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$s'),
278 if (Notice::contentTooLong($content)) {
279 $contentlimit = Notice::maxContent();
280 $content = mb_substr($content, 0, $contentlimit - 4) . ' ...';
283 $repeat = Notice::saveNew($profile->id,
286 array('repeat_of' => $original->id,
288 'is_local' => Notice::GATEWAY));
289 common_log(LOG_INFO, "Saved {$repeat->id} as a repeat of {$original->id}");
290 Notice_to_status::saveNew($repeat->id, $status->id);
295 $notice = new Notice();
297 $notice->profile_id = $profile->id;
298 $notice->uri = $statusUri;
299 $notice->url = $statusUri;
300 $notice->created = strftime(
302 strtotime($status->created_at)
305 $notice->source = 'twitter';
307 $notice->reply_to = null;
309 if (!empty($status->in_reply_to_status_id)) {
310 common_log(LOG_INFO, "Status {$status->id} is a reply to status {$status->in_reply_to_status_id}");
311 $n2s = Notice_to_status::staticGet('status_id', $status->in_reply_to_status_id);
313 common_log(LOG_INFO, "Couldn't find local notice for status {$status->in_reply_to_status_id}");
315 $reply = Notice::staticGet('id', $n2s->notice_id);
317 common_log(LOG_INFO, "Couldn't find local notice for status {$status->in_reply_to_status_id}");
319 common_log(LOG_INFO, "Found local notice {$reply->id} for status {$status->in_reply_to_status_id}");
320 $notice->reply_to = $reply->id;
321 $notice->conversation = $reply->conversation;
326 if (empty($notice->conversation)) {
327 $conv = Conversation::create();
328 $notice->conversation = $conv->id;
329 common_log(LOG_INFO, "No known conversation for status {$status->id} so making a new one {$conv->id}.");
332 $notice->is_local = Notice::GATEWAY;
334 $notice->content = common_shorten_links($status->text);
335 $notice->rendered = common_render_content(
340 if (Event::handle('StartNoticeSave', array(&$notice))) {
342 $id = $notice->insert();
345 common_log_db_error($notice, 'INSERT', __FILE__);
346 common_log(LOG_ERR, $this->name() .
347 ' - Problem saving notice.');
350 Event::handle('EndNoticeSave', array($notice));
353 Notice_to_status::saveNew($notice->id, $status->id);
354 $notice->blowOnInsert();
360 * Make an URI for a status.
362 * @param object $status status object
367 function makeStatusURI($username, $id)
369 return 'http://twitter.com/'
376 * Look up a Profile by profileurl field. Profile::staticGet() was
377 * not working consistently.
379 * @param string $nickname local nickname of the Twitter user
380 * @param string $profileurl the profile url
382 * @return mixed value the first Profile with that url, or null
385 function getProfileByUrl($nickname, $profileurl)
387 $profile = new Profile();
388 $profile->nickname = $nickname;
389 $profile->profileurl = $profileurl;
392 if ($profile->find()) {
401 * Check to see if this Twitter status has already been imported
403 * @param Profile $profile Twitter user's local profile
404 * @param string $statusUri URI of the status on Twitter
406 * @return mixed value a matching Notice or null
409 function checkDupe($profile, $statusUri)
411 $notice = new Notice();
412 $notice->uri = $statusUri;
413 $notice->profile_id = $profile->id;
416 if ($notice->find()) {
424 function ensureProfile($user)
426 // check to see if there's already a profile for this user
428 $profileurl = 'http://twitter.com/' . $user->screen_name;
429 $profile = $this->getProfileByUrl($user->screen_name, $profileurl);
431 if (!empty($profile)) {
432 common_debug($this->name() .
433 " - Profile for $profile->nickname found.");
435 // Check to see if the user's Avatar has changed
437 $this->checkAvatar($user, $profile);
442 common_debug($this->name() . ' - Adding profile and remote profile ' .
443 "for Twitter user: $profileurl.");
445 $profile = new Profile();
446 $profile->query("BEGIN");
448 $profile->nickname = $user->screen_name;
449 $profile->fullname = $user->name;
450 $profile->homepage = $user->url;
451 $profile->bio = $user->description;
452 $profile->location = $user->location;
453 $profile->profileurl = $profileurl;
454 $profile->created = common_sql_now();
457 $id = $profile->insert();
458 } catch(Exception $e) {
459 common_log(LOG_WARNING, $this->name . ' Couldn\'t insert profile - ' . $e->getMessage());
463 common_log_db_error($profile, 'INSERT', __FILE__);
464 $profile->query("ROLLBACK");
468 // check for remote profile
470 $remote_pro = Remote_profile::staticGet('uri', $profileurl);
472 if (empty($remote_pro)) {
474 $remote_pro = new Remote_profile();
476 $remote_pro->id = $id;
477 $remote_pro->uri = $profileurl;
478 $remote_pro->created = common_sql_now();
481 $rid = $remote_pro->insert();
482 } catch (Exception $e) {
483 common_log(LOG_WARNING, $this->name() . ' Couldn\'t save remote profile - ' . $e->getMessage());
487 common_log_db_error($profile, 'INSERT', __FILE__);
488 $profile->query("ROLLBACK");
493 $profile->query("COMMIT");
495 $this->saveAvatars($user, $id);
501 function checkAvatar($twitter_user, $profile)
505 $path_parts = pathinfo($twitter_user->profile_image_url);
507 $newname = 'Twitter_' . $twitter_user->id . '_' .
