# -*- coding: utf-8 -*-
import threading
import logging
import time
import select
import socket
import ssl
import struct
from errors import *
from constants import *
import users
import channels
import blobs
import commands
import messages
import callbacks
import tools
import soundoutput
import mumble_pb2
from pycelt import SUPPORTED_BITSTREAMS
class Mumble(threading.Thread):
"""
Mumble client library main object.
basically a thread
"""
def __init__(self, host=None, port=None, user=None, password=None, client_certif=None, reconnect=False, debug=False):
"""
host=mumble server hostname or address
port=mumble server port
user=user to use for the connection
password=password for the connection
client_certif=client certificate to authenticate the connection (NOT IMPLEMENTED)
reconnect=if True, try to reconnect if disconnected
debug=if True, send debugging messages (lot of...) to the stdout
"""
#TODO: client certificate authentication
#TODO: exit both threads properly
#TODO: use UDP audio
threading.Thread.__init__(self)
self.Log = logging.getLogger("PyMumble") # logging object for errors and debugging
if debug:
self.Log.setLevel(logging.DEBUG)
else:
self.Log.setLevel(logging.ERROR)
ch = logging.StreamHandler()
ch.setLevel(logging.DEBUG)
formatter = logging.Formatter('%(asctime)s-%(name)s-%(levelname)s-%(message)s')
ch.setFormatter(formatter)
self.Log.addHandler(ch)
self.parent_thread = threading.current_thread() # main thread of the calling application
self.mumble_thread = None # thread of the mumble client library
self.host = host
self.port = port
self.user = user
self.password = password
self.client_certif = client_certif
self.reconnect = reconnect
self.receive_sound = False # set to True to treat incoming audio, otherwise it is simply ignored
self.loop_rate = PYMUMBLE_LOOP_RATE
self.application = PYMUMBLE_VERSION_STRING
self.callbacks = callbacks.CallBacks() #callbacks management
self.ready_lock = threading.Lock() # released when the connection is fully established with the server
self.ready_lock.acquire()
def init_connection(self):
"""Initialize variables that are local to a connection, (needed if the client automatically reconnect)"""
self.ready_lock.acquire(False) # reacquire the ready-lock in case of reconnection
self.connected = PYMUMBLE_CONN_STATE_NOT_CONNECTED
self.control_socket = None
self.media_socket = None # Not implemented - for UDP media
self.bandwidth = PYMUMBLE_BANDWIDTH # reset the outgoing bandwidth to it's default before connectiong
self.server_max_bandwidth = None
self.udp_active = False
self.users = users.Users(self, self.callbacks) # contain the server's connected users informations
self.channels = channels.Channels(self, self.callbacks) # contain the server's channels informations
self.blobs = blobs.Blobs(self) # manage the blob objects
self.sound_output = soundoutput.SoundOutput(self, PYMUMBLE_AUDIO_PER_PACKET, self.bandwidth) # manage the outgoing sounds
self.commands = commands.Commands() # manage commands sent between the main and the mumble threads
self.receive_buffer = "" # initialize the control connection input buffer
def run(self):
"""Connect to the server and start the loop in its thread. Retry if requested"""
self.mumble_thread = threading.current_thread()
# loop if auto-reconnect is requested
while True:
self.init_connection() # reset the connection-specific object members
self.connect()
self.loop()
if not self.reconnect or not self.parent_thread.is_alive():
break
time.sleep(PYMUMBLE_CONNECTION_RETRY_INTERVAL)
def connect(self):
"""Connect to the server"""
# Connect the SSL tunnel
self.Log.debug("connecting to %s on port %i.", self.host, self.port)
std_sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
self.control_socket = ssl.wrap_socket(std_sock, certfile=self.client_certif, ssl_version=ssl.PROTOCOL_TLSv1)
self.control_socket.connect((self.host, self.port))
self.control_socket.setblocking(0)
# Perform the Mumble authentication
version = mumble_pb2.Version()
version.version = (PYMUMBLE_PROTOCOL_VERSION[0] << 16) + (PYMUMBLE_PROTOCOL_VERSION[1] << 8) + PYMUMBLE_PROTOCOL_VERSION[2]
version.release = self.application
version.os = PYMUMBLE_OS_STRING
version.os_version = PYMUMBLE_OS_VERSION_STRING
self.Log.debug("sending: version: %s", version)
self.send_message(PYMUMBLE_MSG_TYPES_VERSION, version)
authenticate = mumble_pb2.Authenticate()
authenticate.username = self.user
authenticate.password = self.password
authenticate.celt_versions.extend(SUPPORTED_BITSTREAMS.keys())
# authenticate.celt_versions.extend([-2147483637]) # for debugging - only celt 0.7
authenticate.opus = True
