Index: src/wp-includes/deprecated.php =================================================================== --- src/wp-includes/deprecated.php (revision 25628) +++ src/wp-includes/deprecated.php (working copy) @@ -3317,6 +3317,18 @@ function _save_post_hook() {} /** + * Formerly used internally to tidy up the search terms. + * + * @access private + * @since 2.9.0 + * @deprecated 3.7.0 + */ +function _search_terms_tidy( $t ) { + _deprecated_function( __FUNCTION__, '3.5', '' ); + return trim( $t, "\"'\n\r " ); +} + +/** * Check if the installed version of GD supports particular image type * * @since 2.9.0 Index: src/wp-includes/functions.php =================================================================== --- src/wp-includes/functions.php (revision 25628) +++ src/wp-includes/functions.php (working copy) @@ -3703,19 +3703,6 @@ } /** - * Used internally to tidy up the search terms. - * - * @access private - * @since 2.9.0 - * - * @param string $t - * @return string - */ -function _search_terms_tidy($t) { - return trim($t, "\"'\n\r "); -} - -/** * Returns true. * * Useful for returning true to filters easily. Index: src/wp-includes/query.php =================================================================== --- src/wp-includes/query.php (revision 25628) +++ src/wp-includes/query.php (working copy) @@ -1290,6 +1290,14 @@ var $thumbnails_cached = false; /** + * Cached list of search stopwords. + * + * @since 3.7.0 + * @var array + */ + private $stopwords; + + /** * Resets query flags to false. * * The query flags are what page info WordPress was able to figure out. @@ -1895,6 +1903,166 @@ } /** + * Generate SQL for the WHERE clause based on passed search terms. + * + * @since 3.7.0 + * + * @global type $wpdb + * @param array $q + * @param string $search + */ + function parse_search( &$q, $search ) { + global $wpdb; + + // added slashes screw with quote grouping when done early, so done later + $q['s'] = stripslashes( $q['s'] ); + if ( empty( $_GET['s'] ) && $this->is_main_query() ) + $q['s'] = urldecode( $q['s'] ); + // there are no line breaks in fields + $q['s'] = str_replace( array( "\r", "\n" ), '', $q['s'] ); + $q['num_all_terms'] = 1; + if ( ! empty( $q['sentence'] ) ) { + $q['search_terms'] = array( $q['s'] ); + } else { + if ( preg_match_all( '/".*?("|$)|((?<=[\t ",+])|^)[^\t ",+]+/', $q['s'], $matches ) ) { + $q['num_all_terms'] = count( $matches[0] ); + $q['search_terms'] = $this->parse_search_terms( $matches[0] ); + // if the search string has only short terms or stopwords, or is 10+ terms long, match it as sentence + if ( empty( $q['search_terms'] ) || count( $q['search_terms'] ) > 9 ) + $q['search_terms'] = array( $q['s'] ); + } else { + $q['search_terms'] = array( $q['s'] ); + } + } + + $n = ! empty( $q['exact'] ) ? '' : '%'; + $searchand = ''; + $q['search_orderby_title'] = array(); + foreach ( $q['search_terms'] as $term ) { + $term = esc_sql( like_escape( $term ) ); + if ( $n ) + $q['search_orderby_title'][] = "$wpdb->posts.post_title LIKE '%$term%'"; + + $search .= "{$searchand}(($wpdb->posts.post_title LIKE '{$n}{$term}{$n}') OR ($wpdb->posts.post_content LIKE '{$n}{$term}{$n}'))"; + $searchand = ' AND '; + } + + if ( ! empty( $search ) ) { + $search = " AND ({$search}) "; + if ( ! is_user_logged_in() ) + $search .= " AND ($wpdb->posts.post_password = '') "; + } + + // Allow plugins to contextually add/remove/modify the search section of the database query + return apply_filters( 'posts_search', $search, $this ); + } + + /** + * Check if the terms are suitable for searching. + * + * Includes array of stopwords (terms) that are excluded from the separate term matching when searching for posts. + * The list of English stopwords is the approximate search engines list. MySQL has a much longer default list of full-text stopwords. + * + * @since 3.7.0 + * + * @param array Terms to check + * @return array + */ + protected function parse_search_terms( $terms ) { + $strtolower_func = function_exists( 'mb_strtolower' ) ? 