<?php
/**
 * Text rules of the content checks (F-CONTENT-01, 02, 03).
 *
 * @package ForensixDigital
 */

declare(strict_types=1);

namespace ForensixDigital\Apps\Content;

defined( 'ABSPATH' ) || exit;

/**
 * Pure functions: they look at text and decide, without touching the database.
 */
final class ContentRules {

	public const DEFAULT_TAGLINE = 'Just another WordPress site';
	public const STALE_MONTHS    = 24;
	public const DUPLICATE_MIN   = 120;

	private const PLACEHOLDER_MAIL = '/[\w.+-]+@(?:example\.(?:com|org|net)|domain\.com|yourdomain\.com|yoursite\.com|yourwebsite\.com|email\.com)\b/i';

	/**
	 * Placeholder text found in a piece of content.
	 *
	 * @param string $text Post title or content.
	 * @return array<int, string> "lorem" and/or "email".
	 */
	public static function placeholders( string $text ): array {
		$found = array();
		if ( 1 === preg_match( '/\blorem\s+ipsum\b|\bdolor\s+sit\s+amet\b/i', $text ) ) {
			$found[] = 'lorem';
		}
		if ( 1 === preg_match( self::PLACEHOLDER_MAIL, $text ) ) {
			$found[] = 'email';
		}
		return $found;
	}

	/**
	 * Whether a post is the sample content WordPress creates on installation.
	 *
	 * @param string $type  Post type.
	 * @param string $slug  Post slug.
	 * @param string $title Post title.
	 */
	public static function is_default_post( string $type, string $slug, string $title ): bool {
		return ( 'post' === $type && 'hello-world' === $slug && 'Hello world!' === $title )
			|| ( 'page' === $type && 'sample-page' === $slug && 'Sample Page' === $title );
	}

	/**
	 * Number of empty paragraphs and headings in a piece of content.
	 *
	 * @param string $html Post content.
	 */
	public static function empty_elements( string $html ): int {
		$blank = '(?:\s|&nbsp;|\xC2\xA0|<br\s*\/?>)*';
		$count = (int) preg_match_all( '/<p[^>]*>' . $blank . '<\/p>/i', $html );
		return $count + (int) preg_match_all( '/<h([1-6])[^>]*>' . $blank . '<\/h\1>/i', $html );
	}

	/**
	 * Elementor element types that only hold other elements: classic sections, columns and
	 * containers, and the atomic (V4) div block, flexbox and grid. Any other element type, such
	 * as atomic tabs or a form, counts as content.
	 *
	 * @var string[]
	 */
	public const ELEMENTOR_CONTAINERS = array( 'section', 'column', 'container', 'e-div-block', 'e-flexbox', 'e-grid' );

	/**
	 * Elementor sections, columns and containers that hold no widget at all. An empty one is
	 * counted once, without the empty elements inside it.
	 *
	 * @param array<int|string, mixed> $elements Decoded `_elementor_data` (a list of elements).
	 */
	public static function empty_containers( array $elements ): int {
		$count = 0;
		foreach ( $elements as $element ) {
			if ( ! is_array( $element ) || self::is_content( $element ) ) {
				continue;
			}
			if ( ! self::has_widget( $element ) ) {
				++$count;
				continue;
			}
			$count += self::empty_containers( is_array( $element['elements'] ?? null ) ? $element['elements'] : array() );
		}
		return $count;
	}

	/**
	 * Whether an Elementor element is a widget or holds one.
	 *
	 * @param array<int|string, mixed> $element Element.
	 */
	private static function has_widget( array $element ): bool {
		if ( self::is_content( $element ) ) {
			return true;
		}
		foreach ( is_array( $element['elements'] ?? null ) ? $element['elements'] : array() as $child ) {
			if ( is_array( $child ) && self::has_widget( $child ) ) {
				return true;
			}
		}
		return false;
	}

