Helper functions for rewrting images in body fields from remote to local

drupal-migrate-helpers.php

<?php
/**
 * Updates remote image tags to work locally.
 *
 * @param $node
 */
function _rewrite_images($body)
{
    // quick test before we commit to DOM processing
    // must have images from our old
    // domains 'example.com' or 'example.org'
    // this assumes all images are full URIs so adjust accordingly
    if (false === strpos($body, 'img')
        && false === strpos($body, 'example.com')
        && false === strpos($body, 'example.org')) {
        return $body;
    }

    // look for actual image tags
    if ($images = _extractImages($body)) {
        // keep only images from our domains
        $images = array_filter(
            $images,
            function ($image) {
                $allowed = array('example.com', 'example.org');

                $host = parse_url($image['src'], PHP_URL_HOST);
                return in_array($host, $allowed);
            }
        );

        // import remaining to drupal files
        $map = _import_images($images);

        // prepare url replacements
        $replacements = array();
        foreach ($map as $key => &$file) {
            $file->url = file_create_url($file->uri);
            $file->parts = parse_url($file->url);

            // map old src attributes to new values
            $replacements['src="' . $key . '"']
                = 'src="' . str_replace('//', '/', $file->parts['path']) . '"';
        }

        // update references in body
        $body = preg_replace_callback(
            '/(\<img[^>]+\>)/',
            function($match) use ($replacements) {
                $match[1] = str_replace(
                    array_keys($replacements),
                    array_values($replacements),
                    $match[1]
                );

                return $match[1];
            }, $body);
    }

    return $body;
}

/**
 * extractImages
 *
 * @param $text HTML fragment to update
 * @return array|bool
 */
function _extractImages($text) {
    $header = '<meta http-equiv="Content-Type" '
        . 'content="text/html; charset=UTF-8" />';
    $text = $header . $text;

    $dom = new DOMDocument();
    if (@$dom->loadHTML($text)) {
        $xpath = new DOMXpath($dom);
        if ($images = $xpath->evaluate("//img")) {
            $result = array();

            foreach ($images as $i => $img) {
                // ignore if it doesn't end in an image file extension
                $src = $img->getAttribute('src');

                if (!preg_match('/.(png|jpg|jpeg|gif)$/i', parse_url($src, PHP_URL_PATH))) {
                    continue;
                }

                // fix protocol-relative URLs (start with //)
                if (preg_match('/^\/\//', $src)) {
                    $src = 'http:' . $src;
                }

                // ignore if its not http(s)
                if (!preg_match('/^https?:\/\//', $src)) {
                    continue;
                }

                $result[$i] = array('src' => $src);
            }

            if (!empty($result)) {
                return $result;
            }
        }
    }

    return false;
}

/**
 * Take our array of found images and save them as
 * local files
 * @param array $images
 * @return array
 */
function _import_images($images)
{
    // save them in imags subdir
    $uri_base = 'public://images';
    $perms = FILE_CREATE_DIRECTORY || FILE_MODIFY_PERMISSIONS;

    $map = array();
    foreach ($images as $image) {
        $path = parse_url($image['src']);
        $path = preg_replace(
            '/^\/images\/(.+)/', '$1', $path['path']);

        // make our target directory
        $dir = $uri_base . dirname('/' . $path);

        if (file_prepare_directory($dir, $perms)) {
            if ($file = system_retrieve_file($image['src'],
                $uri_base . '/' . $path, TRUE,
                FILE_EXISTS_REPLACE)) {
                // if saved, add it to our map to track
                // old sources to new location
                $map[$image['src']] = $file;
            }
        }
    }

    return $map;
}