3 * @copyright Copyright (C) 2020, Friendica
5 * @license GNU AGPL version 3 or any later version
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
9 * published by the Free Software Foundation, either version 3 of the
10 * License, or (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 <https://www.gnu.org/licenses/>.
22 namespace Friendica\Protocol;
26 use Friendica\Content\PageInfo;
27 use Friendica\Content\Text\BBCode;
28 use Friendica\Content\Text\HTML;
29 use Friendica\Core\Cache\Duration;
30 use Friendica\Core\Logger;
31 use Friendica\Core\Protocol;
32 use Friendica\Database\DBA;
34 use Friendica\Model\Contact;
35 use Friendica\Model\Item;
36 use Friendica\Model\Post;
37 use Friendica\Model\Tag;
38 use Friendica\Model\User;
39 use Friendica\Util\DateTimeFormat;
40 use Friendica\Util\Network;
41 use Friendica\Util\ParseUrl;
42 use Friendica\Util\Strings;
43 use Friendica\Util\XML;
46 * This class contain functions to import feeds (RSS/RDF/Atom)
51 * consume - process atom feed and update anything/everything we might need to update
53 * $xml = the (atom) feed to consume - RSS isn't as fully supported but may work for simple feeds.
55 * $importer = the contact_record (joined to user_record) of the local user who owns this relationship.
56 * It is this person's stuff that is going to be updated.
57 * $contact = the person who is sending us stuff. If not set, we MAY be processing a "follow" activity
58 * from an external network and MAY create an appropriate contact record. Otherwise, we MUST
59 * have a contact record.
60 * $hub = should we find a hub declation in the feed, pass it back to our calling process, who might (or
61 * might not) try and subscribe to it.
62 * $datedir sorts in reverse order
63 * $pass - by default ($pass = 0) we cannot guarantee that a parent item has been
64 * imported prior to its children being seen in the stream unless we are certain
65 * of how the feed is arranged/ordered.
66 * With $pass = 1, we only pull parent items out of the stream.
67 * With $pass = 2, we only pull children (comments/likes).
69 * So running this twice, first with pass 1 and then with pass 2 will do the right
70 * thing regardless of feed ordering. This won't be adequate in a fully-threaded
71 * model where comments can have sub-threads. That would require some massive sorting
72 * to get all the feed items into a mostly linear ordering, and might still require
76 * @param array $importer
77 * @param array $contact
79 * @throws ImagickException
80 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
82 public static function consume($xml, array $importer, array $contact, &$hub)
84 if ($contact['network'] === Protocol::OSTATUS) {
85 Logger::info('Consume OStatus messages');
86 OStatus::import($xml, $importer, $contact, $hub);
91 if ($contact['network'] === Protocol::FEED) {
92 Logger::info('Consume feeds');
93 self::import($xml, $importer, $contact);
98 if ($contact['network'] === Protocol::DFRN) {
99 Logger::info('Consume DFRN messages');
100 $dfrn_importer = DFRN::getImporter($contact['id'], $importer['uid']);
101 if (!empty($dfrn_importer)) {
102 Logger::info('Now import the DFRN feed');
103 DFRN::import($xml, $dfrn_importer, true);
110 * Read a RSS/RDF/Atom feed and create an item entry for it
112 * @param string $xml The feed data
113 * @param array $importer The user record of the importer
114 * @param array $contact The contact record of the feed
116 * @return array Returns the header and the first item in dry run mode
117 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
119 public static function import($xml, array $importer = [], array $contact = [])
121 $dryRun = empty($importer) && empty($contact);
124 Logger::info("Test Atom/RSS feed");
126 Logger::info("Import Atom/RSS feed '" . $contact["name"] . "' (Contact " . $contact["id"] . ") for user " . $importer["uid"]);
130 Logger::info('XML is empty.');
134 if (!empty($contact['poll'])) {
135 $basepath = $contact['poll'];
136 } elseif (!empty($contact['url'])) {
137 $basepath = $contact['url'];
142 $doc = new DOMDocument();
143 @$doc->loadXML(trim($xml));
144 $xpath = new DOMXPath($doc);
145 $xpath->registerNamespace('atom', ActivityNamespace::ATOM1);
146 $xpath->registerNamespace('dc', "http://purl.org/dc/elements/1.1/");
147 $xpath->registerNamespace('content', "http://purl.org/rss/1.0/modules/content/");
148 $xpath->registerNamespace('rdf', "http://www.w3.org/1999/02/22-rdf-syntax-ns#");
149 $xpath->registerNamespace('rss', "http://purl.org/rss/1.0/");
