gnu-social/classes/Inbox.php

334 lines
11 KiB
PHP
Raw Normal View History

2009-12-30 06:02:46 +00:00
<?php
/**
* StatusNet, the distributed open-source microblogging tool
*
* Data class for user location preferences
*
* PHP version 5
*
* LICENCE: 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 <http://www.gnu.org/licenses/>.
*
* @category Data
* @package StatusNet
* @author Evan Prodromou <evan@status.net>
* @copyright 2009 StatusNet Inc.
* @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
* @link http://status.net/
*/
require_once INSTALLDIR.'/classes/Memcached_DataObject.php';
class Inbox extends Memcached_DataObject
{
2009-12-30 19:06:07 +00:00
const BOXCAR = 128;
const MAX_NOTICES = 1024;
2009-12-30 19:06:07 +00:00
2009-12-30 06:02:46 +00:00
###START_AUTOCODE
/* the code below is auto generated do not remove the above tag */
public $__table = 'inbox'; // table name
public $user_id; // int(4) primary_key not_null
public $notice_ids; // blob
/* Static get */
function staticGet($k,$v=NULL) { return Memcached_DataObject::staticGet('Inbox',$k,$v); }
/* the code above is auto generated do not remove the tag below */
###END_AUTOCODE
function sequenceKey()
{
return array(false, false, false);
}
2009-12-30 19:06:07 +00:00
/**
* Create a new inbox from existing Notice_inbox stuff
*/
static function initialize($user_id)
{
$inbox = Inbox::fromNoticeInbox($user_id);
unset($inbox->fake);
$result = $inbox->insert();
if (!$result) {
common_log_db_error($inbox, 'INSERT', __FILE__);
return null;
}
return $inbox;
}
static function fromNoticeInbox($user_id)
2009-12-30 19:06:07 +00:00
{
$ids = array();
$ni = new Notice_inbox();
$ni->user_id = $user_id;
$ni->selectAdd();
$ni->selectAdd('notice_id');
$ni->orderBy('notice_id DESC');
$ni->limit(0, self::MAX_NOTICES);
if ($ni->find()) {
while($ni->fetch()) {
$ids[] = $ni->notice_id;
}
}
$ni->free();
unset($ni);
2009-12-30 19:06:07 +00:00
$inbox = new Inbox();
$inbox->user_id = $user_id;
$inbox->pack($ids);
$inbox->fake = true;
return $inbox;
}
/**
* Append the given notice to the given user's inbox.
* Caching updates are managed for the inbox itself.
*
* If the notice is already in this inbox, the second
* add will be silently dropped.
*
* @param int @user_id
* @param int $notice_id
* @return boolean success
*/
static function insertNotice($user_id, $notice_id)
{
// Going straight to the DB rather than trusting our caching
// during an update. Note: not using DB_DataObject::staticGet,
// which is unsafe to use directly (in-process caching causes
// memory leaks, which accumulate in queue processes).
$inbox = new Inbox();
if (!$inbox->get('user_id', $user_id)) {
$inbox = Inbox::initialize($user_id);
}
if (empty($inbox)) {
return false;
}
$ids = $inbox->unpack();
if (in_array(intval($notice_id), $ids)) {
// Already in there, we probably re-ran some inbox adds
// due to an error. Skip the dupe silently.
return true;
}
$result = $inbox->query(sprintf('UPDATE inbox '.
'set notice_ids = concat(cast(0x%08x as binary(4)), '.
'substr(notice_ids, 1, %d)) '.
