This repository has been archived on 2023-08-20. You can view files and clone it, but cannot push or open issues or pull requests.
gnu-social-v2-archive-activ.../utils/explorer.php

443 lines
17 KiB
PHP
Raw Normal View History

2018-07-06 11:47:28 +01:00
<?php
/**
* GNU social - a federating social network
*
2018-07-10 00:17:18 +01:00
* ActivityPubPlugin implementation for GNU Social
2018-07-06 11:47:28 +01:00
*
* LICENCE: This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
2018-07-10 00:17:18 +01:00
* @category Plugin
2018-07-06 11:47:28 +01:00
* @package GNUsocial
* @author Diogo Cordeiro <diogo@fc.up.pt>
2018-07-10 01:47:52 +01:00
* @author Daniel Supernault <danielsupernault@gmail.com>
2018-07-10 00:17:18 +01:00
* @copyright 2018 Free Software Foundation http://fsf.org
2018-07-06 11:47:28 +01:00
* @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
2018-07-10 00:17:18 +01:00
* @link https://www.gnu.org/software/social/
2018-07-06 11:47:28 +01:00
*/
if (!defined('GNUSOCIAL')) {
exit(1);
2018-07-10 00:17:18 +01:00
}
2018-07-06 11:47:28 +01:00
2018-07-10 00:17:18 +01:00
/**
2018-07-13 12:32:27 +01:00
* ActivityPub's own Explorer
*
* Allows to discovery new (or the same) Profiles (both local or remote)
2018-07-13 12:32:27 +01:00
*
2018-07-10 00:17:18 +01:00
* @category Plugin
* @package GNUsocial
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
* @link http://www.gnu.org/software/social/
*/
class Activitypub_explorer
2018-07-10 00:17:18 +01:00
{
private $discovered_actor_profiles = [];
/**
* Shortcut function to get a single profile from its URL.
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param string $url
* @return Profile
* @throws Exception
*/
public static function get_profile_from_url($url)
{
$discovery = new Activitypub_explorer;
// Get valid Actor object
$actor_profile = $discovery->lookup($url);
if (!empty($actor_profile)) {
return $actor_profile[0];
}
throw new Exception('Invalid Actor.');
}
2018-07-10 00:17:18 +01:00
/**
* Get every profile from the given URL
* This function cleans the $this->discovered_actor_profiles array
* so that there is no erroneous data
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param string $url User's url
* @return array of Profile objects
*/
public function lookup($url)
{
if (in_array($url, ACTIVITYPUB_PUBLIC_TO)) {
return [];
}
common_debug('ActivityPub Explorer: Started now looking for '.$url);
$this->discovered_actor_profiles = [];
return $this->_lookup($url);
}
2018-07-10 00:17:18 +01:00
/**
* Get every profile from the given URL
* This is a recursive function that will accumulate the results on
* $discovered_actor_profiles array
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param string $url User's url
* @return array of Profile objects
*/
private function _lookup($url)
{
// First check if we already have it locally and, if so, return it
// If the local fetch fails: grab it remotely, store locally and return
if (! ($this->grab_local_user($url) || $this->grab_remote_user($url))) {
throw new Exception('User not found.');
2018-07-09 00:07:14 +01:00
}
2018-07-10 00:17:18 +01:00
return $this->discovered_actor_profiles;
}
/**
* This ensures that we are using a valid ActivityPub URI
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param string $url
* @return boolean success state (related to the response)
* @throws Exception (If the HTTP request fails)
*/
private function ensure_proper_remote_uri($url)
{
$client = new HTTPClient();
$headers = [];
$headers[] = 'Accept: application/ld+json; profile="https://www.w3.org/ns/activitystreams"';
$headers[] = 'User-Agent: GNUSocialBot v0.1 - https://gnu.io/social';
$response = $client->get($url, $headers);
2018-08-02 05:00:35 +01:00
$res = json_decode($response->getBody(), true);
if (self::validate_remote_response($res)) {
$this->temp_res = $res;
return true;
2018-07-31 20:09:47 +01:00
} else {
2018-08-01 01:58:31 +01:00
common_debug('ActivityPub Explorer: Invalid potential remote actor while ensuring URI: '.$url. '. He returned the following: '.json_encode($res, JSON_UNESCAPED_SLASHES));
2018-07-10 00:17:18 +01:00
}
return false;
}
2018-07-10 00:17:18 +01:00
/**
* Get a local user profile from its URL and joins it on
* $this->discovered_actor_profiles
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param string $uri Actor's uri
* @return boolean success state
*/
private function grab_local_user($uri, $online = false)
{
if ($online) {
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Searching locally for '.$uri. ' with online resources.');
} else {
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Searching locally for '.$uri. ' offline.');
}
// Ensure proper remote URI
// If an exception occurs here it's better to just leave everything
// break than to continue processing
if ($online && $this->ensure_proper_remote_uri($uri)) {
$uri = $this->temp_res["id"];
2018-07-09 00:07:14 +01:00
}
// Try standard ActivityPub route
// Is this a known filthy little mudblood?
