* @ingroup Search
*/
+use MediaWiki\MediaWikiServices;
+
/**
* Search engine hook for MySQL 4+
* @ingroup Search
private static $mMinSearchLength;
/**
- * Parse the user's query and transform it into an SQL fragment which will
- * become part of a WHERE clause
+ * Parse the user's query and transform it into two SQL fragments:
+ * a WHERE condition and an ORDER BY expression
*
* @param string $filteredText
* @param string $fulltext
*
- * @return string
+ * @return array
*/
- function parseQuery( $filteredText, $fulltext ) {
- global $wgContLang;
-
+ private function parseQuery( $filteredText, $fulltext ) {
$lc = $this->legalSearchChars( self::CHARS_NO_SYNTAX ); // Minus syntax chars (" and *)
$searchon = '';
$this->searchTerms = [];
// 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 );
+ $contLang = MediaWikiServices::getInstance()->getContentLanguage();
+ $convertedVariants = $contLang->autoConvertToAllVariants( $term );
if ( is_array( $convertedVariants ) ) {
$variants = array_unique( array_values( $convertedVariants ) );
} else {
// 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 );
+ $strippedVariants = array_map( [ $contLang, '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
$searchon = $this->db->addQuotes( $searchon );
$field = $this->getIndexField( $fulltext );
- return " MATCH($field) AGAINST($searchon IN BOOLEAN MODE) ";
+ return [
+ " MATCH($field) AGAINST($searchon IN BOOLEAN MODE) ",
+ " MATCH($field) AGAINST($searchon IN NATURAL LANGUAGE MODE) DESC "
+ ];
}
- function regexTerm( $string, $wildcard ) {
- global $wgContLang;
-
+ private function regexTerm( $string, $wildcard ) {
$regex = preg_quote( $string, '/' );
- if ( $wgContLang->hasWordBreaks() ) {
+ if ( MediaWikiServices::getInstance()->getContentLanguage()->hasWordBreaks() ) {
if ( $wildcard ) {
// Don't cut off the final bit!
$regex = "\b$regex";
} 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.
+ // for Latin chars.
}
return $regex;
}
* @param string $term Raw search term
* @return SqlSearchResultSet
*/
- function searchText( $term ) {
+ protected function doSearchTextInDB( $term ) {
return $this->searchInternal( $term, true );
}
* @param string $term Raw search term
* @return SqlSearchResultSet
*/
- function searchTitle( $term ) {
+ protected function doSearchTitleInDB( $term ) {
return $this->searchInternal( $term, false );
}
* @return array
* @since 1.18 (changed)
*/
- function getQuery( $filteredTerm, $fulltext ) {
+ private function getQuery( $filteredTerm, $fulltext ) {
$query = [
'tables' => [],
'fields' => [],
* @param bool $fulltext
* @return string
*/
- function getIndexField( $fulltext ) {
+ private function getIndexField( $fulltext ) {
return $fulltext ? 'si_text' : 'si_title';
}
* @param bool $fulltext
* @since 1.18 (changed)
*/
- function queryMain( &$query, $filteredTerm, $fulltext ) {
+ private function queryMain( &$query, $filteredTerm, $fulltext ) {
$match = $this->parseQuery( $filteredTerm, $fulltext );
$query['tables'][] = 'page';
$query['tables'][] = 'searchindex';
$query['fields'][] = 'page_namespace';
$query['fields'][] = 'page_title';
$query['conds'][] = 'page_id=si_page';
- $query['conds'][] = $match;
+ $query['conds'][] = $match[0];
+ $query['options']['ORDER BY'] = $match[1];
}
/**
* @param bool $fulltext
* @return array
*/
- function getCountQuery( $filteredTerm, $fulltext ) {
+ private function getCountQuery( $filteredTerm, $fulltext ) {
$match = $this->parseQuery( $filteredTerm, $fulltext );
$query = [
'tables' => [ 'page', 'searchindex' ],
'fields' => [ 'COUNT(*) as c' ],
- 'conds' => [ 'page_id=si_page', $match ],
+ 'conds' => [ 'page_id=si_page', $match[0] ],
'options' => [],
'joins' => [],
];
* @return mixed|string
*/
function normalizeText( $string ) {
- global $wgContLang;
-
$out = parent::normalizeText( $string );
// MySQL fulltext index doesn't grok utf-8, so we
$out = preg_replace_callback(
"/([\\xc0-\\xff][\\x80-\\xbf]*)/",
[ $this, 'stripForSearchCallback' ],
- $wgContLang->lc( $out ) );
+ MediaWikiServices::getInstance()->getContentLanguage()->lc( $out ) );
// And to add insult to injury, the default indexing
// ignores short words... Pad them so we can pass them