	/**
	 * Whether an Elementor element is content itself: a widget, or an element type that is not
	 * a plain container.
	 *
	 * @param array<int|string, mixed> $element Element.
	 */
	public static function is_content( array $element ): bool {
		return ! empty( $element['widgetType'] ) || ! in_array( $element['elType'] ?? '', self::ELEMENTOR_CONTAINERS, true );
	}

	/**
	 * Whether a post was last changed longer ago than the given number of months.
	 *
	 * @param int $modified Unix time of the last change.
	 * @param int $now      Unix time now.
	 * @param int $months   Months.
	 */
	public static function is_stale( int $modified, int $now, int $months ): bool {
		return $modified > 0 && $now - $modified > $months * 30 * 86400;
	}

	/**
	 * Dates in the past that sit in a sentence about something still to come ("join us on",
	 * "sale ends", "register by"). Only full dates are read: 2024-05-12, 12 May 2024 and
	 * May 12, 2024. Plain statements such as "founded on 3 June 2001" are not touched.
	 *
	 * @param string $text Post content.
	 * @param int    $now  Unix time now.
	 * @return array<int, string> The dates as written, at most five.
	 */
	public static function past_event_dates( string $text, int $now ): array {
		$text  = html_entity_decode( wp_strip_all_tags( $text ), ENT_QUOTES, 'UTF-8' );
		$found = array();
		$month = '(?:January|February|March|April|May|June|July|August|September|October|November|December|Jan|Feb|Mar|Apr|Jun|Jul|Aug|Sep|Sept|Oct|Nov|Dec)\.?';
		$date  = '(\d{4}-\d{2}-\d{2}|\d{1,2}(?:st|nd|rd|th)?\s+' . $month . ',?\s+\d{4}|' . $month . '\s+\d{1,2}(?:st|nd|rd|th)?,?\s+\d{4})';
		$cue   = '/\b(?:join us|takes? place|will be held|register(?:ation)?\s+(?:by|until|before)|sign up\s+(?:by|before)|sale ends|offer (?:ends|valid)|valid until|until|deadline|early bird|doors open|starts? on|ends? on|coming soon|save the date|event|tickets?|book (?:now|by|before)|expires?|closes?)\b/i';
		foreach ( (array) preg_split( '/(?<=[.!?])\s+|\n+/', $text ) as $sentence ) {
			if ( 1 !== preg_match( $cue, (string) $sentence ) || 1 !== preg_match_all( '/' . $date . '/i', (string) $sentence, $m ) ) {
				continue;
			}
			foreach ( $m[1] as $written ) {
				$time = strtotime( preg_replace( '/(\d)(?:st|nd|rd|th)/i', '$1', str_replace( 'Sept', 'Sep', $written ) ) . ' 23:59:59 UTC' );
				if ( false !== $time && $time < $now - DAY_IN_SECONDS ) {
					$found[ $written ] = true;
				}
			}
		}
		return array_slice( array_keys( $found ), 0, 5 );
	}

	/**
	 * Hashes of the longer paragraphs of a piece of content, for finding text that appears on
	 * several pages. Short paragraphs are skipped, because short sentences repeat by chance.
	 *
	 * @param string $html Post content.
	 * @return array<int, int> Hash values without repeats.
	 */
	public static function paragraph_hashes( string $html ): array {
		$out = array();
		if ( false === preg_match_all( '/<p[^>]*>(.*?)<\/p>/is', $html, $m ) ) {
			return $out;
		}
		foreach ( $m[1] as $paragraph ) {
			$text = strtolower( trim( (string) preg_replace( '/\s+/', ' ', html_entity_decode( wp_strip_all_tags( $paragraph ), ENT_QUOTES, 'UTF-8' ) ) ) );
			if ( mb_strlen( $text ) >= self::DUPLICATE_MIN ) {
				$out[ crc32( $text ) ] = crc32( $text );
			}
		}
		return array_values( $out );
	}
}