4 * StatusNet - the distributed open-source microblogging tool
5 * Copyright (C) 2008-2010, StatusNet, Inc.
7 * This program is free software: you can redistribute it and/or modify
8 * it under the terms of the GNU Affero General Public License as published by
9 * the Free Software Foundation, either version 3 of the License, or
10 * (at your option) any later version.
12 * This program is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU Affero General Public License for more details.
17 * You should have received a copy of the GNU Affero General Public License
18 * along with this program. If not, see <http://www.gnu.org/licenses/>.
21 define('INSTALLDIR', realpath(dirname(__FILE__) . '/../../..'));
23 // Tune number of processes and how often to poll Twitter
24 // XXX: Should these things be in config.php?
25 define('MAXCHILDREN', 2);
26 define('POLL_INTERVAL', 60); // in seconds
28 $shortoptions = 'di::';
29 $longoptions = array('id::', 'debug');
31 $helptext = <<<END_OF_TRIM_HELP
32 Batch script for retrieving Twitter messages from foreign service.
34 -i --id Identity (default 'generic')
35 -d --debug Debug (lots of log output)
39 require_once INSTALLDIR . '/scripts/commandline.inc';
40 require_once INSTALLDIR . '/lib/common.php';
41 require_once INSTALLDIR . '/lib/daemon.php';
42 require_once INSTALLDIR . '/plugins/TwitterBridge/twitter.php';
43 require_once INSTALLDIR . '/plugins/TwitterBridge/twitterbasicauthclient.php';
44 require_once INSTALLDIR . '/plugins/TwitterBridge/twitteroauthclient.php';
47 * Fetch statuses from Twitter
49 * Fetches statuses from Twitter and inserts them as notices
51 * NOTE: an Avatar path MUST be set in config.php for this
52 * script to work, e.g.:
53 * $config['avatar']['path'] = $config['site']['path'] . '/avatar/';
55 * @todo @fixme @gar Fix the above. For some reason $_path is always empty when
56 * this script is run, so the default avatar path is always set wrong in
57 * default.php. Therefore it must be set explicitly in config.php. --Z
61 * @author Zach Copley <zach@status.net>
62 * @author Evan Prodromou <evan@status.net>
63 * @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
64 * @link http://status.net/
67 class TwitterStatusFetcher extends ParallelizingDaemon
72 * @param string $id the name/id of this daemon
73 * @param int $interval sleep this long before doing everything again
74 * @param int $max_children maximum number of child processes at a time
75 * @param boolean $debug debug output flag
80 function __construct($id = null, $interval = 60,
81 $max_children = 2, $debug = null)
83 parent::__construct($id, $interval, $max_children, $debug);
89 * @return string Name of the daemon.
94 return ('twitterstatusfetcher.'.$this->_id);
98 * Find all the Twitter foreign links for users who have requested
99 * importing of their friends' timelines
101 * @return array flinks an array of Foreign_link objects
104 function getObjects()
106 global $_DB_DATAOBJECT;
108 $flink = new Foreign_link();
109 $conn = &$flink->getDatabaseConnection();
111 $flink->service = TWITTER_SERVICE;
112 $flink->orderBy('last_noticesync');
117 while ($flink->fetch()) {
119 if (($flink->noticesync & FOREIGN_NOTICE_RECV) ==
120 FOREIGN_NOTICE_RECV) {
121 $flinks[] = clone($flink);
122 common_log(LOG_INFO, "sync: foreign id $flink->foreign_id");
124 common_log(LOG_INFO, "nothing to sync");
132 unset($_DB_DATAOBJECT['CONNECTIONS']);
137 function childTask($flink) {
139 // Each child ps needs its own DB connection
141 // Note: DataObject::getDatabaseConnection() creates
142 // a new connection if there isn't one already
144 $conn = &$flink->getDatabaseConnection();
146 $this->getTimeline($flink);
148 $flink->last_friendsync = common_sql_now();
153 // XXX: Couldn't find a less brutal way to blow
154 // away a cached connection
156 global $_DB_DATAOBJECT;
157 unset($_DB_DATAOBJECT['CONNECTIONS']);
160 function getTimeline($flink)
163 common_log(LOG_WARNING, $this->name() .
164 " - Can't retrieve Foreign_link for foreign ID $fid");
168 common_debug($this->name() . ' - Trying to get timeline for Twitter user ' .