'WHERE user_id = %d',
$notice_id,
4 * (self::MAX_NOTICES - 1),
$user_id));
if ($result) {
Offload inbox updates to a queue handler to speed up posting online Moved much of the writing that happens when posting a notice to a new queuehandler, distribqueuehandler. This updates tags, groups, replies and inboxes at queue time (or at Web time, if queues are disabled). To make this work well, I had to break up the monolithic Notice::blowCaches() and make cache blowing happen closer to where data is updated. Squashed commit of the following: commit 5257626c62750ac4ac1db0ce2b71410c5711cfa3 Author: Evan Prodromou <evan@status.net> Date: Mon Jan 25 14:56:41 2010 -0500 slightly better handling of blowing tag memory cache commit 8a22a3cdf6ec28685da129a0313e7b2a0837c9ef Author: Evan Prodromou <evan@status.net> Date: Mon Jan 25 01:42:56 2010 -0500 change 'distribute' to 'distrib' so not too long for dbqueue commit 7a063315b0f7fad27cb6fbd2bdd74e253af83e4f Author: Evan Prodromou <evan@status.net> Date: Mon Jan 25 01:39:15 2010 -0500 change handle_notice() to handle() in distributqueuehandler commit 1a39ccd28b9994137d7bfd21bb4f230546938e77 Author: Evan Prodromou <evan@status.net> Date: Mon Jan 25 16:05:25 2010 -0500 error with queuemanager commit e6b3bb93f305cfd2de71a6340b8aa6fb890049b7 Author: Evan Prodromou <evan@status.net> Date: Mon Jan 25 01:11:34 2010 -0500 Blow memcache at different point rather than one big function for Notice class commit 94d557cdc016187d1d0647ae1794cd94d6fb8ac8 Author: Evan Prodromou <evan@status.net> Date: Mon Jan 25 00:48:44 2010 -0500 Blow memcache at different point rather than one big function for Notice class commit 1c781dd08c88a35dafc5c01230b4872fd6b95182 Author: Evan Prodromou <evan@status.net> Date: Wed Jan 20 08:54:18 2010 -0500 move broadcasting and distributing to new queuehandler commit da3e46d26b84e4f028f34a13fd2ee373e4c1b954 Author: Evan Prodromou <evan@status.net> Date: Wed Jan 20 08:53:12 2010 -0500 Move distribution of notices to new distribute queue handler
2010-01-25 23:08:21 +00:00
self::blow('inbox:user_id:%d', $user_id);
}
return $result;
2009-12-30 19:06:07 +00:00
}
static function bulkInsert($notice_id, $user_ids)
{
foreach ($user_ids as $user_id)
{
Inbox::insertNotice($user_id, $notice_id);
2009-12-30 19:06:07 +00:00
}
}
function stream($user_id, $offset, $limit, $since_id, $max_id, $own=false)
2009-12-30 19:06:07 +00:00
{
$inbox = Inbox::staticGet('user_id', $user_id);
if (empty($inbox)) {
$inbox = Inbox::fromNoticeInbox($user_id);
if (empty($inbox)) {
return array();
} else {
$inbox->encache();
}
2009-12-30 19:06:07 +00:00
}
$ids = $inbox->unpack();
2009-12-30 19:06:07 +00:00
2010-01-14 06:32:17 +00:00
if (!empty($since_id)) {
$newids = array();
foreach ($ids as $id) {
if ($id > $since_id) {
$newids[] = $id;
}
2010-01-14 06:32:17 +00:00
}
$ids = $newids;
2010-01-14 06:32:17 +00:00
}
if (!empty($max_id)) {
$newids = array();
foreach ($ids as $id) {
if ($id <= $max_id) {
$newids[] = $id;
}
2010-01-14 06:32:17 +00:00
}
$ids = $newids;
2010-01-14 06:32:17 +00:00
}
2009-12-30 19:06:07 +00:00
$ids = array_slice($ids, $offset, $limit);
return $ids;
2009-12-30 19:06:07 +00:00
}
/**
* Wrapper for Inbox::stream() and Notice::getStreamByIds() returning
* additional items up to the limit if we were short due to deleted
* notices still being listed in the inbox.
*
* This is meant to assist threaded views, and optimizes paging for
* threadness. Not ideal for very late pages, as we have to bump about
* through all previous items.
*
* Should avoid duplicates in paging, though.
*
* @param int $user_id
* @param int $offset skip past the most recent N notices (after since_id checks)
* @param int $limit
* @param mixed $since_id return only notices after but not including this id
* @param mixed $max_id return only notices up to and including this id
* @param mixed $own ignored?
