<?php
/**
 * Title and description templates with %variable% tokens.
 *
 * @package OrbitstemSEO
 */

namespace Orbitstem\SEO\Content;

use Orbitstem\SEO\Core\Settings;

defined( 'ABSPATH' ) || exit;

/**
 * Replaces %tokens% with values and tidies separators left by empty values,
 * so "%term% %sep% %page% %sep% %sitename%" on page 1 reads "Term – Site".
 */
class TemplateEngine {

	/**
	 * Settings service.
	 *
	 * @var Settings
	 */
	private $settings;

	/**
	 * Constructor.
	 *
	 * @param Settings $settings Settings service.
	 */
	public function __construct( Settings $settings ) {
		$this->settings = $settings;
	}

	/**
	 * Tokens available in templates, for help text.
	 *
	 * @return array<string,string> token => description
	 */
	public static function tokens() {
		return array(
			'%title%'              => __( 'Post or page title', 'orbitstem-seo' ),
			'%excerpt%'            => __( 'Excerpt, or the start of the content', 'orbitstem-seo' ),
			'%sitename%'           => __( 'Site title', 'orbitstem-seo' ),
			'%tagline%'            => __( 'Site tagline', 'orbitstem-seo' ),
			'%sep%'                => __( 'Separator', 'orbitstem-seo' ),
			'%term%'               => __( 'Term name', 'orbitstem-seo' ),
			'%term_description%'   => __( 'Term description', 'orbitstem-seo' ),
			'%type%'               => __( 'Post type name (plural)', 'orbitstem-seo' ),
			'%author%'             => __( 'Author display name', 'orbitstem-seo' ),
			'%author_description%' => __( 'Author biographical info', 'orbitstem-seo' ),
			'%date%'               => __( 'Archive date', 'orbitstem-seo' ),
			'%search%'             => __( 'Search phrase', 'orbitstem-seo' ),
			'%page%'               => __( '"Page 2" on paginated views, empty on page 1', 'orbitstem-seo' ),
			'%not_found%'          => __( '"Page not found"', 'orbitstem-seo' ),
		);
	}

	/**
	 * Renders a template.
	 *
	 * @param string $template Template text.
	 * @param array  $vars     Values keyed by token name without percent signs.
	 * @return string Plain text (not escaped).
	 */
	public function render( $template, array $vars ) {
		$sep  = $this->settings->separator();
		$vars = array_merge(
			array(
				'sitename' => self::plain( get_bloginfo( 'name' ) ),
				'tagline'  => self::plain( get_bloginfo( 'description' ) ),
			),
			$vars
		);

		$out = preg_replace_callback(
			'/%([a-z_]+)%/',
			function ( $m ) use ( $vars ) {
				if ( 'sep' === $m[1] ) {
					return "\x1F"; // Placeholder so a literal separator character in a value is never collapsed.
				}
				return isset( $vars[ $m[1] ] ) ? (string) $vars[ $m[1] ] : '';
			},
			(string) $template
		);

		// Collapse runs of separators and drop separators at either end.
		$out = preg_replace( '/\x1F(?:\s*\x1F)+/', "\x1F", $out );
		$out = preg_replace( '/^[\s\x1F]+|[\s\x1F]+$/', '', $out );
		$out = str_replace( "\x1F", $sep, $out );
		$out = preg_replace( '/\s+/', ' ', $out );

		return trim( $out );
	}

	/**
	 * Strips tags, then decodes entities, so values are plain text before
	 * escaping. Stripping first keeps an encoded "&lt;5" as text instead of
	 * treating it as a tag. Block boundaries and page breaks become spaces so
	 * words from separate paragraphs are not glued together.
	 *
	 * @param string $text Text.
	 * @return string
	 */
	public static function plain( $text ) {
		$text = preg_replace( '#(</(?:p|div|h[1-6]|li|ul|ol|dl|dt|dd|blockquote|pre|figure|figcaption|table|tr|td|th|section|article|aside|header|footer|nav)\s*>|<br\s*/?>|<hr\s*/?>|<!--nextpage-->)#i', '$1 ', (string) $text );
		$text = Settings::strip_tags( $text );
		$text = html_entity_decode( $text, ENT_QUOTES | ENT_HTML5, 'UTF-8' );
		return trim( (string) preg_replace( '/\s+/', ' ', $text ) );
	}

	/**
	 * Shortens text to a length at a word boundary, adding an ellipsis.
	 *
	 * @param string $text   Plain text.
	 * @param int    $length Maximum characters.
	 * @return string
	 */
	public static function truncate( $text, $length = 155 ) {
		$text = self::plain( $text );
		if ( mb_strlen( $text ) <= $length ) {
			return $text;
		}
		$cut   = mb_substr( $text, 0, $length );
		$words = preg_replace( '/\s+\S*$/u', '', $cut ); // Drop the partial last word.
		if ( mb_strlen( $words ) > $length * 0.6 ) {
			$cut = $words;
		}
		return rtrim( $cut, " \t\n\r\0\x0B,.;:-" ) . "\u{2026}";
	}
}