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

class OWWCQC_Scanner {
	private $rules;
	public function __construct( OWWCQC_Rules $rules ) { $this->rules = $rules; }
	public function rules() { return $this->rules; }
	public function total( $issues ) { return array_sum( array_map( 'intval', (array) $issues ) ); }
	public function scan_post( $post_id ) { return $this->scan( get_post_field( 'post_content', $post_id ) ); }
	public function details_post( $post_id ) { return $this->details( get_post_field( 'post_content', $post_id ) ); }

	public function scan( $content ) {
		$content = (string) $content;
		$issues = array();
		$this->match_count( $issues, 'nested_strong', '/<strong\b[^>]*>\s*<strong\b[^>]*>/i', $content );
		$this->match_count( $issues, 'nested_em', '/<em\b[^>]*>\s*<em\b[^>]*>/i', $content );
		$this->match_count( $issues, 'empty_paragraph', '/<p\b[^>]*>\s*(?:&nbsp;|&#160;|\s|<br\s*\/?>)*<\/p>/i', $content );

		preg_match_all( '/<p\b[^>]*>(.*?)<\/p>/is', $content, $paragraphs );
		$previous = null; $count = 0;
		foreach ( $paragraphs[1] as $paragraph ) {
			$current = strtolower( trim( preg_replace( '/\s+/', ' ', wp_strip_all_tags( html_entity_decode( $paragraph, ENT_QUOTES, 'UTF-8' ) ) ) ) );
			if ( '' !== $current && null !== $previous && $current === $previous ) { $count++; }
			$previous = $current;
		}
		$this->set( $issues, 'duplicate_paragraph', $count );

		preg_match_all( '/<h1\b[^>]*>.*?<\/h1>/is', $content, $h1s );
		if ( count( $h1s[0] ) > 1 ) { $issues['multiple_h1'] = count( $h1s[0] ) - 1; }

		preg_match_all( '/<h([1-6])\b[^>]*>.*?<\/h\1>/is', $content, $headings );
		$previous = null; $count = 0;
		foreach ( $headings[1] as $level ) {
			$level = (int) $level;
			if ( null !== $previous && $level > $previous + 1 ) { $count++; }
			$previous = $level;
		}
		$this->set( $issues, 'heading_skip', $count );

		preg_match_all( '/\sid=["\']([^"\']+)["\']/i', $content, $ids );
		$id_values = isset( $ids[1] ) ? $ids[1] : array(); $count = 0;
		foreach ( array_count_values( $id_values ) as $occurrences ) { if ( $occurrences > 1 ) { $count += $occurrences - 1; } }
		$this->set( $issues, 'duplicate_ids', $count );

		preg_match_all( '/href=["\']#([^"\']+)["\']/i', $content, $anchors ); $count = 0;
		foreach ( $anchors[1] as $anchor ) { if ( ! in_array( $anchor, $id_values, true ) ) { $count++; } }
		$this->set( $issues, 'broken_anchor', $count );

		preg_match_all( '/<img\b[^>]*>/i', $content, $images ); $count = 0;
		foreach ( $images[0] as $tag ) { if ( ! preg_match( '/\salt=["\'][^"\']*["\']/i', $tag ) ) { $count++; } }
		$this->set( $issues, 'missing_alt', $count );

		preg_match_all( '/<a\b[^>]*target=["\']_blank["\'][^>]*>/i', $content, $links ); $count = 0;
		foreach ( $links[0] as $tag ) {
			if ( ! preg_match( '/\brel=["\'][^"\']*\b(?:noopener|noreferrer)\b[^"\']*["\']/i', $tag ) ) { $count++; }
		}
		$this->set( $issues, 'unsafe_blank_link', $count );
		return apply_filters( 'owwcqc_scan_results', $issues, $content );
	}

	/**
	 * Return human-readable locations without changing the saved content.
	 * Each item also contains a small needle used by admin.js to locate a
	 * matching Gutenberg block. The needle is never written to the post.
	 */
	public function details( $content ) {
		$content = (string) $content;
		$details = array();
		$this->regex_details( $details, 'nested_strong', '/<strong\b[^>]*>\s*<strong\b[^>]*>/i', $content, 'Nested bold markup' );
		$this->regex_details( $details, 'nested_em', '/<em\b[^>]*>\s*<em\b[^>]*>/i', $content, 'Nested italic markup' );
		$this->empty_paragraph_details( $details, $content );
		$this->regex_details( $details, 'missing_alt', '/<img\b(?![^>]*\salt=["\'])[^>]*>/i', $content, 'Image without an alt attribute' );

