Merge branch 'sessionwork'

IPython/frontend/qt/base_frontend_mixin.py

@@ -1,109 +1,109 b''
 """ Defines a convenient mix-in class for implementing Qt frontends.
 """
 
 class BaseFrontendMixin(object):
     """ A mix-in class for implementing Qt frontends.
 
     To handle messages of a particular type, frontends need only define an
     appropriate handler method. For example, to handle 'stream' messaged, define
     a '_handle_stream(msg)' method.
     """
 
     #---------------------------------------------------------------------------
     # 'BaseFrontendMixin' concrete interface
     #---------------------------------------------------------------------------
     
     def _get_kernel_manager(self):
         """ Returns the current kernel manager.
         """
         return self._kernel_manager
 
     def _set_kernel_manager(self, kernel_manager):
         """ Disconnect from the current kernel manager (if any) and set a new
             kernel manager.
         """
         # Disconnect the old kernel manager, if necessary.
         old_manager = self._kernel_manager
         if old_manager is not None:
             old_manager.started_channels.disconnect(self._started_channels)
             old_manager.stopped_channels.disconnect(self._stopped_channels)
 
             # Disconnect the old kernel manager's channels.
             old_manager.sub_channel.message_received.disconnect(self._dispatch)
             old_manager.shell_channel.message_received.disconnect(self._dispatch)
             old_manager.stdin_channel.message_received.disconnect(self._dispatch)
             old_manager.hb_channel.kernel_died.disconnect(
                 self._handle_kernel_died)
     
             # Handle the case where the old kernel manager is still listening.
             if old_manager.channels_running:
                 self._stopped_channels()
 
         # Set the new kernel manager.
         self._kernel_manager = kernel_manager
         if kernel_manager is None:
             return
 
         # Connect the new kernel manager.
         kernel_manager.started_channels.connect(self._started_channels)
         kernel_manager.stopped_channels.connect(self._stopped_channels)
 
         # Connect the new kernel manager's channels.
         kernel_manager.sub_channel.message_received.connect(self._dispatch)
         kernel_manager.shell_channel.message_received.connect(self._dispatch)
         kernel_manager.stdin_channel.message_received.connect(self._dispatch)
         kernel_manager.hb_channel.kernel_died.connect(self._handle_kernel_died)
 
         # Handle the case where the kernel manager started channels before
         # we connected.
         if kernel_manager.channels_running:
             self._started_channels()
 
     kernel_manager = property(_get_kernel_manager, _set_kernel_manager)
 
     #---------------------------------------------------------------------------
     # 'BaseFrontendMixin' abstract interface
     #---------------------------------------------------------------------------
 
     def _handle_kernel_died(self, since_last_heartbeat):
         """ This is called when the ``kernel_died`` signal is emitted.
 
         This method is called when the kernel heartbeat has not been
         active for a certain amount of time. The typical action will be to
         give the user the option of restarting the kernel.
 
         Parameters
         ----------
         since_last_heartbeat : float
             The time since the heartbeat was last received.
         """
     
     def _started_channels(self):
         """ Called when the KernelManager channels have started listening or 
             when the frontend is assigned an already listening KernelManager.
         """
 
     def _stopped_channels(self):
         """ Called when the KernelManager channels have stopped listening or
             when a listening KernelManager is removed from the frontend.
         """
 
     #---------------------------------------------------------------------------
     # 'BaseFrontendMixin' protected interface
     #---------------------------------------------------------------------------
 
     def _dispatch(self, msg):
         """ Calls the frontend handler associated with the message type of the 
             given message.
         """
-        msg_type = msg['msg_type']
+        msg_type = msg['header']['msg_type']
         handler = getattr(self, '_handle_' + msg_type, None)
         if handler:
             handler(msg)
 
     def _is_from_this_session(self, msg):
         """ Returns whether a reply from the kernel originated from a request
             from this frontend.
         """
         session = self._kernel_manager.session.session
         return msg['parent_header']['session'] == session

IPython/frontend/qt/kernelmanager.py

@@ -1,243 +1,243 b''
 """ Defines a KernelManager that provides signals and slots.
 """
 
 # System library imports.
 from IPython.external.qt import QtCore
 
 # IPython imports.
 from IPython.utils.traitlets import Type
 from IPython.zmq.kernelmanager import KernelManager, SubSocketChannel, \
     ShellSocketChannel, StdInSocketChannel, HBSocketChannel
 from util import MetaQObjectHasTraits, SuperQObject
 
 
 class SocketChannelQObject(SuperQObject):
 
     # Emitted when the channel is started.
     started = QtCore.Signal()
 
     # Emitted when the channel is stopped.
     stopped = QtCore.Signal()
 
     #---------------------------------------------------------------------------
     # 'ZMQSocketChannel' interface
     #---------------------------------------------------------------------------
 
     def start(self):
         """ Reimplemented to emit signal.
         """
         super(SocketChannelQObject, self).start()
         self.started.emit()
 
     def stop(self):
         """ Reimplemented to emit signal.
         """
         super(SocketChannelQObject, self).stop()
         self.stopped.emit()
 
 
 class QtShellSocketChannel(SocketChannelQObject, ShellSocketChannel):
 
     # Emitted when any message is received.
     message_received = QtCore.Signal(object)
 
     # Emitted when a reply has been received for the corresponding request
     # type.
     execute_reply = QtCore.Signal(object)
     complete_reply = QtCore.Signal(object)
     object_info_reply = QtCore.Signal(object)
     history_reply = QtCore.Signal(object)
 
     # Emitted when the first reply comes back.
     first_reply = QtCore.Signal()
 
     # Used by the first_reply signal logic to determine if a reply is the 
     # first.
     _handlers_called = False
 
     #---------------------------------------------------------------------------
     # 'ShellSocketChannel' interface
     #---------------------------------------------------------------------------
     
     def call_handlers(self, msg):
         """ Reimplemented to emit signals instead of making callbacks.
         """
         # Emit the generic signal.
         self.message_received.emit(msg)
         
         # Emit signals for specialized message types.
-        msg_type = msg['msg_type']
+        msg_type = msg['header']['msg_type']
         signal = getattr(self, msg_type, None)
         if signal:
             signal.emit(msg)
 
         if not self._handlers_called:
             self.first_reply.emit()
             self._handlers_called = True
 
     #---------------------------------------------------------------------------
     # 'QtShellSocketChannel' interface
     #---------------------------------------------------------------------------
 
     def reset_first_reply(self):
         """ Reset the first_reply signal to fire again on the next reply.
         """
         self._handlers_called = False
 
 
 class QtSubSocketChannel(SocketChannelQObject, SubSocketChannel):
 
     # Emitted when any message is received.
     message_received = QtCore.Signal(object)
 
     # Emitted when a message of type 'stream' is received.
     stream_received = QtCore.Signal(object)
 
     # Emitted when a message of type 'pyin' is received.
     pyin_received = QtCore.Signal(object)
 
     # Emitted when a message of type 'pyout' is received.
     pyout_received = QtCore.Signal(object)
 
     # Emitted when a message of type 'pyerr' is received.
     pyerr_received = QtCore.Signal(object)
 
     # Emitted when a message of type 'display_data' is received
     display_data_received = QtCore.Signal(object)
 
     # Emitted when a crash report message is received from the kernel's
     # last-resort sys.excepthook.
     crash_received = QtCore.Signal(object)
 
     # Emitted when a shutdown is noticed.
     shutdown_reply_received = QtCore.Signal(object)
 
     #---------------------------------------------------------------------------
     # 'SubSocketChannel' interface
     #---------------------------------------------------------------------------
     
     def call_handlers(self, msg):
         """ Reimplemented to emit signals instead of making callbacks.
         """
         # Emit the generic signal.
         self.message_received.emit(msg)
         # Emit signals for specialized message types.
-        msg_type = msg['msg_type']
+        msg_type = msg['header']['msg_type']
         signal = getattr(self, msg_type + '_received', None)
         if signal:
             signal.emit(msg)
         elif msg_type in ('stdout', 'stderr'):
             self.stream_received.emit(msg)
 
     def flush(self):
         """ Reimplemented to ensure that signals are dispatched immediately.
         """
         super(QtSubSocketChannel, self).flush()
         QtCore.QCoreApplication.instance().processEvents()
 
 
 class QtStdInSocketChannel(SocketChannelQObject, StdInSocketChannel):
 
