3 * @copyright Copyright (C) 2010-2022, the Friendica project
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\Model;
24 use Friendica\Content\Text\BBCode;
25 use Friendica\Core\Cache\Enum\Duration;
26 use Friendica\Core\Logger;
27 use Friendica\Core\Protocol;
28 use Friendica\Core\System;
29 use Friendica\Database\Database;
30 use Friendica\Database\DBA;
32 use Friendica\Protocol\ActivityPub;
33 use Friendica\Util\DateTimeFormat;
34 use Friendica\Util\Strings;
39 * This Model class handles tag table interactions.
40 * This tables stores relevant tags related to posts, like hashtags and mentions.
48 * An implicit mention is a mention in a comment body that is redundant with the threading information.
50 const IMPLICIT_MENTION = 8;
52 * An exclusive mention transmits the post only to the target account without transmitting it to the followers, usually a forum.
54 const EXCLUSIVE_MENTION = 9;
62 const GENERAL_COLLECTION = 2;
63 const FOLLOWER_COLLECTION = 3;
64 const PUBLIC_COLLECTION = 4;
66 const TAG_CHARACTER = [
69 self::EXCLUSIVE_MENTION => '!',
70 self::IMPLICIT_MENTION => '%',
74 * Store tag/mention elements
76 * @param integer $uriid
77 * @param integer $type
80 * @param integer $target
82 public static function store(int $uriid, int $type, string $name, string $url = '', int $target = null)
84 if ($type == self::HASHTAG) {
85 // Trim Unicode non-word characters
86 $name = preg_replace('/(^\W+)|(\W+$)/us', '', $name);
88 $tags = explode(self::TAG_CHARACTER[self::HASHTAG], $name);
89 if (count($tags) > 1) {
90 foreach ($tags as $tag) {
91 self::store($uriid, $type, $tag, $url);
104 if (in_array($type, [self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION, self::TO, self::CC, self::BTO, self::BCC])) {
106 // No mention without a contact url
110 if ((substr($url, 0, 7) == 'https//') || (substr($url, 0, 6) == 'http//')) {
111 Logger::notice('Wrong scheme in url', ['url' => $url, 'callstack' => System::callstack(20)]);
114 $cid = Contact::getIdForURL($url, 0, false);
115 Logger::debug('Got id for contact', ['cid' => $cid, 'url' => $url]);
118 $tag = DBA::selectFirst('tag', ['name', 'type'], ['url' => $url]);
120 if ($tag['name'] != substr($name, 0, 96)) {
121 DBA::update('tag', ['name' => substr($name, 0, 96)], ['url' => $url]);
123 if (!empty($target) && ($tag['type'] != $target)) {
124 DBA::update('tag', ['type' => $target], ['url' => $url]);
131 if (!in_array($type, [self::TO, self::CC, self::BTO, self::BCC])) {
132 if (($type != self::HASHTAG) && !empty($url) && ($url != $name)) {
133 $url = strtolower($url);
139 $tagid = self::getID($name, $url, $target);
145 $fields = ['uri-id' => $uriid, 'type' => $type, 'tid' => $tagid, 'cid' => $cid];
147 if (in_array($type, [self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION])) {
148 $condition = $fields;
149 $condition['type'] = [self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION];
150 if (DBA::exists('post-tag', $condition)) {
151 Logger::info('Tag already exists', $fields);
156 DBA::insert('post-tag', $fields, Database::INSERT_IGNORE);
158 Logger::info('Stored tag/mention', ['uri-id' => $uriid, 'tag-id' => $tagid, 'contact-id' => $cid, 'name' => $name, 'type' => $type, 'callstack' => System::callstack(8)]);
162 * Fetch the target type for the given url
165 * @param bool $fetch Fetch information via network operations
168 public static function getTargetType(string $url, bool $fetch = true)
176 $tag = DBA::selectFirst('tag', ['url', 'type'], ['url' => $url]);
