Import featured images from the post content for all posts on a WordPress site that do not have a featured image thumbnail

rs-import-thumbnails.php

<?php
/*
Plugin Name: RS Import Thumbnails
Description: Import thumbnails to your posts using the first image in the post content. To use this tool add to the end of your site url: ?rs_import_thumbnails_2023318_2330
Version: 1.0
Author: Radley Sustaire
Author URI: https://radleysustaire.com/
*/


// TEST THIS PLUGIN THOROUGHLY
// Results:
// 1. https://radleysustaire.com/s3/bbc95f/chrome
// 2. https://radleysustaire.com/s3/891141/chrome
// 3. https://radleysustaire.com/s3/d81801/chrome


// ?rs_import_thumbnails_2023318_2330
function rs_import_thumbnails() {
	
	// Get all posts that have no featured image (featured image is stored as metadata "_thumbnail_id")
	$args = array(
		'post_type'      => 'post',
		'posts_per_page' => 20,
		'paged'          => 1, // CHANGE THIS TO GO TO THE NEXT PAGE
		'meta_query'     => array(
			array(
				'key'     => '_thumbnail_id',
				'compare' => 'NOT EXISTS',
			),
		),
	);
	
	$query = new WP_Query( $args );
	
	// If no more posts, all done
	if ( ! $query->have_posts() ) {
		echo 'No more posts to process, all done!';
	}

	$processed = 0;
	$found_images = 0;
	
	// Loop through posts
	while ( $query->have_posts() ): $query->the_post();
		
		echo '<h3>Checking post <a src="' . get_the_permalink() . '">' . get_the_title() . '</a> (#'. get_the_ID() .')</h3>';
	
		$content = get_the_content(); // Get the post content
		$pattern = '/<img[^>]+>/i'; // Regular expression pattern to match <img> tags
		
		// Find the first <img> tag in the post content
		preg_match( $pattern, $content, $matches );
		
		echo '<pre>';
		
		if ( ! empty( $matches ) ) {
			$first_image = $matches[0]; // The first <img> tag found
			
			// Extract the src attribute value using another regular expression
			$src_pattern = '/src=[\'"]?([^\'" >]+)/i';
			preg_match( $src_pattern, $first_image, $src_matches );
			
			if ( ! empty( $src_matches ) && isset( $src_matches[1] ) ) {
				
				// Attach the image to the post
				$result = rs_attach_thumbnail_to_post( get_the_ID(), $src_matches[1] );
				
				if ( is_numeric($result)  ) {
					echo 'Image attached successfully! Image id #' . $result;
					echo "\n\n";
					echo wp_get_attachment_image( $result, 'thumbnail' );
				}else{
					echo 'Failed to attach image to post, error:' . "\n\n" . $result;
				}
			} else {
				echo 'Image was found, but src attribute was invalid:' . "\n\n" . esc_html($first_image);
			}
			
		} else {
			// No <img> tag found
			echo 'No image found.';
		}
		
		echo '</pre>';
	
	endwhile;
	
	exit;
}
if ( isset($_GET['rs_import_thumbnails_2023318_2330']) ) add_action( 'init', 'rs_import_thumbnails' );

function rs_attach_thumbnail_to_post( $post_id, $image_url ) {
	if ( ! function_exists( 'wp_generate_attachment_metadata' ) ) {
		require_once( ABSPATH . 'wp-admin/includes/image.php' );
		require_once( ABSPATH . 'wp-admin/includes/file.php' );
		require_once( ABSPATH . 'wp-admin/includes/media.php' );
	}
	
	$filename = rs_get_image_filename_from_url( $image_url, get_the_title($post_id) );
	
	// Upload the image to the media library
	$upload_file = wp_upload_bits( $filename, null, file_get_contents( $image_url ) );
	
	// Check if the upload was successful
	if ( ! $upload_file['error'] ) {
		// Prepare to create a new attachment using this metadata
		$attachment = array(
			'post_mime_type' => $upload_file['type'],
			'post_title'     => sanitize_file_name( $upload_file['file'] ),
			'post_content'   => '',
			'post_status'    => 'inherit'
		);
		
		// Create the attachment
		$attachment_id = wp_insert_attachment( $attachment, $upload_file['file'], $post_id );
		
		// Set the attachment as the featured image
		if ( ! is_wp_error( $attachment_id ) ) {
			
			// Generate metadata and thumbnails for the attachment
			$metadata = wp_generate_attachment_metadata( $attachment_id, get_attached_file( $attachment_id ) );
			
			if ( $metadata ) {
				// Update the attachment's metadata
				wp_update_attachment_metadata( $attachment_id, $metadata );
			}
			
			// Attach the thumbnail
			if ( set_post_thumbnail( $post_id, $attachment_id ) ) {
				return $attachment_id;
			}else{
				return 'attachment was uploaded, but set_post_thumbnail() failed';
			}
			
		}else{
			return 'wp_insert_attachment() failed, attachment_id returned the error: ' . $attachment_id->get_error_message();
		}
	}else{
		return 'wp_upload_bits() failed';
	}
}

function rs_upload_image_to_media_library( $image_url ) {
	// Check if the required function is available
	if ( ! function_exists( 'media_sideload_image' ) ) {
		require_once( ABSPATH . 'wp-admin/includes/image.php' );
		require_once( ABSPATH . 'wp-admin/includes/file.php' );
		require_once( ABSPATH . 'wp-admin/includes/media.php' );
	}
	
	// Download the image from the URL and upload it to the media library
	$attachment_id = media_sideload_image( $image_url, 0 );
	
	if ( ! is_wp_error( $attachment_id ) ) {
		return $attachment_id;
	}
	
	return false;
}


function rs_get_image_filename_from_url( $url, $post_title ) {
	// Extract the path from the URL
	$path = parse_url( $url, PHP_URL_PATH );
	
	// Get the filename from the path
	$filename = basename( $path );
	$extension = pathinfo( $filename, PATHINFO_EXTENSION );
	
	// If the extracted filename is empty, use a generic filename with the ".jpg" extension
	if ( empty( $extension ) ) {
		$filename = sanitize_title_with_dashes($post_title);
		$extension = rs_get_image_file_type( $url );
		
		if ( $extension ) {
			$filename .= '.' . $extension;
		}else{
			$filename .= '.jpg';
		}
	}
	
	return $filename;
}

function rs_get_image_file_type( $url ) {
	// Get the MIME type of the content
	$mime_type = mime_content_type( $url );
	
	// Check if the MIME type indicates a PNG file
	if ( $mime_type === 'image/png' ) {
		return 'png';
	}
	
	// Default to 'jpg' if the MIME type is unknown or doesn't match JPG or PNG
	return 'jpg';
}