     # Emitted when any message is received.
     message_received = QtCore.Signal(object)
 
     # Emitted when an input request is received.
     input_requested = QtCore.Signal(object)
 
     #---------------------------------------------------------------------------
     # 'StdInSocketChannel' interface
     #---------------------------------------------------------------------------
 
     def call_handlers(self, msg):
         """ Reimplemented to emit signals instead of making callbacks.
         """
         # Emit the generic signal.
         self.message_received.emit(msg)
         
         # Emit signals for specialized message types.
-        msg_type = msg['msg_type']
+        msg_type = msg['header']['msg_type']
         if msg_type == 'input_request':
             self.input_requested.emit(msg)
 
 
 class QtHBSocketChannel(SocketChannelQObject, HBSocketChannel):
 
     # Emitted when the kernel has died.
     kernel_died = QtCore.Signal(object)
 
     #---------------------------------------------------------------------------
     # 'HBSocketChannel' interface
     #---------------------------------------------------------------------------
 
     def call_handlers(self, since_last_heartbeat):
         """ Reimplemented to emit signals instead of making callbacks.
         """
         # Emit the generic signal.
         self.kernel_died.emit(since_last_heartbeat)
 
 
 class QtKernelManager(KernelManager, SuperQObject):
     """ A KernelManager that provides signals and slots.
     """
 
     __metaclass__ = MetaQObjectHasTraits
 
     # Emitted when the kernel manager has started listening.
     started_channels = QtCore.Signal()
 
     # Emitted when the kernel manager has stopped listening.
     stopped_channels = QtCore.Signal()
 
     # Use Qt-specific channel classes that emit signals.
     sub_channel_class = Type(QtSubSocketChannel)
     shell_channel_class = Type(QtShellSocketChannel)
     stdin_channel_class = Type(QtStdInSocketChannel)
     hb_channel_class = Type(QtHBSocketChannel)
 
     #---------------------------------------------------------------------------
     # 'KernelManager' interface
     #---------------------------------------------------------------------------
 
     #------ Kernel process management ------------------------------------------
 
     def start_kernel(self, *args, **kw):
         """ Reimplemented for proper heartbeat management.
         """
         if self._shell_channel is not None:
             self._shell_channel.reset_first_reply()
         super(QtKernelManager, self).start_kernel(*args, **kw)
 
     #------ Channel management -------------------------------------------------
     
     def start_channels(self, *args, **kw):
         """ Reimplemented to emit signal.
         """
         super(QtKernelManager, self).start_channels(*args, **kw)
         self.started_channels.emit()
 
     def stop_channels(self):
         """ Reimplemented to emit signal.
         """ 
         super(QtKernelManager, self).stop_channels()
         self.stopped_channels.emit()
 
     @property
     def shell_channel(self):
         """ Reimplemented for proper heartbeat management.
         """
         if self._shell_channel is None:
             self._shell_channel = super(QtKernelManager, self).shell_channel
             self._shell_channel.first_reply.connect(self._first_reply)
         return self._shell_channel
 
     #---------------------------------------------------------------------------
     # Protected interface
     #---------------------------------------------------------------------------
     
     def _first_reply(self):
         """ Unpauses the heartbeat channel when the first reply is received on
             the execute channel. Note that this will *not* start the heartbeat
             channel if it is not already running!
         """
         if self._hb_channel is not None:
             self._hb_channel.unpause()

IPython/parallel/client/client.py

@@ -1,1435 +1,1435 b''
 """A semi-synchronous Client for the ZMQ cluster
 
 Authors:
 
 * MinRK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import os
 import json
 import sys
 import time
 import warnings
 from datetime import datetime
 from getpass import getpass
 from pprint import pprint
 
 pjoin = os.path.join
 
 import zmq
 # from zmq.eventloop import ioloop, zmqstream
 
 from IPython.config.configurable import MultipleInstanceError
 from IPython.core.application import BaseIPythonApplication
 
 from IPython.utils.jsonutil import rekey
 from IPython.utils.localinterfaces import LOCAL_IPS
 from IPython.utils.path import get_ipython_dir
 from IPython.utils.traitlets import (HasTraits, Int, Instance, Unicode,
                                     Dict, List, Bool, Set)
 from IPython.external.decorator import decorator
 from IPython.external.ssh import tunnel
 
 from IPython.parallel import error
 from IPython.parallel import util
 
 from IPython.zmq.session import Session, Message
 
 from .asyncresult import AsyncResult, AsyncHubResult
 from IPython.core.profiledir import ProfileDir, ProfileDirError
 from .view import DirectView, LoadBalancedView
 
 if sys.version_info[0] >= 3:
     # xrange is used in a couple 'isinstance' tests in py2
     # should be just 'range' in 3k
     xrange = range
 
 #--------------------------------------------------------------------------
 # Decorators for Client methods
 #--------------------------------------------------------------------------
 
 @decorator
 def spin_first(f, self, *args, **kwargs):
     """Call spin() to sync state prior to calling the method."""
     self.spin()
     return f(self, *args, **kwargs)
 
 
 #--------------------------------------------------------------------------
 # Classes
 #--------------------------------------------------------------------------
 
 class Metadata(dict):
     """Subclass of dict for initializing metadata values.
     
     Attribute access works on keys.
     
     These objects have a strict set of keys - errors will raise if you try
     to add new keys.
     """
     def __init__(self, *args, **kwargs):
         dict.__init__(self)
         md = {'msg_id' : None,
               'submitted' : None,
               'started' : None,
               'completed' : None,
               'received' : None,
               'engine_uuid' : None,
               'engine_id' : None,
               'follow' : None,
               'after' : None,
               'status' : None,
 
               'pyin' : None,
               'pyout' : None,
               'pyerr' : None,
               'stdout' : '',
               'stderr' : '',
             }
         self.update(md)
         self.update(dict(*args, **kwargs))
     
     def __getattr__(self, key):
         """getattr aliased to getitem"""
         if key in self.iterkeys():
             return self[key]
         else:
             raise AttributeError(key)
 
     def __setattr__(self, key, value):
         """setattr aliased to setitem, with strict"""
         if key in self.iterkeys():
             self[key] = value
         else:
             raise AttributeError(key)
     
     def __setitem__(self, key, value):
         """strict static key enforcement"""
         if key in self.iterkeys():
             dict.__setitem__(self, key, value)
         else:
             raise KeyError(key)
         
 
 class Client(HasTraits):
     """A semi-synchronous client to the IPython ZMQ cluster
     
     Parameters
     ----------
     
     url_or_file : bytes or unicode; zmq url or path to ipcontroller-client.json
         Connection information for the Hub's registration.  If a json connector
         file is given, then likely no further configuration is necessary.
         [Default: use profile]
     profile : bytes
         The name of the Cluster profile to be used to find connector information.
         If run from an IPython application, the default profile will be the same
         as the running application, otherwise it will be 'default'.
     context : zmq.Context
         Pass an existing zmq.Context instance, otherwise the client will create its own.
     debug : bool
         flag for lots of message printing for debug purposes
     timeout : int/float 
         time (in seconds) to wait for connection replies from the Hub
         [Default: 10]
  
     #-------------- session related args ----------------
     
     config : Config object
         If specified, this will be relayed to the Session for configuration
     username : str
         set username for the session object
     packer : str (import_string) or callable
         Can be either the simple keyword 'json' or 'pickle', or an import_string to a
         function to serialize messages. Must support same input as
         JSON, and output must be bytes.
         You can pass a callable directly as `pack`
     unpacker : str (import_string) or callable
         The inverse of packer.  Only necessary if packer is specified as *not* one
         of 'json' or 'pickle'.
     