177 if (!empty($tag['type'])) {
178 $target = $tag['type'];
179 if ($target != self::GENERAL_COLLECTION) {
180 Logger::debug('Found existing type', ['type' => $tag['type'], 'url' => $url]);
185 if ($url == ActivityPub::PUBLIC_COLLECTION) {
186 $target = self::PUBLIC_COLLECTION;
187 Logger::debug('Public collection', ['url' => $url]);
189 if (DBA::exists('apcontact', ['followers' => $url])) {
190 $target = self::FOLLOWER_COLLECTION;
191 Logger::debug('Found collection via existing apcontact', ['url' => $url]);
192 } elseif (Contact::getIdForURL($url, 0, $fetch ? null : false)) {
193 $target = self::ACCOUNT;
194 Logger::debug('URL is an account', ['url' => $url]);
195 } elseif ($fetch && ($target != self::GENERAL_COLLECTION)) {
196 $content = ActivityPub::fetchContent($url);
197 if (!empty($content['type']) && ($content['type'] == 'OrderedCollection')) {
198 $target = self::GENERAL_COLLECTION;
199 Logger::debug('URL is an ordered collection', ['url' => $url]);
204 if (!empty($target) && !empty($tag['url']) && ($tag['type'] != $target)) {
205 DBA::update('tag', ['type' => $target], ['url' => $url]);
208 if (empty($target)) {
209 Logger::debug('No type could be detected', ['url' => $url]);
216 * Get a tag id for a given tag name and url
218 * @param string $name
223 public static function getID(string $name, string $url = '', int $type = null)
225 $fields = ['name' => substr($name, 0, 96), 'url' => $url];
228 $fields['type'] = $type;
231 $tag = DBA::selectFirst('tag', ['id'], $fields);
232 if (DBA::isResult($tag)) {
236 DBA::insert('tag', $fields, Database::INSERT_IGNORE);
237 $tid = DBA::lastInsertId();
242 Logger::error('No tag id created', $fields);
247 * Store tag/mention elements
249 * @param integer $uriid
250 * @param string $hash
251 * @param string $name
253 * @param boolean $probing
255 public static function storeByHash(int $uriid, string $hash, string $name, string $url = '', $probing = true)
257 $type = self::getTypeForHash($hash);
258 if ($type == self::UNKNOWN) {
262 self::store($uriid, $type, $name, $url, $probing);
266 * Get tags and mentions from the body
268 * @param string $body Body of the post
269 * @param string $tags Accepted tags
271 * @return array Tag list
273 public static function getFromBody(string $body, string $tags = null)
275 if (is_null($tags)) {
276 $tags = self::TAG_CHARACTER[self::HASHTAG] . self::TAG_CHARACTER[self::MENTION] . self::TAG_CHARACTER[self::EXCLUSIVE_MENTION];
279 if (!preg_match_all("/([" . $tags . "])\[url\=([^\[\]]*)\]([^\[\]]*)\[\/url\]/ism", $body, $result, PREG_SET_ORDER)) {
287 * Store tags and mentions from the body
289 * @param integer $uriid URI-Id
290 * @param string $body Body of the post
291 * @param string $tags Accepted tags
292 * @param boolean $probing Perform a probing for contacts, adding them if needed
294 public static function storeFromBody(int $uriid, string $body, string $tags = null, $probing = true)
296 Logger::info('Check for tags', ['uri-id' => $uriid, 'hash' => $tags, 'callstack' => System::callstack()]);
298 $result = self::getFromBody($body, $tags);
299 if (empty($result)) {
303 Logger::info('Found tags', ['uri-id' => $uriid, 'hash' => $tags, 'result' => $result]);
305 foreach ($result as $tag) {
306 self::storeByHash($uriid, $tag[1], $tag[3], $tag[2], $probing);
311 * Store raw tags (not encapsulated in links) from the body
312 * This function is needed in the intermediate phase.
313 * Later we can call item::setHashtags in advance to have all tags converted.
