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

/**
 * Finds images across the site (Media Library attachments and images
 * embedded directly in post/page content) that are missing alt text.
 */
class AIATG_Scanner {

	/**
	 * Runs a full scan and returns a flat, JSON-serializable list of items.
	 * Each item has a unique_key so the admin UI/AJAX layer can address it.
	 */
	public static function scan() {
		$items = array();

		foreach ( self::scan_media_library() as $item ) {
			$items[] = $item;
		}

		foreach ( self::scan_content_images() as $item ) {
			$items[] = $item;
		}

		return $items;
	}

	/**
	 * Media Library images whose _wp_attachment_image_alt meta is missing or empty.
	 */
	public static function scan_media_library() {
		$items = array();

		$query = new WP_Query(
			array(
				'post_type'      => 'attachment',
				'post_status'    => 'inherit',
				'post_mime_type' => array( 'image/jpeg', 'image/png', 'image/gif', 'image/webp' ),
				'posts_per_page' => -1,
				'fields'         => 'ids',
				'meta_query'     => array(
					'relation' => 'OR',
					array(
						'key'     => '_wp_attachment_image_alt',
						'compare' => 'NOT EXISTS',
					),
					array(
						'key'     => '_wp_attachment_image_alt',
						'value'   => '',
						'compare' => '=',
					),
				),
			)
		);

		foreach ( $query->posts as $attachment_id ) {
			$thumb = wp_get_attachment_image_src( $attachment_id, 'thumbnail' );
			$full  = wp_get_attachment_image_src( $attachment_id, 'full' );
			$title = get_the_title( $attachment_id );
			$parent_id = wp_get_post_parent_id( $attachment_id );
			$location  = __( 'Media Library', 'fix-experts-ai-image-alt-tag-generator' );

			if ( $parent_id && 'product' === get_post_type( $parent_id ) ) {
				$location = __( 'WooCommerce Product', 'fix-experts-ai-image-alt-tag-generator' );
			}

			$items[] = array(
				'unique_key'   => 'attachment-' . $attachment_id,
				'type'         => 'attachment',
				'id'           => $attachment_id,
				'post_id'      => 0,
				'index'        => 0,
				'src'          => $full ? $full[0] : '',
				'thumb'        => $thumb ? $thumb[0] : ( $full ? $full[0] : '' ),
				'title'        => $title ? $title : __( '(untitled)', 'fix-experts-ai-image-alt-tag-generator' ),
				'location'     => $location,
				'edit_link'    => get_edit_post_link( $parent_id ? $parent_id : $attachment_id, 'raw' ),
				'current_alt'  => '',
			);
		}

		return $items;
	}

	/**
	 * Images embedded directly inside published post/page content that lack
	 * a non-empty alt attribute (independent of any Media Library alt meta,
	 * since WordPress "bakes" alt text into content at insertion time).
	 */
	public static function scan_content_images() {
		$items = array();
		$limit = (int) AIATG_Settings::get( 'content_scan_limit' );

		$query = new WP_Query(
			array(
				'post_type'      => array( 'post', 'page' ),
				'post_status'    => 'publish',
				'posts_per_page' => $limit > 0 ? $limit : 200,
				'fields'         => 'ids',
				's'              => '',
			)
		);

		foreach ( $query->posts as $post_id ) {
			$content = get_post_field( 'post_content', $post_id );

			if ( false === stripos( $content, '<img' ) ) {
				continue;
			}

			$images = self::extract_missing_alt_images( $content );

			if ( empty( $images ) ) {
				continue;
			}

			$title = get_the_title( $post_id );

			foreach ( $images as $index => $src ) {
				$items[] = array(
					'unique_key'  => 'content-' . $post_id . '-' . $index,
					'type'        => 'content',
					'id'          => 0,
					'post_id'     => $post_id,
					'index'       => $index,
					'src'         => $src,
					'thumb'       => $src,
					'title'       => $title ? $title : __( '(untitled)', 'fix-experts-ai-image-alt-tag-generator' ),
					'location'    => get_post_type( $post_id ) === 'page' ? __( 'Page content', 'fix-experts-ai-image-alt-tag-generator' ) : __( 'Post content', 'fix-experts-ai-image-alt-tag-generator' ),
					'edit_link'   => get_edit_post_link( $post_id, 'raw' ),
					'current_alt' => '',
				);
			}
		}