150 $xpath->registerNamespace('media', "http://search.yahoo.com/mrss/");
151 $xpath->registerNamespace('poco', ActivityNamespace::POCO);
157 if ($xpath->query('/rdf:RDF/rss:channel')->length > 0) {
158 $author["author-link"] = XML::getFirstNodeValue($xpath, '/rdf:RDF/rss:channel/rss:link/text()');
159 $author["author-name"] = XML::getFirstNodeValue($xpath, '/rdf:RDF/rss:channel/rss:title/text()');
161 if (empty($author["author-name"])) {
162 $author["author-name"] = XML::getFirstNodeValue($xpath, '/rdf:RDF/rss:channel/rss:description/text()');
164 $entries = $xpath->query('/rdf:RDF/rss:item');
168 if ($xpath->query('/atom:feed')->length > 0) {
169 $alternate = XML::getFirstAttributes($xpath, "atom:link[@rel='alternate']");
170 if (is_object($alternate)) {
171 foreach ($alternate AS $attribute) {
172 if ($attribute->name == "href") {
173 $author["author-link"] = $attribute->textContent;
178 if (empty($author["author-link"])) {
179 $self = XML::getFirstAttributes($xpath, "atom:link[@rel='self']");
180 if (is_object($self)) {
181 foreach ($self AS $attribute) {
182 if ($attribute->name == "href") {
183 $author["author-link"] = $attribute->textContent;
189 if (empty($author["author-link"])) {
190 $author["author-link"] = XML::getFirstNodeValue($xpath, '/atom:feed/atom:id/text()');
192 $author["author-avatar"] = XML::getFirstNodeValue($xpath, '/atom:feed/atom:logo/text()');
194 $author["author-name"] = XML::getFirstNodeValue($xpath, '/atom:feed/atom:title/text()');
196 if (empty($author["author-name"])) {
197 $author["author-name"] = XML::getFirstNodeValue($xpath, '/atom:feed/atom:subtitle/text()');
200 if (empty($author["author-name"])) {
201 $author["author-name"] = XML::getFirstNodeValue($xpath, '/atom:feed/atom:author/atom:name/text()');
204 $value = XML::getFirstNodeValue($xpath, 'atom:author/poco:displayName/text()');
206 $author["author-name"] = $value;
210 $author["author-id"] = XML::getFirstNodeValue($xpath, '/atom:feed/atom:author/atom:id/text()');
212 // See https://tools.ietf.org/html/rfc4287#section-3.2.2
213 $value = XML::getFirstNodeValue($xpath, 'atom:author/atom:uri/text()');
215 $author["author-link"] = $value;
218 $value = XML::getFirstNodeValue($xpath, 'atom:author/poco:preferredUsername/text()');
220 $author["author-nick"] = $value;
223 $value = XML::getFirstNodeValue($xpath, 'atom:author/poco:address/poco:formatted/text()');
225 $author["author-location"] = $value;
228 $value = XML::getFirstNodeValue($xpath, 'atom:author/poco:note/text()');
230 $author["author-about"] = $value;
233 $avatar = XML::getFirstAttributes($xpath, "atom:author/atom:link[@rel='avatar']");
234 if (is_object($avatar)) {
235 foreach ($avatar AS $attribute) {
236 if ($attribute->name == "href") {
237 $author["author-avatar"] = $attribute->textContent;
243 $author["edited"] = $author["created"] = XML::getFirstNodeValue($xpath, '/atom:feed/atom:updated/text()');
245 $author["app"] = XML::getFirstNodeValue($xpath, '/atom:feed/atom:generator/text()');
247 $entries = $xpath->query('/atom:feed/atom:entry');
251 if ($xpath->query('/rss/channel')->length > 0) {
252 $author["author-link"] = XML::getFirstNodeValue($xpath, '/rss/channel/link/text()');
254 $author["author-name"] = XML::getFirstNodeValue($xpath, '/rss/channel/title/text()');
255 $author["author-avatar"] = XML::getFirstNodeValue($xpath, '/rss/channel/image/url/text()');
257 if (empty($author["author-name"])) {
258 $author["author-name"] = XML::getFirstNodeValue($xpath, '/rss/channel/copyright/text()');
261 if (empty($author["author-name"])) {
262 $author["author-name"] = XML::getFirstNodeValue($xpath, '/rss/channel/description/text()');
265 $author["edited"] = $author["created"] = XML::getFirstNodeValue($xpath, '/rss/channel/pubDate/text()');
267 $author["app"] = XML::getFirstNodeValue($xpath, '/rss/channel/generator/text()');
269 $entries = $xpath->query('/rss/channel/item');
273 $author["author-link"] = $contact["url"];
275 if (empty($author["author-name"])) {
276 $author["author-name"] = $contact["name"];
279 $author["author-avatar"] = $contact["thumb"];
281 $author["owner-link"] = $contact["url"];
282 $author["owner-name"] = $contact["name"];
283 $author["owner-avatar"] = $contact["thumb"];
287 $header["uid"] = $importer["uid"] ?? 0;
288 $header["network"] = Protocol::FEED;
290 $header["origin"] = 0;
291 $header["gravity"] = GRAVITY_PARENT;
292 $header["private"] = Item::PUBLIC;
293 $header["verb"] = Activity::POST;
294 $header["object-type"] = Activity\ObjectType::NOTE;
296 $header["contact-id"] = $contact["id"] ?? 0;
298 if (!is_object($entries)) {
