friendica/src/Model/Tag.php

542 lines
15 KiB
PHP
Raw Normal View History

2020-04-15 12:37:09 -04:00
<?php
/**
* @copyright Copyright (C) 2020, Friendica
*
* @license GNU AGPL version 3 or any later version
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as
* published by the Free Software Foundation, either version 3 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
*
*/
namespace Friendica\Model;
use Friendica\Content\Text\BBCode;
2020-05-01 06:57:32 -04:00
use Friendica\Core\Cache\Duration;
2020-04-15 12:37:09 -04:00
use Friendica\Core\Logger;
use Friendica\Core\System;
2020-04-15 12:37:09 -04:00
use Friendica\Database\DBA;
2020-04-26 11:24:58 -04:00
use Friendica\DI;
use Friendica\Util\Strings;
2020-04-15 12:37:09 -04:00
/**
* Class Tag
*
* This Model class handles tag table interactions.
* This tables stores relevant tags related to posts, like hashtags and mentions.
*/
class Tag
{
2020-04-16 00:20:06 -04:00
const UNKNOWN = 0;
const HASHTAG = 1;
const MENTION = 2;
2020-04-15 12:37:09 -04:00
/**
* An implicit mention is a mention in a comment body that is redundant with the threading information.
*/
2020-04-16 00:20:06 -04:00
const IMPLICIT_MENTION = 8;
2020-04-15 12:37:09 -04:00
/**
* An exclusive mention transfers the ownership of the post to the target account, usually a forum.
*/
2020-04-16 00:20:06 -04:00
const EXCLUSIVE_MENTION = 9;
2020-04-15 12:37:09 -04:00
2020-04-16 00:20:06 -04:00
const TAG_CHARACTER = [
self::HASHTAG => '#',
self::MENTION => '@',
self::IMPLICIT_MENTION => '%',
self::EXCLUSIVE_MENTION => '!',
];
2020-04-15 12:37:09 -04:00
2020-04-17 03:58:54 -04:00
/**
* Store tag/mention elements
*
* @param integer $uriid
* @param integer $type
2020-04-20 05:47:26 -04:00
* @param string $name
* @param string $url
* @param boolean $probing
2020-04-17 03:58:54 -04:00
*/
2020-04-20 05:47:26 -04:00
public static function store(int $uriid, int $type, string $name, string $url = '', $probing = true)
{
2020-04-29 06:49:13 -04:00
if ($type == self::HASHTAG) {
2020-04-28 07:52:51 -04:00
// Remove some common "garbarge" from tags
$name = trim($name, "\x00..\x20\xFF#!@,;.:'/?!^°$%".'"');
$tags = explode(self::TAG_CHARACTER[self::HASHTAG], $name);
if (count($tags) > 1) {
foreach ($tags as $tag) {
self::store($uriid, $type, $tag, $url, $probing);
}
return;
}
}
if (empty($name)) {
return;
}
$cid = 0;
$tagid = 0;
2020-04-28 07:52:51 -04:00
if (in_array($type, [self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION])) {
if (empty($url)) {
// No mention without a contact url
return;
}
2020-04-20 05:47:26 -04:00
if (!$probing) {
$condition = ['nurl' => Strings::normaliseLink($url), 'uid' => 0, 'deleted' => false];
$contact = DBA::selectFirst('contact', ['id'], $condition, ['order' => ['id']]);
if (DBA::isResult($contact)) {
$cid = $contact['id'];
Logger::info('Got id for contact url', ['cid' => $cid, 'url' => $url]);
}
if (empty($cid)) {
$ssl_url = str_replace('http://', 'https://', $url);
$condition = ['`alias` IN (?, ?, ?) AND `uid` = ? AND NOT `deleted`', $url, Strings::normaliseLink($url), $ssl_url, 0];
$contact = DBA::selectFirst('contact', ['id'], $condition, ['order' => ['id']]);
if (DBA::isResult($contact)) {
$cid = $contact['id'];
Logger::info('Got id for contact alias', ['cid' => $cid, 'url' => $url]);
}
}
} else {
2020-04-20 05:47:26 -04:00
$cid = Contact::getIdForURL($url, 0, true);
Logger::info('Got id by probing', ['cid' => $cid, 'url' => $url]);
}
if (empty($cid)) {
// The contact wasn't found in the system (most likely some dead account)
