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/twitteroauthclient.php';
46 * Fetch statuses from Twitter
48 * Fetches statuses from Twitter and inserts them as notices
50 * NOTE: an Avatar path MUST be set in config.php for this
51 * script to work, e.g.:
52 * $config['avatar']['path'] = $config['site']['path'] . '/avatar/';
54 * @todo @fixme @gar Fix the above. For some reason $_path is always empty when
55 * this script is run, so the default avatar path is always set wrong in
56 * default.php. Therefore it must be set explicitly in config.php. --Z
60 * @author Zach Copley <zach@status.net>
61 * @author Evan Prodromou <evan@status.net>
62 * @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
63 * @link http://status.net/
66 class TwitterStatusFetcher extends ParallelizingDaemon
71 * @param string $id the name/id of this daemon
72 * @param int $interval sleep this long before doing everything again
73 * @param int $max_children maximum number of child processes at a time
74 * @param boolean $debug debug output flag
79 function __construct($id = null, $interval = 60,
80 $max_children = 2, $debug = null)
82 parent::__construct($id, $interval, $max_children, $debug);
88 * @return string Name of the daemon.
93 return ('twitterstatusfetcher.'.$this->_id);
97 * Find all the Twitter foreign links for users who have requested
98 * importing of their friends' timelines
100 * @return array flinks an array of Foreign_link objects
103 function getObjects()
105 global $_DB_DATAOBJECT;
107 $flink = new Foreign_link();
108 $conn = &$flink->getDatabaseConnection();
110 $flink->service = TWITTER_SERVICE;
111 $flink->orderBy('last_noticesync');
116 while ($flink->fetch()) {
118 if (($flink->noticesync & FOREIGN_NOTICE_RECV) ==
119 FOREIGN_NOTICE_RECV) {
120 $flinks[] = clone($flink);
121 common_log(LOG_INFO, "sync: foreign id $flink->foreign_id");
123 common_log(LOG_INFO, "nothing to sync");
131 unset($_DB_DATAOBJECT['CONNECTIONS']);
136 function childTask($flink) {
138 // Each child ps needs its own DB connection
140 // Note: DataObject::getDatabaseConnection() creates
141 // a new connection if there isn't one already
143 $conn = &$flink->getDatabaseConnection();
145 $this->getTimeline($flink);
147 $flink->last_friendsync = common_sql_now();
152 // XXX: Couldn't find a less brutal way to blow
153 // away a cached connection
155 global $_DB_DATAOBJECT;
156 unset($_DB_DATAOBJECT['CONNECTIONS']);
159 function getTimeline($flink)
162 common_log(LOG_WARNING, $this->name() .
163 " - Can't retrieve Foreign_link for foreign ID $fid");
167 common_debug($this->name() . ' - Trying to get timeline for Twitter user ' .
170 // XXX: Biggest remaining issue - How do we know at which status
171 // to start importing? How many statuses? Right now I'm going
172 // with the default last 20.
176 if (TwitterOAuthClient::isPackedToken($flink->credentials)) {
177 $token = TwitterOAuthClient::unpackToken($flink->credentials);
178 $client = new TwitterOAuthClient($token->key, $token->secret);
179 common_debug($this->name() . ' - Grabbing friends timeline with OAuth.');
181 common_debug("Skipping friends timeline for $flink->foreign_id since not OAuth.");
187 $timeline = $client->statusesHomeTimeline();
188 } catch (Exception $e) {
189 common_log(LOG_WARNING, $this->name() .
190 ' - Twitter client unable to get friends timeline for user ' .
191 $flink->user_id . ' - code: ' .
192 $e->getCode() . 'msg: ' . $e->getMessage());
195 if (empty($timeline)) {
196 common_log(LOG_WARNING, $this->name() . " - Empty timeline.");
200 common_debug(LOG_INFO, $this->name() . ' - Retrieved ' . sizeof($timeline) . ' statuses from Twitter.');
202 // Reverse to preserve order
204 foreach (array_reverse($timeline) as $status) {
206 // Hacktastic: filter out stuff coming from this StatusNet
208 $source = mb_strtolower(common_config('integration', 'source'));
210 if (preg_match("/$source/", mb_strtolower($status->source))) {
211 common_debug($this->name() . ' - Skipping import of status ' .