315 * @param integer $uriid URI-Id
316 * @param string $body Body of the post
318 public static function storeRawTagsFromBody(int $uriid, string $body)
320 Logger::info('Check for tags', ['uri-id' => $uriid, 'callstack' => System::callstack()]);
322 $result = BBCode::getTags($body);
323 if (empty($result)) {
327 Logger::info('Found tags', ['uri-id' => $uriid, 'result' => $result]);
329 foreach ($result as $tag) {
330 if (substr($tag, 0, 1) != self::TAG_CHARACTER[self::HASHTAG]) {
333 self::storeByHash($uriid, substr($tag, 0, 1), substr($tag, 1));
338 * Checks for stored hashtags and mentions for the given post
340 * @param integer $uriid
343 public static function existsForPost(int $uriid)
345 return DBA::exists('post-tag', ['uri-id' => $uriid, 'type' => [self::HASHTAG, self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION]]);
351 * @param integer $uriid
352 * @param integer $type
353 * @param string $name
356 public static function remove(int $uriid, int $type, string $name, string $url = '')
358 $condition = ['uri-id' => $uriid, 'type' => $type, 'url' => $url];
359 if ($type == self::HASHTAG) {
360 $condition['name'] = $name;
363 $tag = DBA::selectFirst('tag-view', ['tid', 'cid'], $condition);
364 if (!DBA::isResult($tag)) {
368 Logger::info('Removing tag/mention', ['uri-id' => $uriid, 'tid' => $tag['tid'], 'name' => $name, 'url' => $url, 'callstack' => System::callstack(8)]);
369 DBA::delete('post-tag', ['uri-id' => $uriid, 'type' => $type, 'tid' => $tag['tid'], 'cid' => $tag['cid']]);
375 * @param integer $uriid
376 * @param string $hash
377 * @param string $name
380 public static function removeByHash(int $uriid, string $hash, string $name, string $url = '')
382 $type = self::getTypeForHash($hash);
383 if ($type == self::UNKNOWN) {
387 self::remove($uriid, $type, $name, $url);
391 * Get the type for the given hash
393 * @param string $hash
394 * @return integer type
396 private static function getTypeForHash(string $hash)
398 if ($hash == self::TAG_CHARACTER[self::MENTION]) {
399 return self::MENTION;
400 } elseif ($hash == self::TAG_CHARACTER[self::EXCLUSIVE_MENTION]) {
401 return self::EXCLUSIVE_MENTION;
402 } elseif ($hash == self::TAG_CHARACTER[self::IMPLICIT_MENTION]) {
403 return self::IMPLICIT_MENTION;
404 } elseif ($hash == self::TAG_CHARACTER[self::HASHTAG]) {
405 return self::HASHTAG;
407 return self::UNKNOWN;
412 * Create implicit mentions for a given post
414 * @param integer $uri_id
415 * @param integer $parent_uri_id
417 public static function createImplicitMentions(int $uri_id, int $parent_uri_id)
419 // Always mention the direct parent author
420 $parent = Post::selectFirst(['author-link', 'author-name'], ['uri-id' => $parent_uri_id]);
421 self::store($uri_id, self::IMPLICIT_MENTION, $parent['author-name'], $parent['author-link']);
423 if (DI::config()->get('system', 'disable_implicit_mentions')) {
427 $tags = DBA::select('tag-view', ['name', 'url'], ['uri-id' => $parent_uri_id, 'type' => [self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION]]);
428 while ($tag = DBA::fetch($tags)) {
429 self::store($uri_id, self::IMPLICIT_MENTION, $tag['name'], $tag['url']);
435 * Retrieves the terms from the provided type(s) associated with the provided item ID.
437 * @param int $item_id
438 * @param int|array $type
442 public static function getByURIId(int $uri_id, array $type = [self::HASHTAG, self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION])
444 $condition = ['uri-id' => $uri_id, 'type' => $type];
445 return DBA::selectToArray('tag-view', ['type', 'name', 'url', 'tag-type'], $condition);
449 * Return a string with all tags and mentions
451 * @param integer $uri_id
453 * @return string tags and mentions
456 public static function getCSVByURIId(int $uri_id, array $type = [self::HASHTAG, self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION])
459 $tags = self::getByURIId($uri_id, $type);
460 foreach ($tags as $tag) {
461 $tag_list[] = self::TAG_CHARACTER[$tag['type']] . '[url=' . $tag['url'] . ']' . $tag['name'] . '[/url]';
464 return implode(',', $tag_list);
468 * Sorts an item's tags into mentions, hashtags and other tags. Generate personalized URLs by user and modify the
469 * provided item's body with them.