299 Logger::info("There are no entries in this feed.");
304 $creation_dates = [];
306 // Limit the number of items that are about to be fetched
307 $total_items = ($entries->length - 1);
308 $max_items = DI::config()->get('system', 'max_feed_items');
309 if (($max_items > 0) && ($total_items > $max_items)) {
310 $total_items = $max_items;
313 // Importing older entries first
314 for ($i = $total_items; $i >= 0; --$i) {
315 $entry = $entries->item($i);
317 $item = array_merge($header, $author);
319 $alternate = XML::getFirstAttributes($xpath, "atom:link[@rel='alternate']", $entry);
320 if (!is_object($alternate)) {
321 $alternate = XML::getFirstAttributes($xpath, "atom:link", $entry);
323 if (is_object($alternate)) {
324 foreach ($alternate AS $attribute) {
325 if ($attribute->name == "href") {
326 $item["plink"] = $attribute->textContent;
331 if (empty($item["plink"])) {
332 $item["plink"] = XML::getFirstNodeValue($xpath, 'link/text()', $entry);
335 if (empty($item["plink"])) {
336 $item["plink"] = XML::getFirstNodeValue($xpath, 'rss:link/text()', $entry);
339 $item["uri"] = XML::getFirstNodeValue($xpath, 'atom:id/text()', $entry);
341 if (empty($item["uri"])) {
342 $item["uri"] = XML::getFirstNodeValue($xpath, 'guid/text()', $entry);
345 if (empty($item["uri"])) {
346 $item["uri"] = $item["plink"];
349 // Add the base path if missing
350 $item["uri"] = Network::addBasePath($item["uri"], $basepath);
351 $item["plink"] = Network::addBasePath($item["plink"], $basepath);
353 $orig_plink = $item["plink"];
355 $item["plink"] = DI::httpRequest()->finalUrl($item["plink"]);
357 $item["parent-uri"] = $item["uri"];
359 $item["title"] = XML::getFirstNodeValue($xpath, 'atom:title/text()', $entry);
361 if (empty($item["title"])) {
362 $item["title"] = XML::getFirstNodeValue($xpath, 'title/text()', $entry);
364 if (empty($item["title"])) {
365 $item["title"] = XML::getFirstNodeValue($xpath, 'rss:title/text()', $entry);
368 $item["title"] = html_entity_decode($item["title"], ENT_QUOTES, 'UTF-8');
370 $published = XML::getFirstNodeValue($xpath, 'atom:published/text()', $entry);
372 if (empty($published)) {
373 $published = XML::getFirstNodeValue($xpath, 'pubDate/text()', $entry);
376 if (empty($published)) {
377 $published = XML::getFirstNodeValue($xpath, 'dc:date/text()', $entry);
380 $updated = XML::getFirstNodeValue($xpath, 'atom:updated/text()', $entry);
382 if (empty($updated) && !empty($published)) {
383 $updated = $published;
386 if (empty($published) && !empty($updated)) {
387 $published = $updated;
390 if ($published != "") {
391 $item["created"] = $published;
394 if ($updated != "") {
395 $item["edited"] = $updated;
399 $condition = ["`uid` = ? AND `uri` = ? AND `network` IN (?, ?)",
400 $importer["uid"], $item["uri"], Protocol::FEED, Protocol::DFRN];
401 $previous = Item::selectFirst(['id', 'created'], $condition);
402 if (DBA::isResult($previous)) {
403 // Use the creation date when the post had been stored. It can happen this date changes in the feed.
404 $creation_dates[] = $previous['created'];
405 Logger::info("Item with uri " . $item["uri"] . " for user " . $importer["uid"] . " already existed under id " . $previous["id"]);
408 $creation_dates[] = DateTimeFormat::utc($item['created']);
411 $creator = XML::getFirstNodeValue($xpath, 'author/text()', $entry);
413 if (empty($creator)) {
414 $creator = XML::getFirstNodeValue($xpath, 'atom:author/atom:name/text()', $entry);
417 if (empty($creator)) {
418 $creator = XML::getFirstNodeValue($xpath, 'dc:creator/text()', $entry);
421 if ($creator != "") {
422 $item["author-name"] = $creator;
425 $creator = XML::getFirstNodeValue($xpath, 'dc:creator/text()', $entry);
427 if ($creator != "") {
428 $item["author-name"] = $creator;
432 // <category>Ausland</category>
433 // <media:thumbnail width="152" height="76" url="http://www.taz.de/picture/667875/192/14388767.jpg"/>
437 $enclosures = $xpath->query("enclosure|atom:link[@rel='enclosure']", $entry);
438 foreach ($enclosures AS $enclosure) {
443 foreach ($enclosure->attributes AS $attribute) {
444 if (in_array($attribute->name, ["url", "href"])) {
445 $href = $attribute->textContent;
446 } elseif ($attribute->name == "length") {
447 $length = $attribute->textContent;
448 } elseif ($attribute->name == "type") {
449 $type = $attribute->textContent;
453 if (!empty($item["attach"])) {
454 $item["attach"] .= ',';
456 $item["attach"] = '';
459 $attachments[] = ["link" => $href, "type" => $type, "length" => $length];
461 $item["attach"] .= Post\Media::getAttachElement($href, $length, $type);
