<?php
/**
 * Content rules.
 *
 * @package A11yMate
 */

namespace A11yMate\Rules\Server;

use A11yMate\Support\Html;

defined( 'ABSPATH' ) || exit;

/**
 * Reading level, long paragraphs, PDFs.
 */
class Content {

	/**
	 * Run the checks.
	 *
	 * @param Context $c Context.
	 * @return void
	 */
	public function run( Context $c ) {
		$main = $c->q( "//main | //*[@role='main']" );
		$root = $main ? $main[0] : null;
		if ( ! $root ) {
			$body = $c->q( '//body' );
			$root = $body ? $body[0] : null;
		}
		if ( ! $root ) {
			return;
		}

		$text_parts = array();
		foreach ( $c->q( './/p', $root ) as $p ) {
			if ( $c->q( "ancestor::nav | ancestor::header | ancestor::footer | ancestor::*[@role='navigation']", $p ) ) {
				continue;
			}
			$t     = Html::clean( Html::text_of( $p ) );
			$words = str_word_count( $t );
			if ( $words > 150 ) {
				$c->add( 'paragraph-long', $p, array( 'words' => $words ) );
			}
			if ( $words >= 5 ) {
				$text_parts[] = $t;
			}
		}

		$text = implode( ' ', $text_parts );
		if ( $c->on( 'reading-level' ) && str_word_count( $text ) >= 150 ) {
			$score = self::flesch( $text );
			if ( $score < 40 ) {
				$c->add( 'reading-level', $root, array( 'flesch' => round( $score ) ) );
			}
		}

		if ( $c->on( 'pdf-heavy' ) ) {
			$pdfs = $c->q( ".//a[contains(translate(@href,'PDF','pdf'),'.pdf')]", $root );
			if ( count( $pdfs ) >= 5 ) {
				$c->add( 'pdf-heavy', $root, array( 'count' => count( $pdfs ) ) );
			}
		}
	}

	/**
	 * Flesch reading ease (English).
	 *
	 * @param string $text Text.
	 * @return float
	 */
	public static function flesch( $text ) {
		$sentences = max( 1, preg_match_all( '/[.!?]+(\s|$)/', $text ) );
		preg_match_all( "/[A-Za-z']+/", $text, $m );
		$words = max( 1, count( $m[0] ) );
		$syl   = 0;
		foreach ( $m[0] as $w ) {
			$syl += self::syllables( $w );
		}
		return 206.835 - 1.015 * ( $words / $sentences ) - 84.6 * ( $syl / $words );
	}

	/**
	 * Rough syllable count.
	 *
	 * @param string $word Word.
	 * @return int
	 */
	public static function syllables( $word ) {
		$word = strtolower( preg_replace( '/[^a-z]/i', '', $word ) );
		if ( strlen( $word ) <= 3 ) {
			return 1;
		}
		$word = preg_replace( '/(?:[^laeiouy]es|ed|[^laeiouy]e)$/', '', $word );
		$word = preg_replace( '/^y/', '', $word );
		$n    = preg_match_all( '/[aeiouy]{1,2}/', $word );
		return max( 1, (int) $n );
	}
}