if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  70                           $filteredText, $m, PREG_SET_ORDER ) ) {
  71                         foreach( $m as $bits ) {
  72                                 @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  73
  74                                 if( $nonQuoted != '' ) {
  75                                         $term = $nonQuoted;
  76                                         $quote = '';
  77                                 } else {
  78                                         $term = str_replace( '"', '', $term );
  79                                         $quote = '"';
  80                                 }
  81
  82                                 if( $searchon !== '' ) $searchon .= ' ';
  83                                 if( $this->strictMatching && ($modifier == '') ) {
  84                                         // If we leave this out, boolean op defaults to OR which is rarely helpful.
  85                                         $modifier = '+';
  86                                 }
  87
  88                                 // Some languages such as Serbian store the input form in the search index,
  89                                 // so we may need to search for matches in multiple writing system variants.
  90                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  91                                 if( is_array( $convertedVariants ) ) {
  92                                         $variants = array_unique( array_values( $convertedVariants ) );
  93                                 } else {
  94                                         $variants = array( $term );
  95                                 }
  96
  97                                 // The low-level search index does some processing on input to work
  98                                 // around problems with minimum lengths and encoding in MySQL's
  99                                 // fulltext engine.
 100                                 // For Chinese this also inserts spaces between adjacent Han characters.
 101                                 $strippedVariants = array_map(
 102                                         array( $wgContLang, 'stripForSearch' ),
 103                                         $variants );
 104
 105                                 // Some languages such as Chinese force all variants to a canonical
 106                                 // form when stripping to the low-level search index, so to be sure
 107                                 // let's check our variants list for unique items after stripping.
 108                                 $strippedVariants = array_unique( $strippedVariants );
 109
 110                                 $searchon .= $modifier;
 111                                 if( count( $strippedVariants) > 1 )
 112                                         $searchon .= '(';
 113                                 foreach( $strippedVariants as $stripped ) {
 114                                         if( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
 115                                                 // Hack for Chinese: we need to toss in quotes for
 116                                                 // multiple-character phrases since stripForSearch()
 117                                                 // added spaces between them to make word breaks.
 118                                                 $stripped = '"' . trim( $stripped ) . '"';
 119                                         }
 120                                         $searchon .= "$quote$stripped$quote$wildcard ";
 121                                 }
 122                                 if( count( $strippedVariants) > 1 )
 123                                         $searchon .= ')';
 124
 125                                 // Match individual terms or quoted phrase in result highlighting...
 126                                 // Note that variants will be introduced in a later stage for highlighting!
 127                                 $regexp = $this->regexTerm( $term, $wildcard );
 128                                 $this->searchTerms[] = $regexp;
 129                         }
 130                         wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
 131                         wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
 132                 } else {
 133                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 134                 }
 135
 136                 $searchon = $this->db->strencode( $searchon );
 137                 $field = $this->getIndexField( $fulltext );
 138                 return " $field MATCH '$searchon' ";
 139         }
 140
 141         function regexTerm( $string, $wildcard ) {
 142                 global $wgContLang;
 143
 144                 $regex = preg_quote( $string, '/' );
 145                 if( $wgContLang->hasWordBreaks() ) {
 146                         if( $wildcard ) {
 147                                 // Don't cut off the final bit!
 148                                 $regex = "\b$regex";
 149                         } else {
 150                                 $regex = "\b$regex\b";
 151                         }
 152                 } else {
 153                         // For Chinese, words may legitimately abut other words in the text literal.
 154                         // Don't add \b boundary checks... note this could cause false positives
 155                         // for latin chars.
 156                 }
 157                 return $regex;
 158         }
 159
 160         public static function legalSearchChars() {
 161                 return "\"*" . parent::legalSearchChars();
 162         }
 163
 164         /**
 165          * Perform a full text search query and return a result set.
 166          *
 167          * @param $term String: raw search term
 168          * @return SqliteSearchResultSet
 169          */
 170         function searchText( $term ) {
 171                 return $this->searchInternal( $term, true );
 172         }
 173
 174         /**
 175          * Perform a title-only search query and return a result set.
 176          *
 177          * @param $term String: raw search term
 178          * @return SqliteSearchResultSet
 179          */
 180         function searchTitle( $term ) {
 181                 return $this->searchInternal( $term, false );
 182         }
 183
 184         protected function searchInternal( $term, $fulltext ) {
 185                 global $wgSearchMySQLTotalHits, $wgContLang;
 186
 187                 if ( !$this->fulltextSearchSupported() ) {
 188                         return null;
 189                 }
 190
 191                 $filteredTerm = $this->filter( $wgContLang->lc( $term ) );
 192                 $resultSet = $this->db->query( $this->getQuery( $filteredTerm, $fulltext ) );
 193
 194                 $total = null;
 195                 if( $wgSearchMySQLTotalHits ) {
 196                         $totalResult = $this->db->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 197                         $row = $totalResult->fetchObject();
 198                         if( $row ) {
 199                                 $total = intval( $row->c );
 200                         }
 201                         $totalResult->free();
 202                 }
 203
 204                 return new SqliteSearchResultSet( $resultSet, $this->searchTerms, $total );
 205         }
 206
 207
 208         /**
 209          * Return a partial WHERE clause to exclude redirects, if so set
 210          * @return String
 211          */
 212         function queryRedirect() {
 213                 if( $this->showRedirects ) {
 214                         return '';
 215                 } else {
 216                         return 'AND page_is_redirect=0';
 217                 }
 218         }
 219
 220         /**
 221          * Return a partial WHERE clause to limit the search to the given namespaces
 222          * @return String
 223          */
 224         function queryNamespaces() {
 225                 if( is_null($this->namespaces) )
 226                         return '';  # search all
 227                 if ( !count( $this->namespaces ) ) {
 228                         $namespaces = '0';
 229                 } else {
 230                         $namespaces = $this->db->makeList( $this->namespaces );
 231                 }
 232                 return 'AND page_namespace IN (' . $namespaces . ')';
 233         }
 234
 235         /**
 236          * Returns a query with limit for number of results set.
