. // }}} /** * OEmbed and OpenGraph implementation for GNU social * * @package GNUsocial * * @author Mikael Nordfeldth * @author Stephen Paul Weber * @author hannes * @author Mikael Nordfeldth * @author Miguel Dantas * @author Diogo Peralta Cordeiro * @authir Hugo Sales * * @copyright 2014-2021 Free Software Foundation, Inc http://www.fsf.org * @license https://www.gnu.org/licenses/agpl.html GNU AGPL v3 or later */ namespace Plugin\Embed; use App\Core\Cache; use App\Core\DB\DB; use App\Core\Event; use App\Core\GSFile; use App\Core\HTTPClient; use App\Core\Log; use App\Core\Modules\Plugin; use App\Core\Router\RouteLoader; use App\Core\Router\Router; use App\Entity\Attachment; use App\Util\Common; use App\Util\Exception\DuplicateFoundException; use App\Util\Exception\NotFoundException; use App\Util\Formatting; use App\Util\TemporaryFile; use Embed\Embed as LibEmbed; use Exception; use Symfony\Component\HttpFoundation\Request; /** * Base class for the Embed plugin that does most of the heavy lifting to get * and display representations for remote content. * * @copyright 2014-2021 Free Software Foundation, Inc http://www.fsf.org * @license https://www.gnu.org/licenses/agpl.html GNU AGPL v3 or later */ class Embed extends Plugin { /** * Settings which can be set in social.local.yaml * WARNING, these are _regexps_ (slashes added later). Always escape your dots and end ('$') your strings */ public $domain_allowlist = [ // hostname => service provider '.*' => '', // Default to allowing any host ]; /** * This code executes when GNU social creates the page routing, and we hook * on this event to add our action handler for Embed. * * @param $m URLMapper the router that was initialized. * * @throws Exception * * @return bool true if successful, the exception object if it isn't. * */ public function onAddRoute(RouteLoader $m): bool { $m->connect('oembed', 'main/oembed', Controller\Embed::class); $m->connect('embed', 'main/embed', Controller\Embed::class); return Event::next; } /** * Insert oembed and opengraph tags in all HTML head elements */ public function onShowHeadElements(Request $request, array &$result) { $matches = []; preg_match(',/?([^/]+)/?(.*),', $request->getPathInfo(), $matches); switch ($matches[1]) { case 'attachment': $url = "{$matches[1]}/{$matches[2]}"; break; } if (isset($url)) { foreach (['xml', 'json'] as $format) { $result[] = [ 'link' => [ 'rel' => 'alternate', 'type' => "application/{$format}+oembed", 'href' => Router::url('embed', ['format' => $format, 'url' => $url]), 'title' => 'oEmbed', ], ]; } } return Event::next; } /** * Save embedding information for an Attachment, if applicable. * * Normally this event is called through File::saveNew() * * @param Attachment $attachment The newly inserted Attachment object. * * @return bool success */ public function onAttachmentStoreNew(Attachment $attachment): bool { try { DB::findOneBy('attachment_embed', ['attachment_id' => $attachment->getId()]); } catch (NotFoundException) { if ($attachment->hasRemoteUrl() && $attachment->hasMimetype()) { $mimetype = $attachment->getMimetype(); if (Formatting::startsWith($mimetype, 'text/html') || Formatting::startsWith($mimetype, 'application/xhtml+xml')) { try { $embed_data = $this->getEmbed($attachment->getRemoteUrl(), $attachment); $embed_data['attachment_id'] = $attachment->getId(); DB::persist(Entity\AttachmentEmbed::create($embed_data)); DB::flush(); } catch (Exception $e) { Log::warning($e); } } } } catch (DuplicateFoundException) { Log::warning("Strangely, an attachment_embed object exists for new file {$attachment->getID()}"); } return Event::next; } /** * Replace enclosure representation of an attachment with the data from embed */ public function onAttachmentFileInfo(int $attachment_id, ?array &$enclosure) { try { $embed = DB::findOneBy('attachment_embed', ['attachment_id' => $attachment_id]); } catch (NotFoundException) { return Event::next; } // We know about this attachment, so we 'own' it, but know // that it doesn't have an image if (!$embed->isImage()) { $enclosure = null; return Event::stop; } $enclosure = [ 'filepath' => $embed->getFilepath(), 'mimetype' => $embed->getMimetype(), 'title' => $embed->getTitle(), 'width' => $embed->getWidth(), 'height' => $embed->getHeight(), 'url' => $embed->getMediaUrl(), ]; return Event::stop; } /** * Show this attachment enhanced with the corresponding Embed data, if available * @param array $vars * @param array $res * @return bool */ public function onViewRemoteAttachment(array $vars, array &$res): bool { $attachment = $vars['attachment']; try { $embed = Cache::get('attachment-embed-' . $attachment->getId(), fn () => DB::findOneBy('attachment_embed', ['attachment_id' => $attachment->getId()])); } catch (DuplicateFoundException $e) { Log::warning($e); return Event::next; } catch (NotFoundException) { return Event::next; } if (is_null($embed) && empty($embed->getAuthorName()) && empty($embed->getProvider())) { Log::debug('Embed doesn\'t have a representation for the attachment #' . $attachment->getId()); return Event::next; } $attributes = $embed->getImageHTMLAttributes(['class' => 'u-photo embed']); $res[] = Formatting::twigRenderString(<<