		return $items;
	}

	/**
	 * Parses HTML content and returns a 0-indexed map of img-tag-index => src
	 * for every <img> tag whose alt attribute is missing or empty.
	 *
	 * The index refers to the Nth <img> tag found in document order and is
	 * the same numbering used by save_content_alt() to write the value back.
	 */
	public static function extract_missing_alt_images( $content ) {
		$missing = array();

		$dom = self::load_dom( $content );
		if ( ! $dom ) {
			return $missing;
		}

		$imgs = $dom->getElementsByTagName( 'img' );

		foreach ( $imgs as $index => $img ) {
			$alt = $img->getAttribute( 'alt' );
			$src = $img->getAttribute( 'src' );

			if ( '' === trim( $alt ) && ! empty( $src ) ) {
				$missing[ $index ] = $src;
			}
		}

		return $missing;
	}

	/**
	 * Writes $alt_text into the Nth <img> tag of $content and returns the
	 * updated HTML. Returns WP_Error if the index can't be found, or if
	 * expected_src is given and no longer matches (content changed since
	 * the scan, so the index could now point at a different image).
	 */
	public static function set_content_image_alt( $content, $index, $alt_text, $expected_src = '' ) {
		$dom = self::load_dom( $content );
		if ( ! $dom ) {
			return new WP_Error( 'aiatg_dom_error', __( 'Could not parse post content.', 'fix-experts-ai-image-alt-tag-generator' ) );
		}

		$imgs = $dom->getElementsByTagName( 'img' );

		if ( ! isset( $imgs[ $index ] ) ) {
			return new WP_Error( 'aiatg_image_not_found', __( 'That image could not be found in the post content anymore (content may have changed).', 'fix-experts-ai-image-alt-tag-generator' ) );
		}

		if ( '' !== $expected_src && $imgs[ $index ]->getAttribute( 'src' ) !== $expected_src ) {
			return new WP_Error( 'aiatg_image_mismatch', __( 'The post content changed since this was scanned, so this image could not be safely updated. Please rescan.', 'fix-experts-ai-image-alt-tag-generator' ) );
		}

		$imgs[ $index ]->setAttribute( 'alt', $alt_text );

		return self::dom_to_html( $dom );
	}

	/**
	 * Loads HTML content into a DOMDocument wrapped in a neutral <div> so
	 * DOMDocument doesn't inject <html>/<body> wrappers around it.
	 */
	private static function load_dom( $content ) {
		if ( '' === trim( $content ) ) {
			return false;
		}

		$internal_errors = libxml_use_internal_errors( true );

		$dom = new DOMDocument( '1.0', 'UTF-8' );
		$dom->loadHTML(
			'<?xml encoding="utf-8" ?><div id="aiatg-root">' . $content . '</div>',
			LIBXML_HTML_NOIMPLIED | LIBXML_HTML_NODEFDTD
		);

		libxml_clear_errors();
		libxml_use_internal_errors( $internal_errors );

		return $dom;
	}

	/**
	 * Serializes the DOMDocument back to an HTML string, unwrapping the
	 * neutral wrapper div added by load_dom().
	 */
	private static function dom_to_html( DOMDocument $dom ) {
		$wrapper = $dom->getElementById( 'aiatg-root' );

		if ( ! $wrapper ) {
			return new WP_Error( 'aiatg_dom_error', __( 'Could not reassemble post content.', 'fix-experts-ai-image-alt-tag-generator' ) );
		}

		$html = '';
		foreach ( $wrapper->childNodes as $child ) {
			$html .= $dom->saveHTML( $child );
		}

		return $html;
	}
}