// We ensure that we only store a single entry by overwriting the previous name
2020-04-20 05:47:26 -04:00
Logger::info('Contact not found, updating tag', ['url' => $url, 'name' => $name]);
DBA::update('tag', ['name' => substr($name, 0, 96)], ['url' => $url]);
}
}
if (empty($cid)) {
2020-04-28 07:52:51 -04:00
if (($type != self::HASHTAG) && !empty($url) && ($url != $name)) {
$url = strtolower($url);
} else {
$url = '';
}
2020-04-20 05:47:26 -04:00
$tagid = self::getID($name, $url);
if (empty($tagid)) {
return;
}
}
2020-04-20 01:43:13 -04:00
$fields = ['uri-id' => $uriid, 'type' => $type, 'tid' => $tagid, 'cid' => $cid];
2020-04-28 07:52:51 -04:00
if (in_array($type, [self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION])) {
2020-04-20 01:43:13 -04:00
$condition = $fields;
2020-04-28 07:52:51 -04:00
$condition['type'] = [self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION];
2020-04-20 01:43:13 -04:00
if (DBA::exists('post-tag', $condition)) {
Logger::info('Tag already exists', $fields);
return;
}
}
DBA::insert('post-tag', $fields, true);
2020-04-20 05:47:26 -04:00
Logger::info('Stored tag/mention', ['uri-id' => $uriid, 'tag-id' => $tagid, 'contact-id' => $cid, 'name' => $name, 'type' => $type, 'callstack' => System::callstack(8)]);
}
/**
* Get a tag id for a given tag name and url
*
* @param string $name
* @param string $url
* @return void
*/
public static function getID(string $name, string $url = '')
{
$fields = ['name' => substr($name, 0, 96), 'url' => $url];
$tag = DBA::selectFirst('tag', ['id'], $fields);
if (DBA::isResult($tag)) {
return $tag['id'];
}
DBA::insert('tag', $fields, true);
$tid = DBA::lastInsertId();
if (!empty($tid)) {
return $tid;
}
Logger::error('No tag id created', $fields);
return 0;
}
2020-04-17 03:58:54 -04:00
/**
* Store tag/mention elements
*
* @param integer $uriid
* @param string $hash
* @param string $name
* @param string $url
2020-04-20 05:47:26 -04:00
* @param boolean $probing
2020-04-17 03:58:54 -04:00
*/
2020-04-20 05:47:26 -04:00
public static function storeByHash(int $uriid, string $hash, string $name, string $url = '', $probing = true)
{
$type = self::getTypeForHash($hash);
if ($type == self::UNKNOWN) {
return;
}
2020-04-20 05:47:26 -04:00
self::store($uriid, $type, $name, $url, $probing);
}
2020-04-15 12:37:09 -04:00
/**
* Store tags and mentions from the body
*
2020-04-20 05:47:26 -04:00
* @param integer $uriid URI-Id
* @param string $body Body of the post
* @param string $tags Accepted tags
* @param boolean $probing Perform a probing for contacts, adding them if needed
2020-04-15 12:37:09 -04:00
*/
2020-04-20 05:47:26 -04:00
public static function storeFromBody(int $uriid, string $body, string $tags = null, $probing = true)
2020-04-15 12:37:09 -04:00
{
2020-04-18 12:14:38 -04:00
if (is_null($tags)) {
$tags = self::TAG_CHARACTER[self::HASHTAG] . self::TAG_CHARACTER[self::MENTION] . self::TAG_CHARACTER[self::EXCLUSIVE_MENTION];
}
Logger::info('Check for tags', ['uri-id' => $uriid, 'hash' => $tags, 'callstack' => System::callstack()]);
if (!preg_match_all("/([" . $tags . "])\[url\=([^\[\]]*)\]([^\[\]]*)\[\/url\]/ism", $body, $result, PREG_SET_ORDER)) {
2020-04-15 12:37:09 -04:00
return;
}
Logger::info('Found tags', ['uri-id' => $uriid, 'hash' => $tags, 'result' => $result]);
foreach ($result as $tag) {
2020-04-20 05:47:26 -04:00
self::storeByHash($uriid, $tag[1], $tag[3], $tag[2], $probing);
2020-04-15 12:37:09 -04:00
}
}
/**
* Store raw tags (not encapsulated in links) from the body
* This function is needed in the intermediate phase.