self.Log.debug("sending: authenticate: %s", authenticate)
self.send_message(PYMUMBLE_MSG_TYPES_AUTHENTICATE, authenticate)
self.connected = PYMUMBLE_CONN_STATE_AUTHENTICATING
def loop(self):
"""
Main loop
waiting for a message from the server for maximum self.loop_rate time
take care of sending the ping
take care of sending the queued commands to the server
check on every iteration for outgoing sound
check for disconnection
"""
self.Log.debug("entering loop")
last_ping = time.time() # keep track of the last ping time
# loop as long as the connection and the parent thread are alive
while self.connected != PYMUMBLE_CONN_STATE_NOT_CONNECTED and self.parent_thread.is_alive():
if last_ping + PYMUMBLE_PING_DELAY <= time.time(): # when it is time, send the ping
self.ping()
last_ping = time.time()
if self.connected == PYMUMBLE_CONN_STATE_CONNECTED:
while self.commands.is_cmd():
self.treat_command(self.commands.pop_cmd()) # send the commands coming from the application to the server
self.sound_output.send_audio() # send outgoing audio if available
(rlist, wlist, xlist) = select.select([self.control_socket], [], [self.control_socket], self.loop_rate) # wait for a socket activity
if self.control_socket in rlist: # something to be read on the control socket
self.read_control_messages()
elif self.control_socket in xlist: # socket was closed
self.control_socket.close()
self.connected = PYMUMBLE_CONN_STATE_NOT_CONNECTED
def ping(self):
"""Send the keepalive through available channels"""
#TODO: Ping counters
ping = mumble_pb2.Ping()
ping.timestamp=int(time.time())
self.Log.debug("sending: ping: %s", ping)
self.send_message(PYMUMBLE_MSG_TYPES_PING, ping)
def send_message(self, type, message):
"""Send a control message to the server"""
packet=struct.pack("!HL", type, message.ByteSize()) + message.SerializeToString()
while len(packet)>0:
self.Log.debug("sending message")
sent=self.control_socket.send(packet)
if sent < 0:
raise socket.error("Server socket error")
packet=packet[sent:]
def read_control_messages(self):
"""Read control messages coming from the server"""
# from tools import toHex # for debugging
buffer = self.control_socket.recv(PYMUMBLE_READ_BUFFER_SIZE)
self.receive_buffer += buffer
while len(self.receive_buffer) >= 6: # header is present (type + length)
self.Log.debug("read control connection")
header = self.receive_buffer[0:6]
(type, size) = struct.unpack("!HL", header) # decode header
if len(self.receive_buffer) < size+6: # if not length data, read further
break
# self.Log.debug("message received : " + toHex(self.receive_buffer[0:size+6])) # for debugging
message = self.receive_buffer[6:size+6] # get the control message
self.receive_buffer = self.receive_buffer[size+6:] # remove from the buffer the read part
self.dispatch_control_message(type, message)
def dispatch_control_message(self, type, message):
"""Dispatch control messages based on their type"""
self.Log.debug("dispatch control message")
if type == PYMUMBLE_MSG_TYPES_UDPTUNNEL: # audio encapsulated in control message
self.sound_received(message)
elif type == PYMUMBLE_MSG_TYPES_VERSION:
mess = mumble_pb2.Version()
mess.ParseFromString(message)
self.Log.debug("message: Version : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_AUTHENTICATE:
mess = mumble_pb2.Authenticate()
mess.ParseFromString(message)
self.Log.debug("message: Authenticate : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_PING:
mess = mumble_pb2.Ping()
mess.ParseFromString(message)
self.Log.debug("message: Ping : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_REJECT:
mess = mumble_pb2.Reject()
mess.ParseFromString(message)
self.Log.debug("message: reject : %s", mess)
self.ready_lock.release()
raise ConnectionRejectedError(mess.reason)
elif type == PYMUMBLE_MSG_TYPES_SERVERSYNC: # this message finish the connection process
mess = mumble_pb2.ServerSync()
mess.ParseFromString(message)
self.Log.debug("message: serversync : %s", mess)
self.users.set_myself(mess.session)
self.server_max_bandwidth = mess.max_bandwidth
self.set_bandwidth(mess.max_bandwidth)
if self.connected == PYMUMBLE_CONN_STATE_AUTHENTICATING:
self.connected = PYMUMBLE_CONN_STATE_CONNECTED
self.callbacks(PYMUMBLE_CLBK_CONNECTED)
self.ready_lock.release() # release the ready-lock
elif type == PYMUMBLE_MSG_TYPES_CHANNELREMOVE:
mess = mumble_pb2.ChannelRemove()
mess.ParseFromString(message)
self.Log.debug("message: ChannelRemove : %s", mess)
self.channels.remove(mess.channel_id)
elif type == PYMUMBLE_MSG_TYPES_CHANNELSTATE:
mess = mumble_pb2.ChannelState()
mess.ParseFromString(message)
self.Log.debug("message: channelstate : %s", mess)
self.channels.update(mess)
elif type == PYMUMBLE_MSG_TYPES_USERREMOVE:
mess = mumble_pb2.UserRemove()
mess.ParseFromString(message)
self.Log.debug("message: UserRemove : %s", mess)
self.users.remove(mess)
elif type == PYMUMBLE_MSG_TYPES_USERSTATE:
mess = mumble_pb2.UserState()
mess.ParseFromString(message)
self.Log.debug("message: userstate : %s", mess)
self.users.update(mess)
elif type == PYMUMBLE_MSG_TYPES_BANLIST:
mess = mumble_pb2.BanList()
mess.ParseFromString(message)
self.Log.debug("message: BanList : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_TEXTMESSAGE:
mess = mumble_pb2.TextMessage()
mess.ParseFromString(message)
self.Log.debug("message: TextMessage : %s", mess)
self.callbacks(PYMUMBLE_CLBK_TEXTMESSAGERECEIVED, mess.message)
elif type == PYMUMBLE_MSG_TYPES_PERMISSIONDENIED:
mess = mumble_pb2.PermissionDenied()
mess.ParseFromString(message)
self.Log.debug("message: PermissionDenied : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_ACL:
mess = mumble_pb2.ACL()
mess.ParseFromString(message)
self.Log.debug("message: ACL : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_QUERYUSERS:
mess = mumble_pb2.QueryUsers()
mess.ParseFromString(message)
self.Log.debug("message: QueryUsers : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_CRYPTSETUP:
mess = mumble_pb2.CryptSetup()
mess.ParseFromString(message)
self.Log.debug("message: CryptSetup : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_CONTEXTACTIONADD:
mess = mumble_pb2.ContextActionAdd()
mess.ParseFromString(message)
self.Log.debug("message: ContextActionAdd : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_CONTEXTACTION:
mess = mumble_pb2.ContextActionAdd()
mess.ParseFromString(message)
self.Log.debug("message: ContextActionAdd : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_USERLIST:
mess = mumble_pb2.UserList()
mess.ParseFromString(message)
self.Log.debug("message: UserList : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_VOICETARGET:
mess = mumble_pb2.VoiceTarget()
mess.ParseFromString(message)
self.Log.debug("message: VoiceTarget : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_PERMISSIONQUERY:
mess = mumble_pb2.PermissionQuery()
mess.ParseFromString(message)
self.Log.debug("message: PermissionQuery : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_CODECVERSION:
mess = mumble_pb2.CodecVersion()
mess.ParseFromString(message)
self.Log.debug("message: CodecVersion : %s", mess)
self.sound_output.set_default_codec(mess)
elif type == PYMUMBLE_MSG_TYPES_USERSTATS:
mess = mumble_pb2.UserStats()
mess.ParseFromString(message)
self.Log.debug("message: UserStats : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_REQUESTBLOB:
mess = mumble_pb2.RequestBlob()
mess.ParseFromString(message)
self.Log.debug("message: RequestBlob : %s", mess)
elif type == PYMUMBLE_MSG_TYPES_SERVERCONFIG:
mess = mumble_pb2.ServerConfig()
mess.ParseFromString(message)
self.Log.debug("message: ServerConfig : %s", mess)
def set_bandwidth(self, bandwidth):
"""set the total allowed outgoing bandwidth"""
if self.server_max_bandwidth is not None and bandwidth > self.server_max_bandwidth:
self.bandwidth = self.server_max_bandwidth
else:
self.bandwidth = bandwidth
self.sound_output.set_bandwidth(self.bandwidth) # communicate the update to the outgoing audio manager
def sound_received(self, message):
"""Manage a received sound message"""
# from tools import toHex # for debugging
pos = 0
# self.Log.debug("sound packet : " + toHex(message)) # for debugging
(header, ) = struct.unpack("!B", message[pos]) # extract the header
type = ( header & 0b11100000 ) >> 5
target = header & 0b00011111
pos += 1
if type == PYMUMBLE_AUDIO_TYPE_PING:
return
session = tools.VarInt() # decode session id
pos += session.decode(message[pos:pos+10])
sequence = tools.VarInt() # decode sequence number
pos += sequence.decode(message[pos:pos+10])
self.Log.debug("audio packet received from %i, sequence %i, type:%i, target:%i, lenght:%i", session.value, sequence.value, type, target, len(message))
terminator = False # set to true if it's the last 10 ms audio frame for the packet (used with CELT codec)
while ( pos < len(message)) and not terminator: # get the audio frames one by one
if type == PYMUMBLE_AUDIO_TYPE_OPUS:
size = tools.VarInt() # OPUS use varint for the frame length
pos += size.decode(message[pos:pos+10])
size = size.value
if not (size & 0x2000): # terminator is 0x2000 in the resulting int.
