SearchMssql.php

Go to the documentation of this file.

<?php
use MediaWiki\MediaWikiServices;
use Wikimedia\Rdbms\IResultWrapper;
 
class SearchMssql extends SearchDatabase {
    protected function doSearchTextInDB( $term ) {
        $resultSet = $this->db->query( $this->getQuery( $this->filter( $term ), true ) );
        return new SqlSearchResultSet( $resultSet, $this->searchTerms );
    }
 
    protected function doSearchTitleInDB( $term ) {
        $resultSet = $this->db->query( $this->getQuery( $this->filter( $term ), false ) );
        return new SqlSearchResultSet( $resultSet, $this->searchTerms );
    }
 
    private function queryNamespaces() {
        $namespaces = implode( ',', $this->namespaces );
        if ( $namespaces == '' ) {
            $namespaces = '0';
        }
        return 'AND page_namespace IN (' . $namespaces . ')';
    }
 
    private function queryLimit( $sql ) {
        return $this->db->limitResult( $sql, $this->limit, $this->offset );
    }
 
    function queryRanking( $filteredTerm, $fulltext ) {
        return ' ORDER BY ftindex.[RANK] DESC'; // return ' ORDER BY score(1)';
    }
 
    private function getQuery( $filteredTerm, $fulltext ) {
        return $this->queryLimit( $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
            $this->queryNamespaces() . ' ' .
            $this->queryRanking( $filteredTerm, $fulltext ) . ' ' );
    }
 
    function getIndexField( $fulltext ) {
        return $fulltext ? 'si_text' : 'si_title';
    }
 
    private function queryMain( $filteredTerm, $fulltext ) {
        $match = $this->parseQuery( $filteredTerm, $fulltext );
        $page = $this->db->tableName( 'page' );
        $searchindex = $this->db->tableName( 'searchindex' );
 
        return 'SELECT page_id, page_namespace, page_title, ftindex.[RANK]' .
            "FROM $page,FREETEXTTABLE($searchindex , $match, LANGUAGE 'English') as ftindex " .
            'WHERE page_id=ftindex.[KEY] ';
    }
 
    private function parseQuery( $filteredText, $fulltext ) {
        $lc = $this->legalSearchChars( self::CHARS_NO_SYNTAX );
        $this->searchTerms = [];
 
        # @todo FIXME: This doesn't handle parenthetical expressions.
        $m = [];
        $q = [];
 
        if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
            $filteredText, $m, PREG_SET_ORDER ) ) {
            foreach ( $m as $terms ) {
                $q[] = $terms[1] . MediaWikiServices::getInstance()->getContentLanguage()->
                    normalizeForSearch( $terms[2] );
 
                if ( !empty( $terms[3] ) ) {
                    $regexp = preg_quote( $terms[3], '/' );
                    if ( $terms[4] ) {
                        $regexp .= "[0-9A-Za-z_]+";
                    }
                } else {
                    $regexp = preg_quote( str_replace( '"', '', $terms[2] ), '/' );
                }
                $this->searchTerms[] = $regexp;
            }
        }
 
        $searchon = $this->db->addQuotes( implode( ',', $q ) );
        $field = $this->getIndexField( $fulltext );
        return "$field, $searchon";
    }
 
    function update( $id, $title, $text ) {
        // We store the column data as UTF-8 byte order marked binary stream
        // because we are invoking the plain text IFilter on it so that, and we want it
        // to properly decode the stream as UTF-8.  SQL doesn't support UTF8 as a data type
        // but the indexer will correctly handle it by this method.  Since all we are doing
        // is passing this data to the indexer and never retrieving it via PHP, this will save space
        $table = $this->db->tableName( 'searchindex' );
        $utf8bom = '0xEFBBBF';
        $si_title = $utf8bom . bin2hex( $title );
        $si_text = $utf8bom . bin2hex( $text );
        $sql = "DELETE FROM $table WHERE si_page = $id;";
        $sql .= "INSERT INTO $table (si_page, si_title, si_text) VALUES ($id, $si_title, $si_text)";
        return $this->db->query( $sql, 'SearchMssql::update' );
    }
 
    function updateTitle( $id, $title ) {
        $table = $this->db->tableName( 'searchindex' );
 
        // see update for why we are using the utf8bom
        $utf8bom = '0xEFBBBF';
        $si_title = $utf8bom . bin2hex( $title );
        $sql = "DELETE FROM $table WHERE si_page = $id;";
        $sql .= "INSERT INTO $table (si_page, si_title, si_text) VALUES ($id, $si_title, 0x00)";
        return $this->db->query( $sql, 'SearchMssql::updateTitle' );
    }
}