* @return array of Notice objects
*
* @todo consider repacking the inbox when this happens?
* @fixme reimplement $own if we need it?
*/
function streamNoticesThreaded($user_id, $offset, $limit, $since_id, $max_id, $own=false)
{
// So what we want is:
// * slurp in the beginning of the notice list
// * filter out deleted notices
// * replace any reply notices with their conversation roots
// * filter out any duplicate conversations
// * return $limit notices after skipping $offset from the most recent
$ids = self::stream($user_id, 0, self::MAX_NOTICES, $since_id, $max_id, $own);
// Do a bulk lookup for the first $limit items
// Fast path when nothing's deleted.
$firstChunk = array_slice($ids, 0, $offset + $limit);
$notices = NoticeStream::getStreamByIds($firstChunk);
assert($notices instanceof ArrayWrapper);
$items = $notices->_items;
// Extract the latest non-deleted item in each convo
$noticeByConvo = array();
foreach ($items as $notice) {
if (empty($noticeByConvo[$notice->conversation])) {
$noticeByConvo[$notice->conversation] = $notice;
}
}
$wanted = count($firstChunk); // raw entry count in the inbox up to our $limit
// There were deleted notices, we'll need to look for more.
$remainder = array_slice($ids, $limit);
for ($i = $offset + $limit; count($noticeByConvo) < $wanted && $i < count($ids); $i++) {
$notice = Notice::staticGet($ids[$i]);
if ($notice && empty($noticeByConvo[$notice->conversation])) {
$noticeByConvo[$notice->conversation] = $notice;
}
}
$slice = array_slice($noticeByConvo, $offset, $limit, false);
return new ArrayWrapper($slice);
}
/**
* Wrapper for Inbox::stream() and Notice::getStreamByIds() returning
* additional items up to the limit if we were short due to deleted
* notices still being listed in the inbox.
*
* The fast path (when no items are deleted) should be just as fast; the
* offset parameter is applied *before* lookups for maximum efficiency.
*
* This means offset-based paging may show duplicates, but similar behavior
* already exists when new notices are posted between page views, so we
* think people will be ok with this until id-based paging is introduced
* to the user interface.
*
* @param int $user_id
* @param int $offset skip past the most recent N notices (after since_id checks)
* @param int $limit
* @param mixed $since_id return only notices after but not including this id
* @param mixed $max_id return only notices up to and including this id
* @param mixed $own ignored?
* @return array of Notice objects
*
* @todo consider repacking the inbox when this happens?
* @fixme reimplement $own if we need it?
*/
function streamNotices($user_id, $offset, $limit, $since_id, $max_id, $own=false)
{
$ids = self::stream($user_id, $offset, self::MAX_NOTICES, $since_id, $max_id, $own);
// Do a bulk lookup for the first $limit items
// Fast path when nothing's deleted.
$firstChunk = array_slice($ids, 0, $limit);
$notices = NoticeStream::getStreamByIds($firstChunk);
$wanted = count($firstChunk); // raw entry count in the inbox up to our $limit
if ($notices->N >= $wanted) {
return $notices;
}
// There were deleted notices, we'll need to look for more.
assert($notices instanceof ArrayWrapper);
$items = $notices->_items;
$remainder = array_slice($ids, $limit);
while (count($items) < $wanted && count($remainder) > 0) {
$notice = Notice::staticGet(array_shift($remainder));
if ($notice) {
$items[] = $notice;
} else {
}
}
return new ArrayWrapper($items);
}
/**
* Saves a list of integer notice_ids into a packed blob in this object.
* @param array $ids list of integer notice_ids
*/
protected function pack(array $ids)
{
$this->notice_ids = call_user_func_array('pack', array_merge(array('N*'), $ids));
}
/**
* @return array of integer notice_ids
*/
protected function unpack()
{
return unpack('N*', $this->notice_ids);
}
2009-12-30 06:02:46 +00:00
}