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