Two WordPress functions for maybe side loading URLs to the media library. Useful for content migrations that need to be run multiple times without producing duplicate downloads.

maybe-sideload.php

<?php

/**
 * Example useage:
 *
 * maybe_sideload_image( 'https://dummyimage.com/600x400/000/fff.png' ); // Downloads image to the media library and returns an attachment ID
 * maybe_sideload_image( 'https://dummyimage.com/600x400/000/fff.png' ); // Returns an attachment ID as the image has already been downloaded and added to the media library
 */

/**
 * Given an image URL return an attachment_id. Image will be sideloaded into the media library if it doesn't exist.
 *
 * @param string $url The image URL to maybe sideload
 * @uses media_sideload_image
 */
function maybe_sideload_image( $url = '' ) {
	global $wpdb;
	require_once ABSPATH . 'wp-admin/includes/media.php';
	require_once ABSPATH . 'wp-admin/includes/file.php';
	require_once ABSPATH . 'wp-admin/includes/image.php';

	// Check to see if the URL has already been fetched, if so return the attachment ID
	$attachment_id = $wpdb->get_var(
		$wpdb->prepare( "SELECT `post_id` FROM {$wpdb->postmeta} WHERE `meta_key` = '_source_url' AND `meta_value` = %s", $url )
	);
	if ( ! empty( $attachment_id ) ) {
		return $attachment_id;
	}

	$attachment_id = $wpdb->get_var(
		$wpdb->prepare( "SELECT `ID` FROM {$wpdb->posts} WHERE guid=%s", $url )
	);
	if ( ! empty( $attachment_id ) ) {
		return $attachment_id;
	}

	// If the URL doesn't exist, sideload it to the media library
	return media_sideload_image( $url, $post_id = 0, $desc = null, $return = 'id' );
}

/**
 * Given a file URL return an attachment ID. If the URL doesn't already exist then it gets sideloaded into the media library.
 *
 * @param string $url URL of the file to maybe sideload
 * @uses media_handle_sideload
 */
function maybe_sideload_url( $url = '' ) {
	global $wpdb;
	require_once ABSPATH . 'wp-admin/includes/media.php';
	require_once ABSPATH . 'wp-admin/includes/file.php';
	require_once ABSPATH . 'wp-admin/includes/image.php';

	// Check to see if the URL has already been fetched, if so return the attachment ID
	$attachment_id = $wpdb->get_var(
		$wpdb->prepare( "SELECT `post_id` FROM {$wpdb->postmeta} WHERE `meta_key` = '_source_url' AND `meta_value` = %s", $url )
	);
	if ( ! empty( $attachment_id ) ) {
		return $attachment_id;
	}

	$attachment_id = $wpdb->get_var(
		$wpdb->prepare( "SELECT `ID` FROM {$wpdb->posts} WHERE guid=%s", $url )
	);
	if ( ! empty( $attachment_id ) ) {
		return $attachment_id;
	}

	// If the URL doesn't exist, sideload it to the media library
	$file_array         = array();
	$file_array['name'] = wp_basename( $url );

	// Download file to temp location.
	$file_array['tmp_name'] = download_url( $url );

	// If error storing temporarily, return
	if ( is_wp_error( $file_array['tmp_name'] ) ) {
		return 0;
	}

	// Do the validation and storage stuff.
	$id = media_handle_sideload( $file_array, $post_id, $desc );

	// If error storing permanently, unlink.
	if ( is_wp_error( $id ) ) {
		@unlink( $file_array['tmp_name'] );
		return $id;
	}

	// Store the original attachment source in meta. Mirrors media_sideload_image
	add_post_meta( $id, '_source_url', $url );

	return $id;
}

Line 7 seems like it should read:

• maybe_sideload_url(