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 SearchDatabase {
  29         /**
  30          * Whether fulltext search is supported by current schema
  31          * @return Boolean
  32          */
  33         function fulltextSearchSupported() {
  34                 return $this->db->checkForEnabledSearch();
  35         }
  36
  37         /**
  38          * Parse the user's query and transform it into an SQL fragment which will
  39          * become part of a WHERE clause
  40          *
  41          * @return string
  42          */
  43         function parseQuery( $filteredText, $fulltext ) {
  44                 global $wgContLang;
  45                 $lc = SearchEngine::legalSearchChars(); // Minus format chars
  46                 $searchon = '';
  47                 $this->searchTerms = array();
  48
  49                 $m = array();
  50                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  51                                 $filteredText, $m, PREG_SET_ORDER ) ) {
  52                         foreach ( $m as $bits ) {
  53                                 @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  54
  55                                 if ( $nonQuoted != '' ) {
  56                                         $term = $nonQuoted;
  57                                         $quote = '';
  58                                 } else {
  59                                         $term = str_replace( '"', '', $term );
  60                                         $quote = '"';
  61                                 }
  62
  63                                 if ( $searchon !== '' ) {
  64                                         $searchon .= ' ';
  65                                 }
  66
  67                                 // Some languages such as Serbian store the input form in the search index,
  68                                 // so we may need to search for matches in multiple writing system variants.
  69                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  70                                 if ( is_array( $convertedVariants ) ) {
  71                                         $variants = array_unique( array_values( $convertedVariants ) );
  72                                 } else {
  73                                         $variants = array( $term );
  74                                 }
  75
  76                                 // The low-level search index does some processing on input to work
  77                                 // around problems with minimum lengths and encoding in MySQL's
  78                                 // fulltext engine.
  79                                 // For Chinese this also inserts spaces between adjacent Han characters.
  80                                 $strippedVariants = array_map(
  81                                         array( $wgContLang, 'normalizeForSearch' ),
  82                                         $variants );
  83
  84                                 // Some languages such as Chinese force all variants to a canonical
  85                                 // form when stripping to the low-level search index, so to be sure
  86                                 // let's check our variants list for unique items after stripping.
  87                                 $strippedVariants = array_unique( $strippedVariants );
  88
  89                                 $searchon .= $modifier;
  90                                 if ( count( $strippedVariants ) > 1 ) {
  91                                         $searchon .= '(';
  92                                 }
  93                                 foreach ( $strippedVariants as $stripped ) {
  94                                         if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
  95                                                 // Hack for Chinese: we need to toss in quotes for
  96                                                 // multiple-character phrases since normalizeForSearch()
  97                                                 // added spaces between them to make word breaks.
  98                                                 $stripped = '"' . trim( $stripped ) . '"';
  99                                         }
 100                                         $searchon .= "$quote$stripped$quote$wildcard ";
 101                                 }
 102                                 if ( count( $strippedVariants ) > 1 ) {
 103                                         $searchon .= ')';
 104                                 }
 105
 106                                 // Match individual terms or quoted phrase in result highlighting...
 107                                 // Note that variants will be introduced in a later stage for highlighting!
 108                                 $regexp = $this->regexTerm( $term, $wildcard );
 109                                 $this->searchTerms[] = $regexp;
 110                         }
 111
 112                 } else {
 113                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 114                 }
 115
 116                 $searchon = $this->db->strencode( $searchon );
 117                 $field = $this->getIndexField( $fulltext );
 118                 return " $field MATCH '$searchon' ";
 119         }
 120
 121         function regexTerm( $string, $wildcard ) {
 122                 global $wgContLang;
 123
 124                 $regex = preg_quote( $string, '/' );
 125                 if ( $wgContLang->hasWordBreaks() ) {
 126                         if ( $wildcard ) {
 127                                 // Don't cut off the final bit!
 128                                 $regex = "\b$regex";
 129                         } else {
 130                                 $regex = "\b$regex\b";
 131                         }
 132                 } else {
 133                         // For Chinese, words may legitimately abut other words in the text literal.
 134                         // Don't add \b boundary checks... note this could cause false positives
 135                         // for latin chars.
 136                 }
 137                 return $regex;
 138         }
 139
 140         public static function legalSearchChars() {
 141                 return "\"*" . parent::legalSearchChars();
 142         }
 143
 144         /**
 145          * Perform a full text search query and return a result set.
 146          *
 147          * @param string $term raw search term
 148          * @return SqliteSearchResultSet
 149          */
 150         function searchText( $term ) {
 151                 return $this->searchInternal( $term, true );
 152         }
 153
 154         /**
 155          * Perform a title-only search query and return a result set.