465 $categories = $xpath->query("category", $entry);
466 foreach ($categories AS $category) {
467 $taglist[] = $category->nodeValue;
470 $body = trim(XML::getFirstNodeValue($xpath, 'atom:content/text()', $entry));
473 $body = trim(XML::getFirstNodeValue($xpath, 'content:encoded/text()', $entry));
476 $summary = trim(XML::getFirstNodeValue($xpath, 'atom:summary/text()', $entry));
478 if (empty($summary)) {
479 $summary = trim(XML::getFirstNodeValue($xpath, 'description/text()', $entry));
487 if ($body == $summary) {
491 // remove the content of the title if it is identically to the body
492 // This helps with auto generated titles e.g. from tumblr
493 if (self::titleIsBody($item["title"], $body)) {
496 $item["body"] = HTML::toBBCode($body, $basepath);
498 // Remove tracking pixels
499 $item["body"] = preg_replace("/\[img=1x1\]([^\[\]]*)\[\/img\]/Usi", '', $item["body"]);
501 if (($item["body"] == '') && ($item["title"] != '')) {
502 $item["body"] = $item["title"];
509 } elseif (!Item::isValid($item)) {
510 Logger::info('Feed is invalid', ['created' => $item['created'], 'uid' => $item['uid'], 'uri' => $item['uri']]);
515 if (!empty($contact["fetch_further_information"]) && ($contact["fetch_further_information"] < 3)) {
516 // Handle enclosures and treat them as preview picture
517 foreach ($attachments AS $attachment) {
518 if ($attachment["type"] == "image/jpeg") {
519 $preview = $attachment["link"];
523 // Remove a possible link to the item itself
524 $item["body"] = str_replace($item["plink"], '', $item["body"]);
525 $item["body"] = trim(preg_replace('/\[url\=\](\w+.*?)\[\/url\]/i', '', $item["body"]));
527 // Replace the content when the title is longer than the body
528 $replace = (strlen($item["title"]) > strlen($item["body"]));
530 // Replace it, when there is an image in the body
531 if (strstr($item["body"], '[/img]')) {
535 // Replace it, when there is a link in the body
536 if (strstr($item["body"], '[/url]')) {
540 $saved_body = $item["body"];
541 $saved_title = $item["title"];
544 $item["body"] = trim($item["title"]);
547 $data = ParseUrl::getSiteinfoCached($item['plink'], true);
548 if (!empty($data['text']) && !empty($data['title']) && (mb_strlen($item['body']) < mb_strlen($data['text']))) {
549 // When the fetched page info text is longer than the body, we do try to enhance the body
550 if (!empty($item['body']) && (strpos($data['title'], $item['body']) === false) && (strpos($data['text'], $item['body']) === false)) {
551 // The body is not part of the fetched page info title or page info text. So we add the text to the body
552 $item['body'] .= "\n\n" . $data['text'];
554 // Else we replace the body with the page info text
555 $item['body'] = $data['text'];
559 $data = PageInfo::queryUrl($item["plink"], false, $preview, ($contact["fetch_further_information"] == 2), $contact["ffi_keyword_denylist"] ?? '');
561 // Take the data that was provided by the feed if the query is empty
562 if (($data['type'] == 'link') && empty($data['title']) && empty($data['text'])) {
563 $data['title'] = $saved_title;
564 $item["body"] = $saved_body;
567 $data_text = strip_tags(trim($data['text'] ?? ''));
568 $item_body = strip_tags(trim($item['body'] ?? ''));
570 if (!empty($data_text) && (($data_text == $item_body) || strstr($item_body, $data_text))) {
574 // We always strip the title since it will be added in the page information
576 $item["body"] = $item["body"] . "\n" . PageInfo::getFooterFromData($data, false);
577 $taglist = $contact["fetch_further_information"] == 2 ? PageInfo::getTagsFromUrl($item["plink"], $preview, $contact["ffi_keyword_denylist"] ?? '') : [];
578 $item["object-type"] = Activity\ObjectType::BOOKMARK;
579 unset($item["attach"]);
581 if (!empty($summary)) {
582 $item["body"] = '[abstract]' . HTML::toBBCode($summary, $basepath) . "[/abstract]\n" . $item["body"];
585 if (!empty($contact["fetch_further_information"]) && ($contact["fetch_further_information"] == 3)) {
586 if (empty($taglist)) {
587 $taglist = PageInfo::getTagsFromUrl($item["plink"], $preview, $contact["ffi_keyword_denylist"] ?? '');
589 $item["body"] .= "\n" . self::tagToString($taglist);
594 // Add the link to the original feed entry if not present in feed
595 if (($item['plink'] != '') && !strstr($item["body"], $item['plink'])) {
596 $item["body"] .= "[hr][url]" . $item['plink'] . "[/url]";
600 Logger::info('Stored feed', ['item' => $item]);
602 $notify = Item::isRemoteSelf($contact, $item);
604 // Distributed items should have a well formatted URI.