473 * @throws \Friendica\Network\HTTPException\InternalServerErrorException
474 * @throws \ImagickException
476 public static function populateFromItem(&$item)
482 'implicit_mentions' => [],
485 $searchpath = DI::baseUrl() . "/search?tag=";
487 $taglist = DBA::select('tag-view', ['type', 'name', 'url', 'cid'],
488 ['uri-id' => $item['uri-id'], 'type' => [self::HASHTAG, self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION]]);
489 while ($tag = DBA::fetch($taglist)) {
490 if ($tag['url'] == '') {
491 $tag['url'] = $searchpath . rawurlencode($tag['name']);
494 $orig_tag = $tag['url'];
496 $prefix = self::TAG_CHARACTER[$tag['type']];
497 switch($tag['type']) {
499 if ($orig_tag != $tag['url']) {
500 $item['body'] = str_replace($orig_tag, $tag['url'], $item['body']);
503 $return['hashtags'][] = '<bdi>' . $prefix . '<a href="' . $tag['url'] . '" target="_blank" rel="noopener noreferrer">' . htmlspecialchars($tag['name']) . '</a></bdi>';
504 $return['tags'][] = '<bdi>' . $prefix . '<a href="' . $tag['url'] . '" target="_blank" rel="noopener noreferrer">' . htmlspecialchars($tag['name']) . '</a></bdi>';
507 case self::EXCLUSIVE_MENTION:
508 if (!empty($tag['cid'])) {
509 $tag['url'] = Contact::magicLinkById($tag['cid']);
511 $tag['url'] = Contact::magicLink($tag['url']);
513 $return['mentions'][] = '<bdi>' . $prefix . '<a href="' . $tag['url'] . '" target="_blank" rel="noopener noreferrer">' . htmlspecialchars($tag['name']) . '</a></bdi>';
514 $return['tags'][] = '<bdi>' . $prefix . '<a href="' . $tag['url'] . '" target="_blank" rel="noopener noreferrer">' . htmlspecialchars($tag['name']) . '</a></bdi>';
516 case self::IMPLICIT_MENTION:
517 $return['implicit_mentions'][] = $prefix . $tag['name'];
521 DBA::close($taglist);
527 * Counts posts for given tag
529 * @param string $search
530 * @param integer $uid
531 * @return integer number of posts
533 public static function countByTag(string $search, int $uid = 0)
535 $condition = ["`name` = ? AND (`uid` = ? OR (`uid` = ? AND NOT `global`))
536 AND (`network` IN (?, ?, ?, ?) OR (`uid` = ? AND `uid` != ?))",
537 $search, 0, $uid, Protocol::ACTIVITYPUB, Protocol::DFRN, Protocol::DIASPORA, Protocol::OSTATUS, $uid, 0];
539 return DBA::count('tag-search-view', $condition);
543 * Search posts for given tag
545 * @param string $search
546 * @param integer $uid
547 * @param integer $start
548 * @param integer $limit
549 * @param integer $last_uriid
550 * @return array with URI-ID
552 public static function getURIIdListByTag(string $search, int $uid = 0, int $start = 0, int $limit = 100, int $last_uriid = 0)
554 $condition = ["`name` = ? AND (`uid` = ? OR (`uid` = ? AND NOT `global`))
555 AND (`network` IN (?, ?, ?, ?) OR (`uid` = ? AND `uid` != ?))",
556 $search, 0, $uid, Protocol::ACTIVITYPUB, Protocol::DFRN, Protocol::DIASPORA, Protocol::OSTATUS, $uid, 0];
558 if (!empty($last_uriid)) {
559 $condition = DBA::mergeConditions($condition, ["`uri-id` < ?", $last_uriid]);
563 'order' => ['uri-id' => true],
564 'limit' => [$start, $limit]
567 $tags = DBA::select('tag-search-view', ['uri-id'], $condition, $params);
570 while ($tag = DBA::fetch($tags)) {
571 $uriids[] = $tag['uri-id'];
579 * Returns a list of the most frequent global hashtags over the given period
581 * @param int $period Period in hours to consider posts
582 * @param int $limit Number of returned tags
586 public static function getGlobalTrendingHashtags(int $period, $limit = 10)
588 $tags = DI::cache()->get('global_trending_tags-' . $period . '-' . $limit);
592 return self::setGlobalTrendingHashtags($period, $limit);
597 * Fetch the blocked tags as SQL
601 private static function getBlockedSQL()
603 $blocked_txt = DI::config()->get('system', 'blocked_tags');
604 if (empty($blocked_txt)) {
608 $blocked = explode(',', $blocked_txt);
609 array_walk($blocked, function(&$value) { $value = "'" . DBA::escape(trim($value)) . "'";});
610 return " AND NOT `name` IN (" . implode(',', $blocked) . ")";
614 * Creates a list of the most frequent global hashtags over the given period
616 * @param int $period Period in hours to consider posts
617 * @param int $limit Number of returned tags
621 public static function setGlobalTrendingHashtags(int $period, int $limit = 10)
623 // Get a uri-id that is at least X hours old.