{% if attributes != false %} {% endif %}
{{embed.getTitle() | escape}}
{% if embed.getAuthorName() is not null %}
{% if embed.getAuthorUrl() is null %}

{{embed.getAuthorName()}}

{% else %} {{embed.getAuthorName()}} {% endif %}
{% endif %} {% if embed.getProvider() is not null %}
{% if embed.getProviderUrl() is null %}

{{embed.getProvider()}}

{% else %} {{embed.getProvider()}} {% endif %}
{% endif %}
{{ embed.getHtml() | escape }}
END, ['embed' => $embed, 'attributes' => $attributes, 'attachment' => $attachment]); return Event::stop; } /** * @throws ServerException if check is made but fails * * @return bool false on no check made, provider name on success * @return false|string on no check made, provider name on success * * */ protected function checkAllowlist(string $url): string | bool { if ($this->check_allowlist ?? false) { return false; // indicates "no check made" } $host = parse_url($url, PHP_URL_HOST); foreach ($this->domain_allowlist as $regex => $provider) { if (preg_match("/{$regex}/", $host)) { return $provider; // we trust this source, return provider name } } throw new ServerException(_m('Domain not in remote thumbnail source allowlist: {host}', ['host' => $host])); } /** * Check the file size of a remote file using a HEAD request and checking * the content-length variable returned. This isn't 100% foolproof but is * reliable enough for our purposes. * * @param string $url * @param null|array $headers - if we already made a request * * @return null|int the file size if it succeeds, false otherwise. */ private function getRemoteFileSize(string $url, ?array $headers = null): ?int { try { if ($headers === null) { if (!Common::isValidHttpUrl($url)) { Log::error('Invalid URL in Embed::getRemoteFileSize()'); return false; } $head = HTTPClient::head($url); $headers = $head->getHeaders(); $headers = array_change_key_case($headers, CASE_LOWER); } return $headers['content-length'][0] ?? false; } catch (Exception $e) { Loog::error($e); return false; } } /** * A private helper function that uses a HEAD request to check the mime type * of a remote URL to see it it's an image. * * @param mixed $url * @param null|mixed $headers * * @return bool true if the remote URL is an image, or false otherwise. */ private function isRemoteImage(string $url, ?array $headers = null): bool { try { if ($headers === null) { if (!Common::isValidHttpUrl($url)) { Log::error('Invalid URL in Embed::getRemoteFileSize()'); return false; } $head = HTTPClient::head($url); $headers = $head->getHeaders(); $headers = array_change_key_case($headers, CASE_LOWER); } return !empty($headers['content-type']) && GSFile::mimetypeMajor($headers['content-type'][0]) === 'image'; } catch (Exception $e) { Log::error($e); return false; } } /** * Validate that $imgData is a valid image, place it in it's folder and resize * * @param $imgData - The image data to validate * @param null|array $headers - The headers possible previous request to $url */ protected function validateAndWriteImage($imgData, ?array $headers = null): array { $file = new TemporaryFile(); $file->write($imgData); Event::handle('HashFile', [$file->getRealPath(), &$hash]); $filepath = Common::config('storage', 'dir') . "embed/{$hash}" . Common::config('thumbnail', 'extension'); $width = Common::config('plugin_embed', 'width'); $height = Common::config('plugin_embed', 'height'); $smart_crop = Common::config('plugin_embed', 'smart_crop'); Event::handle('ResizeImagePath', [$file->getRealPath(), $filepath, &$width, &$height, $smart_crop, &$mimetype]); unset($file); if (!is_null($headers) && array_key_exists('content-disposition', $headers) && preg_match('/^.+; filename="(.+?)"