605 // Additionally we have to avoid conflicts with identical URI between imported feeds and these items.
607 $item['guid'] = Item::guidFromUri($orig_plink, DI::baseUrl()->getHostname());
609 unset($item['parent-uri']);
611 // Set the delivery priority for "remote self" to "medium"
612 $notify = PRIORITY_MEDIUM;
615 $id = Item::insert($item, $notify);
617 Logger::info("Feed for contact " . $contact["url"] . " stored under id " . $id);
619 if (!empty($id) && !empty($taglist)) {
620 $feeditem = Item::selectFirst(['uri-id'], ['id' => $id]);
621 foreach ($taglist as $tag) {
622 Tag::store($feeditem['uri-id'], Tag::HASHTAG, $tag);
627 if (!$dryRun && DI::config()->get('system', 'adjust_poll_frequency')) {
628 self::adjustPollFrequency($contact, $creation_dates);
631 return ["header" => $author, "items" => $items];
635 * Automatically adjust the poll frequency according to the post frequency
637 * @param array $contact
638 * @param array $creation_dates
641 private static function adjustPollFrequency(array $contact, array $creation_dates)
643 if ($contact['network'] != Protocol::FEED) {
644 Logger::info('Contact is no feed, skip.', ['id' => $contact['id'], 'uid' => $contact['uid'], 'url' => $contact['url'], 'network' => $contact['network']]);
648 if (!empty($creation_dates)) {
649 // Count the post frequency and the earliest and latest post date
653 $oldest_date = $newest_date = '';
655 foreach ($creation_dates as $date) {
656 $timestamp = strtotime($date);
657 $day = intdiv($timestamp, 86400);
658 $hour = $timestamp % 86400;
660 // Only have a look at values from the last seven days
661 if (((time() / 86400) - $day) < 7) {
662 if (empty($frequency[$day])) {
663 $frequency[$day] = ['count' => 1, 'low' => $hour, 'high' => $hour];
665 ++$frequency[$day]['count'];
666 if ($frequency[$day]['low'] > $hour) {
667 $frequency[$day]['low'] = $hour;
669 if ($frequency[$day]['high'] < $hour) {
670 $frequency[$day]['high'] = $hour;
674 if ($oldest > $day) {
676 $oldest_date = $date;
679 if ($newest < $day) {
681 $newest_date = $date;
685 if (count($creation_dates) == 1) {
686 Logger::info('Feed had posted a single time, switching to daily polling', ['newest' => $newest_date, 'id' => $contact['id'], 'uid' => $contact['uid'], 'url' => $contact['url']]);
687 $priority = 8; // Poll once a day
690 if (empty($priority) && (((time() / 86400) - $newest) > 730)) {
691 Logger::info('Feed had not posted for two years, switching to monthly polling', ['newest' => $newest_date, 'id' => $contact['id'], 'uid' => $contact['uid'], 'url' => $contact['url']]);
692 $priority = 10; // Poll every month
695 if (empty($priority) && (((time() / 86400) - $newest) > 365)) {
696 Logger::info('Feed had not posted for a year, switching to weekly polling', ['newest' => $newest_date, 'id' => $contact['id'], 'uid' => $contact['uid'], 'url' => $contact['url']]);
697 $priority = 9; // Poll every week
700 if (empty($priority) && empty($frequency)) {
701 Logger::info('Feed had not posted for at least a week, switching to daily polling', ['newest' => $newest_date, 'id' => $contact['id'], 'uid' => $contact['uid'], 'url' => $contact['url']]);
702 $priority = 8; // Poll once a day
705 if (empty($priority)) {
706 // Calculate the highest "posts per day" value
708 foreach ($frequency as $entry) {
709 if (($entry['count'] == 1) || ($entry['high'] == $entry['low'])) {
713 // We take the earliest and latest post day and interpolate the number of post per day
714 // that would had been created with this post frequency
716 // Assume at least four hours between oldest and newest post per day - should be okay for news outlets
717 $duration = max($entry['high'] - $entry['low'], 14400);
718 $ppd = (86400 / $duration) * $entry['count'];
724 $priority = 1; // Poll every quarter hour
725 } elseif ($max > 24) {
726 $priority = 2; // Poll half an hour
727 } elseif ($max > 12) {
728 $priority = 3; // Poll hourly
729 } elseif ($max > 8) {
730 $priority = 4; // Poll every two hours
731 } elseif ($max > 4) {
732 $priority = 5; // Poll every three hours
733 } elseif ($max > 2) {
734 $priority = 6; // Poll every six hours
736 $priority = 7; // Poll twice a day
738 Logger::info('Calculated priority by the posts per day', ['priority' => $priority, 'max' => round($max, 2), 'id' => $contact['id'], 'uid' => $contact['uid'], 'url' => $contact['url']]);
741 Logger::info('No posts, switching to daily polling', ['id' => $contact['id'], 'uid' => $contact['uid'], 'url' => $contact['url']]);
742 $priority = 8; // Poll once a day