		preg_match_all( '/<p\b[^>]*>(.*?)<\/p>/is', $content, $paragraphs, PREG_OFFSET_CAPTURE );
		$previous = null;
		foreach ( isset( $paragraphs[1] ) ? $paragraphs[1] : array() as $index => $paragraph ) {
			$text = $this->plain( $paragraph[0] );
			$normal = strtolower( $text );
			if ( '' !== $normal && null !== $previous && $normal === $previous ) {
				$selection = $this->expand_paragraph_block( $content, $paragraphs[0][ $index ][1], strlen( $paragraphs[0][ $index ][0] ) );
				$this->add_detail(
					$details,
					'duplicate_paragraph',
					'Repeated paragraph: “' . $this->shorten( $text ) . '”',
					$selection['text'],
					$content,
					$selection['offset'],
					array( 'editor_action' => 'remove' )
				);
			}
			$previous = $normal;
		}

		preg_match_all( '/<h([1-6])\b[^>]*>(.*?)<\/h\1>/is', $content, $headings, PREG_OFFSET_CAPTURE );
		$previous_level = null;
		foreach ( isset( $headings[1] ) ? $headings[1] : array() as $index => $level_match ) {
			$level = (int) $level_match[0];
			$text = $this->plain( $headings[2][ $index ][0] );
			$offset = $headings[0][ $index ][1];
			if ( null !== $previous_level && $level > $previous_level + 1 ) {
				$this->add_detail(
					$details,
					'heading_skip',
					'Heading jumps from H' . $previous_level . ' to H' . $level . ': “' . $this->shorten( $text ) . '”',
					$headings[0][ $index ][0],
					$content,
					$offset,
					array(
						'editor_action' => 'heading',
						'current_level' => $level,
						'target_level'  => $previous_level + 1,
					)
				);
			}
			$previous_level = $level;
		}

		if ( ! empty( $headings[1] ) ) {
			$h1_seen = 0;
			foreach ( $headings[1] as $index => $level_match ) {
				if ( 1 === (int) $level_match[0] ) {
					$h1_seen++;
					if ( $h1_seen > 1 ) {
						$text = $this->plain( $headings[2][ $index ][0] );
						$this->add_detail(
							$details,
							'multiple_h1',
							'Additional H1: “' . $this->shorten( $text ) . '”',
							$headings[0][ $index ][0],
							$content,
							$headings[0][ $index ][1],
							array(
								'editor_action' => 'heading',
								'current_level' => 1,
								'target_level'  => 2,
							)
						);
					}
				}
			}
		}

		preg_match_all( '/\sid=["\']([^"\']+)["\']/i', $content, $ids, PREG_OFFSET_CAPTURE );
		$id_counts = array_count_values( array_map( function( $item ) { return $item[0]; }, isset( $ids[1] ) ? $ids[1] : array() ) );
		$id_seen = array();
		foreach ( isset( $ids[1] ) ? $ids[1] : array() as $index => $id ) {
			$id_seen[ $id[0] ] = isset( $id_seen[ $id[0] ] ) ? $id_seen[ $id[0] ] + 1 : 1;
			if ( isset( $id_counts[ $id[0] ] ) && $id_counts[ $id[0] ] > 1 && $id_seen[ $id[0] ] > 1 ) {
				$attribute = $ids[0][ $index ];
				$this->add_detail( $details, 'duplicate_ids', 'Duplicate ID: #' . $id[0] . '. Keep the first occurrence; rename this one.', trim( $attribute[0] ), $content, $attribute[1], array( 'editor_action' => 'rename_id' ) );
			}
		}

		preg_match_all( '/href=["\']#([^"\']+)["\']/i', $content, $anchors, PREG_OFFSET_CAPTURE );
		$id_values = array_map( function( $item ) { return $item[0]; }, isset( $ids[1] ) ? $ids[1] : array() );
		foreach ( isset( $anchors[1] ) ? $anchors[1] : array() as $anchor ) {
			if ( ! in_array( $anchor[0], $id_values, true ) ) {
				$this->add_detail( $details, 'broken_anchor', 'Link points to missing #' . $anchor[0], 'href="#' . $anchor[0] . '"', $content, $anchor[1] );
			}
		}