508 $path_parts['basename'];
510 $oldname = $profile->getAvatar(48)->filename;
512 if ($newname != $oldname) {
513 common_debug($this->name() . ' - Avatar for Twitter user ' .
514 "$profile->nickname has changed.");
515 common_debug($this->name() . " - old: $oldname new: $newname");
517 $this->updateAvatars($twitter_user, $profile);
520 if ($this->missingAvatarFile($profile)) {
521 common_debug($this->name() . ' - Twitter user ' .
523 ' is missing one or more local avatars.');
524 common_debug($this->name() ." - old: $oldname new: $newname");
526 $this->updateAvatars($twitter_user, $profile);
530 function updateAvatars($twitter_user, $profile) {
534 $path_parts = pathinfo($twitter_user->profile_image_url);
536 $img_root = substr($path_parts['basename'], 0, -11);
537 $ext = $path_parts['extension'];
538 $mediatype = $this->getMediatype($ext);
540 foreach (array('mini', 'normal', 'bigger') as $size) {
541 $url = $path_parts['dirname'] . '/' .
542 $img_root . '_' . $size . ".$ext";
543 $filename = 'Twitter_' . $twitter_user->id . '_' .
544 $img_root . "_$size.$ext";
546 $this->updateAvatar($profile->id, $size, $mediatype, $filename);
547 $this->fetchAvatar($url, $filename);
551 function missingAvatarFile($profile) {
552 foreach (array(24, 48, 73) as $size) {
553 $filename = $profile->getAvatar($size)->filename;
554 $avatarpath = Avatar::path($filename);
555 if (file_exists($avatarpath) == FALSE) {
562 function getMediatype($ext)
566 switch (strtolower($ext)) {
568 $mediatype = 'image/jpg';
571 $mediatype = 'image/gif';
574 $mediatype = 'image/png';
580 function saveAvatars($user, $id)
584 $path_parts = pathinfo($user->profile_image_url);
585 $ext = $path_parts['extension'];
586 $end = strlen('_normal' . $ext);
587 $img_root = substr($path_parts['basename'], 0, -($end+1));
588 $mediatype = $this->getMediatype($ext);
590 foreach (array('mini', 'normal', 'bigger') as $size) {
591 $url = $path_parts['dirname'] . '/' .
592 $img_root . '_' . $size . ".$ext";
593 $filename = 'Twitter_' . $user->id . '_' .
594 $img_root . "_$size.$ext";
596 if ($this->fetchAvatar($url, $filename)) {
597 $this->newAvatar($id, $size, $mediatype, $filename);
599 common_log(LOG_WARNING, $id() .
600 " - Problem fetching Avatar: $url");
605 function updateAvatar($profile_id, $size, $mediatype, $filename) {
607 common_debug($this->name() . " - Updating avatar: $size");
609 $profile = Profile::staticGet($profile_id);
611 if (empty($profile)) {
612 common_debug($this->name() . " - Couldn't get profile: $profile_id!");
616 $sizes = array('mini' => 24, 'normal' => 48, 'bigger' => 73);
617 $avatar = $profile->getAvatar($sizes[$size]);
619 // Delete the avatar, if present
625 $this->newAvatar($profile->id, $size, $mediatype, $filename);
628 function newAvatar($profile_id, $size, $mediatype, $filename)
632 $avatar = new Avatar();
633 $avatar->profile_id = $profile_id;
638 $avatar->height = 24;
642 $avatar->height = 48;
646 // Note: Twitter's big avatars are a different size than
647 // StatusNet's (StatusNet's = 96)
650 $avatar->height = 73;
653 $avatar->original = 0; // we don't have the original
654 $avatar->mediatype = $mediatype;
655 $avatar->filename = $filename;
656 $avatar->url = Avatar::url($filename);
658 $avatar->created = common_sql_now();
661 $id = $avatar->insert();
662 } catch (Exception $e) {
663 common_log(LOG_WARNING, $this->name() . ' Couldn\'t insert avatar - ' . $e->getMessage());
667 common_log_db_error($avatar, 'INSERT', __FILE__);
671 common_debug($this->name() .
672 " - Saved new $size avatar for $profile_id.");
678 * Fetch a remote avatar image and save to local storage.
680 * @param string $url avatar source URL
681 * @param string $filename bare local filename for download
682 * @return bool true on success, false on failure
684 function fetchAvatar($url, $filename)
686 common_debug($this->name() . " - Fetching Twitter avatar: $url");
688 $request = HTTPClient::start();
689 $response = $request->get($url);
690 if ($response->isOk()) {
691 $avatarfile = Avatar::path($filename);
692 $ok = file_put_contents($avatarfile, $response->getBody());
694 common_log(LOG_WARNING, $this->name() .
695 " - Couldn't open file $filename");
709 if (have_option('i')) {
710 $id = get_option_value('i');
711 } else if (have_option('--id')) {
712 $id = get_option_value('--id');
713 } else if (count($args) > 0) {
719 if (have_option('d') || have_option('debug')) {
723 $fetcher = new TwitterStatusFetcher($id, 60, 2, $debug);