<?php
/**
 * Resources loaded from other websites (F-PERF-02, F-PERF-04).
 *
 * @package ForensixDigital
 */

declare(strict_types=1);

namespace ForensixDigital\Apps\Perf;

defined( 'ABSPATH' ) || exit;

use ForensixDigital\Services\Findings\Observation;
use ForensixDigital\Services\Findings\Priority;
use ForensixDigital\Services\Scan\Check;
use ForensixDigital\Services\Scan\ParsedPage;
use ForensixDigital\Services\Scan\Reference;
use ForensixDigital\Services\Scan\Url;

/**
 * Lists the scripts and stylesheets a page loads from another host, one finding per host and
 * place. Google Fonts get their own finding id, because loading them from Google sends the
 * visitor's address to Google. All of these are suggestions: many sites need some of them.
 */
final class ThirdPartyCheck implements Check {

	public const ID = 'perf.third_party';

	/**
	 * Creates the check.
	 *
	 * @param string[] $hosts Hosts of the site itself.
	 */
	public function __construct( private array $hosts ) {}

	/**
	 * Check id.
	 */
	public function id(): string {
		return self::ID;
	}

	/**
	 * The check looks only at the page itself.
	 *
	 * @param ParsedPage $page Parsed page.
	 * @return string[]
	 */
	public function targets( ParsedPage $page ): array {
		return array();
	}

	/**
	 * Which group a host belongs to: fonts, tracking, cdn or other.
	 *
	 * @param string $host Lower-case host name.
	 */
	public static function group( string $host ): string {
		foreach ( self::groups() as $group => $list ) {
			foreach ( $list as $known ) {
				if ( $host === $known || str_ends_with( $host, '.' . $known ) ) {
					return $group;
				}
			}
		}
		return 'other';
	}

	/**
	 * The known hosts by group, from data/third-party-hosts.json.
	 *
	 * @return array<string, array<int, string>>
	 */
	private static function groups(): array {
		static $groups = null;
		if ( null === $groups ) {
			$groups = array();
			$json   = file_get_contents( dirname( __DIR__, 3 ) . '/data/third-party-hosts.json' ); // phpcs:ignore WordPress.WP.AlternativeFunctions.file_get_contents_file_get_contents -- Reads a file shipped with the plugin.
			$data   = is_string( $json ) ? json_decode( $json, true ) : null;
			foreach ( is_array( $data ) ? $data : array() as $name => $list ) {
				$groups[ (string) $name ] = is_array( $list ) ? array_values( array_map( 'strval', $list ) ) : array();
			}
		}
		return $groups;
	}

	/**
	 * Problems on this page.
	 *
	 * @param ParsedPage $page    Parsed page.
	 * @param array      $results Not used.
	 * @phpstan-param array<string, \ForensixDigital\Services\Scan\TargetResult> $results
	 * @return array<int, array{observation: Observation, context: string}>
	 */
	public function inspect( ParsedPage $page, array $results ): array {
		$hits = array();
		$seen = array();
		foreach ( $page->references as $reference ) {
			if ( ! in_array( $reference->kind, array( Reference::SCRIPT, Reference::STYLE ), true ) || ! str_starts_with( $reference->url, 'http' ) ) {
				continue;
			}
			$host = strtolower( (string) wp_parse_url( $reference->url, PHP_URL_HOST ) );
			if ( '' === $host || Url::is_internal( $reference->url, $this->hosts ) ) {
				continue;
			}
			$key = $host . '|' . $reference->source->type . '|' . $reference->source->ref;
			if ( isset( $seen[ $key ] ) ) {
				continue;
			}
			$seen[ $key ] = true;
			$group        = self::group( $host );
			$hits[]       = array(
				'observation' => new Observation(
					'fonts' === $group ? 'perf.google_fonts' : self::ID,
					Priority::LOW,
					Priority::SUGGESTION,
					'fonts' === $group
						? __( 'Fonts are loaded from an outside server', 'forensix-digital' )
						/* translators: %s: host name. */
						: sprintf( __( 'Files are loaded from %s', 'forensix-digital' ), $host ),
					$host,
					$reference->source->type,
					$reference->source->ref,
					'',
					array(
						'host'  => $host,
						'group' => $group,
					)
				),
				'context'     => $reference->url,
			);
		}
		return $hits;
	}
}