<?php
/**
 * Content maintenance check (F-CONTENT-01, 02, 03).
 *
 * @package ForensixDigital
 */

declare(strict_types=1);

namespace ForensixDigital\Apps\Content;

defined( 'ABSPATH' ) || exit;

use ForensixDigital\Apps\Business\BusinessConfig;
use ForensixDigital\Apps\Business\ContactConsistency;
use ForensixDigital\Services\Findings\FindingStore;
use ForensixDigital\Services\Findings\Observation;
use ForensixDigital\Services\Findings\Priority;
use ForensixDigital\Services\Scan\StoredContentCheck;
use ForensixDigital\Services\Sources\Source;
use ForensixDigital\Services\Sources\SourceLabels;

/**
 * Reads published posts and pages, 40 per step, and reports placeholder text, the sample
 * content of a new WordPress site, the default tagline, empty paragraphs and headings, and
 * posts that have not been changed for a long time. It only reads; stale posts are
 * suggestions to review, not errors.
 */
final class ContentCheck implements StoredContentCheck {

	private const PER_STEP = 40;

	private const DUPLICATE_CAP = 1500;

	/**
	 * Creates the check.
	 *
	 * @param string       $app   App id the findings belong to.
	 * @param FindingStore $store Findings.
	 */
	public function __construct( private string $app, private FindingStore $store ) {}

	/**
	 * Handles one batch of posts.
	 *
	 * @param int                  $scan_id Scan id.
	 * @param array<string, mixed> $cursor  Cursor with the last post id seen.
	 * @return array<string, mixed>|null
	 */
	public function step( int $scan_id, array $cursor ): ?array {
		global $wpdb;
		$last = (int) ( $cursor['last'] ?? 0 );
		if ( 0 === $last ) {
			$this->check_tagline( $scan_id );
		}
		$rows = (array) $wpdb->get_results( // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery, WordPress.DB.DirectDatabaseQuery.NoCaching -- Reads posts in batches.
			$wpdb->prepare(
				"SELECT ID, post_type, post_name, post_title, post_content, post_modified_gmt FROM {$wpdb->posts} WHERE post_status = 'publish' AND post_type IN ('post','page') AND ID > %d ORDER BY ID ASC LIMIT %d", // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared -- Core table name.
				$last,
				self::PER_STEP
			)
		);

		$profile = self::business_details();
		$seen    = array();
		foreach ( (array) ( $cursor['seen'] ?? array() ) as $hash => $post ) {
			$seen[ (int) $hash ] = (int) $post;
		}
		foreach ( $rows as $row ) {
			$this->inspect( $scan_id, $row, $profile, $seen );
			$last = (int) $row->ID;
		}
		return count( $rows ) < self::PER_STEP ? null : array(
			'last' => $last,
			'seen' => $seen,
		);
	}

	/**
	 * The business details, when the Business details app is active and has a phone number or
	 * an email address to compare with.
	 *
	 * @return array<string, mixed>|null
	 */
	private static function business_details(): ?array {
		/**
		 * Filters the business details. The Business details app provides them while it is active.
		 *
		 * @param array<string, mixed>|null $details Details, null when none.
		 */
		$details = apply_filters( 'forensix_business_details', null );
		return is_array( $details ) && BusinessConfig::has_contacts( $details ) ? $details : null;
	}

	/**
	 * Months after which a post counts as not updated for a long time.
	 */
	public static function stale_months(): int {
		/**
		 * Filters the number of months after which a post is suggested for review.
		 *
		 * @param int $months Months.
		 */
		return max( 1, (int) apply_filters( 'forensix_content_stale_months', ContentRules::STALE_MONTHS ) );
	}

	/**
	 * Reports the tagline WordPress sets on installation.
	 *
	 * @param int $scan_id Scan id.
	 */
	private function check_tagline( int $scan_id ): void {
		if ( ContentRules::DEFAULT_TAGLINE !== (string) get_option( 'blogdescription' ) ) {
			return;
		}
		$source = new Source( Source::OPTION, 'blogdescription' );
		$this->store->record(
			$scan_id,
			$this->app,
			new Observation(
				'content.default_tagline',
				Priority::LOW,
				Priority::ERROR,
				__( 'The site still has the default tagline', 'forensix-digital' ),
				home_url( '/' ),
				$source->type,
				$source->ref,
				__( 'Site tagline', 'forensix-digital' ),
				array()
			),
			admin_url( 'options-general.php' ),
			ContentRules::DEFAULT_TAGLINE
		);
	}

	/**
	 * The Elementor elements of a post, or none when it is not built with Elementor.
	 *
	 * @param int $id Post id.
	 * @return array<int|string, mixed>
	 */
	private function elementor_elements( int $id ): array {
		$data = get_post_meta( $id, '_elementor_data', true );
		$data = is_string( $data ) && '' !== $data ? json_decode( $data, true ) : $data;
		return is_array( $data ) ? $data : array();
	}