* Later we can call item::setHashtags in advance to have all tags converted.
*
* @param integer $uriid URI-Id
* @param string $body Body of the post
*/
public static function storeRawTagsFromBody(int $uriid, string $body)
{
Logger::info('Check for tags', ['uri-id' => $uriid, 'callstack' => System::callstack()]);
$result = BBCode::getTags($body);
if (empty($result)) {
return;
}
Logger::info('Found tags', ['uri-id' => $uriid, 'result' => $result]);
foreach ($result as $tag) {
if (substr($tag, 0, 1) != self::TAG_CHARACTER[self::HASHTAG]) {
continue;
}
self::storeByHash($uriid, substr($tag, 0, 1), substr($tag, 1));
}
}
2020-04-26 11:24:58 -04:00
/**
* Checks for stored hashtags and mentions for the given post
*
* @param integer $uriid
* @return bool
*/
public static function existsForPost(int $uriid)
{
return DBA::exists('post-tag', ['uri-id' => $uriid, 'type' => [self::HASHTAG, self::MENTION, self::IMPLICIT_MENTION, self::EXCLUSIVE_MENTION]]);
}
/**
* Remove tag/mention
*
* @param integer $uriid
* @param integer $type
* @param string $name
* @param string $url
*/
public static function remove(int $uriid, int $type, string $name, string $url = '')
{
2020-04-24 07:04:50 -04:00
$condition = ['uri-id' => $uriid, 'type' => $type, 'url' => $url];
if ($type == self::HASHTAG) {
$condition['name'] = $name;
}
$tag = DBA::selectFirst('tag-view', ['tid', 'cid'], $condition);
if (!DBA::isResult($tag)) {
return;
}
2020-04-24 07:04:50 -04:00
Logger::info('Removing tag/mention', ['uri-id' => $uriid, 'tid' => $tag['tid'], 'name' => $name, 'url' => $url, 'callstack' => System::callstack(8)]);
DBA::delete('post-tag', ['uri-id' => $uriid, 'type' => $type, 'tid' => $tag['tid'], 'cid' => $tag['cid']]);
}
/**
* Remove tag/mention
*
* @param integer $uriid
* @param string $hash
* @param string $name
* @param string $url
*/
public static function removeByHash(int $uriid, string $hash, string $name, string $url = '')
{
$type = self::getTypeForHash($hash);
if ($type == self::UNKNOWN) {
return;
}
self::remove($uriid, $type, $name, $url);
}
/**
* Get the type for the given hash
*
* @param string $hash
* @return integer type
*/
private static function getTypeForHash(string $hash)
{
if ($hash == self::TAG_CHARACTER[self::MENTION]) {
return self::MENTION;
} elseif ($hash == self::TAG_CHARACTER[self::EXCLUSIVE_MENTION]) {
return self::EXCLUSIVE_MENTION;
} elseif ($hash == self::TAG_CHARACTER[self::IMPLICIT_MENTION]) {
return self::IMPLICIT_MENTION;
} elseif ($hash == self::TAG_CHARACTER[self::HASHTAG]) {
return self::HASHTAG;
} else {
return self::UNKNOWN;
}
2020-04-26 11:24:58 -04:00
}
/**
* Create implicit mentions for a given post
*
* @param integer $uri_id
* @param integer $parent_uri_id
*/
public static function createImplicitMentions(int $uri_id, int $parent_uri_id)
{
if ($uri_id == $parent_uri_id) {
return;
}
2020-05-09 04:08:33 -04:00
// Always mention the direct parent author
$parent = Item::selectFirst(['author-link', 'author-name'], ['uri-id' => $parent_uri_id]);
self::store($uri_id, self::IMPLICIT_MENTION, $parent['author-name'], $parent['author-link']);
if (DI::config()->get('system', 'disable_implicit_mentions')) {
return;
}
$tags = DBA::select('tag-view', ['name', 'url'], ['uri-id' => $parent_uri_id]);
while ($tag = DBA::fetch($tags)) {
self::store($uri_id, self::IMPLICIT_MENTION, $tag['name'], $tag['url']);
}
}
2020-04-26 11:24:58 -04:00
/**
* Retrieves the terms from the provided type(s) associated with the provided item ID.