745 if ($contact['rating'] != $priority) {
746 Logger::notice('Adjusting priority', ['old' => $contact['rating'], 'new' => $priority, 'id' => $contact['id'], 'uid' => $contact['uid'], 'url' => $contact['url']]);
747 DBA::update('contact', ['rating' => $priority], ['id' => $contact['id']]);
752 * Convert a tag array to a tag string
755 * @return string tag string
757 private static function tagToString(array $tags)
761 foreach ($tags as $tag) {
766 $tagstr .= "#[url=" . DI::baseUrl() . "/search?tag=" . urlencode($tag) . "]" . $tag . "[/url]";
772 private static function titleIsBody($title, $body)
774 $title = strip_tags($title);
775 $title = trim($title);
776 $title = html_entity_decode($title, ENT_QUOTES, 'UTF-8');
777 $title = str_replace(["\n", "\r", "\t", " "], ["", "", "", ""], $title);
779 $body = strip_tags($body);
781 $body = html_entity_decode($body, ENT_QUOTES, 'UTF-8');
782 $body = str_replace(["\n", "\r", "\t", " "], ["", "", "", ""], $body);
784 if (strlen($title) < strlen($body)) {
785 $body = substr($body, 0, strlen($title));
788 if (($title != $body) && (substr($title, -3) == "...")) {
789 $pos = strrpos($title, "...");
791 $title = substr($title, 0, $pos);
792 $body = substr($body, 0, $pos);
795 return ($title == $body);
799 * Creates the Atom feed for a given nickname
802 * - activity (default): all the public posts
803 * - posts: all the public top-level posts
804 * - comments: all the public replies
806 * Updates the provided last_update parameter if the result comes from the
807 * cache or it is empty
809 * @param string $owner_nick Nickname of the feed owner
810 * @param string $last_update Date of the last update
811 * @param integer $max_items Number of maximum items to fetch
812 * @param string $filter Feed items filter (activity, posts or comments)
813 * @param boolean $nocache Wether to bypass caching
815 * @return string Atom feed
816 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
817 * @throws \ImagickException
819 public static function atom($owner_nick, $last_update, $max_items = 300, $filter = 'activity', $nocache = false)
821 $stamp = microtime(true);
823 $owner = User::getOwnerDataByNick($owner_nick);
828 $cachekey = "feed:feed:" . $owner_nick . ":" . $filter . ":" . $last_update;
830 $previous_created = $last_update;
832 // Don't cache when the last item was posted less then 15 minutes ago (Cache duration)
833 if ((time() - strtotime($owner['last-item'])) < 15*60) {
834 $result = DI::cache()->get($cachekey);
835 if (!$nocache && !is_null($result)) {
836 Logger::info('Cached feed duration', ['seconds' => number_format(microtime(true) - $stamp, 3), 'nick' => $owner_nick, 'filter' => $filter, 'created' => $previous_created]);
837 return $result['feed'];
841 $check_date = empty($last_update) ? '' : DateTimeFormat::utc($last_update);
842 $authorid = Contact::getIdForURL($owner["url"]);
844 $condition = ["`uid` = ? AND `received` > ? AND NOT `deleted` AND `gravity` IN (?, ?)
845 AND `private` != ? AND `visible` AND `wall` AND `parent-network` IN (?, ?, ?, ?)",
846 $owner["uid"], $check_date, GRAVITY_PARENT, GRAVITY_COMMENT,
847 Item::PRIVATE, Protocol::ACTIVITYPUB,
848 Protocol::OSTATUS, Protocol::DFRN, Protocol::DIASPORA];
850 if ($filter === 'comments') {
851 $condition[0] .= " AND `object-type` = ? ";
852 $condition[] = Activity\ObjectType::COMMENT;
855 if ($owner['account-type'] != User::ACCOUNT_TYPE_COMMUNITY) {
856 $condition[0] .= " AND `contact-id` = ? AND `author-id` = ?";
857 $condition[] = $owner["id"];
858 $condition[] = $authorid;
861 $params = ['order' => ['received' => true], 'limit' => $max_items];
863 if ($filter === 'posts') {
864 $ret = Item::selectThread([], $condition, $params);
866 $ret = Item::select([], $condition, $params);
869 $items = Item::inArray($ret);
871 $doc = new DOMDocument('1.0', 'utf-8');
872 $doc->formatOutput = true;
874 $root = self::addHeader($doc, $owner, $filter);
876 foreach ($items as $item) {
877 $entry = self::entry($doc, $item, $owner);
878 $root->appendChild($entry);
880 if ($last_update < $item['created']) {
881 $last_update = $item['created'];
885 $feeddata = trim($doc->saveXML());
887 $msg = ['feed' => $feeddata, 'last_update' => $last_update];
888 DI::cache()->set($cachekey, $msg, Duration::QUARTER_HOUR);
890 Logger::info('Feed duration', ['seconds' => number_format(microtime(true) - $stamp, 3), 'nick' => $owner_nick, 'filter' => $filter, 'created' => $previous_created]);
896 * Adds the header elements to the XML document
898 * @param DOMDocument $doc XML document
