The livechat plugin stores some data on the server,
in the /var/www/peertube/storage/plugins/data/peertube-plugin-livechat/
folder.
This page describes these data.
The prosody
folder is used by the Prosody XMPP server to store its data and logs.
This plugin uses the Prosoxy XMPP server in background. This server code is embedded as an AppImage.
When the plugin starts, it deflate this AppImage in the prosodyAppImage
folder.
To handle federation between Peertube instances, the plugin needs to store some information concerning remote instances (available protocols, …).
The plugin stores these data in the serverInfos/instance_uri
folder
(where instance_uri
is replaced by the instance uri).
In each instance’s folder, there can be these files:
last-update
: json file containing the timestamp of the last information update. So we can avoid refreshing too often.ws-s2s
: if the server allows XMPP S2S Websocket connections, here are the endpoint informations2s
: if the server allows direct XMPP S2S connections, here are the port and url informationTo handle federation the plugin needs to store some information about remote videos.
So, each time we open a new remote chat, a file videoInfos/remote/instance_uri/video_uuid.json
is created
(where instance_uri
is the origin instance uri, and video_uuid
is the video uuid).
This JSON files contain some data about the remote chat (is it enabled, are anonymous users authorized, which protocol can we use, …). These data can then be read by the Prosody server to connect to the remote chat.
Moreover, when the current instance builds such data for local videos,
it stores it in videoInfos/local/video_uuid.json
(where video_uuid
is the video uuid).
The channelConfigurationOptions
folder contains JSON files describing channels advanced configuration.
Filenames are like 1.json
where 1
is the channel id.
The content of the files are similar to the content sent by the front-end when saving these configuration.
Some parts of the plugin need a quick way to get the channel id from the room Jabber ID, or the all room Jabber ID from a channel id. We won’t use SQL queries, because we only want such information for video that have a chatroom.
So we will store in the room-channel/muc_domain.json
file (where muc_domain
is the current MUC domain,
something like room.instance.tld
) a JSON object representing these relations).
In the JSON object, keys are the channel ID (as string), values are arrays of strings representing the rooms JIDs local part (without the MUC domain).
When a chatroom is created, the corresponding entry will be added.
Here is a sample file:
{
"1": [
"8df24108-6e70-4fc8-b1cc-f2db7fcdd535"
]
}
This file is loaded at the plugin startup into an object that can manipulate these data.
So we can easily list all rooms for a given channel id or get the channel id for a room JID (Jabber ID).
Note: we include the MUC domain (room.instance.tld
) in the filename in case the instance domain changes.
In such case, existing rooms could get lost, and we want a way to ignore them to avoid gettings errors.
Note: there could be some inconsistencies, when video or rooms are deleted. The code must take this into account, and always double check room or channel existence. There will be some cleaning batch, to delete deprecated files.
The bot/muc_domain
(where muc_domain is the current MUC domain) folder contains configuration files that are read by the moderation bot.
This bot uses the xmppjs-chat-bot package.
Note: we include the MUC domain (room.instance.tld
) in the dirname in case the instance domain changes.
In such case, existing rooms could get lost, and we want a way to ignore them to avoid gettings errors.
The bot/muc_domain/moderation.json
file contains the moderation bot global configuration.
This bot uses the xmppjs-chat-bot package, see it’s README file for more information.
Note: this includes the bot username and password. Don’t let it leak.
The bot/muc_domain/rooms
folder contains room configuration files.
See the xmppjs-chat-bot package help for more information.