'mb_strtolower' : 'strtolower'; + $checked = $stopwords = array(); + + if ( ! isset( $this->stopwords ) ) { + // @todo translator comment + $_words = explode( ',', _x( 'about,an,are,as,at,be,by,com,for,from,how,in,is,it,of,on,or,that,the,this,to,was,what,when,where,who,will,with,www', + 'Comma separated list of common words to exclude when searching (stopwords).' ) ); + + foreach( $_words as $word ) { + $word = trim( $word, "\r\n\t " ); + + if ( ! $word ) + continue; + + $stopwords[] = $word; + } + + $this->stopwords = apply_filters( 'wp_search_stopwords', $stopwords ); + } + + foreach ( $terms as $term ) { + // keep before/after spaces when term is for exact match + if ( preg_match( '/^".+"$/', $term ) ) + $term = trim( $term, "\"'" ); + else + $term = trim( $term, "\"' " ); + + // \p{L} matches a single letter that is not a Chinese, Japanese, etc. char + if ( ! $term || preg_match( '/^\p{L}$/u', $term ) ) + continue; + + if ( in_array( $strtolower_func( $term ), $this->stopwords, true ) ) + continue; + + $checked[] = $term; + } + + return $checked; + } + + /** + * Generate SQL for the ORDER BY condition based on passed search terms. + * + * @global wpdb $wpdb + * @param array $q + * @return string + */ + function parse_search_order( &$q ) { + global $wpdb; + + $search_orderby = ''; + // Allow plugins to override sorting on 'post_title' and/or 'post_content'. + // Passing an empty array would disable sorting. + $search_orderby_on = apply_filters( 'posts_search_orderby_on', array( 'post_title', 'post_content' ), $this ); + + if ( ! empty( $search_orderby_on ) ) { + if ( $q['num_all_terms'] > 1 ) { + $num_terms = count( $q['search_orderby_title'] ); + $search_orderby_s = esc_sql( like_escape( $q['s'] ) ); + + if ( in_array( 'post_title', $search_orderby_on, true ) ) { + $search_orderby = '(CASE '; + // sentence match in 'post_title' + $search_orderby .= "WHEN $wpdb->posts.post_title LIKE '%{$search_orderby_s}%' THEN 1 "; + + // sanity limit, sort as sentence when more than 6 terms + // (few searches are longer than 6 terms and most titles are not) + if ( $num_terms < 7 ) { + // all words in title + $search_orderby .= 'WHEN ' . implode( ' AND ', $q['search_orderby_title'] ) . ' THEN 2 '; + // any word in title, not needed when $num_terms == 1 + if ( $num_terms > 1 ) + $search_orderby .= 'WHEN ' . implode( ' OR ', $q['search_orderby_title'] ) . ' THEN 3 '; + } + + // sentence match in 'post_content' + if ( in_array( 'post_content', $search_orderby_on, true ) ) + $search_orderby .= "WHEN $wpdb->posts.post_content LIKE '%{$search_orderby_s}%' THEN 4 "; + + $search_orderby .= 'ELSE 5 END)'; + } elseif ( in_array( 'post_content', $search_orderby_on, true ) ) { + // Not sorting on 'post_title', order by sentence matches in 'post_content' + $search_orderby = "$wpdb->posts.post_content LIKE '%{$search_orderby_s}%' DESC"; + } + } elseif ( in_array( 'post_title', $search_orderby_on, true ) ) { + // single word or sentence search + $search_orderby = reset( $q['search_orderby_title'] ) . ' DESC'; + } + } + // Allow plugins to add/remove/modify the 'order by' for the search section of the database query + return apply_filters( 'posts_search_orderby', $search_orderby, $this ); + } + + /** * Sets the 404 property and saves whether query is feed. * * @since 2.0.0 @@ -2236,35 +2404,11 @@ } // If a search pattern is specified, load the posts that match - if ( !empty($q['s']) ) { - // added slashes screw with quote grouping when done early, so done later - $q['s'] = stripslashes($q['s']); - if ( empty( $_GET['s'] ) && $this->is_main_query() ) - $q['s'] = urldecode($q['s']); - if ( !empty($q['sentence']) ) { - $q['search_terms'] = array($q['s']); - } else { - preg_match_all('/".*?("|$)|((?<=[\r\n\t ",+])|^)[^\r\n\t ",+]+/', $q['s'], $matches); - $q['search_terms'] = array_map('_search_terms_tidy', $matches[0]); - } - $n = !empty($q['exact']) ? '' : '%'; - $searchand = ''; - foreach( (array) $q['search_terms'] as $term ) { - $term = esc_sql( like_escape( $term ) ); - $search .