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 use MediaWiki\MediaWikiServices;
  25 use Wikimedia\Rdbms\DatabaseSqlite;
  26
  27 /**
  28  * Search engine hook for SQLite
  29  * @ingroup Search
  30  */
  31 class SearchSqlite extends SearchDatabase {
  32         /**
  33          * Whether fulltext search is supported by current schema
  34          * @return bool
  35          */
  36         function fulltextSearchSupported() {
  37                 /** @var DatabaseSqlite $dbr */
  38                 $dbr = $this->lb->getConnection( DB_REPLICA );
  39
  40                 return $dbr->checkForEnabledSearch();
  41         }
  42
  43         /**
  44          * Parse the user's query and transform it into an SQL fragment which will
  45          * become part of a WHERE clause
  46          *
  47          * @param string $filteredText
  48          * @param bool $fulltext
  49          * @return string
  50          */
  51         private function parseQuery( $filteredText, $fulltext ) {
  52                 $lc = $this->legalSearchChars( self::CHARS_NO_SYNTAX ); // Minus syntax chars (" and *)
  53                 $searchon = '';
  54                 $this->searchTerms = [];
  55
  56                 $m = [];
  57                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  58                                 $filteredText, $m, PREG_SET_ORDER ) ) {
  59                         foreach ( $m as $bits ) {
  60                                 Wikimedia\suppressWarnings();
  61                                 list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  62                                 Wikimedia\restoreWarnings();
  63
  64                                 if ( $nonQuoted != '' ) {
  65                                         $term = $nonQuoted;
  66                                         $quote = '';
  67                                 } else {
  68                                         $term = str_replace( '"', '', $term );
  69                                         $quote = '"';
  70                                 }
  71
  72                                 if ( $searchon !== '' ) {
  73                                         $searchon .= ' ';
  74                                 }
  75
  76                                 // Some languages such as Serbian store the input form in the search index,
  77                                 // so we may need to search for matches in multiple writing system variants.
  78                                 $convertedVariants = MediaWikiServices::getInstance()->getContentLanguage()->
  79                                         autoConvertToAllVariants( $term );
  80                                 if ( is_array( $convertedVariants ) ) {
  81                                         $variants = array_unique( array_values( $convertedVariants ) );
  82                                 } else {
  83                                         $variants = [ $term ];
  84                                 }
  85
  86                                 // The low-level search index does some processing on input to work
  87                                 // around problems with minimum lengths and encoding in MySQL's
  88                                 // fulltext engine.
  89                                 // For Chinese this also inserts spaces between adjacent Han characters.
  90                                 $strippedVariants = array_map(
  91                                         [ MediaWikiServices::getInstance()->getContentLanguage(),
  92                                                 'normalizeForSearch' ],
  93                                         $variants );
  94
  95                                 // Some languages such as Chinese force all variants to a canonical
  96                                 // form when stripping to the low-level search index, so to be sure
  97                                 // let's check our variants list for unique items after stripping.
  98                                 $strippedVariants = array_unique( $strippedVariants );
  99
 100                                 $searchon .= $modifier;
 101                                 if ( count( $strippedVariants ) > 1 ) {
 102                                         $searchon .= '(';
 103                                 }
 104                                 foreach ( $strippedVariants as $stripped ) {
 105                                         if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
 106                                                 // Hack for Chinese: we need to toss in quotes for
 107                                                 // multiple-character phrases since normalizeForSearch()
 108                                                 // added spaces between them to make word breaks.
 109                                                 $stripped = '"' . trim( $stripped ) . '"';
 110                                         }
 111                                         $searchon .= "$quote$stripped$quote$wildcard ";
 112                                 }
 113                                 if ( count( $strippedVariants ) > 1 ) {
 114                                         $searchon .= ')';
 115                                 }
 116
 117                                 // Match individual terms or quoted phrase in result highlighting...
 118                                 // Note that variants will be introduced in a later stage for highlighting!
 119                                 $regexp = $this->regexTerm( $term, $wildcard );
 120                                 $this->searchTerms[] = $regexp;
 121                         }
 122
 123                 } else {
 124                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 125                 }
 126
 127                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 128                 $searchon = $dbr->addQuotes( $searchon );
 129                 $field = $this->getIndexField( $fulltext );
 130
 131                 return " $field MATCH $searchon ";
 132         }
 133
 134         private function regexTerm( $string, $wildcard ) {
 135                 $regex = preg_quote( $string, '/' );
 136                 if ( MediaWikiServices::getInstance()->getContentLanguage()->hasWordBreaks() ) {
 137                         if ( $wildcard ) {
 138                                 // Don't cut off the final bit!
 139                                 $regex = "\b$regex";
 140                         } else {
 141                                 $regex = "\b$regex\b";
 142                         }
 143                 } else {
 144                         // For Chinese, words may legitimately abut other words in the text literal.
 145                         // Don't add \b boundary checks... note this could cause false positives
 146                         // for Latin chars.
 147                 }
 148                 return $regex;
 149         }
 150
 151         public function legalSearchChars( $type = self::CHARS_ALL ) {
 152                 $searchChars = parent::legalSearchChars( $type );
 153                 if ( $type === self::CHARS_ALL ) {
 154                         // " for phrase, * for wildcard
 155                         $searchChars = "\"*" . $searchChars;
 156                 }
 157                 return $searchChars;
 158         }
 159
 160         /**
 161          * Perform a full text search query and return a result set.