		preg_match_all( '/<a\b[^>]*target=["\']_blank["\'][^>]*>/i', $content, $links, PREG_OFFSET_CAPTURE );
		foreach ( isset( $links[0] ) ? $links[0] : array() as $link ) {
			if ( ! preg_match( '/\brel=["\'][^"\']*\b(?:noopener|noreferrer)\b[^"\']*["\']/i', $link[0] ) ) {
				$this->add_detail( $details, 'unsafe_blank_link', 'New-tab link needs rel="noopener noreferrer"', $link[0], $content, $link[1] );
			}
		}
		return apply_filters( 'owwcqc_issue_details', $details, $content );
	}

	private function regex_details( &$details, $key, $pattern, $content, $label ) {
		preg_match_all( $pattern, $content, $matches, PREG_OFFSET_CAPTURE );
		foreach ( isset( $matches[0] ) ? $matches[0] : array() as $match ) {
			$this->add_detail( $details, $key, $label . ': ' . $this->html_clue( $content, $match[1] ), $match[0], $content, $match[1] );
		}
	}

	private function add_detail( &$details, $key, $message, $needle, $content, $offset, $extra = array() ) {
		if ( ! isset( $details[ $key ] ) ) { $details[ $key ] = array(); }
		$item = array(
			'message' => $message,
			'needle'  => (string) $needle,
			'line'    => substr_count( substr( $content, 0, max( 0, (int) $offset ) ), "\n" ) + 1,
			'offset'  => max( 0, (int) $offset ),
			'length'  => strlen( (string) $needle ),
		);
		$details[ $key ][] = array_merge( $item, (array) $extra );
	}

	private function empty_paragraph_details( &$details, $content ) {
		preg_match_all( '/<p\b[^>]*>\s*(?:&nbsp;|&#160;|\s|<br\s*\/?>)*<\/p>/i', $content, $matches, PREG_OFFSET_CAPTURE );
		foreach ( isset( $matches[0] ) ? $matches[0] : array() as $match ) {
			$selection = $this->expand_paragraph_block( $content, $match[1], strlen( $match[0] ) );
			$this->add_detail(
				$details,
				'empty_paragraph',
				'Empty paragraph block: ' . $this->html_clue( $content, $match[1] ),
				$selection['text'],
				$content,
				$selection['offset'],
				array( 'editor_action' => 'remove' )
			);
		}
	}

	/**
	 * Include a Gutenberg paragraph wrapper only when the matched paragraph is
	 * the wrapper's complete content. This avoids removing a parent block.
	 */
	private function expand_paragraph_block( $content, $offset, $length ) {
		$selection = array(
			'text'   => substr( $content, $offset, $length ),
			'offset' => (int) $offset,
		);
		$before = substr( $content, 0, $offset );
		$open = strrpos( $before, '<!-- wp:paragraph' );
		if ( false === $open ) { return $selection; }
		$open_end = strpos( $content, '-->', $open );
		if ( false === $open_end || '' !== trim( substr( $content, $open_end + 3, $offset - ( $open_end + 3 ) ) ) ) { return $selection; }
		$match_end = $offset + $length;
		$close = strpos( $content, '<!-- /wp:paragraph -->', $match_end );
		if ( false === $close || '' !== trim( substr( $content, $match_end, $close - $match_end ) ) ) { return $selection; }
		$close_end = $close + strlen( '<!-- /wp:paragraph -->' );
		$selection['text'] = substr( $content, $open, $close_end - $open );
		$selection['offset'] = $open;
		return $selection;
	}

	private function plain( $html ) { return trim( preg_replace( '/\s+/', ' ', wp_strip_all_tags( html_entity_decode( (string) $html, ENT_QUOTES, 'UTF-8' ) ) ) ); }
	private function shorten( $text, $limit = 100 ) { $text = trim( preg_replace( '/\s+/', ' ', (string) $text ) ); return strlen( $text ) > $limit ? substr( $text, 0, $limit - 1 ) . '…' : $text; }
	private function html_clue( $content, $offset ) {
		$start = max( 0, (int) $offset - 45 );
		return 'HTML near line ' . ( substr_count( substr( $content, 0, $offset ), "\n" ) + 1 ) . ' — ' . $this->shorten( substr( $content, $start, 130 ), 120 );
	}

	private function match_count( &$issues, $key, $pattern, $content ) {
		preg_match_all( $pattern, $content, $matches ); $this->set( $issues, $key, count( $matches[0] ) );
	}
	private function set( &$issues, $key, $count ) { if ( $count > 0 ) { $issues[ $key ] = (int) $count; } }
}