	/**
	 * Checks one post.
	 *
	 * @param int                       $scan_id Scan id.
	 * @param \stdClass                 $row     Post row.
	 * @param array<string, mixed>|null $profile Business details, null when not used.
	 * @param array<int, int>           $seen    Paragraph hash => id of the first post that has it; extended here.
	 */
	private function inspect( int $scan_id, \stdClass $row, ?array $profile, array &$seen ): void {
		$id    = (int) $row->ID;
		$title = (string) $row->post_title;
		$html  = (string) $row->post_content;

		if ( ContentRules::is_default_post( (string) $row->post_type, (string) $row->post_name, $title ) ) {
			$this->record( $scan_id, $id, 'default_content', Priority::MEDIUM, Priority::ERROR, __( 'This is the sample content WordPress adds to a new site', 'forensix-digital' ), array(), $title );
		}
		$placeholders = array_unique( array_merge( ContentRules::placeholders( $title ), ContentRules::placeholders( $html ) ) );
		if ( array() !== $placeholders ) {
			$this->record(
				$scan_id,
				$id,
				'placeholder_text',
				Priority::MEDIUM,
				Priority::ERROR,
				__( 'This page contains placeholder text', 'forensix-digital' ),
				array( 'kinds' => array_values( $placeholders ) ),
				implode( ', ', array_map( static fn( string $kind ): string => 'lorem' === $kind ? __( 'Lorem ipsum text', 'forensix-digital' ) : __( 'placeholder email address', 'forensix-digital' ), $placeholders ) )
			);
		}
		$empty = ContentRules::empty_elements( $html ) + ContentRules::empty_containers( $this->elementor_elements( $id ) );
		if ( $empty > 0 ) {
			$this->record(
				$scan_id,
				$id,
				'empty_element',
				Priority::LOW,
				Priority::SUGGESTION,
				__( 'This page has empty paragraphs, headings or Elementor sections', 'forensix-digital' ),
				array( 'count' => $empty ),
				/* translators: %d: number of empty elements. */
				sprintf( _n( '%d empty element', '%d empty elements', $empty, 'forensix-digital' ), $empty )
			);
		}
		if ( null !== $profile ) {
			$wrong = ContactConsistency::mismatches( $html, $profile, (string) wp_parse_url( home_url(), PHP_URL_HOST ) );
			$found = array_merge( $wrong['phones'], $wrong['emails'] );
			if ( array() !== $found ) {
				$this->record(
					$scan_id,
					$id,
					'contact_mismatch',
					Priority::MEDIUM,
					Priority::ERROR,
					__( 'This page uses a phone number or email address that is not in your business details', 'forensix-digital' ),
					$wrong,
					implode( ', ', $found )
				);
			}
		}
		$repeated = $this->repeated_paragraphs( $id, $html, $seen );
		if ( array() !== $repeated ) {
			$other = (int) reset( $repeated );
			$this->record(
				$scan_id,
				$id,
				'duplicate_text',
				Priority::LOW,
				Priority::SUGGESTION,
				__( 'Paragraphs on this page also appear on another page', 'forensix-digital' ),
				array(
					'count' => count( $repeated ),
					'other' => $other,
				),
				/* translators: 1: number of paragraphs, 2: title of the other page. */
				sprintf( _n( '%1$d paragraph also appears on "%2$s"', '%1$d paragraphs also appear on "%2$s"', count( $repeated ), 'forensix-digital' ), count( $repeated ), get_the_title( $other ) )
			);
		}
		$dates = ContentRules::past_event_dates( $html, time() );
		if ( array() !== $dates ) {
			$this->record(
				$scan_id,
				$id,
				'past_date',
				Priority::LOW,
				Priority::SUGGESTION,
				__( 'This page mentions a date that has passed', 'forensix-digital' ),
				array( 'dates' => $dates ),
				implode( ', ', $dates )
			);
		}
		$modified = (int) strtotime( (string) $row->post_modified_gmt . ' UTC' );
		if ( 'post' === $row->post_type && ContentRules::is_stale( $modified, time(), self::stale_months() ) ) {
			$this->record(
				$scan_id,
				$id,
				'stale',
				Priority::LOW,
				Priority::SUGGESTION,
				__( 'This post has not been updated for a long time', 'forensix-digital' ),
				array( 'modified' => $modified ),
				/* translators: %s: date of the last change. */
				sprintf( __( 'Last changed %s', 'forensix-digital' ), wp_date( get_option( 'date_format' ), $modified ) )
			);
		}
	}

	/**
	 * Paragraphs of a post that an earlier post already has. The first post to have a paragraph
	 * keeps it; later posts are the ones reported.
	 *
	 * @param int             $id   Post id.
	 * @param string          $html Post content.
	 * @param array<int, int> $seen Paragraph hash => id of the first post; extended here.
	 * @return array<int, int> Hash => id of the other post, for the repeated paragraphs.
	 */
	private function repeated_paragraphs( int $id, string $html, array &$seen ): array {
		$repeated = array();
		foreach ( ContentRules::paragraph_hashes( $html ) as $hash ) {
			if ( isset( $seen[ $hash ] ) && $seen[ $hash ] !== $id ) {
				$repeated[ $hash ] = $seen[ $hash ];
			} elseif ( ! isset( $seen[ $hash ] ) && count( $seen ) < self::DUPLICATE_CAP ) {
				$seen[ $hash ] = $id;
			}
		}
		return $repeated;
	}

	/**
	 * Records one finding.
	 *
	 * @param int                  $scan_id Scan id.
	 * @param int                  $id      Post id.
	 * @param string               $name    Check name after "content.".
	 * @param string               $level   Severity.
	 * @param string               $type    Finding type.
	 * @param string               $title   Plain sentence.
	 * @param array<string, mixed> $details Facts for the explanation.
	 * @param string               $context Short note shown with the finding.
	 */
	private function record( int $scan_id, int $id, string $name, string $level, string $type, string $title, array $details, string $context ): void {
		$source = new Source( Source::POST_CONTENT, (string) $id );
		$link   = get_permalink( $id );
		$this->store->record(
			$scan_id,
			$this->app,
			new Observation(
				'content.' . $name,
				$level,
				$type,
				$title,
				false !== $link ? $link : home_url( '/' ),
				$source->type,
				$source->ref,
				( new SourceLabels() )->label( $source ),
				$details
			),
			(string) get_edit_post_link( $id, 'raw' ),
			$context
		);
	}
}