 162          *
 163          * @param string $term Raw search term
 164          * @return SqlSearchResultSet
 165          */
 166         protected function doSearchTextInDB( $term ) {
 167                 return $this->searchInternal( $term, true );
 168         }
 169
 170         /**
 171          * Perform a title-only search query and return a result set.
 172          *
 173          * @param string $term Raw search term
 174          * @return SqlSearchResultSet
 175          */
 176         protected function doSearchTitleInDB( $term ) {
 177                 return $this->searchInternal( $term, false );
 178         }
 179
 180         protected function searchInternal( $term, $fulltext ) {
 181                 if ( !$this->fulltextSearchSupported() ) {
 182                         return null;
 183                 }
 184
 185                 $filteredTerm =
 186                         $this->filter( MediaWikiServices::getInstance()->getContentLanguage()->lc( $term ) );
 187                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 188                 $resultSet = $dbr->query( $this->getQuery( $filteredTerm, $fulltext ) );
 189
 190                 $total = null;
 191                 $totalResult = $dbr->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 192                 $row = $totalResult->fetchObject();
 193                 if ( $row ) {
 194                         $total = intval( $row->c );
 195                 }
 196                 $totalResult->free();
 197
 198                 return new SqlSearchResultSet( $resultSet, $this->searchTerms, $total );
 199         }
 200
 201         /**
 202          * Return a partial WHERE clause to limit the search to the given namespaces
 203          * @return string
 204          */
 205         private function queryNamespaces() {
 206                 if ( is_null( $this->namespaces ) ) {
 207                         return '';  # search all
 208                 }
 209                 if ( $this->namespaces === [] ) {
 210                         $namespaces = '0';
 211                 } else {
 212                         $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 213                         $namespaces = $dbr->makeList( $this->namespaces );
 214                 }
 215                 return 'AND page_namespace IN (' . $namespaces . ')';
 216         }
 217
 218         /**
 219          * Returns a query with limit for number of results set.
 220          * @param string $sql
 221          * @return string
 222          */
 223         private function limitResult( $sql ) {
 224                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 225
 226                 return $dbr->limitResult( $sql, $this->limit, $this->offset );
 227         }
 228
 229         /**
 230          * Construct the full SQL query to do the search.
 231          * The guts shoulds be constructed in queryMain()
 232          * @param string $filteredTerm
 233          * @param bool $fulltext
 234          * @return string
 235          */
 236         private function getQuery( $filteredTerm, $fulltext ) {
 237                 return $this->limitResult(
 238                         $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 239                         $this->queryNamespaces()
 240                 );
 241         }
 242
 243         /**
 244          * Picks which field to index on, depending on what type of query.
 245          * @param bool $fulltext
 246          * @return string
 247          */
 248         private function getIndexField( $fulltext ) {
 249                 return $fulltext ? 'si_text' : 'si_title';
 250         }
 251
 252         /**
 253          * Get the base part of the search query.
 254          *
 255          * @param string $filteredTerm
 256          * @param bool $fulltext
 257          * @return string
 258          */
 259         private function queryMain( $filteredTerm, $fulltext ) {
 260                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 261                 $dbr = $this->lb->getMaintenanceConnectionRef( DB_REPLICA );
 262                 $page = $dbr->tableName( 'page' );
 263                 $searchindex = $dbr->tableName( 'searchindex' );
 264                 return "SELECT $searchindex.rowid, page_namespace, page_title " .
 265                         "FROM $page,$searchindex " .
 266                         "WHERE page_id=$searchindex.rowid AND $match";
 267         }
 268
 269         private function getCountQuery( $filteredTerm, $fulltext ) {
 270                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 271                 $dbr = $this->lb->getMaintenanceConnectionRef( DB_REPLICA );
 272                 $page = $dbr->tableName( 'page' );
 273                 $searchindex = $dbr->tableName( 'searchindex' );
 274                 return "SELECT COUNT(*) AS c " .
 275                         "FROM $page,$searchindex " .
 276                         "WHERE page_id=$searchindex.rowid AND $match " .
 277                         $this->queryNamespaces();
 278         }
 279
 280         /**
 281          * Create or update the search index record for the given page.
 282          * Title and text should be pre-processed.
 283          *
 284          * @param int $id
 285          * @param string $title
 286          * @param string $text
 287          */
 288         function update( $id, $title, $text ) {
 289                 if ( !$this->fulltextSearchSupported() ) {
 290                         return;
 291                 }
 292                 // @todo find a method to do it in a single request,
 293                 // couldn't do it so far due to typelessness of FTS3 tables.
 294                 $dbw = $this->lb->getConnectionRef( DB_MASTER );
 295                 $dbw->delete( 'searchindex', [ 'rowid' => $id ], __METHOD__ );
 296                 $dbw->insert( 'searchindex',
 297                         [
 298                                 'rowid' => $id,
 299                                 'si_title' => $title,
 300                                 'si_text' => $text
 301                         ], __METHOD__ );
 302         }
 303
 304         /**
 305          * Update a search index record's title only.
 306          * Title should be pre-processed.
 307          *
 308          * @param int $id
 309          * @param string $title
 310          */
 311         function updateTitle( $id, $title ) {
 312                 if ( !$this->fulltextSearchSupported() ) {
 313                         return;
 314                 }
 315
 316                 $dbw = $this->lb->getConnectionRef( DB_MASTER );
 317                 $dbw->update( 'searchindex',
 318                         [ 'si_title' => $title ],
 319                         [ 'rowid' => $id ],
 320                         __METHOD__ );
 321         }
 322 }