SearchMySQL.php

Go to the documentation of this file.

<?php
class SearchMySQL extends SearchDatabase {
    protected $strictMatching = true;

    private static $mMinSearchLength;

    function parseQuery( $filteredText, $fulltext ) {
        global $wgContLang;

        $lc = $this->legalSearchChars(); // Minus format chars
        $searchon = '';
        $this->searchTerms = [];

        # @todo FIXME: This doesn't handle parenthetical expressions.
        $m = [];
        if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
                $filteredText, $m, PREG_SET_ORDER ) ) {
            foreach ( $m as $bits ) {
                MediaWiki\suppressWarnings();
                list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
                MediaWiki\restoreWarnings();

                if ( $nonQuoted != '' ) {
                    $term = $nonQuoted;
                    $quote = '';
                } else {
                    $term = str_replace( '"', '', $term );
                    $quote = '"';
                }

                if ( $searchon !== '' ) {
                    $searchon .= ' ';
                }
                if ( $this->strictMatching && ( $modifier == '' ) ) {
                    // If we leave this out, boolean op defaults to OR which is rarely helpful.
                    $modifier = '+';
                }

                // Some languages such as Serbian store the input form in the search index,
                // so we may need to search for matches in multiple writing system variants.
                $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
                if ( is_array( $convertedVariants ) ) {
                    $variants = array_unique( array_values( $convertedVariants ) );
                } else {
                    $variants = [ $term ];
                }

                // The low-level search index does some processing on input to work
                // around problems with minimum lengths and encoding in MySQL's
                // fulltext engine.
                // For Chinese this also inserts spaces between adjacent Han characters.
                $strippedVariants = array_map(
                    [ $wgContLang, 'normalizeForSearch' ],
                    $variants );

                // Some languages such as Chinese force all variants to a canonical
                // form when stripping to the low-level search index, so to be sure
                // let's check our variants list for unique items after stripping.
                $strippedVariants = array_unique( $strippedVariants );

                $searchon .= $modifier;
                if ( count( $strippedVariants ) > 1 ) {
                    $searchon .= '(';
                }
                foreach ( $strippedVariants as $stripped ) {
                    $stripped = $this->normalizeText( $stripped );
                    if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
                        // Hack for Chinese: we need to toss in quotes for
                        // multiple-character phrases since normalizeForSearch()
                        // added spaces between them to make word breaks.
                        $stripped = '"' . trim( $stripped ) . '"';
                    }
                    $searchon .= "$quote$stripped$quote$wildcard ";
                }
                if ( count( $strippedVariants ) > 1 ) {
                    $searchon .= ')';
                }

                // Match individual terms or quoted phrase in result highlighting...
                // Note that variants will be introduced in a later stage for highlighting!
                $regexp = $this->regexTerm( $term, $wildcard );
                $this->searchTerms[] = $regexp;
            }
            wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
            wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
        } else {
            wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
        }

        $searchon = $this->db->addQuotes( $searchon );
        $field = $this->getIndexField( $fulltext );
        return " MATCH($field) AGAINST($searchon IN BOOLEAN MODE) ";
    }

    function regexTerm( $string, $wildcard ) {
        global $wgContLang;

        $regex = preg_quote( $string, '/' );
        if ( $wgContLang->hasWordBreaks() ) {
            if ( $wildcard ) {
                // Don't cut off the final bit!
                $regex = "\b$regex";
            } else {
                $regex = "\b$regex\b";
            }
        } else {
            // For Chinese, words may legitimately abut other words in the text literal.
            // Don't add \b boundary checks... note this could cause false positives
            // for latin chars.
        }
        return $regex;
    }

    public static function legalSearchChars() {
        return "\"*" . parent::legalSearchChars();
    }

    function searchText( $term ) {
        return $this->searchInternal( $term, true );
    }

    function searchTitle( $term ) {
        return $this->searchInternal( $term, false );
    }

    protected function searchInternal( $term, $fulltext ) {
        // This seems out of place, why is this called with empty term?
        if ( trim( $term ) === '' ) {
            return null;
        }

        $filteredTerm = $this->filter( $term );
        $query = $this->getQuery( $filteredTerm, $fulltext );
        $resultSet = $this->db->select(
            $query['tables'], $query['fields'], $query['conds'],
            __METHOD__, $query['options'], $query['joins']
        );

        $total = null;
        $query = $this->getCountQuery( $filteredTerm, $fulltext );
        $totalResult = $this->db->select(
            $query['tables'], $query['fields'], $query['conds'],
            __METHOD__, $query['options'], $query['joins']
        );