 156          *
 157          * @param string $term raw search term
 158          * @return SqliteSearchResultSet
 159          */
 160         function searchTitle( $term ) {
 161                 return $this->searchInternal( $term, false );
 162         }
 163
 164         protected function searchInternal( $term, $fulltext ) {
 165                 global $wgCountTotalSearchHits, $wgContLang;
 166
 167                 if ( !$this->fulltextSearchSupported() ) {
 168                         return null;
 169                 }
 170
 171                 $filteredTerm = $this->filter( $wgContLang->lc( $term ) );
 172                 $resultSet = $this->db->query( $this->getQuery( $filteredTerm, $fulltext ) );
 173
 174                 $total = null;
 175                 if ( $wgCountTotalSearchHits ) {
 176                         $totalResult = $this->db->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 177                         $row = $totalResult->fetchObject();
 178                         if ( $row ) {
 179                                 $total = intval( $row->c );
 180                         }
 181                         $totalResult->free();
 182                 }
 183
 184                 return new SqliteSearchResultSet( $resultSet, $this->searchTerms, $total );
 185         }
 186
 187         /**
 188          * Return a partial WHERE clause to exclude redirects, if so set
 189          * @return String
 190          */
 191         function queryRedirect() {
 192                 if ( $this->showRedirects ) {
 193                         return '';
 194                 } else {
 195                         return 'AND page_is_redirect=0';
 196                 }
 197         }
 198
 199         /**
 200          * Return a partial WHERE clause to limit the search to the given namespaces
 201          * @return String
 202          */
 203         function queryNamespaces() {
 204                 if ( is_null( $this->namespaces ) ) {
 205                         return '';  # search all
 206                 }
 207                 if ( !count( $this->namespaces ) ) {
 208                         $namespaces = '0';
 209                 } else {
 210                         $namespaces = $this->db->makeList( $this->namespaces );
 211                 }
 212                 return 'AND page_namespace IN (' . $namespaces . ')';
 213         }
 214
 215         /**
 216          * Returns a query with limit for number of results set.
 217          * @param $sql String:
 218          * @return String
 219          */
 220         function limitResult( $sql ) {
 221                 return $this->db->limitResult( $sql, $this->limit, $this->offset );
 222         }
 223
 224         /**
 225          * Construct the full SQL query to do the search.
 226          * The guts shoulds be constructed in queryMain()
 227          * @param $filteredTerm String
 228          * @param $fulltext Boolean
 229          * @return String
 230          */
 231         function getQuery( $filteredTerm, $fulltext ) {
 232                 return $this->limitResult(
 233                         $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 234                         $this->queryRedirect() . ' ' .
 235                         $this->queryNamespaces()
 236                 );
 237         }
 238
 239         /**
 240          * Picks which field to index on, depending on what type of query.
 241          * @param $fulltext Boolean
 242          * @return String
 243          */
 244         function getIndexField( $fulltext ) {
 245                 return $fulltext ? 'si_text' : 'si_title';
 246         }
 247
 248         /**
 249          * Get the base part of the search query.
 250          *
 251          * @param $filteredTerm String
 252          * @param $fulltext Boolean
 253          * @return String
 254          */
 255         function queryMain( $filteredTerm, $fulltext ) {
 256                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 257                 $page = $this->db->tableName( 'page' );
 258                 $searchindex = $this->db->tableName( 'searchindex' );
 259                 return "SELECT $searchindex.rowid, page_namespace, page_title " .
 260                         "FROM $page,$searchindex " .
 261                         "WHERE page_id=$searchindex.rowid AND $match";
 262         }
 263
 264         function getCountQuery( $filteredTerm, $fulltext ) {
 265                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 266                 $page = $this->db->tableName( 'page' );
 267                 $searchindex = $this->db->tableName( 'searchindex' );
 268                 return "SELECT COUNT(*) AS c " .
 269                         "FROM $page,$searchindex " .
 270                         "WHERE page_id=$searchindex.rowid AND $match" .
 271                         $this->queryRedirect() . ' ' .
 272                         $this->queryNamespaces();
 273         }
 274
 275         /**
 276          * Create or update the search index record for the given page.
 277          * Title and text should be pre-processed.
 278          *
 279          * @param $id Integer
 280          * @param $title String
 281          * @param $text String
 282          */
 283         function update( $id, $title, $text ) {
 284                 if ( !$this->fulltextSearchSupported() ) {
 285                         return;
 286                 }
 287                 // @todo find a method to do it in a single request,
 288                 // couldn't do it so far due to typelessness of FTS3 tables.
 289                 $dbw = wfGetDB( DB_MASTER );
 290
 291                 $dbw->delete( 'searchindex', array( 'rowid' => $id ), __METHOD__ );
 292
 293                 $dbw->insert( 'searchindex',
 294                         array(
 295                                 'rowid' => $id,
 296                                 'si_title' => $title,
 297                                 'si_text' => $text
 298                         ), __METHOD__ );
 299         }
 300
 301         /**
 302          * Update a search index record's title only.
 303          * Title should be pre-processed.
 304          *
 305          * @param $id Integer
 306          * @param $title String
 307          */
 308         function updateTitle( $id, $title ) {
 309                 if ( !$this->fulltextSearchSupported() ) {
 310                         return;
 311                 }
 312                 $dbw = wfGetDB( DB_MASTER );
 313
 314                 $dbw->update( 'searchindex',
 315                         array( 'si_title' => $title ),
 316                         array( 'rowid' => $id ),
 317                         __METHOD__ );
 318         }
 319 }
 320
 321 /**
 322  * @ingroup Search
 323  */
 324 class SqliteSearchResultSet extends SqlSearchResultSet {
 325         function __construct( $resultSet, $terms, $totalHits = null ) {
 326                 parent::__construct( $resultSet, $terms );
 327                 $this->mTotalHits = $totalHits;
 328         }
 329
 330         function getTotalHits() {
 331                 return $this->mTotalHits;
 332         }
 333 }