<?php
if ( ! defined( 'ABSPATH' ) ) { exit; }

class OWWCQC_Fixer {
	private $safe_rules = array( 'nested_strong', 'nested_em', 'missing_alt', 'unsafe_blank_link' );
	private $reviewed_rules = array( 'empty_paragraph', 'heading_skip', 'multiple_h1', 'duplicate_ids' );

	public function safe_rules() { return $this->safe_rules; }
	public function can_fix_one( $rule ) { return in_array( $rule, array_merge( $this->safe_rules, $this->reviewed_rules ), true ); }

	public function fixable_issues( $issues ) {
		return array_intersect_key( (array) $issues, array_flip( $this->safe_rules ) );
	}

	public function fixable_total( $issues ) {
		return array_sum( array_map( 'intval', $this->fixable_issues( $issues ) ) );
	}

	public function fix_one( $content, $rule, $occurrence ) {
		$content = (string) $content;
		$occurrence = max( 0, (int) $occurrence );
		if ( ! $this->can_fix_one( $rule ) ) { return array( 'content' => $content, 'changed' => false ); }

		if ( 'empty_paragraph' === $rule ) {
			$pattern = '/(?:<!--\s+wp:paragraph(?:\s+\{.*?\})?\s+-->\s*)?<p\b[^>]*>\s*(?:&nbsp;|&#160;|\s|<br\s*\/?>)*<\/p>\s*(?:<!--\s+\/wp:paragraph\s+-->)?/is';
			return $this->replace_occurrence( $content, $pattern, $occurrence, function() { return ''; } );
		}

		if ( 'duplicate_ids' === $rule ) { return $this->rename_duplicate_id( $content, $occurrence ); }
		if ( 'heading_skip' === $rule || 'multiple_h1' === $rule ) { return $this->repair_heading( $content, $rule, $occurrence ); }

		if ( 'nested_strong' === $rule || 'nested_em' === $rule ) {
			$tag = 'nested_strong' === $rule ? 'strong' : 'em';
			$pattern = '/<' . $tag . '\\b([^>]*)>\\s*<' . $tag . '\\b[^>]*>(.*?)<\\/' . $tag . '>\\s*<\\/' . $tag . '>/is';
			return $this->replace_occurrence( $content, $pattern, $occurrence, function( $match ) use ( $tag ) { return '<' . $tag . $match[1] . '>' . $match[2] . '</' . $tag . '>'; } );
		}

		if ( 'missing_alt' === $rule ) {
			return $this->replace_occurrence( $content, '/<img\\b(?![^>]*\\salt=["\'])[^>]*>/i', $occurrence, function( $match ) { return preg_replace( '/\\s*\\/?\\s*>$/', ' alt=""$0', $match[0], 1 ); } );
		}

		$current = 0; $changed = false;
		$updated = preg_replace_callback( '/<a\\b[^>]*target=["\']_blank["\'][^>]*>/i', function( $match ) use ( &$current, &$changed, $occurrence ) {
			$tag = $match[0];
			if ( preg_match( '/\\brel=["\'][^"\']*\\b(?:noopener|noreferrer)\\b[^"\']*["\']/i', $tag ) ) { return $tag; }
			if ( $current++ !== $occurrence ) { return $tag; }
			$changed = true;
			if ( preg_match( '/\\brel=(["\'])(.*?)\\1/i', $tag, $rel ) ) {
				$tokens = array_filter( preg_split( '/\\s+/', strtolower( trim( $rel[2] ) ) ) );
				foreach ( array( 'noopener', 'noreferrer' ) as $needed ) { if ( ! in_array( $needed, $tokens, true ) ) { $tokens[] = $needed; } }
				return preg_replace( '/\\brel=(["\'])(.*?)\\1/i', 'rel="' . implode( ' ', $tokens ) . '"', $tag, 1 );
			}
			return preg_replace( '/>$/', ' rel="noopener noreferrer">', $tag, 1 );
		}, $content );
		return array( 'content' => $updated, 'changed' => $changed );
	}

	private function rename_duplicate_id( $content, $wanted ) {
		preg_match_all( '/\sid=(["\'])([^"\']+)\1/i', $content, $matches, PREG_OFFSET_CAPTURE );
		$counts = array_count_values( array_map( function( $item ) { return $item[0]; }, $matches[2] ) );
		$seen = array(); $candidate = 0; $changed = false;
		$updated = preg_replace_callback( '/\sid=(["\'])([^"\']+)\1/i', function( $match ) use ( &$seen, &$candidate, &$changed, $wanted, $counts, $content ) {
			$id = $match[2];
			$seen[ $id ] = isset( $seen[ $id ] ) ? $seen[ $id ] + 1 : 1;
			if ( empty( $counts[ $id ] ) || $counts[ $id ] < 2 || 1 === $seen[ $id ] ) { return $match[0]; }
			if ( $candidate++ !== $wanted ) { return $match[0]; }
			if ( preg_match( '/(?:url\(\s*#|xlink:href=["\']#)' . preg_quote( $id, '/' ) . '/i', $content ) ) { return $match[0]; }
			$suffix = $seen[ $id ];
			do { $new_id = $id . '-' . $suffix++; } while ( preg_match( '/\sid=["\']' . preg_quote( $new_id, '/' ) . '["\']/i', $content ) );
			$changed = true;
			return ' id=' . $match[1] . $new_id . $match[1];
		}, $content );
		return array( 'content' => $updated, 'changed' => $changed );
	}