171 // XXX: Biggest remaining issue - How do we know at which status
172 // to start importing? How many statuses? Right now I'm going
173 // with the default last 20.
177 if (TwitterOAuthClient::isPackedToken($flink->credentials)) {
178 $token = TwitterOAuthClient::unpackToken($flink->credentials);
179 $client = new TwitterOAuthClient($token->key, $token->secret);
180 common_debug($this->name() . ' - Grabbing friends timeline with OAuth.');
182 common_debug("Skipping friends timeline for $flink->foreign_id since not OAuth.");
188 $timeline = $client->statusesHomeTimeline();
189 } catch (Exception $e) {
190 common_log(LOG_WARNING, $this->name() .
191 ' - Twitter client unable to get friends timeline for user ' .
192 $flink->user_id . ' - code: ' .
193 $e->getCode() . 'msg: ' . $e->getMessage());
196 if (empty($timeline)) {
197 common_log(LOG_WARNING, $this->name() . " - Empty timeline.");
201 common_debug(LOG_INFO, $this->name() . ' - Retrieved ' . sizeof($timeline) . ' statuses from Twitter.');
203 // Reverse to preserve order
205 foreach (array_reverse($timeline) as $status) {
207 // Hacktastic: filter out stuff coming from this StatusNet
209 $source = mb_strtolower(common_config('integration', 'source'));
211 if (preg_match("/$source/", mb_strtolower($status->source))) {
212 common_debug($this->name() . ' - Skipping import of status ' .
213 $status->id . ' with source ' . $source);
217 // Don't save it if the user is protected
218 // FIXME: save it but treat it as private
220 if ($status->user->protected) {
224 $notice = $this->saveStatus($status);
226 if (!empty($notice)) {
227 Inbox::insertNotice($flink->user_id, $notice->id);
231 // Okay, record the time we synced with Twitter for posterity
233 $flink->last_noticesync = common_sql_now();
237 function saveStatus($status)
239 $profile = $this->ensureProfile($status->user);
241 if (empty($profile)) {
242 common_log(LOG_ERR, $this->name() .
243 ' - Problem saving notice. No associated Profile.');
247 $statusUri = $this->makeStatusURI($status->user->screen_name, $status->id);
249 // check to see if we've already imported the status
251 $dupe = $this->checkDupe($profile, $statusUri);
257 " - Ignoring duplicate import: $statusUri"
262 common_debug("Saving status {$status->id} with data " . print_r($status, true));
264 // If it's a retweet, save it as a repeat!
266 if (!empty($status->retweeted_status)) {
267 common_log(LOG_INFO, "Status {$status->id} is a retweet of {$status->retweeted_status->id}.");
268 $original = $this->saveStatus($status->retweeted_status);
269 $repeat = $original->repeat($profile->id, 'twitter');
270 common_log(LOG_INFO, "Saved {$repeat->id} as a repeat of {$original->id}");
274 $notice = new Notice();
276 $notice->profile_id = $profile->id;
277 $notice->uri = $statusUri;
278 $notice->url = $statusUri;
279 $notice->created = strftime(
281 strtotime($status->created_at)
284 $notice->source = 'twitter';
286 $notice->reply_to = null;
288 if (!empty($status->in_reply_to_status_id)) {
289 common_log(LOG_INFO, "Status {$status->id} is a reply to status {$status->in_reply_to_status_id}");
290 $replyUri = $this->makeStatusURI($status->in_reply_to_screen_name, $status->in_reply_to_status_id);
291 $reply = Notice::staticGet('uri', $replyUri);
293 common_log(LOG_INFO, "Couldn't find local notice for status {$status->in_reply_to_status_id}");
295 common_log(LOG_INFO, "Found local notice {$reply->id} for status {$status->in_reply_to_status_id}");
296 $notice->reply_to = $reply->id;
297 $notice->conversation = $reply->conversation;
301 if (empty($notice->conversation)) {
302 $conv = Conversation::create();
303 $notice->conversation = $conv->id;
304 common_log(LOG_INFO, "No known conversation for status {$status->id} so making a new one {$conv->id}.");
307 $notice->is_local = Notice::GATEWAY;
309 $notice->content = common_shorten_links($status->text);
310 $notice->rendered = common_render_content(
315 if (Event::handle('StartNoticeSave', array(&$notice))) {
317 $id = $notice->insert();
320 common_log_db_error($notice, 'INSERT', __FILE__);
321 common_log(LOG_ERR, $this->name() .