*
* @param int $item_id
* @param int|array $type
* @return array
* @throws \Exception
*/
2020-05-01 02:01:22 -04:00
public static function getByURIId(int $uri_id, array $type = [self::HASHTAG, self::MENTION, self::IMPLICIT_MENTION, self::EXCLUSIVE_MENTION])
2020-04-26 11:24:58 -04:00
{
$condition = ['uri-id' => $uri_id, 'type' => $type];
return DBA::selectToArray('tag-view', ['type', 'name', 'url'], $condition);
}
2020-04-26 11:24:58 -04:00
/**
* Return a string with all tags and mentions
*
* @param integer $uri_id
* @return string tags and mentions
*/
2020-05-05 17:58:25 -04:00
public static function getCSVByURIId(int $uri_id)
{
$tag_list = [];
$tags = self::getByURIId($uri_id);
foreach ($tags as $tag) {
$tag_list[] = self::TAG_CHARACTER[$tag['type']] . '[url=' . $tag['url'] . ']' . $tag['name'] . '[/url]';
}
return implode(',', $tag_list);
}
2020-05-01 02:01:22 -04:00
/**
2020-04-26 11:24:58 -04:00
* Sorts an item's tags into mentions, hashtags and other tags. Generate personalized URLs by user and modify the
* provided item's body with them.
*
* @param array $item
* @return array
* @throws \Friendica\Network\HTTPException\InternalServerErrorException
* @throws \ImagickException
*/
2020-05-01 02:01:22 -04:00
public static function populateFromItem(&$item)
2020-04-26 11:24:58 -04:00
{
$return = [
'tags' => [],
'hashtags' => [],
'mentions' => [],
'implicit_mentions' => [],
];
$searchpath = DI::baseUrl() . "/search?tag=";
$taglist = DBA::select('tag-view', ['type', 'name', 'url'],
['uri-id' => $item['uri-id'], 'type' => [self::HASHTAG, self::MENTION, self::EXCLUSIVE_MENTION, self::IMPLICIT_MENTION]]);
while ($tag = DBA::fetch($taglist)) {
if ($tag['url'] == '') {
$tag['url'] = $searchpath . rawurlencode($tag['name']);
}
$orig_tag = $tag['url'];
$prefix = self::TAG_CHARACTER[$tag['type']];
switch($tag['type']) {
case self::HASHTAG:
if ($orig_tag != $tag['url']) {
$item['body'] = str_replace($orig_tag, $tag['url'], $item['body']);
}
$return['hashtags'][] = $prefix . '<a href="' . $tag['url'] . '" target="_blank" rel="noopener noreferrer">' . htmlspecialchars($tag['name']) . '</a>';
$return['tags'][] = $prefix . '<a href="' . $tag['url'] . '" target="_blank" rel="noopener noreferrer">' . htmlspecialchars($tag['name']) . '</a>';
break;
case self::MENTION:
case self::EXCLUSIVE_MENTION:
$tag['url'] = Contact::magicLink($tag['url']);
$return['mentions'][] = $prefix . '<a href="' . $tag['url'] . '" target="_blank" rel="noopener noreferrer">' . htmlspecialchars($tag['name']) . '</a>';
$return['tags'][] = $prefix . '<a href="' . $tag['url'] . '" target="_blank" rel="noopener noreferrer">' . htmlspecialchars($tag['name']) . '</a>';
break;
case self::IMPLICIT_MENTION:
$return['implicit_mentions'][] = $prefix . $tag['name'];
break;
}
}
DBA::close($taglist);
return $return;
}
2020-04-28 07:52:51 -04:00
/**
* Search posts for given tag
*
* @param string $search
* @param integer $uid
* @param integer $start
* @param integer $limit
* @return array with URI-ID
*/
2020-05-01 09:13:58 -04:00
public static function getURIIdListByTag(string $search, int $uid = 0, int $start = 0, int $limit = 100)
2020-04-28 07:52:51 -04:00
{
$condition = ["`name` = ? AND (NOT `private` OR (`private` AND `uid` = ?))", $search, $uid];
$params = [
'order' => ['uri-id' => true],
'group_by' => ['uri-id'],
'limit' => [$start, $limit]
];
$tags = DBA::select('tag-search-view', ['uri-id'], $condition, $params);
$uriids = [];
while ($tag = DBA::fetch($tags)) {
$uriids[] = $tag['uri-id'];
}
DBA::close($tags);
return $uriids;
}
2020-05-01 06:57:32 -04:00
/**
* Returns a list of the most frequent global hashtags over the given period
*
* @param int $period Period in hours to consider posts
* @return array
* @throws \Exception
*/
public static function getGlobalTrendingHashtags(int $period, $limit = 10)
{
$tags = DI::cache()->get('global_trending_tags');
if (empty($tags)) {
$tagsStmt = DBA::p("SELECT `name` AS `term`, COUNT(*) AS `score`
FROM `tag-search-view`
WHERE `private` = ? AND `received` > DATE_SUB(NOW(), INTERVAL ? HOUR)
GROUP BY `term` ORDER BY `score` DESC LIMIT ?",
Item::PUBLIC, $period, $limit);
if (DBA::isResult($tagsStmt)) {
$tags = DBA::toArray($tagsStmt);
DI::cache()->set('global_trending_tags', $tags, Duration::HOUR);
}
}
return $tags ?: [];
}
/**
* Returns a list of the most frequent local hashtags over the given period
*
* @param int $period Period in hours to consider posts
* @return array
* @throws \Exception
*/
public static function getLocalTrendingHashtags(int $period, $limit = 10)
{
$tags = DI::cache()->get('local_trending_tags');
if (empty($tags)) {
$tagsStmt = DBA::p("SELECT `name` AS `term`, COUNT(*) AS `score`
FROM `tag-search-view`
WHERE `private` = ? AND `wall` AND `origin` AND `received` > DATE_SUB(NOW(), INTERVAL ? HOUR)
GROUP BY `term` ORDER BY `score` DESC LIMIT ?",
Item::PUBLIC, $period, $limit);
if (DBA::isResult($tagsStmt)) {
$tags = DBA::toArray($tagsStmt);
DI::cache()->set('local_trending_tags', $tags, Duration::HOUR);
}
}
return $tags ?: [];
}
2020-05-01 08:39:41 -04:00
/**
* Check if the provided tag is of one of the provided term types.
*
* @param string $tag
* @param int ...$types
* @return bool
*/
public static function isType($tag, ...$types)
{
$tag_chars = [];
foreach ($types as $type) {
if (array_key_exists($type, self::TAG_CHARACTER)) {
$tag_chars[] = self::TAG_CHARACTER[$type];
}
}
return Strings::startsWith($tag, $tag_chars);
}
2020-04-15 12:37:09 -04:00
}