899 * @param array $owner Contact data of the poster
900 * @param string $filter The related feed filter (activity, posts or comments)
902 * @return object header root element
903 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
905 private static function addHeader(DOMDocument $doc, array $owner, $filter)
907 $root = $doc->createElementNS(ActivityNamespace::ATOM1, 'feed');
908 $doc->appendChild($root);
911 $selfUri = '/feed/' . $owner["nick"] . '/';
914 $title = DI::l10n()->t('%s\'s timeline', $owner['name']);
918 $title = DI::l10n()->t('%s\'s posts', $owner['name']);
921 $title = DI::l10n()->t('%s\'s comments', $owner['name']);
926 $attributes = ["uri" => "https://friendi.ca", "version" => FRIENDICA_VERSION . "-" . DB_UPDATE_VERSION];
927 XML::addElement($doc, $root, "generator", FRIENDICA_PLATFORM, $attributes);
928 XML::addElement($doc, $root, "id", DI::baseUrl() . "/profile/" . $owner["nick"]);
929 XML::addElement($doc, $root, "title", $title);
930 XML::addElement($doc, $root, "subtitle", sprintf("Updates from %s on %s", $owner["name"], DI::config()->get('config', 'sitename')));
931 XML::addElement($doc, $root, "logo", $owner["photo"]);
932 XML::addElement($doc, $root, "updated", DateTimeFormat::utcNow(DateTimeFormat::ATOM));
934 $author = self::addAuthor($doc, $owner);
935 $root->appendChild($author);
937 $attributes = ["href" => $owner["url"], "rel" => "alternate", "type" => "text/html"];
938 XML::addElement($doc, $root, "link", "", $attributes);
940 OStatus::hublinks($doc, $root, $owner["nick"]);
942 $attributes = ["href" => DI::baseUrl() . $selfUri, "rel" => "self", "type" => "application/atom+xml"];
943 XML::addElement($doc, $root, "link", "", $attributes);
949 * Adds the author element to the XML document
951 * @param DOMDocument $doc XML document
952 * @param array $owner Contact data of the poster
954 * @return \DOMElement author element
955 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
957 private static function addAuthor(DOMDocument $doc, array $owner)
959 $author = $doc->createElement("author");
960 XML::addElement($doc, $author, "uri", $owner["url"]);
961 XML::addElement($doc, $author, "name", $owner["nick"]);
962 XML::addElement($doc, $author, "email", $owner["addr"]);
968 * Adds an entry element to the XML document
970 * @param DOMDocument $doc XML document
971 * @param array $item Data of the item that is to be posted
972 * @param array $owner Contact data of the poster
973 * @param bool $toplevel optional default false
975 * @return \DOMElement Entry element
976 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
977 * @throws \ImagickException
979 private static function entry(DOMDocument $doc, array $item, array $owner)
983 $repeated_guid = OStatus::getResharedGuid($item);
984 if ($repeated_guid != "") {
985 $xml = self::reshareEntry($doc, $item, $owner, $repeated_guid);
992 return self::noteEntry($doc, $item, $owner);
996 * Adds an entry element with reshared content
998 * @param DOMDocument $doc XML document
999 * @param array $item Data of the item that is to be posted
1000 * @param array $owner Contact data of the poster
1001 * @param string $repeated_guid guid
1002 * @param bool $toplevel Is it for en entry element (false) or a feed entry (true)?
1004 * @return bool Entry element
1005 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
1006 * @throws \ImagickException
1008 private static function reshareEntry(DOMDocument $doc, array $item, array $owner, $repeated_guid)
1010 if (($item['gravity'] != GRAVITY_PARENT) && (Strings::normaliseLink($item["author-link"]) != Strings::normaliseLink($owner["url"]))) {
1011 Logger::info('Feed entry author does not match feed owner', ['owner' => $owner["url"], 'author' => $item["author-link"]]);
1014 $entry = OStatus::entryHeader($doc, $owner, $item, false);
1016 $condition = ['uid' => $owner["uid"], 'guid' => $repeated_guid, 'private' => [Item::PUBLIC, Item::UNLISTED],
1017 'network' => Protocol::FEDERATED];
1018 $repeated_item = Item::selectFirst([], $condition);
1019 if (!DBA::isResult($repeated_item)) {
1023 self::entryContent($doc, $entry, $item, self::getTitle($repeated_item), Activity::SHARE, false);
1025 self::entryFooter($doc, $entry, $item, $owner);
1031 * Adds a regular entry element
1033 * @param DOMDocument $doc XML document
1034 * @param array $item Data of the item that is to be posted
1035 * @param array $owner Contact data of the poster
1036 * @param bool $toplevel Is it for en entry element (false) or a feed entry (true)?