= "{$searchand}(($wpdb->posts.post_title LIKE '{$n}{$term}{$n}') OR ($wpdb->posts.post_content LIKE '{$n}{$term}{$n}'))"; - $searchand = ' AND '; - } + // Sanity check: search string shouldn't be more than 1600 characters. + // See ticket #21688 for more info. + if ( ! empty( $q['s'] ) && strlen( $q['s'] ) < 1600 ) + $search = $this->parse_search( $q, $search ); - if ( !empty($search) ) { - $search = " AND ({$search}) "; - if ( !is_user_logged_in() ) - $search .= " AND ($wpdb->posts.post_password = '') "; - } - } - - // Allow plugins to contextually add/remove/modify the search section of the database query - $search = apply_filters_ref_array('posts_search', array( $search, &$this ) ); - // Taxonomies if ( !$this->is_singular ) { $this->parse_tax_query( $q ); @@ -2461,6 +2605,13 @@ $orderby .= " {$q['order']}"; } + // Order search results by relevance when another "orderby" is not specified in the query + if ( ! empty( $q['search_orderby_title'] ) && empty( $q['orderby'] ) ) { + $search_orderby = $this->parse_search_order( $q ); + if ( $search_orderby ) + $orderby = $orderby ? $search_orderby . ', ' . $orderby : $search_orderby; + } + if ( is_array( $post_type ) && count( $post_type ) > 1 ) { $post_type_cap = 'multiple_post_type'; } else { Index: tests/phpunit/tests/query/search.php =================================================================== --- tests/phpunit/tests/query/search.php (revision 0) +++ tests/phpunit/tests/query/search.php (working copy) @@ -0,0 +1,80 @@ +post_type = rand_str( 12 ); + register_post_type( $this->post_type ); + + $this->q = new WP_Query(); + } + + function tearDown() { + parent::tearDown(); + + _unregister_post_type( $this->post_type ); + unset( $this->q ); + } + + function get_search_results( $terms ) { + $args = http_build_query( array( 's' => $terms, 'post_type' => $this->post_type ) ); + return $this->q->query( $args ); + } + + function test_search_order_title_relevance() { + foreach ( range( 1, 7 ) as $i ) + $this->factory->post->create( array( 'post_content' => $i . rand_str() . ' about', 'post_type' => $this->post_type ) ); + $post_id = $this->factory->post->create( array( 'post_title' => 'About', 'post_type' => $this->post_type ) ); + + $posts = $this->get_search_results( 'About' ); + $this->assertEquals( $post_id, reset( $posts )->ID ); + } + + function test_search_order_content_relevance() { + $post_id1 = $this->factory->post->create( array( 'post_title' => 'About', 'post_type' => $this->post_type ) ); + $post_id2 = $this->factory->post->create( array( 'post_content' => 'About', 'post_type' => $this->post_type ) ); + + $posts1 = $this->get_search_results( 'About' ); + $this->assertEquals( $post_id1, reset( $posts1 )->ID ); + + add_filter( 'posts_search_orderby_on', array( $this, 'filter_posts_search_orderby_on' ) ); + + $posts2 = $this->get_search_results( 'About' ); + $this->assertEquals( $post_id2, reset( $posts2 )->ID ); + + remove_filter( 'posts_search_orderby_on', array( $this, 'filter_posts_search_orderby_on' ) ); + } + + function filter_posts_search_orderby_on() { + return array( 'post_content' ); + } + + function test_search_terms_query_var() { + $terms = 'This is a search term'; + $query = new WP_Query( array( 's' => 'This is a search term' ) ); + $this->assertNotEquals( explode( ' ', $terms ), $query->get( 'search_terms' ) ); + $this->assertEquals( array( 'search', 'term' ), $query->get( 'search_terms' ) ); + } + + function test_filter_stopwords() { + $terms = 'This is a search term'; + add_filter( 'wp_search_stopwords', array( $this, 'filter_wp_search_stopwords' ) ); + $query = new WP_Query( array( 's' => $terms ) ); + remove_filter( 'wp_search_stopwords', array( $this, 'filter_wp_search_stopwords' ) ); + + $this->assertNotEquals( array( 'search', 'term' ), $query->get( 'search_terms' ) ); + $this->assertEquals( array( 'This', 'is', 'search', 'term' ), $query->get( 'search_terms' ) ); + } + + function filter_wp_search_stopwords() { + return array(); + } +} \ No newline at end of file Property changes on: tests/phpunit/tests/query/search.php ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property