     #-------------- ssh related args ----------------
     # These are args for configuring the ssh tunnel to be used
     # credentials are used to forward connections over ssh to the Controller
     # Note that the ip given in `addr` needs to be relative to sshserver
     # The most basic case is to leave addr as pointing to localhost (127.0.0.1),
     # and set sshserver as the same machine the Controller is on. However, 
     # the only requirement is that sshserver is able to see the Controller
     # (i.e. is within the same trusted network).
     
     sshserver : str
         A string of the form passed to ssh, i.e. 'server.tld' or '[e-mail]:port'
         If keyfile or password is specified, and this is not, it will default to
         the ip given in addr.
     sshkey : str; path to public ssh key file
         This specifies a key to be used in ssh login, default None.
         Regular default ssh keys will be used without specifying this argument.
     password : str 
         Your ssh password to sshserver. Note that if this is left None,
         you will be prompted for it if passwordless key based login is unavailable.
     paramiko : bool
         flag for whether to use paramiko instead of shell ssh for tunneling.
         [default: True on win32, False else]
     
     ------- exec authentication args -------
     If even localhost is untrusted, you can have some protection against
     unauthorized execution by signing messages with HMAC digests.  
     Messages are still sent as cleartext, so if someone can snoop your 
     loopback traffic this will not protect your privacy, but will prevent
     unauthorized execution.
     
     exec_key : str
         an authentication key or file containing a key
         default: None
     
     
     Attributes
     ----------
     
     ids : list of int engine IDs
         requesting the ids attribute always synchronizes
         the registration state. To request ids without synchronization,
         use semi-private _ids attributes.
     
     history : list of msg_ids
         a list of msg_ids, keeping track of all the execution
         messages you have submitted in order.
     
     outstanding : set of msg_ids
         a set of msg_ids that have been submitted, but whose
         results have not yet been received.
     
     results : dict
         a dict of all our results, keyed by msg_id
     
     block : bool
         determines default behavior when block not specified
         in execution methods
     
     Methods
     -------
     
     spin
         flushes incoming results and registration state changes
         control methods spin, and requesting `ids` also ensures up to date
     
     wait
         wait on one or more msg_ids
     
     execution methods
         apply
         legacy: execute, run
     
     data movement
         push, pull, scatter, gather
     
     query methods
         queue_status, get_result, purge, result_status
     
     control methods
         abort, shutdown
     
     """
     
     
     block = Bool(False)
     outstanding = Set()
     results = Instance('collections.defaultdict', (dict,))
     metadata = Instance('collections.defaultdict', (Metadata,))
     history = List()
     debug = Bool(False)
     
     profile=Unicode()
     def _profile_default(self):
         if BaseIPythonApplication.initialized():
             # an IPython app *might* be running, try to get its profile
             try:
                 return BaseIPythonApplication.instance().profile
             except (AttributeError, MultipleInstanceError):
                 # could be a *different* subclass of config.Application,
                 # which would raise one of these two errors.
                 return u'default'
         else:
             return u'default'
         
     
     _outstanding_dict = Instance('collections.defaultdict', (set,))
     _ids = List()
     _connected=Bool(False)
     _ssh=Bool(False)
     _context = Instance('zmq.Context')
     _config = Dict()
     _engines=Instance(util.ReverseDict, (), {})
     # _hub_socket=Instance('zmq.Socket')
     _query_socket=Instance('zmq.Socket')
     _control_socket=Instance('zmq.Socket')
     _iopub_socket=Instance('zmq.Socket')
     _notification_socket=Instance('zmq.Socket')
     _mux_socket=Instance('zmq.Socket')
     _task_socket=Instance('zmq.Socket')
     _task_scheme=Unicode()
     _closed = False
     _ignored_control_replies=Int(0)
     _ignored_hub_replies=Int(0)
     
     def __new__(self, *args, **kw):
         # don't raise on positional args
         return HasTraits.__new__(self, **kw)
     
     def __init__(self, url_or_file=None, profile=None, profile_dir=None, ipython_dir=None,
             context=None, debug=False, exec_key=None,
             sshserver=None, sshkey=None, password=None, paramiko=None,
             timeout=10, **extra_args
             ):
         if profile:
             super(Client, self).__init__(debug=debug, profile=profile)
         else:
             super(Client, self).__init__(debug=debug)
         if context is None:
             context = zmq.Context.instance()
         self._context = context
         
         self._setup_profile_dir(self.profile, profile_dir, ipython_dir)
         if self._cd is not None:
             if url_or_file is None:
                 url_or_file = pjoin(self._cd.security_dir, 'ipcontroller-client.json')
         assert url_or_file is not None, "I can't find enough information to connect to a hub!"\
             " Please specify at least one of url_or_file or profile."
         
         try:
             util.validate_url(url_or_file)
         except AssertionError:
             if not os.path.exists(url_or_file):
                 if self._cd:
                     url_or_file = os.path.join(self._cd.security_dir, url_or_file)
                 assert os.path.exists(url_or_file), "Not a valid connection file or url: %r"%url_or_file
             with open(url_or_file) as f:
                 cfg = json.loads(f.read())
         else:
             cfg = {'url':url_or_file}
         
         # sync defaults from args, json:
         if sshserver:
             cfg['ssh'] = sshserver
         if exec_key:
             cfg['exec_key'] = exec_key
         exec_key = cfg['exec_key']
         location = cfg.setdefault('location', None)
         cfg['url'] = util.disambiguate_url(cfg['url'], location)
         url = cfg['url']
         proto,addr,port = util.split_url(url)
         if location is not None and addr == '127.0.0.1':
             # location specified, and connection is expected to be local
             if location not in LOCAL_IPS and not sshserver:
                 # load ssh from JSON *only* if the controller is not on
                 # this machine
                 sshserver=cfg['ssh']
             if location not in LOCAL_IPS and not sshserver:
                 # warn if no ssh specified, but SSH is probably needed
                 # This is only a warning, because the most likely cause
                 # is a local Controller on a laptop whose IP is dynamic
                 warnings.warn("""
             Controller appears to be listening on localhost, but not on this machine.
             If this is true, you should specify Client(...,sshserver='you@%s')
             or instruct your controller to listen on an external IP."""%location,
                 RuntimeWarning)
         elif not sshserver:
             # otherwise sync with cfg
             sshserver = cfg['ssh']
         
         self._config = cfg
         
         self._ssh = bool(sshserver or sshkey or password)
         if self._ssh and sshserver is None:
             # default to ssh via localhost
             sshserver = url.split('://')[1].split(':')[0]
         if self._ssh and password is None:
             if tunnel.try_passwordless_ssh(sshserver, sshkey, paramiko):
                 password=False
             else:
                 password = getpass("SSH Password for %s: "%sshserver)
         ssh_kwargs = dict(keyfile=sshkey, password=password, paramiko=paramiko)
         
         # configure and construct the session
         if exec_key is not None:
             if os.path.isfile(exec_key):
                 extra_args['keyfile'] = exec_key
             else:
                 exec_key = util.asbytes(exec_key)
                 extra_args['key'] = exec_key
         self.session = Session(**extra_args)
         
         self._query_socket = self._context.socket(zmq.XREQ)
         self._query_socket.setsockopt(zmq.IDENTITY, util.asbytes(self.session.session))
         if self._ssh:
             tunnel.tunnel_connection(self._query_socket, url, sshserver, **ssh_kwargs)
         else:
             self._query_socket.connect(url)
         
         self.session.debug = self.debug
         
         self._notification_handlers = {'registration_notification' : self._register_engine,
                                     'unregistration_notification' : self._unregister_engine,
                                     'shutdown_notification' : lambda msg: self.close(),
                                     }
         self._queue_handlers = {'execute_reply' : self._handle_execute_reply,
                                 'apply_reply' : self._handle_apply_reply}
         self._connect(sshserver, ssh_kwargs, timeout)
         
     def __del__(self):
         """cleanup sockets, but _not_ context."""
         self.close()
     
     def _setup_profile_dir(self, profile, profile_dir, ipython_dir):
         if ipython_dir is None:
             ipython_dir = get_ipython_dir()
         if profile_dir is not None:
             try:
                 self._cd = ProfileDir.find_profile_dir(profile_dir)
                 return
             except ProfileDirError:
                 pass
         elif profile is not None:
             try:
                 self._cd = ProfileDir.find_profile_dir_by_name(
                     ipython_dir, profile)
                 return
             except ProfileDirError:
                 pass
         self._cd = None
     
     def _update_engines(self, engines):
         """Update our engines dict and _ids from a dict of the form: {id:uuid}."""
         for k,v in engines.iteritems():
             eid = int(k)
             self._engines[eid] = v
             self._ids.append(eid)
         self._ids = sorted(self._ids)
         if sorted(self._engines.keys()) != range(len(self._engines)) and \
                         self._task_scheme == 'pure' and self._task_socket:
             self._stop_scheduling_tasks()
     
     def _stop_scheduling_tasks(self):
         """Stop scheduling tasks because an engine has been unregistered
         from a pure ZMQ scheduler.
         """
         self._task_socket.close()
         self._task_socket = None
         msg = "An engine has been unregistered, and we are using pure " +\
               "ZMQ task scheduling.  Task farming will be disabled."
         if self.outstanding:
             msg += " If you were running tasks when this happened, " +\
                    "some `outstanding` msg_ids may never resolve."
         warnings.warn(msg, RuntimeWarning)
     
     def _build_targets(self, targets):
         """Turn valid target IDs or 'all' into two lists:
         (int_ids, uuids).
         """
         if not self._ids:
             # flush notification socket if no engines yet, just in case
             if not self.ids:
                 raise error.NoEnginesRegistered("Can't build targets without any engines")
         
         if targets is None:
             targets = self._ids
         elif isinstance(targets, basestring):
             if targets.lower() == 'all':
                 targets = self._ids
             else:
                 raise TypeError("%r not valid str target, must be 'all'"%(targets))
         elif isinstance(targets, int):
             if targets < 0:
                 targets = self.ids[targets]
             if targets not in self._ids:
                 raise IndexError("No such engine: %i"%targets)
             targets = [targets]
         
         if isinstance(targets, slice):
             indices = range(len(self._ids))[targets]
             ids = self.ids
             targets = [ ids[i] for i in indices ]
         
         if not isinstance(targets, (tuple, list, xrange)):
             raise TypeError("targets by int/slice/collection of ints only, not %s"%(type(targets)))
             
         return [util.asbytes(self._engines[t]) for t in targets], list(targets)
     
     def _connect(self, sshserver, ssh_kwargs, timeout):
         """setup all our socket connections to the cluster. This is called from
         __init__."""
         