        $row = $totalResult->fetchObject();
        if ( $row ) {
            $total = intval( $row->c );
        }
        $totalResult->free();

        return new SqlSearchResultSet( $resultSet, $this->searchTerms, $total );
    }

    public function supports( $feature ) {
        switch ( $feature ) {
        case 'title-suffix-filter':
            return true;
        default:
            return parent::supports( $feature );
        }
    }

    protected function queryFeatures( &$query ) {
        foreach ( $this->features as $feature => $value ) {
            if ( $feature === 'title-suffix-filter' && $value ) {
                $query['conds'][] = 'page_title' . $this->db->buildLike( $this->db->anyString(), $value );
            }
        }
    }

    function queryNamespaces( &$query ) {
        if ( is_array( $this->namespaces ) ) {
            if ( count( $this->namespaces ) === 0 ) {
                $this->namespaces[] = '0';
            }
            $query['conds']['page_namespace'] = $this->namespaces;
        }
    }

    protected function limitResult( &$query ) {
        $query['options']['LIMIT'] = $this->limit;
        $query['options']['OFFSET'] = $this->offset;
    }

    function getQuery( $filteredTerm, $fulltext ) {
        $query = [
            'tables' => [],
            'fields' => [],
            'conds' => [],
            'options' => [],
            'joins' => [],
        ];

        $this->queryMain( $query, $filteredTerm, $fulltext );
        $this->queryFeatures( $query );
        $this->queryNamespaces( $query );
        $this->limitResult( $query );

        return $query;
    }

    function getIndexField( $fulltext ) {
        return $fulltext ? 'si_text' : 'si_title';
    }

    function queryMain( &$query, $filteredTerm, $fulltext ) {
        $match = $this->parseQuery( $filteredTerm, $fulltext );
        $query['tables'][] = 'page';
        $query['tables'][] = 'searchindex';
        $query['fields'][] = 'page_id';
        $query['fields'][] = 'page_namespace';
        $query['fields'][] = 'page_title';
        $query['conds'][] = 'page_id=si_page';
        $query['conds'][] = $match;
    }

    function getCountQuery( $filteredTerm, $fulltext ) {
        $match = $this->parseQuery( $filteredTerm, $fulltext );

        $query = [
            'tables' => [ 'page', 'searchindex' ],
            'fields' => [ 'COUNT(*) as c' ],
            'conds' => [ 'page_id=si_page', $match ],
            'options' => [],
            'joins' => [],
        ];

        $this->queryFeatures( $query );
        $this->queryNamespaces( $query );

        return $query;
    }

    function update( $id, $title, $text ) {
        $dbw = wfGetDB( DB_MASTER );
        $dbw->replace( 'searchindex',
            [ 'si_page' ],
            [
                'si_page' => $id,
                'si_title' => $this->normalizeText( $title ),
                'si_text' => $this->normalizeText( $text )
            ], __METHOD__ );
    }

    function updateTitle( $id, $title ) {
        $dbw = wfGetDB( DB_MASTER );

        $dbw->update( 'searchindex',
            [ 'si_title' => $this->normalizeText( $title ) ],
            [ 'si_page' => $id ],
            __METHOD__,
            [ $dbw->lowPriorityOption() ] );
    }

    function delete( $id, $title ) {
        $dbw = wfGetDB( DB_MASTER );

        $dbw->delete( 'searchindex', [ 'si_page' => $id ], __METHOD__ );
    }

    function normalizeText( $string ) {
        global $wgContLang;

        $out = parent::normalizeText( $string );

        // MySQL fulltext index doesn't grok utf-8, so we
        // need to fold cases and convert to hex
        $out = preg_replace_callback(
            "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
            [ $this, 'stripForSearchCallback' ],
            $wgContLang->lc( $out ) );

        // And to add insult to injury, the default indexing
        // ignores short words... Pad them so we can pass them
        // through without reconfiguring the server...
        $minLength = $this->minSearchLength();
        if ( $minLength > 1 ) {
            $n = $minLength - 1;
            $out = preg_replace(
                "/\b(\w{1,$n})\b/",
                "$1u800",
                $out );
        }

        // Periods within things like hostnames and IP addresses
        // are also important -- we want a search for "example.com"
        // or "192.168.1.1" to work sanely.
        // MySQL's search seems to ignore them, so you'd match on
        // "example.wikipedia.com" and "192.168.83.1" as well.
        $out = preg_replace(
            "/(\w)\.(\w|\*)/u",
            "$1u82e$2",
            $out );

        return $out;
    }

    protected function stripForSearchCallback( $matches ) {
        return 'u8' . bin2hex( $matches[1] );
    }

    protected function minSearchLength() {
        if ( is_null( self::$mMinSearchLength ) ) {
            $sql = "SHOW GLOBAL VARIABLES LIKE 'ft\\_min\\_word\\_len'";

            $dbr = wfGetDB( DB_SLAVE );
            $result = $dbr->query( $sql, __METHOD__ );
            $row = $result->fetchObject();
            $result->free();

            if ( $row && $row->Variable_name == 'ft_min_word_len' ) {
                self::$mMinSearchLength = intval( $row->Value );
            } else {
                self::$mMinSearchLength = 0;
            }
        }
        return self::$mMinSearchLength;
    }
}