$aprofile = self::get_aprofile_by_url($uri);
if ($aprofile instanceof Activitypub_profile) {
$profile = $aprofile->local_profile();
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Found a local Aprofile for '.$uri);
// We found something!
$this->discovered_actor_profiles[]= $profile;
unset($this->temp_res); // IMPORTANT to avoid _dangerous_ noise in the Explorer system
return true;
} else {
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Unable to find a local Aprofile for '.$uri.' - looking for a Profile instead.');
// Well, maybe it is a pure blood?
// Iff, we are in the same instance:
$ACTIVITYPUB_BASE_ACTOR_URI_length = strlen(ACTIVITYPUB_BASE_ACTOR_URI);
if (substr($uri, 0, $ACTIVITYPUB_BASE_ACTOR_URI_length) == ACTIVITYPUB_BASE_ACTOR_URI) {
try {
$profile = Profile::getByID(intval(substr($uri, $ACTIVITYPUB_BASE_ACTOR_URI_length)));
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Found a Profile for '.$uri);
// We found something!
$this->discovered_actor_profiles[]= $profile;
unset($this->temp_res); // IMPORTANT to avoid _dangerous_ noise in the Explorer system
return true;
} catch (Exception $e) {
// Let the exception go on its merry way.
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Unable to find a Profile for '.$uri);
}
}
2018-07-10 00:17:18 +01:00
}
// If offline grabbing failed, attempt again with online resources
if (!$online) {
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Will try everything again with online resources against: '.$uri);
return $this->grab_local_user($uri, true);
}
return false;
}
2018-07-10 00:17:18 +01:00
/**
* Get a remote user(s) profile(s) from its URL and joins it on
* $this->discovered_actor_profiles
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param string $url User's url
* @return boolean success state
*/
private function grab_remote_user($url)
{
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Trying to grab a remote actor for '.$url);
if (!isset($this->temp_res)) {
$client = new HTTPClient();
$headers = array();
$headers[] = 'Accept: application/ld+json; profile="https://www.w3.org/ns/activitystreams"';
$headers[] = 'User-Agent: GNUSocialBot v0.1 - https://gnu.io/social';
$response = $client->get($url, $headers);
2018-08-02 05:00:35 +01:00
$res = json_decode($response->getBody(), true);
} else {
$res = $this->temp_res;
unset($this->temp_res);
}
if (isset($res['type']) && $res['type'] === 'OrderedCollection' && isset($res['first'])) { // It's a potential collection of actors!!!