322 ' - Problem saving notice.');
325 Event::handle('EndNoticeSave', array($notice));
328 $notice->blowOnInsert();
334 * Make an URI for a status.
336 * @param object $status status object
341 function makeStatusURI($username, $id)
343 return 'http://twitter.com/'
350 * Look up a Profile by profileurl field. Profile::staticGet() was
351 * not working consistently.
353 * @param string $nickname local nickname of the Twitter user
354 * @param string $profileurl the profile url
356 * @return mixed value the first Profile with that url, or null
359 function getProfileByUrl($nickname, $profileurl)
361 $profile = new Profile();
362 $profile->nickname = $nickname;
363 $profile->profileurl = $profileurl;
366 if ($profile->find()) {
375 * Check to see if this Twitter status has already been imported
377 * @param Profile $profile Twitter user's local profile
378 * @param string $statusUri URI of the status on Twitter
380 * @return mixed value a matching Notice or null
383 function checkDupe($profile, $statusUri)
385 $notice = new Notice();
386 $notice->uri = $statusUri;
387 $notice->profile_id = $profile->id;
390 if ($notice->find()) {
398 function ensureProfile($user)
400 // check to see if there's already a profile for this user
402 $profileurl = 'http://twitter.com/' . $user->screen_name;
403 $profile = $this->getProfileByUrl($user->screen_name, $profileurl);
405 if (!empty($profile)) {
406 common_debug($this->name() .
407 " - Profile for $profile->nickname found.");
409 // Check to see if the user's Avatar has changed
411 $this->checkAvatar($user, $profile);
416 common_debug($this->name() . ' - Adding profile and remote profile ' .
417 "for Twitter user: $profileurl.");
419 $profile = new Profile();
420 $profile->query("BEGIN");
422 $profile->nickname = $user->screen_name;
423 $profile->fullname = $user->name;
424 $profile->homepage = $user->url;
425 $profile->bio = $user->description;
426 $profile->location = $user->location;
427 $profile->profileurl = $profileurl;
428 $profile->created = common_sql_now();
431 $id = $profile->insert();
432 } catch(Exception $e) {
433 common_log(LOG_WARNING, $this->name . ' Couldn\'t insert profile - ' . $e->getMessage());
437 common_log_db_error($profile, 'INSERT', __FILE__);
438 $profile->query("ROLLBACK");
442 // check for remote profile
444 $remote_pro = Remote_profile::staticGet('uri', $profileurl);
446 if (empty($remote_pro)) {
448 $remote_pro = new Remote_profile();
450 $remote_pro->id = $id;
451 $remote_pro->uri = $profileurl;
452 $remote_pro->created = common_sql_now();
455 $rid = $remote_pro->insert();
456 } catch (Exception $e) {
457 common_log(LOG_WARNING, $this->name() . ' Couldn\'t save remote profile - ' . $e->getMessage());
461 common_log_db_error($profile, 'INSERT', __FILE__);
462 $profile->query("ROLLBACK");
467 $profile->query("COMMIT");
469 $this->saveAvatars($user, $id);
475 function checkAvatar($twitter_user, $profile)
479 $path_parts = pathinfo($twitter_user->profile_image_url);
481 $newname = 'Twitter_' . $twitter_user->id . '_' .
482 $path_parts['basename'];
484 $oldname = $profile->getAvatar(48)->filename;
486 if ($newname != $oldname) {
487 common_debug($this->name() . ' - Avatar for Twitter user ' .
488 "$profile->nickname has changed.");
489 common_debug($this->name() . " - old: $oldname new: $newname");
491 $this->updateAvatars($twitter_user, $profile);
494 if ($this->missingAvatarFile($profile)) {
495 common_debug($this->name() . ' - Twitter user ' .
497 ' is missing one or more local avatars.');
498 common_debug($this->name() ." - old: $oldname new: $newname");
500 $this->updateAvatars($twitter_user, $profile);
504 function updateAvatars($twitter_user, $profile) {
508 $path_parts = pathinfo($twitter_user->profile_image_url);
510 $img_root = substr($path_parts['basename'], 0, -11);
511 $ext = $path_parts['extension'];
512 $mediatype = $this->getMediatype($ext);
514 foreach (array('mini', 'normal', 'bigger') as $size) {
515 $url = $path_parts['dirname'] . '/' .