1038 * @return \DOMElement Entry element
1039 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
1040 * @throws \ImagickException
1042 private static function noteEntry(DOMDocument $doc, array $item, array $owner)
1044 if (($item['gravity'] != GRAVITY_PARENT) && (Strings::normaliseLink($item["author-link"]) != Strings::normaliseLink($owner["url"]))) {
1045 Logger::info('Feed entry author does not match feed owner', ['owner' => $owner["url"], 'author' => $item["author-link"]]);
1048 $entry = OStatus::entryHeader($doc, $owner, $item, false);
1050 self::entryContent($doc, $entry, $item, self::getTitle($item), '', true);
1052 self::entryFooter($doc, $entry, $item, $owner);
1058 * Adds elements to the XML document
1060 * @param DOMDocument $doc XML document
1061 * @param \DOMElement $entry Entry element where the content is added
1062 * @param array $item Data of the item that is to be posted
1063 * @param array $owner Contact data of the poster
1064 * @param string $title Title for the post
1065 * @param string $verb The activity verb
1066 * @param bool $complete Add the "status_net" element?
1067 * @param bool $feed_mode Behave like a regular feed for users if true
1069 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
1071 private static function entryContent(DOMDocument $doc, \DOMElement $entry, array $item, $title, $verb = "", $complete = true)
1074 $verb = OStatus::constructVerb($item);
1077 XML::addElement($doc, $entry, "id", $item["uri"]);
1078 XML::addElement($doc, $entry, "title", html_entity_decode($title, ENT_QUOTES, 'UTF-8'));
1080 $body = OStatus::formatPicturePost($item['body']);
1082 $body = BBCode::convert($body, false, BBCode::OSTATUS);
1084 XML::addElement($doc, $entry, "content", $body, ["type" => "html"]);
1086 XML::addElement($doc, $entry, "link", "", ["rel" => "alternate", "type" => "text/html",
1087 "href" => DI::baseUrl()."/display/".$item["guid"]]
1090 XML::addElement($doc, $entry, "published", DateTimeFormat::utc($item["created"]."+00:00", DateTimeFormat::ATOM));
1091 XML::addElement($doc, $entry, "updated", DateTimeFormat::utc($item["edited"]."+00:00", DateTimeFormat::ATOM));
1095 * Adds the elements at the foot of an entry to the XML document
1097 * @param DOMDocument $doc XML document
1098 * @param object $entry The entry element where the elements are added
1099 * @param array $item Data of the item that is to be posted
1100 * @param array $owner Contact data of the poster
1101 * @param bool $complete default true
1103 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
1105 private static function entryFooter(DOMDocument $doc, $entry, array $item, array $owner)
1109 if ($item['gravity'] != GRAVITY_PARENT) {
1110 $parent = Item::selectFirst(['guid', 'author-link', 'owner-link'], ['id' => $item['parent']]);
1111 $parent_item = (($item['thr-parent']) ? $item['thr-parent'] : $item['parent-uri']);
1113 $thrparent = Item::selectFirst(['guid', 'author-link', 'owner-link', 'plink'], ['uid' => $owner["uid"], 'uri' => $parent_item]);
1115 if (DBA::isResult($thrparent)) {
1116 $mentioned[$thrparent["author-link"]] = $thrparent["author-link"];
1117 $mentioned[$thrparent["owner-link"]] = $thrparent["owner-link"];
1118 $parent_plink = $thrparent["plink"];
1120 $mentioned[$parent["author-link"]] = $parent["author-link"];
1121 $mentioned[$parent["owner-link"]] = $parent["owner-link"];
1122 $parent_plink = DI::baseUrl()."/display/".$parent["guid"];
1126 "ref" => $parent_item,
1127 "href" => $parent_plink];
1128 XML::addElement($doc, $entry, "thr:in-reply-to", "", $attributes);
1132 "href" => $parent_plink];
1133 XML::addElement($doc, $entry, "link", "", $attributes);
1136 // uri-id isn't present for follow entry pseudo-items
1137 $tags = Tag::getByURIId($item['uri-id'] ?? 0);
1138 foreach ($tags as $tag) {
1139 $mentioned[$tag['url']] = $tag['url'];
1142 foreach ($tags as $tag) {
1143 if ($tag['type'] == Tag::HASHTAG) {
1144 XML::addElement($doc, $entry, "category", "", ["term" => $tag['name']]);
1148 OStatus::getAttachment($doc, $entry, $item);
1152 * Fetch or create title for feed entry
1154 * @param array $item
1155 * @return string title
1157 private static function getTitle(array $item)
1159 if ($item['title'] != '') {
1160 return BBCode::convert($item['title'], false, BBCode::OSTATUS);
1163 // Fetch information about the post
1164 $siteinfo = BBCode::getAttachedData($item["body"]);
1165 if (isset($siteinfo["title"])) {
1166 return $siteinfo["title"];
1169 // If no bookmark is found then take the first line
1170 // Remove the share element before fetching the first line
1171 $title = trim(preg_replace("/\[share.*?\](.*?)\[\/share\]/ism","\n$1\n",$item['body']));
1173 $title = HTML::toPlaintext(BBCode::convert($title, false), 0, true)."\n";
1174 $pos = strpos($title, "\n");
1176 if (($pos == 0) || ($pos > 100)) {
1181 return substr($title, 0, $pos) . $trailer;