We are no longer offering accounts on this server. Consider https://gitlab.freedesktop.org/ as a place to host projects.

Commit 1ceaa50c authored by Brion Vibber's avatar Brion Vibber

Merge branch 'twitstream' into 0.9.x

parents ec795729 9cbda327
......@@ -200,8 +200,15 @@ class TwitterBridgePlugin extends Plugin
return false;
case 'TwitterOAuthClient':
case 'TwitterQueueHandler':
case 'TwitterImport':
case 'JsonStreamReader':
case 'TwitterStreamReader':
include_once $dir . '/' . strtolower($cls) . '.php';
return false;
case 'TwitterSiteStream':
case 'TwitterUserStream':
include_once $dir . '/twitterstreamreader.php';
return false;
case 'Notice_to_status':
case 'Twitter_synch_status':
include_once $dir . '/' . $cls . '.php';
......@@ -267,7 +274,11 @@ class TwitterBridgePlugin extends Plugin
function onEndInitializeQueueManager($manager)
{
if (self::hasKeys()) {
// Outgoing notices -> twitter
$manager->connect('twitter', 'TwitterQueueHandler');
// Incoming statuses <- twitter
$manager->connect('tweetin', 'TweetInQueueHandler');
}
return true;
}
......
#!/usr/bin/env php
<?php
/*
* StatusNet - the distributed open-source microblogging tool
* Copyright (C) 2008-2010, StatusNet, Inc.
*
* 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/>.
*/
define('INSTALLDIR', realpath(dirname(__FILE__) . '/../../..'));
$shortoptions = 'fi::a';
$longoptions = array('id::', 'foreground', 'all');
$helptext = <<<END_OF_XMPP_HELP
Daemon script for receiving new notices from Twitter users.
-i --id Identity (default none)
-a --all Handle Twitter for all local sites
(requires Stomp queue handler, status_network setup)
-f --foreground Stay in the foreground (default background)
END_OF_XMPP_HELP;
require_once INSTALLDIR.'/scripts/commandline.inc';
require_once INSTALLDIR . '/lib/jabber.php';
class TwitterDaemon extends SpawningDaemon
{
protected $allsites = false;
function __construct($id=null, $daemonize=true, $threads=1, $allsites=false)
{
if ($threads != 1) {
// This should never happen. :)
throw new Exception("TwitterDaemon must run single-threaded");
}
parent::__construct($id, $daemonize, $threads);
$this->allsites = $allsites;
}
function runThread()
{
common_log(LOG_INFO, 'Waiting to listen to Twitter and queues');
$master = new TwitterMaster($this->get_id(), $this->processManager());
$master->init($this->allsites);
$master->service();
common_log(LOG_INFO, 'terminating normally');
return $master->respawn ? self::EXIT_RESTART : self::EXIT_SHUTDOWN;
}
}
class TwitterMaster extends IoMaster
{
protected $processManager;
function __construct($id, $processManager)
{
parent::__construct($id);
$this->processManager = $processManager;
}
/**
* Initialize IoManagers for the currently configured site
* which are appropriate to this instance.
*/
function initManagers()
{
$qm = QueueManager::get();
$qm->setActiveGroup('twitter');
$this->instantiate($qm);
$this->instantiate(new TwitterManager());
$this->instantiate($this->processManager);
}
}
class TwitterManager extends IoManager
{
// Recommended resource limits from http://dev.twitter.com/pages/site_streams
const MAX_STREAMS = 1000;
const USERS_PER_STREAM = 100;
const STREAMS_PER_SECOND = 20;
protected $streams;
protected $users;
/**
* Pull the site's active Twitter-importing users and start spawning
* some data streams for them!
*
* @fixme check their last-id and check whether we'll need to do a manual pull.
* @fixme abstract out the fetching so we can work over multiple sites.
*/
protected function initStreams()
{
common_log(LOG_INFO, 'init...');
// Pull Twitter user IDs for all users we want to pull data for
$flink = new Foreign_link();
$flink->service = TWITTER_SERVICE;
// @fixme probably should do the bitfield check in a whereAdd but it's ugly :D
$flink->find();
$userIds = array();
while ($flink->fetch()) {
if (($flink->noticesync & FOREIGN_NOTICE_RECV) ==
FOREIGN_NOTICE_RECV) {
$userIds[] = $flink->foreign_id;
if (count($userIds) >= self::USERS_PER_STREAM) {
$this->spawnStream($userIds);
$userIds = array();
}
}
}
if (count($userIds)) {
$this->spawnStream($userIds);
}
}
/**
* Prepare a Site Stream connection for the given chunk of users.
* The actual connection will be opened later.