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Found a collection of actors for '.$url);
2018-08-02 07:17:04 +01:00
$this->travel_collection($res['first']);
return true;
} elseif (self::validate_remote_response($res)) {
2018-07-31 20:09:47 +01:00
common_debug('ActivityPub Explorer: Found a valid remote actor for '.$url);
$this->discovered_actor_profiles[]= $this->store_profile($res);
return true;
2018-07-31 20:09:47 +01:00
} else {
2018-08-01 01:58:31 +01:00
common_debug('ActivityPub Explorer: Invalid potential remote actor while grabbing remotely: '.$url. '. He returned the following: '.json_encode($res, JSON_UNESCAPED_SLASHES));
}
// TODO: Fallback to OStatus
return false;
}
/**
* Save remote user profile in local instance
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param array $res remote response
* @return Profile remote Profile object
*/
private function store_profile($res)
{
// ActivityPub Profile
$aprofile = new Activitypub_profile;
2018-07-27 21:45:43 +01:00
$aprofile->uri = $res['id'];
$aprofile->nickname = $res['preferredUsername'];
2018-07-31 20:09:47 +01:00
$aprofile->fullname = isset($res['name']) ? $res['name'] : null;
2018-08-01 14:20:27 +01:00
$aprofile->bio = isset($res['summary']) ? substr(strip_tags($res['summary']), 0, 1000) : null;
2018-07-27 21:45:43 +01:00
$aprofile->inboxuri = $res['inbox'];
$aprofile->sharedInboxuri = isset($res['endpoints']['sharedInbox']) ? $res['endpoints']['sharedInbox'] : $res['inbox'];
$aprofile->do_insert();
$profile = $aprofile->local_profile();
// Public Key
$apRSA = new Activitypub_rsa();
$apRSA->profile_id = $profile->getID();
$apRSA->public_key = $res['publicKey']['publicKeyPem'];
$apRSA->store_keys();
2018-08-01 21:24:36 +01:00
// Avatar
if (isset($res['icon']['url'])) {
try {
$this->_store_avatar($profile, $res['icon']['url']);
} catch (Exception $e) {
// Let the exception go, it isn't a serious issue
common_debug('An error ocurred while grabbing remote avatar'.$e->getMessage());
}
2018-08-01 21:24:36 +01:00
}
return $profile;
}
2018-08-01 21:24:36 +01:00
/**
* Download and update given avatar image
*
* @author GNU Social
* @param Profile $profile
* @param string $url
* @return Avatar The Avatar we have on disk.
2018-08-01 21:24:36 +01:00
* @throws Exception in various failure cases
*/
private function _store_avatar(Profile $profile, $url)
2018-08-01 21:24:36 +01:00
{
common_debug('ActivityPub Explorer: Started grabbing remote avatar from: '.$url);
if (!filter_var($url, FILTER_VALIDATE_URL)) {
2018-08-01 21:24:36 +01:00
// TRANS: Server exception. %s is a URL.
common_debug('ActivityPub Explorer: Failed because it is an invalid url: '.$url);
2018-08-01 21:24:36 +01:00
throw new ServerException(sprintf('Invalid avatar URL %s.'), $url);
}
// @todo FIXME: This should be better encapsulated
// ripped from oauthstore.php (for old OMB client)
$temp_filename = tempnam(sys_get_temp_dir(), 'listener_avatar');
try {
$imgData = HTTPClient::quickGet($url);
// Make sure it's at least an image file. ImageFile can do the rest.
if (false === getimagesizefromstring($imgData)) {
common_debug('ActivityPub Explorer: Failed because the downloaded avatar: '.$url. 'is not a valid image.');
throw new UnsupportedMediaException('Downloaded avatar was not an image.');
2018-08-01 21:24:36 +01:00
}
file_put_contents($temp_filename, $imgData);
unset($imgData); // No need to carry this in memory.