         # Maybe allow reconnecting?
         if self._connected:
             return
         self._connected=True
 
         def connect_socket(s, url):
             url = util.disambiguate_url(url, self._config['location'])
             if self._ssh:
                 return tunnel.tunnel_connection(s, url, sshserver, **ssh_kwargs)
             else:
                 return s.connect(url)
             
         self.session.send(self._query_socket, 'connection_request')
         # use Poller because zmq.select has wrong units in pyzmq 2.1.7
         poller = zmq.Poller()
         poller.register(self._query_socket, zmq.POLLIN)
         # poll expects milliseconds, timeout is seconds
         evts = poller.poll(timeout*1000)
         if not evts:
             raise error.TimeoutError("Hub connection request timed out")
         idents,msg = self.session.recv(self._query_socket,mode=0)
         if self.debug:
             pprint(msg)
         msg = Message(msg)
         content = msg.content
         self._config['registration'] = dict(content)
         if content.status == 'ok':
             ident = util.asbytes(self.session.session)
             if content.mux:
                 self._mux_socket = self._context.socket(zmq.XREQ)
                 self._mux_socket.setsockopt(zmq.IDENTITY, ident)
                 connect_socket(self._mux_socket, content.mux)
             if content.task:
                 self._task_scheme, task_addr = content.task
                 self._task_socket = self._context.socket(zmq.XREQ)
                 self._task_socket.setsockopt(zmq.IDENTITY, ident)
                 connect_socket(self._task_socket, task_addr)
             if content.notification:
                 self._notification_socket = self._context.socket(zmq.SUB)
                 connect_socket(self._notification_socket, content.notification)
                 self._notification_socket.setsockopt(zmq.SUBSCRIBE, b'')
             # if content.query:
             #     self._query_socket = self._context.socket(zmq.XREQ)
             #     self._query_socket.setsockopt(zmq.IDENTITY, self.session.session)
             #     connect_socket(self._query_socket, content.query)
             if content.control:
                 self._control_socket = self._context.socket(zmq.XREQ)
                 self._control_socket.setsockopt(zmq.IDENTITY, ident)
                 connect_socket(self._control_socket, content.control)
             if content.iopub:
                 self._iopub_socket = self._context.socket(zmq.SUB)
                 self._iopub_socket.setsockopt(zmq.SUBSCRIBE, b'')
                 self._iopub_socket.setsockopt(zmq.IDENTITY, ident)
                 connect_socket(self._iopub_socket, content.iopub)
             self._update_engines(dict(content.engines))
         else:
             self._connected = False
             raise Exception("Failed to connect!")
     
     #--------------------------------------------------------------------------
     # handlers and callbacks for incoming messages
     #--------------------------------------------------------------------------
     
     def _unwrap_exception(self, content):
         """unwrap exception, and remap engine_id to int."""
         e = error.unwrap_exception(content)
         # print e.traceback
         if e.engine_info:
             e_uuid = e.engine_info['engine_uuid']
             eid = self._engines[e_uuid]
             e.engine_info['engine_id'] = eid
         return e
     
     def _extract_metadata(self, header, parent, content):
         md = {'msg_id' : parent['msg_id'],
               'received' : datetime.now(),
               'engine_uuid' : header.get('engine', None),
               'follow' : parent.get('follow', []),
               'after' : parent.get('after', []),
               'status' : content['status'],
             }
         
         if md['engine_uuid'] is not None:
             md['engine_id'] = self._engines.get(md['engine_uuid'], None)
         
         if 'date' in parent:
             md['submitted'] = parent['date']
         if 'started' in header:
             md['started'] = header['started']
         if 'date' in header:
             md['completed'] = header['date']
         return md
     
     def _register_engine(self, msg):
         """Register a new engine, and update our connection info."""
         content = msg['content']
         eid = content['id']
         d = {eid : content['queue']}
         self._update_engines(d)
 
     def _unregister_engine(self, msg):
         """Unregister an engine that has died."""
         content = msg['content']
         eid = int(content['id'])
         if eid in self._ids:
             self._ids.remove(eid)
             uuid = self._engines.pop(eid)
             
             self._handle_stranded_msgs(eid, uuid)
                 
         if self._task_socket and self._task_scheme == 'pure':
             self._stop_scheduling_tasks()
     
     def _handle_stranded_msgs(self, eid, uuid):
         """Handle messages known to be on an engine when the engine unregisters.
         
         It is possible that this will fire prematurely - that is, an engine will
         go down after completing a result, and the client will be notified
         of the unregistration and later receive the successful result.
         """
         
         outstanding = self._outstanding_dict[uuid]
         
         for msg_id in list(outstanding):
             if msg_id in self.results:
                 # we already 
                 continue
             try:
                 raise error.EngineError("Engine %r died while running task %r"%(eid, msg_id))
             except:
                 content = error.wrap_exception()
             # build a fake message:
             parent = {}
             header = {}
             parent['msg_id'] = msg_id
             header['engine'] = uuid
             header['date'] = datetime.now()
             msg = dict(parent_header=parent, header=header, content=content)
             self._handle_apply_reply(msg)
     
     def _handle_execute_reply(self, msg):
         """Save the reply to an execute_request into our results.
         
         execute messages are never actually used. apply is used instead.
         """
         
         parent = msg['parent_header']
         msg_id = parent['msg_id']
         if msg_id not in self.outstanding:
             if msg_id in self.history:
                 print ("got stale result: %s"%msg_id)
             else:
                 print ("got unknown result: %s"%msg_id)
         else:
             self.outstanding.remove(msg_id)
         self.results[msg_id] = self._unwrap_exception(msg['content'])
     
     def _handle_apply_reply(self, msg):
         """Save the reply to an apply_request into our results."""
         parent = msg['parent_header']
         msg_id = parent['msg_id']
         if msg_id not in self.outstanding:
             if msg_id in self.history:
                 print ("got stale result: %s"%msg_id)
                 print self.results[msg_id]
                 print msg
             else:
                 print ("got unknown result: %s"%msg_id)
         else:
             self.outstanding.remove(msg_id)
         content = msg['content']
         header = msg['header']
         
         # construct metadata:
         md = self.metadata[msg_id]
         md.update(self._extract_metadata(header, parent, content))
         # is this redundant?
         self.metadata[msg_id] = md
         
         e_outstanding = self._outstanding_dict[md['engine_uuid']]
         if msg_id in e_outstanding:
             e_outstanding.remove(msg_id)
         
         # construct result:
         if content['status'] == 'ok':
             self.results[msg_id] = util.unserialize_object(msg['buffers'])[0]
         elif content['status'] == 'aborted':
             self.results[msg_id] = error.TaskAborted(msg_id)
         elif content['status'] == 'resubmitted':
             # TODO: handle resubmission
             pass
         else:
             self.results[msg_id] = self._unwrap_exception(content)
     
     def _flush_notifications(self):
         """Flush notifications of engine registrations waiting
         in ZMQ queue."""
         idents,msg = self.session.recv(self._notification_socket, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
-            msg_type = msg['msg_type']
+            msg_type = msg['header']['msg_type']
             handler = self._notification_handlers.get(msg_type, None)
             if handler is None:
                 raise Exception("Unhandled message type: %s"%msg.msg_type)
             else:
                 handler(msg)
             idents,msg = self.session.recv(self._notification_socket, mode=zmq.NOBLOCK)
     
     def _flush_results(self, sock):
         """Flush task or queue results waiting in ZMQ queue."""
         idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
-            msg_type = msg['msg_type']
+            msg_type = msg['header']['msg_type']
             handler = self._queue_handlers.get(msg_type, None)
             if handler is None:
                 raise Exception("Unhandled message type: %s"%msg.msg_type)
             else:
                 handler(msg)
             idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
     
     def _flush_control(self, sock):
         """Flush replies from the control channel waiting
         in the ZMQ queue.
         