624 // We use the uri-id in the query for the hash tags since this is much faster
625 $post = Post::selectFirstThread(['uri-id'], ["`uid` = ? AND `received` < ?", 0, DateTimeFormat::utc('now - ' . $period . ' hour')],
626 ['order' => ['received' => true]]);
627 if (empty($post['uri-id'])) {
631 $block_sql = self::getBlockedSQL();
633 $tagsStmt = DBA::p("SELECT `name` AS `term`, COUNT(*) AS `score`, COUNT(DISTINCT(`author-id`)) as `authors`
634 FROM `tag-search-view`
635 WHERE `private` = ? AND `uid` = ? AND `uri-id` > ? $block_sql
636 GROUP BY `term` ORDER BY `authors` DESC, `score` DESC LIMIT ?",
637 Item::PUBLIC, 0, $post['uri-id'], $limit);
639 if (DBA::isResult($tagsStmt)) {
640 $tags = DBA::toArray($tagsStmt);
641 DI::cache()->set('global_trending_tags-' . $period . '-' . $limit, $tags, Duration::DAY);
649 * Returns a list of the most frequent local hashtags over the given period
651 * @param int $period Period in hours to consider posts
652 * @param int $limit Number of returned tags
656 public static function getLocalTrendingHashtags(int $period, $limit = 10)
658 $tags = DI::cache()->get('local_trending_tags-' . $period . '-' . $limit);
662 return self::setLocalTrendingHashtags($period, $limit);
667 * Returns a list of the most frequent local hashtags over the given period
669 * @param int $period Period in hours to consider posts
670 * @param int $limit Number of returned tags
674 public static function setLocalTrendingHashtags(int $period, int $limit = 10)
676 // Get a uri-id that is at least X hours old.
677 // We use the uri-id in the query for the hash tags since this is much faster
678 $post = Post::selectFirstThread(['uri-id'], ["`uid` = ? AND `received` < ?", 0, DateTimeFormat::utc('now - ' . $period . ' hour')],
679 ['order' => ['received' => true]]);
680 if (empty($post['uri-id'])) {
684 $block_sql = self::getBlockedSQL();
686 $tagsStmt = DBA::p("SELECT `name` AS `term`, COUNT(*) AS `score`, COUNT(DISTINCT(`author-id`)) as `authors`
687 FROM `tag-search-view`
688 WHERE `private` = ? AND `wall` AND `origin` AND `uri-id` > ? $block_sql
689 GROUP BY `term` ORDER BY `authors` DESC, `score` DESC LIMIT ?",
690 Item::PUBLIC, $post['uri-id'], $limit);
692 if (DBA::isResult($tagsStmt)) {
693 $tags = DBA::toArray($tagsStmt);
694 DI::cache()->set('local_trending_tags-' . $period . '-' . $limit, $tags, Duration::DAY);
702 * Check if the provided tag is of one of the provided term types.
705 * @param int ...$types
708 public static function isType($tag, ...$types)
711 foreach ($types as $type) {
712 if (array_key_exists($type, self::TAG_CHARACTER)) {
713 $tag_chars[] = self::TAG_CHARACTER[$type];
717 return Strings::startsWithChars($tag, $tag_chars);
721 * Fetch user who subscribed to the given tag
724 * @return array User list
726 private static function getUIDListByTag(string $tag)
729 $searches = DBA::select('search', ['uid'], ['term' => $tag]);
730 while ($search = DBA::fetch($searches)) {
731 $uids[] = $search['uid'];
733 DBA::close($searches);
739 * Fetch user who subscribed to the tags of the given item
741 * @param integer $uri_id
742 * @return array User list
744 public static function getUIDListByURIId(int $uri_id)
747 $tags = self::getByURIId($uri_id, [self::HASHTAG]);
749 foreach ($tags as $tag) {
750 $uids = array_merge($uids, self::getUIDListByTag(self::TAG_CHARACTER[self::HASHTAG] . $tag['name']));
753 return array_unique($uids);