516 $img_root . '_' . $size . ".$ext";
517 $filename = 'Twitter_' . $twitter_user->id . '_' .
518 $img_root . "_$size.$ext";
520 $this->updateAvatar($profile->id, $size, $mediatype, $filename);
521 $this->fetchAvatar($url, $filename);
525 function missingAvatarFile($profile) {
526 foreach (array(24, 48, 73) as $size) {
527 $filename = $profile->getAvatar($size)->filename;
528 $avatarpath = Avatar::path($filename);
529 if (file_exists($avatarpath) == FALSE) {
536 function getMediatype($ext)
540 switch (strtolower($ext)) {
542 $mediatype = 'image/jpg';
545 $mediatype = 'image/gif';
548 $mediatype = 'image/png';
554 function saveAvatars($user, $id)
558 $path_parts = pathinfo($user->profile_image_url);
559 $ext = $path_parts['extension'];
560 $end = strlen('_normal' . $ext);
561 $img_root = substr($path_parts['basename'], 0, -($end+1));
562 $mediatype = $this->getMediatype($ext);
564 foreach (array('mini', 'normal', 'bigger') as $size) {
565 $url = $path_parts['dirname'] . '/' .
566 $img_root . '_' . $size . ".$ext";
567 $filename = 'Twitter_' . $user->id . '_' .
568 $img_root . "_$size.$ext";
570 if ($this->fetchAvatar($url, $filename)) {
571 $this->newAvatar($id, $size, $mediatype, $filename);
573 common_log(LOG_WARNING, $id() .
574 " - Problem fetching Avatar: $url");
579 function updateAvatar($profile_id, $size, $mediatype, $filename) {
581 common_debug($this->name() . " - Updating avatar: $size");
583 $profile = Profile::staticGet($profile_id);
585 if (empty($profile)) {
586 common_debug($this->name() . " - Couldn't get profile: $profile_id!");
590 $sizes = array('mini' => 24, 'normal' => 48, 'bigger' => 73);
591 $avatar = $profile->getAvatar($sizes[$size]);
593 // Delete the avatar, if present
599 $this->newAvatar($profile->id, $size, $mediatype, $filename);
602 function newAvatar($profile_id, $size, $mediatype, $filename)
606 $avatar = new Avatar();
607 $avatar->profile_id = $profile_id;
612 $avatar->height = 24;
616 $avatar->height = 48;
620 // Note: Twitter's big avatars are a different size than
621 // StatusNet's (StatusNet's = 96)
624 $avatar->height = 73;
627 $avatar->original = 0; // we don't have the original
628 $avatar->mediatype = $mediatype;
629 $avatar->filename = $filename;
630 $avatar->url = Avatar::url($filename);
632 $avatar->created = common_sql_now();
635 $id = $avatar->insert();
636 } catch (Exception $e) {
637 common_log(LOG_WARNING, $this->name() . ' Couldn\'t insert avatar - ' . $e->getMessage());
641 common_log_db_error($avatar, 'INSERT', __FILE__);
645 common_debug($this->name() .
646 " - Saved new $size avatar for $profile_id.");
652 * Fetch a remote avatar image and save to local storage.
654 * @param string $url avatar source URL
655 * @param string $filename bare local filename for download
656 * @return bool true on success, false on failure
658 function fetchAvatar($url, $filename)
660 common_debug($this->name() . " - Fetching Twitter avatar: $url");
662 $request = HTTPClient::start();
663 $response = $request->get($url);
664 if ($response->isOk()) {
665 $avatarfile = Avatar::path($filename);
666 $ok = file_put_contents($avatarfile, $response->getBody());
668 common_log(LOG_WARNING, $this->name() .
669 " - Couldn't open file $filename");
683 if (have_option('i')) {
684 $id = get_option_value('i');
685 } else if (have_option('--id')) {
686 $id = get_option_value('--id');
687 } else if (count($args) > 0) {
693 if (have_option('d') || have_option('debug')) {
697 $fetcher = new TwitterStatusFetcher($id, 60, 2, $debug);