terminator = True # should actually always be 0 as OPUS can use variable length audio frames
size = size & 0x1fff # isolate the size from the terminator
else:
(header, ) = struct.unpack("!B", message[pos]) # CELT length and terminator is encoded in a 1 byte int
if not (header & 0b10000000):
terminator = True
size = header & 0b01111111
pos += 1
self.Log.debug("Audio frame : time:%f, last:%s, size:%i, type:%i, target:%i, pos:%i",time.time(), str(terminator), size, type, target, pos-1)
if size > 0 and self.receive_sound: # if audio must be treated
try:
newsound = self.users[session.value].sound.add(message[pos:pos+size],
sequence.value,
type,
target) # add the sound to the user's sound queue
self.callbacks(PYMUMBLE_CLBK_SOUNDRECEIVED, self.users[session.value], newsound)
self.Log.debug("Audio frame : time:%f last:%s, size:%i, uncompressed:%i, type:%i, target:%i",time.time(), str(terminator), size, newsound.size, type, target)
except CodecNotSupportedError as msg:
print msg
except KeyError: # sound received after user removed
pass
sequence.value += int(round(newsound.duration / 1000 * 10)) # add 1 sequence per 10ms of audio
# if len(message) - pos < size:
# raise InvalidFormatError("Invalid audio frame size")
pos += size # go further in the packet, after the audio frame
#TODO: get position info
def set_application_string(self, string):
"""Set the application name, that can be viewed by other clients on the server"""
self.application = string
def set_loop_rate(self, rate):
"""set the current main loop rate (pause per iteration)"""
self.loop_rate = rate
def get_loop_rate(self):
"""get the current main loop rate (pause per iteration)"""
return(self.loop_rate)
def set_receive_sound(self, value):
"""Enable or disable the management of incoming sounds"""
if value:
self.receive_sound = True
else:
self.receive_sound = False
def is_ready(self):
"""Wait for the connection to be fully completed. To be used in the main thread"""
self.ready_lock.acquire()
self.ready_lock.release()
def execute_command(self, cmd, blocking=True):
"""Create a command to be sent to the server. To be userd in the main thread"""
self.is_ready()
lock = self.commands.new_cmd(cmd)
if blocking and self.mumble_thread is not threading.current_thread():
lock.acquire()
lock.release()
return lock
#TODO: manage a timeout for blocking commands. Currently, no command actually waits for the server to execute
# The result of these commands should actually be checked against incoming server updates
def treat_command(self, cmd):
"""Send the awaiting commands to the server. Used in the pymumble thread."""
if cmd.cmd == PYMUMBLE_CMD_MOVE:
userstate = mumble_pb2.UserState()
userstate.session = cmd.parameters["session"]
userstate.channel_id = cmd.parameters["channel_id"]
self.Log.debug("Moving to channel")
self.send_message(PYMUMBLE_MSG_TYPES_USERSTATE, userstate)
cmd.response = True
self.commands.answer(cmd)
elif cmd.cmd == PYMUMBLE_CMD_MODUSERSTATE:
userstate = mumble_pb2.UserState()
userstate.session = cmd.parameters["session"]
if "mute" in cmd.parameters:
userstate.mute = cmd.parameters["mute"]
if "self_mute" in cmd.parameters:
userstate.self_mute = cmd.parameters["self_mute"]
if "deaf" in cmd.parameters:
userstate.deaf = cmd.parameters["deaf"]
if "self_deaf" in cmd.parameters:
userstate.self_deaf = cmd.parameters["self_deaf"]
if "suppress" in cmd.parameters:
userstate.suppress = cmd.parameters["suppress"]
if "recording" in cmd.parameters:
userstate.recording = cmd.parameters["recording"]
if "comment" in cmd.parameters:
userstate.comment = cmd.parameters["comment"]
if "texture" in cmd.parameters:
userstate.texture = cmd.parameters["texture"]
self.send_message(PYMUMBLE_MSG_TYPES_USERSTATE, userstate)
cmd.response = True
self.commands.answer(cmd)