212 $status->id . ' with source ' . $source);
216 // Don't save it if the user is protected
217 // FIXME: save it but treat it as private
219 if ($status->user->protected) {
223 $notice = $this->saveStatus($status);
225 if (!empty($notice)) {
226 Inbox::insertNotice($flink->user_id, $notice->id);
230 // Okay, record the time we synced with Twitter for posterity
232 $flink->last_noticesync = common_sql_now();
236 function saveStatus($status)
238 $profile = $this->ensureProfile($status->user);
240 if (empty($profile)) {
241 common_log(LOG_ERR, $this->name() .
242 ' - Problem saving notice. No associated Profile.');
246 $statusUri = $this->makeStatusURI($status->user->screen_name, $status->id);
248 // check to see if we've already imported the status
250 $n2s = Notice_to_status::staticGet('status_id', $status->id);
256 " - Ignoring duplicate import: {$status->id}"
258 return Notice::staticGet('id', $n2s->notice_id);
261 common_debug("Saving status {$status->id} with data " . print_r($status, true));
263 // If it's a retweet, save it as a repeat!
265 if (!empty($status->retweeted_status)) {
266 common_log(LOG_INFO, "Status {$status->id} is a retweet of {$status->retweeted_status->id}.");
267 $original = $this->saveStatus($status->retweeted_status);
268 if (empty($original)) {
271 $author = $original->getProfile();
272 // TRANS: Message used to repeat a notice. RT is the abbreviation of 'retweet'.
273 // TRANS: %1$s is the repeated user's name, %2$s is the repeated notice.
274 $content = sprintf(_('RT @%1$s %2$s'),
277 $repeat = Notice::saveNew($profile->id,
280 array('repeat_of' => $original->id,
281 'uri' => $statusUri));
282 common_log(LOG_INFO, "Saved {$repeat->id} as a repeat of {$original->id}");
283 Notice_to_status::saveNew($repeat->id, $status->id);
288 $notice = new Notice();
290 $notice->profile_id = $profile->id;
291 $notice->uri = $statusUri;
292 $notice->url = $statusUri;
293 $notice->created = strftime(
295 strtotime($status->created_at)
298 $notice->source = 'twitter';
300 $notice->reply_to = null;
302 if (!empty($status->in_reply_to_status_id)) {
303 common_log(LOG_INFO, "Status {$status->id} is a reply to status {$status->in_reply_to_status_id}");
304 $n2s = Notice_to_status::staticGet('status_id', $status->in_reply_to_status_id);
306 common_log(LOG_INFO, "Couldn't find local notice for status {$status->in_reply_to_status_id}");
308 $reply = Notice::staticGet('id', $n2s->notice_id);
310 common_log(LOG_INFO, "Couldn't find local notice for status {$status->in_reply_to_status_id}");
312 common_log(LOG_INFO, "Found local notice {$reply->id} for status {$status->in_reply_to_status_id}");
313 $notice->reply_to = $reply->id;
314 $notice->conversation = $reply->conversation;
319 if (empty($notice->conversation)) {
320 $conv = Conversation::create();
321 $notice->conversation = $conv->id;
322 common_log(LOG_INFO, "No known conversation for status {$status->id} so making a new one {$conv->id}.");
325 $notice->is_local = Notice::GATEWAY;
327 $notice->content = common_shorten_links($status->text);
328 $notice->rendered = common_render_content(
333 if (Event::handle('StartNoticeSave', array(&$notice))) {
335 $id = $notice->insert();
338 common_log_db_error($notice, 'INSERT', __FILE__);
339 common_log(LOG_ERR, $this->name() .
340 ' - Problem saving notice.');
343 Event::handle('EndNoticeSave', array($notice));
346 Notice_to_status::saveNew($notice->id, $status->id);
347 $notice->blowOnInsert();
353 * Make an URI for a status.