$/', $headers['content-disposition'][0], $matches) === 1) { $original_name = $matches[1]; } return [$filepath, $width, $height, $original_name ?? null, $mimetype]; } /** * Fetch, Validate and Write a remote image from url to temporary file * * @param Attachment $attachment * @param string $media_url URL for the actual media representation * * @throws Exception * * @return array|bool */ protected function fetchValidateWriteRemoteImage(Attachment $attachment, string $media_url): array | bool { if ($attachment->hasFilename() && file_exists($attachment->getPath())) { throw new AlreadyFulfilledException(_m('A thumbnail seems to already exist for remote file with id=={id}', ['id' => $attachment->getId()])); } if (Formatting::startsWith($media_url, 'file://')) { $filename = Formatting::removePrefix($media_url, 'file://'); $info = getimagesize($filename); $filename = basename($filename); $width = $info[0]; $height = $info[1]; } else { $this->checkAllowlist($media_url); $head = HTTPClient::head($media_url); $headers = $head->getHeaders(); $headers = array_change_key_case($headers, CASE_LOWER); try { $is_image = $this->isRemoteImage($media_url, $headers); if ($is_image == true) { $file_size = $this->getRemoteFileSize($media_url, $headers); $max_size = Common::config('attachments', 'file_quota'); if (($file_size != false) && ($file_size > $max_size)) { throw new \Exception("Wanted to store remote thumbnail of size {$file_size} but the upload limit is {$max_size} so we aborted."); } } else { return false; } } catch (Exception $err) { Log::debug('Could not determine size of remote image, aborted local storage.'); throw $err; } // First we download the file to memory and test whether it's actually an image file Log::debug('Downloading remote thumbnail for file id==' . $attachment->getId() . " with thumbnail URL: {$media_url}"); try { $imgData = HTTPClient::get($media_url)->getContent(); if (isset($imgData)) { [$filepath, $width, $height, $original_name, $mimetype] = $this->validateAndWriteImage($imgData, $headers); } else { throw new UnsupportedMediaException(_m('HTTPClient returned an empty result')); } } catch (UnsupportedMediaException $e) { // Couldn't find anything that looks like an image, nothing to do Log::debug($e); return false; } } return [$filepath, $width, $height, $original_name, $mimetype]; } /** * Perform an oEmbed or OpenGraph lookup for the given $url. * * Some known hosts are allowlisted with API endpoints where we * know they exist but autodiscovery data isn't available. * * Throws exceptions on failure. * * @param string $url * @param Attachment $attachment * * @return array */ public function getEmbed(string $url, Attachment $attachment): array { Log::info('Checking for remote URL metadata for ' . $url); try { Log::info("Trying to find Embed data for {$url} with 'oscarotero/Embed'"); $embed = new LibEmbed(); $info = $embed->get($url); $metadata['title'] = $info->title; $metadata['html'] = $info->description; $metadata['author_name'] = $info->authorName; $metadata['author_url'] = $info->authorUrl; $metadata['provider_name'] = $info->providerName; $metadata['provider_url'] = $info->providerUrl; if (!is_null($info->image)) { $image_url = (string) $info->image; if (Formatting::startsWith($image_url, 'data')) { // Inline image $imgData = base64_decode(substr($info->image, stripos($info->image, 'base64,') + 7)); [$filepath, $width, $height, $original_name, $mimetype] = $this->validateAndWriteImage($imgData); } else { [$filepath, $width, $height, $original_name, $mimetype] = $this->fetchValidateWriteRemoteImage($attachment, $image_url); } $metadata['width'] = $width; $metadata['height'] = $height; $metadata['mimetype'] = $mimetype; $metadata['media_url'] = $image_url; $metadata['filename'] = Formatting::removePrefix($filepath, Common::config('storage', 'dir')); } } catch (Exception $e) { Log::info("Failed to find Embed data for {$url} with 'oscarotero/Embed', got exception: " . $e->getMessage()); } $metadata = self::normalize($metadata); $attachment->setTitle($metadata['title']); return $metadata; } /** * Normalize fetched info. */ public static function normalize(array $data): array { if (isset($metadata['url'])) { // sometimes sites serve the path, not the full URL, for images // let's "be liberal in what you accept from others"! // add protocol and host if the thumbnail_url starts with / if ($metadata['url'][0] == '/') { $thumbnail_url_parsed = parse_url($metadata['url']); $metadata['url'] = "{$thumbnail_url_parsed['scheme']}://{$thumbnail_url_parsed['host']}{$metadata['url']}"; } // Some wordpress opengraph implementations sometimes return a white blank image // no need for us to save that! if ($metadata['url'] == 'https://s0.wp.com/i/blank.jpg') { $metadata['url'] = null; } if (!isset($data['width'])) { $data['width'] = Common::config('plugin_embed', 'width'); $data['height'] = Common::config('plugin_embed', 'height'); } } return $data; } }