         Currently: ignore them."""
         if self._ignored_control_replies <= 0:
             return
         idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
         while msg is not None:
             self._ignored_control_replies -= 1
             if self.debug:
                 pprint(msg)
             idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
     
     def _flush_ignored_control(self):
         """flush ignored control replies"""
         while self._ignored_control_replies > 0:
             self.session.recv(self._control_socket)
             self._ignored_control_replies -= 1
     
     def _flush_ignored_hub_replies(self):
         ident,msg = self.session.recv(self._query_socket, mode=zmq.NOBLOCK)
         while msg is not None:
             ident,msg = self.session.recv(self._query_socket, mode=zmq.NOBLOCK)
     
     def _flush_iopub(self, sock):
         """Flush replies from the iopub channel waiting
         in the ZMQ queue.
         """
         idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
             parent = msg['parent_header']
             msg_id = parent['msg_id']
             content = msg['content']
             header = msg['header']
-            msg_type = msg['msg_type']
+            msg_type = msg['header']['msg_type']
             
             # init metadata:
             md = self.metadata[msg_id]
             
             if msg_type == 'stream':
                 name = content['name']
                 s = md[name] or ''
                 md[name] = s + content['data']
             elif msg_type == 'pyerr':
                 md.update({'pyerr' : self._unwrap_exception(content)})
             elif msg_type == 'pyin':
                 md.update({'pyin' : content['code']})
             else:
                 md.update({msg_type : content.get('data', '')})
             
             # reduntant?
             self.metadata[msg_id] = md
             
             idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
     
     #--------------------------------------------------------------------------
     # len, getitem
     #--------------------------------------------------------------------------
     
     def __len__(self):
         """len(client) returns # of engines."""
         return len(self.ids)
     
     def __getitem__(self, key):
         """index access returns DirectView multiplexer objects
         
         Must be int, slice, or list/tuple/xrange of ints"""
         if not isinstance(key, (int, slice, tuple, list, xrange)):
             raise TypeError("key by int/slice/iterable of ints only, not %s"%(type(key)))
         else:
             return self.direct_view(key)
     
     #--------------------------------------------------------------------------
     # Begin public methods
     #--------------------------------------------------------------------------
     
     @property
     def ids(self):
         """Always up-to-date ids property."""
         self._flush_notifications()
         # always copy:
         return list(self._ids)
         
     def close(self):
         if self._closed:
             return
         snames = filter(lambda n: n.endswith('socket'), dir(self))
         for socket in map(lambda name: getattr(self, name), snames):
             if isinstance(socket, zmq.Socket) and not socket.closed:
                 socket.close()
         self._closed = True
     
     def spin(self):
         """Flush any registration notifications and execution results
         waiting in the ZMQ queue.
         """
         if self._notification_socket:
             self._flush_notifications()
         if self._mux_socket:
             self._flush_results(self._mux_socket)
         if self._task_socket:
             self._flush_results(self._task_socket)
         if self._control_socket:
             self._flush_control(self._control_socket)
         if self._iopub_socket:
             self._flush_iopub(self._iopub_socket)
         if self._query_socket:
             self._flush_ignored_hub_replies()
     
     def wait(self, jobs=None, timeout=-1):
         """waits on one or more `jobs`, for up to `timeout` seconds.
         
         Parameters
         ----------
         
         jobs : int, str, or list of ints and/or strs, or one or more AsyncResult objects
                 ints are indices to self.history
                 strs are msg_ids
                 default: wait on all outstanding messages
         timeout : float
                 a time in seconds, after which to give up.
                 default is -1, which means no timeout
         
         Returns
         -------
         
         True : when all msg_ids are done
         False : timeout reached, some msg_ids still outstanding
         """
         tic = time.time()
         if jobs is None:
             theids = self.outstanding
         else:
             if isinstance(jobs, (int, basestring, AsyncResult)):
                 jobs = [jobs]
             theids = set()
             for job in jobs:
                 if isinstance(job, int):
                     # index access
                     job = self.history[job]
                 elif isinstance(job, AsyncResult):
                     map(theids.add, job.msg_ids)
                     continue
                 theids.add(job)
         if not theids.intersection(self.outstanding):
             return True
         self.spin()
         while theids.intersection(self.outstanding):
             if timeout >= 0 and ( time.time()-tic ) > timeout:
                 break
             time.sleep(1e-3)
             self.spin()
         return len(theids.intersection(self.outstanding)) == 0
     
     #--------------------------------------------------------------------------
     # Control methods
     #--------------------------------------------------------------------------
     
     @spin_first
     def clear(self, targets=None, block=None):
         """Clear the namespace in target(s)."""
         block = self.block if block is None else block
         targets = self._build_targets(targets)[0]
         for t in targets:
             self.session.send(self._control_socket, 'clear_request', content={}, ident=t)
         error = False
         if block:
             self._flush_ignored_control()
             for i in range(len(targets)):
                 idents,msg = self.session.recv(self._control_socket,0)
                 if self.debug:
                     pprint(msg)
                 if msg['content']['status'] != 'ok':
                     error = self._unwrap_exception(msg['content'])
         else:
             self._ignored_control_replies += len(targets)
         if error:
             raise error
         
     
     @spin_first
     def abort(self, jobs=None, targets=None, block=None):
         """Abort specific jobs from the execution queues of target(s).
         
         This is a mechanism to prevent jobs that have already been submitted
         from executing.
         
         Parameters
         ----------
         
         jobs : msg_id, list of msg_ids, or AsyncResult
             The jobs to be aborted
         
         
         """
         block = self.block if block is None else block
         targets = self._build_targets(targets)[0]
         msg_ids = []
         if isinstance(jobs, (basestring,AsyncResult)):
             jobs = [jobs]
         bad_ids = filter(lambda obj: not isinstance(obj, (basestring, AsyncResult)), jobs)
         if bad_ids:
             raise TypeError("Invalid msg_id type %r, expected str or AsyncResult"%bad_ids[0])
         for j in jobs:
             if isinstance(j, AsyncResult):
                 msg_ids.extend(j.msg_ids)
             else:
                 msg_ids.append(j)
         content = dict(msg_ids=msg_ids)
         for t in targets:
             self.session.send(self._control_socket, 'abort_request', 
                     content=content, ident=t)
         error = False
         if block:
             self._flush_ignored_control()
             for i in range(len(targets)):
                 idents,msg = self.session.recv(self._control_socket,0)
                 if self.debug:
                     pprint(msg)
                 if msg['content']['status'] != 'ok':
                     error = self._unwrap_exception(msg['content'])
         else:
             self._ignored_control_replies += len(targets)
         if error:
             raise error
     
     @spin_first
     def shutdown(self, targets=None, restart=False, hub=False, block=None):
         """Terminates one or more engine processes, optionally including the hub."""
         block = self.block if block is None else block
         if hub:
             targets = 'all'
         targets = self._build_targets(targets)[0]
         for t in targets:
             self.session.send(self._control_socket, 'shutdown_request', 
                         content={'restart':restart},ident=t)
         error = False
         if block or hub:
             self._flush_ignored_control()
             for i in range(len(targets)):
                 idents,msg = self.session.recv(self._control_socket, 0)
                 if self.debug:
                     pprint(msg)
                 if msg['content']['status'] != 'ok':
                     error = self._unwrap_exception(msg['content'])
         else:
             self._ignored_control_replies += len(targets)
         
         if hub:
             time.sleep(0.25)
             self.session.send(self._query_socket, 'shutdown_request')
             idents,msg = self.session.recv(self._query_socket, 0)
             if self.debug:
                 pprint(msg)
             if msg['content']['status'] != 'ok':
                 error = self._unwrap_exception(msg['content'])
         
         if error:
             raise error
     
     #--------------------------------------------------------------------------
     # Execution related methods
     #--------------------------------------------------------------------------
     
     def _maybe_raise(self, result):
         """wrapper for maybe raising an exception if apply failed."""
         if isinstance(result, error.RemoteError):
             raise result
         
         return result
     
     def send_apply_message(self, socket, f, args=None, kwargs=None, subheader=None, track=False,
                             ident=None):
         """construct and send an apply message via a socket.
         