355 * @param object $status status object
360 function makeStatusURI($username, $id)
362 return 'http://twitter.com/'
369 * Look up a Profile by profileurl field. Profile::staticGet() was
370 * not working consistently.
372 * @param string $nickname local nickname of the Twitter user
373 * @param string $profileurl the profile url
375 * @return mixed value the first Profile with that url, or null
378 function getProfileByUrl($nickname, $profileurl)
380 $profile = new Profile();
381 $profile->nickname = $nickname;
382 $profile->profileurl = $profileurl;
385 if ($profile->find()) {
394 * Check to see if this Twitter status has already been imported
396 * @param Profile $profile Twitter user's local profile
397 * @param string $statusUri URI of the status on Twitter
399 * @return mixed value a matching Notice or null
402 function checkDupe($profile, $statusUri)
404 $notice = new Notice();
405 $notice->uri = $statusUri;
406 $notice->profile_id = $profile->id;
409 if ($notice->find()) {
417 function ensureProfile($user)
419 // check to see if there's already a profile for this user
421 $profileurl = 'http://twitter.com/' . $user->screen_name;
422 $profile = $this->getProfileByUrl($user->screen_name, $profileurl);
424 if (!empty($profile)) {
425 common_debug($this->name() .
426 " - Profile for $profile->nickname found.");
428 // Check to see if the user's Avatar has changed
430 $this->checkAvatar($user, $profile);
435 common_debug($this->name() . ' - Adding profile and remote profile ' .
436 "for Twitter user: $profileurl.");
438 $profile = new Profile();
439 $profile->query("BEGIN");
441 $profile->nickname = $user->screen_name;
442 $profile->fullname = $user->name;
443 $profile->homepage = $user->url;
444 $profile->bio = $user->description;
445 $profile->location = $user->location;
446 $profile->profileurl = $profileurl;
447 $profile->created = common_sql_now();
450 $id = $profile->insert();
451 } catch(Exception $e) {
452 common_log(LOG_WARNING, $this->name . ' Couldn\'t insert profile - ' . $e->getMessage());
456 common_log_db_error($profile, 'INSERT', __FILE__);
457 $profile->query("ROLLBACK");
461 // check for remote profile
463 $remote_pro = Remote_profile::staticGet('uri', $profileurl);
465 if (empty($remote_pro)) {
467 $remote_pro = new Remote_profile();
469 $remote_pro->id = $id;
470 $remote_pro->uri = $profileurl;
471 $remote_pro->created = common_sql_now();
474 $rid = $remote_pro->insert();
475 } catch (Exception $e) {
476 common_log(LOG_WARNING, $this->name() . ' Couldn\'t save remote profile - ' . $e->getMessage());
480 common_log_db_error($profile, 'INSERT', __FILE__);
481 $profile->query("ROLLBACK");
486 $profile->query("COMMIT");
488 $this->saveAvatars($user, $id);
494 function checkAvatar($twitter_user, $profile)
498 $path_parts = pathinfo($twitter_user->profile_image_url);
500 $newname = 'Twitter_' . $twitter_user->id . '_' .
501 $path_parts['basename'];
503 $oldname = $profile->getAvatar(48)->filename;
505 if ($newname != $oldname) {
506 common_debug($this->name() . ' - Avatar for Twitter user ' .
507 "$profile->nickname has changed.");
508 common_debug($this->name() . " - old: $oldname new: $newname");
510 $this->updateAvatars($twitter_user, $profile);
513 if ($this->missingAvatarFile($profile)) {
514 common_debug($this->name() . ' - Twitter user ' .
516 ' is missing one or more local avatars.');
517 common_debug($this->name() ." - old: $oldname new: $newname");
519 $this->updateAvatars($twitter_user, $profile);
523 function updateAvatars($twitter_user, $profile) {
527 $path_parts = pathinfo($twitter_user->profile_image_url);
529 $img_root = substr($path_parts['basename'], 0, -11);
530 $ext = $path_parts['extension'];
531 $mediatype = $this->getMediatype($ext);
533 foreach (array('mini', 'normal', 'bigger') as $size) {
534 $url = $path_parts['dirname'] . '/' .