	private function repair_heading( $content, $rule, $wanted ) {
		$previous = null; $candidate = 0; $h1_seen = 0; $changed = false;
		$updated = preg_replace_callback( '/<h([1-6])\b([^>]*)>(.*?)<\/h\1>/is', function( $match ) use ( &$previous, &$candidate, &$h1_seen, &$changed, $rule, $wanted ) {
			$level = (int) $match[1]; $target = 0;
			if ( 'heading_skip' === $rule && null !== $previous && $level > $previous + 1 ) { $target = $previous + 1; }
			if ( 'multiple_h1' === $rule && 1 === $level && ++$h1_seen > 1 ) { $target = 2; }
			if ( 'multiple_h1' === $rule && 1 !== $level ) { /* Keep the H1 counter unchanged. */ }
			$previous = $level;
			if ( ! $target || $candidate++ !== $wanted ) { return $match[0]; }
			$changed = true;
			return '<h' . $target . $match[2] . '>' . $match[3] . '</h' . $target . '>';
		}, $content );
		return array( 'content' => $updated, 'changed' => $changed );
	}

	private function replace_occurrence( $content, $pattern, $wanted, $callback ) {
		$current = 0; $changed = false;
		$updated = preg_replace_callback( $pattern, function( $match ) use ( &$current, &$changed, $wanted, $callback ) {
			if ( $current++ !== $wanted ) { return $match[0]; }
			$replacement = call_user_func( $callback, $match );
			$changed = $replacement !== $match[0];
			return $replacement;
		}, $content );
		return array( 'content' => $updated, 'changed' => $changed );
	}

	/**
	 * Applies only attribute/inline-wrapper fixes. WordPress block comments,
	 * element order, paragraph blocks, headings, IDs and anchors are untouched.
	 */
	public function fix( $content ) {
		$content = (string) $content;
		$changes = array();

		$content = $this->collapse_nested_tag( $content, 'strong', 'nested_strong', $changes );
		$content = $this->collapse_nested_tag( $content, 'em', 'nested_em', $changes );

		$content = preg_replace_callback( '/<img\b[^>]*>/i', function ( $match ) use ( &$changes ) {
			if ( preg_match( '/\salt\s*=/i', $match[0] ) ) { return $match[0]; }
			$changes['missing_alt'] = isset( $changes['missing_alt'] ) ? $changes['missing_alt'] + 1 : 1;
			return preg_replace( '/\s*\/?\s*>$/', ' alt=""$0', $match[0], 1 );
		}, $content );

		$content = preg_replace_callback( '/<a\b[^>]*target=["\']_blank["\'][^>]*>/i', function ( $match ) use ( &$changes ) {
			$tag = $match[0];
			if ( preg_match( '/\brel=(["\'])(.*?)\1/i', $tag, $rel ) ) {
				$tokens = preg_split( '/\s+/', trim( $rel[2] ) );
				$tokens = array_filter( array_map( 'strtolower', $tokens ) );
				$updated = false;
				foreach ( array( 'noopener', 'noreferrer' ) as $needed ) {
					if ( ! in_array( $needed, $tokens, true ) ) { $tokens[] = $needed; $updated = true; }
				}
				if ( ! $updated ) { return $tag; }
				$tag = preg_replace( '/\brel=(["\'])(.*?)\1/i', 'rel="' . implode( ' ', $tokens ) . '"', $tag, 1 );
			} else {
				$tag = preg_replace( '/>$/', ' rel="noopener noreferrer">', $tag, 1 );
			}
			$changes['unsafe_blank_link'] = isset( $changes['unsafe_blank_link'] ) ? $changes['unsafe_blank_link'] + 1 : 1;
			return $tag;
		}, $content );

		return array( 'content' => $content, 'changes' => $changes );
	}

	private function collapse_nested_tag( $content, $tag, $rule, &$changes ) {
		$pattern = '/<' . $tag . '\b([^>]*)>\s*<' . $tag . '\b[^>]*>(.*?)<\/' . $tag . '>\s*<\/' . $tag . '>/is';
		do {
			$content = preg_replace_callback( $pattern, function ( $match ) use ( $tag, $rule, &$changes ) {
				$changes[ $rule ] = isset( $changes[ $rule ] ) ? $changes[ $rule ] + 1 : 1;
				return '<' . $tag . $match[1] . '>' . $match[2] . '</' . $tag . '>';
			}, $content, -1, $count );
		} while ( $count > 0 );
		return $content;
	}
}