*
* @param $userIds array of Twitter-side user IDs
*/
protected function spawnStream($userIds)
{
$stream = $this->initSiteStream();
$stream->followUsers($userIds);
// Slip the stream reader into our list of active streams.
// We'll manage its actual connection on the next go-around.
$this->streams[] = $stream;
// Record the user->stream mappings; this makes it easier for us to know
// later if we need to kill something.
foreach ($userIds as $id) {
$this->users[$id] = $stream;
}
}
/**
* Initialize a generic site streams connection object.
* All our connections will look like this, then we'll add users to them.
*
* @return TwitterStreamReader
*/
protected function initSiteStream()
{
$auth = $this->siteStreamAuth();
$stream = new TwitterSiteStream($auth);
// Add our event handler callbacks. Whee!
$this->setupEvents($stream);
return $stream;
}
/**
* Fetch the Twitter OAuth credentials to use to connect to the Site Streams API.
*
* This will use the locally-stored credentials for the applictation's owner account
* from the site configuration. These should be configured through the administration
* panels or manually in the config file.
*
* Will throw an exception if no credentials can be found -- but beware that invalid
* credentials won't cause breakage until later.
*
* @return TwitterOAuthClient
*/
protected function siteStreamAuth()
{
$token = common_config('twitter', 'stream_token');
$secret = common_config('twitter', 'stream_secret');
if (empty($token) || empty($secret)) {
throw new ServerException('Twitter site streams have not been correctly configured. Configure the app owner account via the admin panel.');
}
return new TwitterOAuthClient($token, $secret);
}
/**
* Collect the sockets for all active connections for i/o monitoring.
*
* @return array of resources
*/
public function getSockets()
{
$sockets = array();
foreach ($this->streams as $stream) {
foreach ($stream->getSockets() as $socket) {
$sockets[] = $socket;
}
}
return $sockets;
}
/**
* We're ready to process input from one of our data sources! Woooooo!
* @fixme is there an easier way to map from socket back to owning module? :(
*
* @param resource $socket
* @return boolean success
*/
public function handleInput($socket)
{
foreach ($this->streams as $stream) {
foreach ($stream->getSockets() as $aSocket) {
if ($socket === $aSocket) {
$stream->handleInput($socket);
}
}
}
return true;
}
/**
* Start the i/o system up! Prepare our connections and start opening them.
*
* @fixme do some rate-limiting on the stream setup
* @fixme do some sensible backoff on failure etc
*/
public function start()
{
$this->initStreams();
foreach ($this->streams as $stream) {
$stream->connect();
}
return true;
}
/**
* Close down our connections when the daemon wraps up for business.
*/
public function finish()
{
foreach ($this->streams as $index => $stream) {
$stream->close();
unset($this->streams[$index]);
}
return true;
}
public static function get()
{
throw new Exception('not a singleton');
}
/**
* Set up event handlers on the streaming interface.
*
* @fixme add more event types as we add handling for them
*/
protected function setupEvents(TwitterStreamReader $stream)
{
$handlers = array(
'status',
);
foreach ($handlers as $event) {
$stream->hookEvent($event, array($this, 'onTwitter' . ucfirst($event)));
}
}
/**
* Event callback notifying that a user has a new message in their home timeline.
* We store the incoming message into the queues for processing, keeping our own
* daemon running as shiny-fast as possible.
*
* @param object $status JSON data: Twitter status update
* @fixme in all-sites mode we may need to route queue items into another site's
* destination queues, or multiple sites.
*/
protected function onTwitterStatus($status, $context)
{
$data = array(
'status' => $status,
'for_user' => $context->for_user,
);
$qm = QueueManager::get();
$qm->enqueue($data, 'tweetin');
}
}
if (have_option('i', 'id')) {
$id = get_option_value('i', 'id');
} else if (count($args) > 0) {
$id = $args[0];
} else {
$id = null;
}
$foreground = have_option('f', 'foreground');
$all = have_option('a') || have_option('--all');
$daemon = new TwitterDaemon($id, !$foreground, 1, $all);
$daemon->runOnce();
<?php
/**
* StatusNet, the distributed open-source microblogging tool
*
* PHP version 5
*
* 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/>.
*
* @category Plugin
* @package StatusNet
* @author Brion Vibber <brion@status.net>
* @copyright 2010 StatusNet, Inc.
* @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
* @link http://status.net/
*/
class OAuthData
{
public $consumer_key, $consumer_secret, $token, $token_secret;
}
/**
*
*/
abstract class JsonStreamReader
{
const CRLF = "\r\n";
public $id;
protected $socket = null;
protected $state = 'init'; // 'init', 'connecting', 'waiting', 'headers', 'active'
public function __construct()
{
$this->id = get_class($this) . '.' . substr(md5(mt_rand()), 0, 8);
}
/**
* Starts asynchronous connect operation...