         This is the principal method with which all engine execution is performed by views.
         """
                             
         assert not self._closed, "cannot use me anymore, I'm closed!"
         # defaults:
         args = args if args is not None else []
         kwargs = kwargs if kwargs is not None else {}
         subheader = subheader if subheader is not None else {}
         
         # validate arguments
         if not callable(f):
             raise TypeError("f must be callable, not %s"%type(f))
         if not isinstance(args, (tuple, list)):
             raise TypeError("args must be tuple or list, not %s"%type(args))
         if not isinstance(kwargs, dict):
             raise TypeError("kwargs must be dict, not %s"%type(kwargs))
         if not isinstance(subheader, dict):
             raise TypeError("subheader must be dict, not %s"%type(subheader))
         
         bufs = util.pack_apply_message(f,args,kwargs)
         
         msg = self.session.send(socket, "apply_request", buffers=bufs, ident=ident,
                             subheader=subheader, track=track)
         
-        msg_id = msg['msg_id']
+        msg_id = msg['header']['msg_id']
         self.outstanding.add(msg_id)
         if ident:
             # possibly routed to a specific engine
             if isinstance(ident, list):
                 ident = ident[-1]
             if ident in self._engines.values():
                 # save for later, in case of engine death
                 self._outstanding_dict[ident].add(msg_id)
         self.history.append(msg_id)
         self.metadata[msg_id]['submitted'] = datetime.now()
         
         return msg
 
     #--------------------------------------------------------------------------
     # construct a View object
     #--------------------------------------------------------------------------
     
     def load_balanced_view(self, targets=None):
         """construct a DirectView object.
         
         If no arguments are specified, create a LoadBalancedView
         using all engines.
         
         Parameters
         ----------
         
         targets: list,slice,int,etc. [default: use all engines]
             The subset of engines across which to load-balance
         """
         if targets == 'all':
             targets = None
         if targets is not None:
             targets = self._build_targets(targets)[1]
         return LoadBalancedView(client=self, socket=self._task_socket, targets=targets)
     
     def direct_view(self, targets='all'):
         """construct a DirectView object.
         
         If no targets are specified, create a DirectView
         using all engines.
         
         Parameters
         ----------
         
         targets: list,slice,int,etc. [default: use all engines]
             The engines to use for the View
         """
         single = isinstance(targets, int)
         # allow 'all' to be lazily evaluated at each execution
         if targets != 'all':
             targets = self._build_targets(targets)[1]
         if single:
             targets = targets[0]
         return DirectView(client=self, socket=self._mux_socket, targets=targets)
     
     #--------------------------------------------------------------------------
     # Query methods
     #--------------------------------------------------------------------------
     
     @spin_first
     def get_result(self, indices_or_msg_ids=None, block=None):
         """Retrieve a result by msg_id or history index, wrapped in an AsyncResult object.
         
         If the client already has the results, no request to the Hub will be made.
         
         This is a convenient way to construct AsyncResult objects, which are wrappers
         that include metadata about execution, and allow for awaiting results that
         were not submitted by this Client.
         
         It can also be a convenient way to retrieve the metadata associated with
         blocking execution, since it always retrieves
         
         Examples
         --------
         ::
         
             In [10]: r = client.apply()
         
         Parameters
         ----------
         
         indices_or_msg_ids : integer history index, str msg_id, or list of either
             The indices or msg_ids of indices to be retrieved
         
         block : bool
             Whether to wait for the result to be done
         
         Returns
         -------
         
         AsyncResult
             A single AsyncResult object will always be returned.
         
         AsyncHubResult
             A subclass of AsyncResult that retrieves results from the Hub
         
         """
         block = self.block if block is None else block
         if indices_or_msg_ids is None:
             indices_or_msg_ids = -1
         
         if not isinstance(indices_or_msg_ids, (list,tuple)):
             indices_or_msg_ids = [indices_or_msg_ids]
         
         theids = []
         for id in indices_or_msg_ids:
             if isinstance(id, int):
                 id = self.history[id]
             if not isinstance(id, basestring):
                 raise TypeError("indices must be str or int, not %r"%id)
             theids.append(id)
         
         local_ids = filter(lambda msg_id: msg_id in self.history or msg_id in self.results, theids)
         remote_ids = filter(lambda msg_id: msg_id not in local_ids, theids)
         
         if remote_ids:
             ar = AsyncHubResult(self, msg_ids=theids)
         else:
             ar = AsyncResult(self, msg_ids=theids)
         
         if block:
             ar.wait()
         
         return ar
 
     @spin_first
     def resubmit(self, indices_or_msg_ids=None, subheader=None, block=None):
         """Resubmit one or more tasks.
 
         in-flight tasks may not be resubmitted.
 
         Parameters
         ----------
 
         indices_or_msg_ids : integer history index, str msg_id, or list of either
             The indices or msg_ids of indices to be retrieved
 
         block : bool
             Whether to wait for the result to be done
 
         Returns
         -------
 
         AsyncHubResult
             A subclass of AsyncResult that retrieves results from the Hub
 
         """
         block = self.block if block is None else block
         if indices_or_msg_ids is None:
             indices_or_msg_ids = -1
 
         if not isinstance(indices_or_msg_ids, (list,tuple)):
             indices_or_msg_ids = [indices_or_msg_ids]
 
         theids = []
         for id in indices_or_msg_ids:
             if isinstance(id, int):
                 id = self.history[id]
             if not isinstance(id, basestring):
                 raise TypeError("indices must be str or int, not %r"%id)
             theids.append(id)
 
         for msg_id in theids:
             self.outstanding.discard(msg_id)
             if msg_id in self.history:
                 self.history.remove(msg_id)
             self.results.pop(msg_id, None)
             self.metadata.pop(msg_id, None)
         content = dict(msg_ids = theids)
 
         self.session.send(self._query_socket, 'resubmit_request', content)
 
         zmq.select([self._query_socket], [], [])
         idents,msg = self.session.recv(self._query_socket, zmq.NOBLOCK)
         if self.debug:
             pprint(msg)
         content = msg['content']
         if content['status'] != 'ok':
             raise self._unwrap_exception(content)
 
         ar = AsyncHubResult(self, msg_ids=theids)
 
         if block:
             ar.wait()
 
         return ar
     
     @spin_first
     def result_status(self, msg_ids, status_only=True):
         """Check on the status of the result(s) of the apply request with `msg_ids`.
         
         If status_only is False, then the actual results will be retrieved, else
         only the status of the results will be checked.
         
         Parameters
         ----------
         
         msg_ids : list of msg_ids
             if int:
                 Passed as index to self.history for convenience.
         status_only : bool (default: True)
             if False:
                 Retrieve the actual results of completed tasks.
         
         Returns
         -------
         
         results : dict
             There will always be the keys 'pending' and 'completed', which will
             be lists of msg_ids that are incomplete or complete. If `status_only`
             is False, then completed results will be keyed by their `msg_id`.
         """
         if not isinstance(msg_ids, (list,tuple)):
             msg_ids = [msg_ids]
             
         theids = []
         for msg_id in msg_ids:
             if isinstance(msg_id, int):
                 msg_id = self.history[msg_id]
             if not isinstance(msg_id, basestring):
                 raise TypeError("msg_ids must be str, not %r"%msg_id)
             theids.append(msg_id)
         
         completed = []
         local_results = {}
         
         # comment this block out to temporarily disable local shortcut:
         for msg_id in theids:
             if msg_id in self.results:
                 completed.append(msg_id)
                 local_results[msg_id] = self.results[msg_id]
                 theids.remove(msg_id)
         
         if theids: # some not locally cached
             content = dict(msg_ids=theids, status_only=status_only)
             msg = self.session.send(self._query_socket, "result_request", content=content)
             zmq.select([self._query_socket], [], [])
             idents,msg = self.session.recv(self._query_socket, zmq.NOBLOCK)
             if self.debug:
                 pprint(msg)
             content = msg['content']
             if content['status'] != 'ok':
                 raise self._unwrap_exception(content)
             buffers = msg['buffers']
         else:
             content = dict(completed=[],pending=[])
         
         content['completed'].extend(completed)
         
         if status_only:
             return content
         
         failures = []
         # load cached results into result:
         content.update(local_results)
         
         # update cache with results:
         for msg_id in sorted(theids):
             if msg_id in content['completed']:
                 rec = content[msg_id]
                 parent = rec['header']
                 header = rec['result_header']
                 rcontent = rec['result_content']
                 iodict = rec['io']
                 if isinstance(rcontent, str):
                     rcontent = self.session.unpack(rcontent)
                 
                 md = self.metadata[msg_id]
                 md.update(self._extract_metadata(header, parent, rcontent))
                 md.update(iodict)
                 
                 if rcontent['status'] == 'ok':
                     res,buffers = util.unserialize_object(buffers)
                 else:
                     print rcontent
                     res = self._unwrap_exception(rcontent)
                     failures.append(res)
                 
                 self.results[msg_id] = res
                 content[msg_id] = res
         
         if len(theids) == 1 and failures:
                 raise failures[0]
         
         error.collect_exceptions(failures, "result_status")
         return content
 
     @spin_first
     def queue_status(self, targets='all', verbose=False):
         """Fetch the status of engine queues.
         