common_debug('ActivityPub Explorer: Stored dowloaded avatar in: '.$temp_filename);
2018-08-01 21:24:36 +01:00
$id = $profile->getID();
$imagefile = new ImageFile(null, $temp_filename);
$filename = Avatar::filename(
$id,
image_type_to_extension($imagefile->type),
null,
common_timestamp()
);
2018-08-01 21:24:36 +01:00
rename($temp_filename, Avatar::path($filename));
common_debug('ActivityPub Explorer: Moved avatar from: '.$temp_filename.' to '.$filename);
2018-08-01 21:24:36 +01:00
} catch (Exception $e) {
common_debug('ActivityPub Explorer: Something went wrong while processing the avatar from: '.$url.' details: '.$e->getMessage());
2018-08-01 21:24:36 +01:00
unlink($temp_filename);
throw $e;
}
// @todo FIXME: Hardcoded chmod is lame, but seems to be necessary to
// keep from accidentally saving images from command-line (queues)
// that can't be read from web server, which causes hard-to-notice
// problems later on:
//
// http://status.net/open-source/issues/2663
chmod(Avatar::path($filename), 0644);
$profile->setOriginal($filename);
$orig = clone($profile);
$profile->avatar = $url;
$profile->update($orig);
common_debug('ActivityPub Explorer: Seted Avatar from: '.$url.' to profile.');
2018-08-01 21:24:36 +01:00
return Avatar::getUploaded($profile);
}
/**
* Validates a remote response in order to determine whether this
* response is a valid profile or not
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param array $res remote response
* @return boolean success state
*/
public static function validate_remote_response($res)
{
2018-07-31 20:09:47 +01:00
if (!isset($res['id'], $res['preferredUsername'], $res['inbox'], $res['publicKey']['publicKeyPem'])) {
return false;
}
return true;
}
/**
* Get a ActivityPub Profile from it's uri
* Unfortunately GNU Social cache is not truly reliable when handling
* potential ActivityPub remote profiles, as so it is important to use
* this hacky workaround (at least for now)
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param string $v URL
* @return boolean|Activitypub_profile false if fails | Aprofile object if successful
*/
public static function get_aprofile_by_url($v)
{
$i = Managed_DataObject::getcached("Activitypub_profile", "uri", $v);
if (empty($i)) { // false = cache miss
$i = new Activitypub_profile;
$result = $i->get("uri", $v);
if ($result) {
// Hit!
$i->encache();
} else {
return false;
}
}
return $i;
}
/**
* Given a valid actor profile url returns its inboxes
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param string $url of Actor profile
* @return boolean|array false if fails | array with inbox and shared inbox if successful
*/
public static function get_actor_inboxes_uri($url)
{
$client = new HTTPClient();
$headers = array();
$headers[] = 'Accept: application/ld+json; profile="https://www.w3.org/ns/activitystreams"';
$headers[] = 'User-Agent: GNUSocialBot v0.1 - https://gnu.io/social';
$response = $client->get($url, $headers);
if (!$response->isOk()) {
2018-07-27 21:45:43 +01:00
throw new Exception('Invalid Actor URL.');
}
2018-08-02 05:00:35 +01:00
$res = json_decode($response->getBody(), true);
if (self::validate_remote_response($res)) {
2018-07-27 21:45:43 +01:00
return [
'inbox' => $res['inbox'],
'sharedInbox' => isset($res['endpoints']['sharedInbox']) ? $res['endpoints']['sharedInbox'] : $res['inbox']
];
}
return false;
}
2018-08-02 07:01:39 +01:00
/**
* Allows the Explorer to transverse a collection of persons.
*
* @author Diogo Cordeiro <diogo@fc.up.pt>
* @param type $url
* @return boolean
*/
private function travel_collection($url)
{
$client = new HTTPClient();
$headers = array();
$headers[] = 'Accept: application/ld+json; profile="https://www.w3.org/ns/activitystreams"';
$headers[] = 'User-Agent: GNUSocialBot v0.1 - https://gnu.io/social';
$response = $client->get($url, $headers);
$res = json_decode($response->getBody(), true);
if (!isset($res['orderedItems'])) {
2018-08-02 07:01:39 +01:00
return false;
}
2018-08-02 07:05:50 +01:00
2018-08-02 07:01:39 +01:00
foreach ($res["orderedItems"] as $profile) {
if ($this->_lookup($profile) == false) {
common_debug('ActivityPub Explorer: Found an invalid actor for '.$profile);
// TODO: Invalid actor found, fallback to OStatus
}
}
// Go through entire collection
if (!is_null($res["next"])) {
$this->_lookup($res["next"]);
}
return true;
}
2018-07-06 11:47:28 +01:00
}