*
* @fixme Can we do the open-socket fully async to? (need write select infrastructure)
*
* @param string $url
*/
public function connect($url)
{
common_log(LOG_DEBUG, "$this->id opening connection to $url");
$scheme = parse_url($url, PHP_URL_SCHEME);
if ($scheme == 'http') {
$rawScheme = 'tcp';
} else if ($scheme == 'https') {
$rawScheme = 'ssl';
} else {
throw new ServerException('Invalid URL scheme for HTTP stream reader');
}
$host = parse_url($url, PHP_URL_HOST);
$port = parse_url($url, PHP_URL_PORT);
if (!$port) {
if ($scheme == 'https') {
$port = 443;
} else {
$port = 80;
}
}
$path = parse_url($url, PHP_URL_PATH);
$query = parse_url($url, PHP_URL_QUERY);
if ($query) {
$path .= '?' . $query;
}
$errno = $errstr = null;
$timeout = 5;
//$flags = STREAM_CLIENT_CONNECT | STREAM_CLIENT_ASYNC_CONNECT;
$flags = STREAM_CLIENT_CONNECT;
// @fixme add SSL params
$this->socket = stream_socket_client("$rawScheme://$host:$port", $errno, $errstr, $timeout, $flags);
$this->send($this->httpOpen($host, $path));
stream_set_blocking($this->socket, false);
$this->state = 'waiting';
}
/**
* Send some fun data off to the server.
*
* @param string $buffer
*/
function send($buffer)
{
fwrite($this->socket, $buffer);
}
/**
* Read next packet of data from the socket.
*
* @return string
*/
function read()
{
$buffer = fread($this->socket, 65536);
return $buffer;
}
/**
* Build HTTP request headers.
*
* @param string $host
* @param string $path
* @return string
*/
protected function httpOpen($host, $path)
{
$lines = array(
"GET $path HTTP/1.1",
"Host: $host",
"User-Agent: StatusNet/" . STATUSNET_VERSION . " (TwitterBridgePlugin)",
"Connection: close",
"",
""
);
return implode(self::CRLF, $lines);
}
/**
* Close the current connection, if open.
*/
public function close()
{
if ($this->isConnected()) {
common_log(LOG_DEBUG, "$this->id closing connection.");
fclose($this->socket);
$this->socket = null;
}
}
/**
* Are we currently connected?
*
* @return boolean
*/
public function isConnected()
{
return $this->socket !== null;
}
/**
* Send any sockets we're listening on to the IO manager
* to wait for input.
*
* @return array of resources
*/
public function getSockets()
{
if ($this->isConnected()) {
return array($this->socket);
}
return array();
}
/**
* Take a chunk of input over the horn and go go go! :D
*
* @param string $buffer
*/
public function handleInput($socket)
{
if ($this->socket !== $socket) {
throw new Exception('Got input from unexpected socket!');
}
try {
$buffer = $this->read();
$lines = explode(self::CRLF, $buffer);
foreach ($lines as $line) {
$this->handleLine($line);
}
} catch (Exception $e) {
common_log(LOG_ERR, "$this->id aborting connection due to error: " . $e->getMessage());
fclose($this->socket);
throw $e;
}
}
protected function handleLine($line)
{
switch ($this->state)
{
case 'waiting':
$this->handleLineWaiting($line);
break;
case 'headers':
$this->handleLineHeaders($line);
break;
case 'active':
$this->handleLineActive($line);
break;
default:
throw new Exception('Invalid state in handleLine: ' . $this->state);
}
}
/**
*
* @param <type> $line
*/
protected function handleLineWaiting($line)
{
$bits = explode(' ', $line, 3);
if (count($bits) != 3) {
throw new Exception("Invalid HTTP response line: $line");
}
list($http, $status, $text) = $bits;
if (substr($http, 0, 5) != 'HTTP/') {
throw new Exception("Invalid HTTP response line chunk '$http': $line");
}
if ($status != '200') {
throw new Exception("Bad HTTP response code $status: $line");
}
common_log(LOG_DEBUG, "$this->id $line");
$this->state = 'headers';
}
protected function handleLineHeaders($line)
{
if ($line == '') {
$this->state = 'active';
common_log(LOG_DEBUG, "$this->id connection is active!");
} else {
common_log(LOG_DEBUG, "$this->id read HTTP header: $line");
$this->responseHeaders[] = $line;
}
}
protected function handleLineActive($line)
{
if ($line == "") {
// Server sends empty lines as keepalive.
return;
}
$data = json_decode($line);
if ($data) {
$this->handleJson($data);
} else {
common_log(LOG_ERR, "$this->id received bogus JSON data: " . var_export($line, true));
}
}
abstract protected function handleJson(stdClass $data);
}