<?php
/**
 * Audit check: what bytes are actually served at /robots.txt?
 *
 * @package AIBlackout
 */

namespace AIBlackout\Audit\Checks;

use AIBlackout\Audit\Check;
use AIBlackout\Audit\Loopback;
use AIBlackout\Audit\Result;
use AIBlackout\Enforcement\RobotsTxt;

defined( 'ABSPATH' ) || exit;

/**
 * Fetches /robots.txt and reports the exact bytes a crawler receives.
 *
 * @since 1.0.0
 */
class ServedRobotsCheck implements Check {

	/**
	 * Loopback client.
	 *
	 * @var Loopback
	 */
	private $loopback;

	/**
	 * The robots.txt enforcement layer.
	 *
	 * @var RobotsTxt
	 */
	private $robots;

	/**
	 * Constructor.
	 *
	 * @param Loopback  $loopback Loopback client.
	 * @param RobotsTxt $robots   robots.txt layer.
	 */
	public function __construct( Loopback $loopback, RobotsTxt $robots ) {
		$this->loopback = $loopback;
		$this->robots   = $robots;
	}

	/**
	 * Returns the check identifier.
	 *
	 * @return string Identifier.
	 */
	public function get_id(): string {
		return 'served_robots';
	}

	/**
	 * Whether a loopback request is required.
	 *
	 * @return bool True.
	 */
	public function needs_loopback(): bool {
		return true;
	}

	/**
	 * Runs the check.
	 *
	 * @return Result The finding.
	 */
	public function run(): Result {
		$label    = __( 'What crawlers actually receive at /robots.txt', 'ai-blackout' );
		$response = $this->loopback->get( '/robots.txt' );

		if ( is_wp_error( $response ) ) {
			return new Result(
				$this->get_id(),
				$label,
				Result::UNKNOWN,
				__( 'Your robots.txt could not be fetched, so we cannot show what crawlers receive.', 'ai-blackout' ),
				__( 'Some hosts stop a site from requesting its own pages. This does not mean your robots.txt is wrong, only that we could not read it back.', 'ai-blackout' ),
				__( 'Open /robots.txt in your browser to see it yourself.', 'ai-blackout' ),
				array( 'error' => $response->get_error_message() )
			);
		}

		$body       = $response['body'];
		$has_block  = false !== strpos( $body, RobotsTxt::BEGIN_MARKER );
		$expected   = $this->robots->build_block();
		$want_block = '' !== $expected;

		$detail = array(
			'url'         => $response['url'],
			'status_code' => $response['code'],
			'bytes'       => strlen( $body ),
			'served'      => $body,
			'has_block'   => $has_block,
		);

		if ( 200 !== $response['code'] ) {
			return new Result(
				$this->get_id(),
				$label,
				Result::BROKEN,
				sprintf(
					/* translators: %d: HTTP status code. */
					__( 'Requesting /robots.txt returned %d instead of 200.', 'ai-blackout' ),
					$response['code']
				),
				__( 'Crawlers cannot read your robots.txt, so none of your rules are reaching them, whatever the settings screen shows.', 'ai-blackout' ),
				__( 'Check whether a redirect, security rule or server setting is blocking access to /robots.txt.', 'ai-blackout' ),
				$detail
			);
		}

		if ( $want_block && ! $has_block ) {
			return new Result(
				$this->get_id(),
				$label,
				Result::BROKEN,
				__( 'Your robots.txt is being served, but your AI crawler rules are not in it.', 'ai-blackout' ),
				__( 'This is what crawlers actually receive, and your rules are missing from it. Something between WordPress and the visitor is replacing the file. Usually that is a robots.txt on disk, a CDN, or a server rule.', 'ai-blackout' ),
				__( 'Check whether a robots.txt file exists on disk, and whether a CDN or caching service is serving an older copy.', 'ai-blackout' ),
				$detail
			);
		}

		if ( ! $want_block ) {
			return new Result(
				$this->get_id(),
				$label,
				Result::INFO,
				__( 'This is exactly what crawlers receive at /robots.txt.', 'ai-blackout' ),
				__( 'No AI crawler rules are set, so none appear in the file.', 'ai-blackout' ),
				__( 'Choose one of the blocking answers at the top of this screen.', 'ai-blackout' ),
				$detail
			);
		}

		return new Result(
			$this->get_id(),
			$label,
			Result::DECLARED,
			__( 'This is exactly what crawlers receive, and your AI rules are in it.', 'ai-blackout' ),
			__( 'Confirmed by fetching the file the same way a crawler would. Remember this is a request that crawlers can ignore, not a barrier that stops them.', 'ai-blackout' ),
			'',
			$detail
		);
	}
}