535 $img_root . '_' . $size . ".$ext";
536 $filename = 'Twitter_' . $twitter_user->id . '_' .
537 $img_root . "_$size.$ext";
539 $this->updateAvatar($profile->id, $size, $mediatype, $filename);
540 $this->fetchAvatar($url, $filename);
544 function missingAvatarFile($profile) {
545 foreach (array(24, 48, 73) as $size) {
546 $filename = $profile->getAvatar($size)->filename;
547 $avatarpath = Avatar::path($filename);
548 if (file_exists($avatarpath) == FALSE) {
555 function getMediatype($ext)
559 switch (strtolower($ext)) {
561 $mediatype = 'image/jpg';
564 $mediatype = 'image/gif';
567 $mediatype = 'image/png';
573 function saveAvatars($user, $id)
577 $path_parts = pathinfo($user->profile_image_url);
578 $ext = $path_parts['extension'];
579 $end = strlen('_normal' . $ext);
580 $img_root = substr($path_parts['basename'], 0, -($end+1));
581 $mediatype = $this->getMediatype($ext);
583 foreach (array('mini', 'normal', 'bigger') as $size) {
584 $url = $path_parts['dirname'] . '/' .
585 $img_root . '_' . $size . ".$ext";
586 $filename = 'Twitter_' . $user->id . '_' .
587 $img_root . "_$size.$ext";
589 if ($this->fetchAvatar($url, $filename)) {
590 $this->newAvatar($id, $size, $mediatype, $filename);
592 common_log(LOG_WARNING, $id() .
593 " - Problem fetching Avatar: $url");
598 function updateAvatar($profile_id, $size, $mediatype, $filename) {
600 common_debug($this->name() . " - Updating avatar: $size");
602 $profile = Profile::staticGet($profile_id);
604 if (empty($profile)) {
605 common_debug($this->name() . " - Couldn't get profile: $profile_id!");
609 $sizes = array('mini' => 24, 'normal' => 48, 'bigger' => 73);
610 $avatar = $profile->getAvatar($sizes[$size]);
612 // Delete the avatar, if present
618 $this->newAvatar($profile->id, $size, $mediatype, $filename);
621 function newAvatar($profile_id, $size, $mediatype, $filename)
625 $avatar = new Avatar();
626 $avatar->profile_id = $profile_id;
631 $avatar->height = 24;
635 $avatar->height = 48;
639 // Note: Twitter's big avatars are a different size than
640 // StatusNet's (StatusNet's = 96)
643 $avatar->height = 73;
646 $avatar->original = 0; // we don't have the original
647 $avatar->mediatype = $mediatype;
648 $avatar->filename = $filename;
649 $avatar->url = Avatar::url($filename);
651 $avatar->created = common_sql_now();
654 $id = $avatar->insert();
655 } catch (Exception $e) {
656 common_log(LOG_WARNING, $this->name() . ' Couldn\'t insert avatar - ' . $e->getMessage());
660 common_log_db_error($avatar, 'INSERT', __FILE__);
664 common_debug($this->name() .
665 " - Saved new $size avatar for $profile_id.");
671 * Fetch a remote avatar image and save to local storage.
673 * @param string $url avatar source URL
674 * @param string $filename bare local filename for download
675 * @return bool true on success, false on failure
677 function fetchAvatar($url, $filename)
679 common_debug($this->name() . " - Fetching Twitter avatar: $url");
681 $request = HTTPClient::start();
682 $response = $request->get($url);
683 if ($response->isOk()) {
684 $avatarfile = Avatar::path($filename);
685 $ok = file_put_contents($avatarfile, $response->getBody());
687 common_log(LOG_WARNING, $this->name() .
688 " - Couldn't open file $filename");
702 if (have_option('i')) {
703 $id = get_option_value('i');
704 } else if (have_option('--id')) {
705 $id = get_option_value('--id');
706 } else if (count($args) > 0) {
712 if (have_option('d') || have_option('debug')) {
716 $fetcher = new TwitterStatusFetcher($id, 60, 2, $debug);