         Parameters
         ----------
         
         targets : int/str/list of ints/strs
                 the engines whose states are to be queried.
                 default : all
         verbose : bool
                 Whether to return lengths only, or lists of ids for each element
         """
         engine_ids = self._build_targets(targets)[1]
         content = dict(targets=engine_ids, verbose=verbose)
         self.session.send(self._query_socket, "queue_request", content=content)
         idents,msg = self.session.recv(self._query_socket, 0)
         if self.debug:
             pprint(msg)
         content = msg['content']
         status = content.pop('status')
         if status != 'ok':
             raise self._unwrap_exception(content)
         content = rekey(content)
         if isinstance(targets, int):
             return content[targets]
         else:
             return content
         
     @spin_first
     def purge_results(self, jobs=[], targets=[]):
         """Tell the Hub to forget results.
         
         Individual results can be purged by msg_id, or the entire
         history of specific targets can be purged.
         
         Use `purge_results('all')` to scrub everything from the Hub's db.
         
         Parameters
         ----------
         
         jobs : str or list of str or AsyncResult objects
                 the msg_ids whose results should be forgotten.
         targets : int/str/list of ints/strs
                 The targets, by int_id, whose entire history is to be purged.
                 
                 default : None
         """
         if not targets and not jobs:
             raise ValueError("Must specify at least one of `targets` and `jobs`")
         if targets:
             targets = self._build_targets(targets)[1]
         
         # construct msg_ids from jobs
         if jobs == 'all':
             msg_ids = jobs
         else:
             msg_ids = []
             if isinstance(jobs, (basestring,AsyncResult)):
                 jobs = [jobs]
             bad_ids = filter(lambda obj: not isinstance(obj, (basestring, AsyncResult)), jobs)
             if bad_ids:
                 raise TypeError("Invalid msg_id type %r, expected str or AsyncResult"%bad_ids[0])
             for j in jobs:
                 if isinstance(j, AsyncResult):
                     msg_ids.extend(j.msg_ids)
                 else:
                     msg_ids.append(j)
 
         content = dict(engine_ids=targets, msg_ids=msg_ids)
         self.session.send(self._query_socket, "purge_request", content=content)
         idents, msg = self.session.recv(self._query_socket, 0)
         if self.debug:
             pprint(msg)
         content = msg['content']
         if content['status'] != 'ok':
             raise self._unwrap_exception(content)
 
     @spin_first
     def hub_history(self):
         """Get the Hub's history
         
         Just like the Client, the Hub has a history, which is a list of msg_ids.
         This will contain the history of all clients, and, depending on configuration,
         may contain history across multiple cluster sessions.
         
         Any msg_id returned here is a valid argument to `get_result`.
         
         Returns
         -------
         
         msg_ids : list of strs
                 list of all msg_ids, ordered by task submission time.
         """
         
         self.session.send(self._query_socket, "history_request", content={})
         idents, msg = self.session.recv(self._query_socket, 0)
         
         if self.debug:
             pprint(msg)
         content = msg['content']
         if content['status'] != 'ok':
             raise self._unwrap_exception(content)
         else:
             return content['history']
 
     @spin_first
     def db_query(self, query, keys=None):
         """Query the Hub's TaskRecord database
         
         This will return a list of task record dicts that match `query`
         
         Parameters
         ----------
         
         query : mongodb query dict
             The search dict. See mongodb query docs for details.
         keys : list of strs [optional]
             The subset of keys to be returned.  The default is to fetch everything but buffers.
             'msg_id' will *always* be included.
         """
         if isinstance(keys, basestring):
             keys = [keys]
         content = dict(query=query, keys=keys)
         self.session.send(self._query_socket, "db_request", content=content)
         idents, msg = self.session.recv(self._query_socket, 0)
         if self.debug:
             pprint(msg)
         content = msg['content']
         if content['status'] != 'ok':
             raise self._unwrap_exception(content)
         
         records = content['records']
         
         buffer_lens = content['buffer_lens']
         result_buffer_lens = content['result_buffer_lens']
         buffers = msg['buffers']
         has_bufs = buffer_lens is not None
         has_rbufs = result_buffer_lens is not None
         for i,rec in enumerate(records):
             # relink buffers
             if has_bufs:
                 blen = buffer_lens[i]
                 rec['buffers'], buffers = buffers[:blen],buffers[blen:]
             if has_rbufs:
                 blen = result_buffer_lens[i]
                 rec['result_buffers'], buffers = buffers[:blen],buffers[blen:]
             
         return records
 
 __all__ = [ 'Client' ]

IPython/parallel/client/view.py

@@ -1,1048 +1,1048 b''
 """Views of remote engines.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import imp
 import sys
 import warnings
 from contextlib import contextmanager
 from types import ModuleType
 
 import zmq
 
 from IPython.testing.skipdoctest import skip_doctest
 from IPython.utils.traitlets import HasTraits, Any, Bool, List, Dict, Set, Int, Instance, CFloat, CInt
 from IPython.external.decorator import decorator
 
 from IPython.parallel import util
 from IPython.parallel.controller.dependency import Dependency, dependent
 
 from . import map as Map
 from .asyncresult import AsyncResult, AsyncMapResult
 from .remotefunction import ParallelFunction, parallel, remote
 
 #-----------------------------------------------------------------------------
 # Decorators
 #-----------------------------------------------------------------------------
 
 @decorator
 def save_ids(f, self, *args, **kwargs):
     """Keep our history and outstanding attributes up to date after a method call."""
     n_previous = len(self.client.history)
     try:
         ret = f(self, *args, **kwargs)
     finally:
         nmsgs = len(self.client.history) - n_previous
         msg_ids = self.client.history[-nmsgs:]
         self.history.extend(msg_ids)
         map(self.outstanding.add, msg_ids)
     return ret
 
 @decorator
 def sync_results(f, self, *args, **kwargs):
     """sync relevant results from self.client to our results attribute."""
     ret = f(self, *args, **kwargs)
     delta = self.outstanding.difference(self.client.outstanding)
     completed = self.outstanding.intersection(delta)
     self.outstanding = self.outstanding.difference(completed)
     for msg_id in completed:
         self.results[msg_id] = self.client.results[msg_id]
     return ret
 
 @decorator
 def spin_after(f, self, *args, **kwargs):
     """call spin after the method."""
     ret = f(self, *args, **kwargs)
     self.spin()
     return ret
 
 #-----------------------------------------------------------------------------
 # Classes
 #-----------------------------------------------------------------------------
 
 @skip_doctest
 class View(HasTraits):
     """Base View class for more convenint apply(f,*args,**kwargs) syntax via attributes.
     
     Don't use this class, use subclasses.
     
     Methods
     -------
     
     spin
         flushes incoming results and registration state changes
         control methods spin, and requesting `ids` also ensures up to date
     
     wait
         wait on one or more msg_ids
     
     execution methods
         apply
         legacy: execute, run
     
     data movement
         push, pull, scatter, gather
     
     query methods
         get_result, queue_status, purge_results, result_status
     
     control methods
         abort, shutdown
     
     """
     # flags
     block=Bool(False)
     track=Bool(True)
     targets = Any()
     
     history=List()
     outstanding = Set()
     results = Dict()
     client = Instance('IPython.parallel.Client')
     
     _socket = Instance('zmq.Socket')
     _flag_names = List(['targets', 'block', 'track'])
     _targets = Any()
     _idents = Any()
     
     def __init__(self, client=None, socket=None, **flags):
         super(View, self).__init__(client=client, _socket=socket)
         self.block = client.block
         
         self.set_flags(**flags)
         
         assert not self.__class__ is View, "Don't use base View objects, use subclasses"
         
 
     def __repr__(self):
         strtargets = str(self.targets)
         if len(strtargets) > 16:
             strtargets = strtargets[:12]+'...]'
         return "<%s %s>"%(self.__class__.__name__, strtargets)
 
     def set_flags(self, **kwargs):
         """set my attribute flags by keyword.
         