 237          * @param $sql String:
 238          * @return String
 239          */
 240         function limitResult( $sql ) {
 241                 return $this->db->limitResult( $sql, $this->limit, $this->offset );
 242         }
 243
 244         /**
 245          * Does not do anything for generic search engine
 246          * subclasses may define this though
 247          * @return String
 248          */
 249         function queryRanking( $filteredTerm, $fulltext ) {
 250                 return '';
 251         }
 252
 253         /**
 254          * Construct the full SQL query to do the search.
 255          * The guts shoulds be constructed in queryMain()
 256          * @param $filteredTerm String
 257          * @param $fulltext Boolean
 258          */
 259         function getQuery( $filteredTerm, $fulltext ) {
 260                 return $this->limitResult(
 261                         $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 262                         $this->queryRedirect() . ' ' .
 263                         $this->queryNamespaces() . ' ' .
 264                         $this->queryRanking( $filteredTerm, $fulltext )
 265                 );
 266         }
 267
 268         /**
 269          * Picks which field to index on, depending on what type of query.
 270          * @param $fulltext Boolean
 271          * @return String
 272          */
 273         function getIndexField( $fulltext ) {
 274                 return $fulltext ? 'si_text' : 'si_title';
 275         }
 276
 277         /**
 278          * Get the base part of the search query.
 279          *
 280          * @param $filteredTerm String
 281          * @param $fulltext Boolean
 282          * @return String
 283          */
 284         function queryMain( $filteredTerm, $fulltext ) {
 285                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 286                 $page        = $this->db->tableName( 'page' );
 287                 $searchindex = $this->db->tableName( 'searchindex' );
 288                 return "SELECT $searchindex.rowid, page_namespace, page_title " .
 289                         "FROM $page,$searchindex " .
 290                         "WHERE page_id=$searchindex.rowid AND $match";
 291         }
 292
 293         function getCountQuery( $filteredTerm, $fulltext ) {
 294                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 295                 $page        = $this->db->tableName( 'page' );
 296                 $searchindex = $this->db->tableName( 'searchindex' );
 297                 return "SELECT COUNT(*) AS c " .
 298                         "FROM $page,$searchindex " .
 299                         "WHERE page_id=$searchindex.rowid AND $match" .
 300                         $this->queryRedirect() . ' ' .
 301                         $this->queryNamespaces();
 302         }
 303
 304         /**
 305          * Create or update the search index record for the given page.
 306          * Title and text should be pre-processed.
 307          *
 308          * @param $id Integer
 309          * @param $title String
 310          * @param $text String
 311          */
 312         function update( $id, $title, $text ) {
 313                 if ( !$this->fulltextSearchSupported() ) {
 314                         return;
 315                 }
 316                 // @todo: find a method to do it in a single request,
 317                 // couldn't do it so far due to typelessness of FTS3 tables.
 318                 $dbw = wfGetDB( DB_MASTER );
 319
 320                 $dbw->delete( 'searchindex', array( 'rowid' => $id ), __METHOD__ );
 321
 322                 $dbw->insert( 'searchindex',
 323                         array(
 324                                 'rowid' => $id,
 325                                 'si_title' => $title,
 326                                 'si_text' => $text
 327                         ), __METHOD__ );
 328         }
 329
 330         /**
 331          * Update a search index record's title only.
 332          * Title should be pre-processed.
 333          *
 334          * @param $id Integer
 335          * @param $title String
 336          */
 337     function updateTitle( $id, $title ) {
 338                 if ( !$this->fulltextSearchSupported() ) {
 339                         return;
 340                 }
 341                 $dbw = wfGetDB( DB_MASTER );
 342
 343                 $dbw->update( 'searchindex',
 344                         array( 'rowid'  => $id ),
 345                         array( 'si_title' => $title ),
 346                         __METHOD__ );
 347         }
 348 }
 349
 350 /**
 351  * @ingroup Search
 352  */
 353 class SqliteSearchResultSet extends SearchResultSet {
 354         function SqliteSearchResultSet( $resultSet, $terms, $totalHits=null ) {
 355                 $this->mResultSet = $resultSet;
 356                 $this->mTerms = $terms;
 357                 $this->mTotalHits = $totalHits;
 358         }
 359
 360         function termMatches() {
 361                 return $this->mTerms;
 362         }
 363
 364         function numRows() {
 365                 return $this->mResultSet->numRows();
 366         }
 367
 368         function next() {
 369                 $row = $this->mResultSet->fetchObject();
 370                 if( $row === false ) {
 371                         return false;
 372                 } else {
 373                         return new SearchResult( $row );
 374                 }
 375         }
 376
 377         function free() {
 378                 $this->mResultSet->free();
 379         }
 380
 381
 382         function getTotalHits() {
 383                 return $this->mTotalHits;
 384         }
 385 }