         Views determine behavior with a few attributes (`block`, `track`, etc.).
         These attributes can be set all at once by name with this method.
         
         Parameters
         ----------
         
         block : bool
             whether to wait for results
         track : bool
             whether to create a MessageTracker to allow the user to 
             safely edit after arrays and buffers during non-copying
             sends.
         """
         for name, value in kwargs.iteritems():
             if name not in self._flag_names:
                 raise KeyError("Invalid name: %r"%name)
             else:
                 setattr(self, name, value)
     
     @contextmanager
     def temp_flags(self, **kwargs):
         """temporarily set flags, for use in `with` statements.
         
         See set_flags for permanent setting of flags
         
         Examples
         --------
         
         >>> view.track=False
         ...
         >>> with view.temp_flags(track=True):
         ...    ar = view.apply(dostuff, my_big_array)
         ...    ar.tracker.wait() # wait for send to finish
         >>> view.track
         False
         
         """
         # preflight: save flags, and set temporaries
         saved_flags = {}
         for f in self._flag_names:
             saved_flags[f] = getattr(self, f)
         self.set_flags(**kwargs)
         # yield to the with-statement block
         try:
             yield
         finally:
             # postflight: restore saved flags
             self.set_flags(**saved_flags)
         
         
     #----------------------------------------------------------------
     # apply
     #----------------------------------------------------------------
     
     @sync_results
     @save_ids
     def _really_apply(self, f, args, kwargs, block=None, **options):
         """wrapper for client.send_apply_message"""
         raise NotImplementedError("Implement in subclasses")
     
     def apply(self, f, *args, **kwargs):
         """calls f(*args, **kwargs) on remote engines, returning the result.
         
         This method sets all apply flags via this View's attributes.
         
         if self.block is False:
             returns AsyncResult
         else:
             returns actual result of f(*args, **kwargs)
         """
         return self._really_apply(f, args, kwargs)
 
     def apply_async(self, f, *args, **kwargs):
         """calls f(*args, **kwargs) on remote engines in a nonblocking manner.
         
         returns AsyncResult
         """
         return self._really_apply(f, args, kwargs, block=False)
 
     @spin_after
     def apply_sync(self, f, *args, **kwargs):
         """calls f(*args, **kwargs) on remote engines in a blocking manner,
          returning the result.
         
         returns: actual result of f(*args, **kwargs)
         """
         return self._really_apply(f, args, kwargs, block=True)
 
     #----------------------------------------------------------------
     # wrappers for client and control methods
     #----------------------------------------------------------------
     @sync_results
     def spin(self):
         """spin the client, and sync"""
         self.client.spin()
     
     @sync_results
     def wait(self, jobs=None, timeout=-1):
         """waits on one or more `jobs`, for up to `timeout` seconds.
         
         Parameters
         ----------
         
         jobs : int, str, or list of ints and/or strs, or one or more AsyncResult objects
                 ints are indices to self.history
                 strs are msg_ids
                 default: wait on all outstanding messages
         timeout : float
                 a time in seconds, after which to give up.
                 default is -1, which means no timeout
         
         Returns
         -------
         
         True : when all msg_ids are done
         False : timeout reached, some msg_ids still outstanding
         """
         if jobs is None:
             jobs = self.history
         return self.client.wait(jobs, timeout)
     
     def abort(self, jobs=None, targets=None, block=None):
         """Abort jobs on my engines.
         
         Parameters
         ----------
         
         jobs : None, str, list of strs, optional
             if None: abort all jobs.
             else: abort specific msg_id(s).
         """
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         return self.client.abort(jobs=jobs, targets=targets, block=block)
 
     def queue_status(self, targets=None, verbose=False):
         """Fetch the Queue status of my engines"""
         targets = targets if targets is not None else self.targets
         return self.client.queue_status(targets=targets, verbose=verbose)
     
     def purge_results(self, jobs=[], targets=[]):
         """Instruct the controller to forget specific results."""
         if targets is None or targets == 'all':
             targets = self.targets
         return self.client.purge_results(jobs=jobs, targets=targets)
     
     def shutdown(self, targets=None, restart=False, hub=False, block=None):
         """Terminates one or more engine processes, optionally including the hub.
         """
         block = self.block if block is None else block
         if targets is None or targets == 'all':
             targets = self.targets
         return self.client.shutdown(targets=targets, restart=restart, hub=hub, block=block)
     
     @spin_after
     def get_result(self, indices_or_msg_ids=None):
         """return one or more results, specified by history index or msg_id.
         
         See client.get_result for details.
         
         """
         
         if indices_or_msg_ids is None:
             indices_or_msg_ids = -1
         if isinstance(indices_or_msg_ids, int):
             indices_or_msg_ids = self.history[indices_or_msg_ids]
         elif isinstance(indices_or_msg_ids, (list,tuple,set)):
             indices_or_msg_ids = list(indices_or_msg_ids)
             for i,index in enumerate(indices_or_msg_ids):
                 if isinstance(index, int):
                     indices_or_msg_ids[i] = self.history[index]
         return self.client.get_result(indices_or_msg_ids)
     
     #-------------------------------------------------------------------
     # Map
     #-------------------------------------------------------------------
     
     def map(self, f, *sequences, **kwargs):
         """override in subclasses"""
         raise NotImplementedError
     
     def map_async(self, f, *sequences, **kwargs):
         """Parallel version of builtin `map`, using this view's engines.
         
         This is equivalent to map(...block=False)
         
         See `self.map` for details.
         """
         if 'block' in kwargs:
             raise TypeError("map_async doesn't take a `block` keyword argument.")
         kwargs['block'] = False
         return self.map(f,*sequences,**kwargs)
     
     def map_sync(self, f, *sequences, **kwargs):
         """Parallel version of builtin `map`, using this view's engines.
         
         This is equivalent to map(...block=True)
         
         See `self.map` for details.
         """
         if 'block' in kwargs:
             raise TypeError("map_sync doesn't take a `block` keyword argument.")
         kwargs['block'] = True
         return self.map(f,*sequences,**kwargs)
     
     def imap(self, f, *sequences, **kwargs):
         """Parallel version of `itertools.imap`.
         
         See `self.map` for details.
         
         """
         
         return iter(self.map_async(f,*sequences, **kwargs))
     
     #-------------------------------------------------------------------
     # Decorators
     #-------------------------------------------------------------------
     
     def remote(self, block=True, **flags):
         """Decorator for making a RemoteFunction"""
         block = self.block if block is None else block
         return remote(self, block=block, **flags)
     
     def parallel(self, dist='b', block=None, **flags):
         """Decorator for making a ParallelFunction"""
         block = self.block if block is None else block
         return parallel(self, dist=dist, block=block, **flags)
 
 @skip_doctest
 class DirectView(View):
     """Direct Multiplexer View of one or more engines.
     
     These are created via indexed access to a client:
     
     >>> dv_1 = client[1]
     >>> dv_all = client[:]
     >>> dv_even = client[::2]
     >>> dv_some = client[1:3]
     
     This object provides dictionary access to engine namespaces:
     
     # push a=5:
     >>> dv['a'] = 5 
     # pull 'foo':
     >>> db['foo']
     
     """
     
     def __init__(self, client=None, socket=None, targets=None):
         super(DirectView, self).__init__(client=client, socket=socket, targets=targets)
     
     @property
     def importer(self):
         """sync_imports(local=True) as a property.
         
         See sync_imports for details.
         
         """
         return self.sync_imports(True)
     
     @contextmanager
     def sync_imports(self, local=True):
         """Context Manager for performing simultaneous local and remote imports.
         
         'import x as y' will *not* work.  The 'as y' part will simply be ignored.
         
         >>> with view.sync_imports():
         ...    from numpy import recarray
         importing recarray from numpy on engine(s)
         
         """
         import __builtin__
         local_import = __builtin__.__import__
         modules = set()
         results = []
         @util.interactive
         def remote_import(name, fromlist, level):
             """the function to be passed to apply, that actually performs the import
             on the engine, and loads up the user namespace.
             """
             import sys
             user_ns = globals()
             mod = __import__(name, fromlist=fromlist, level=level)
             if fromlist:
                 for key in fromlist:
                     user_ns[key] = getattr(mod, key)
             else:
                 user_ns[name] = sys.modules[name]
         
         def view_import(name, globals={}, locals={}, fromlist=[], level=-1):
             """the drop-in replacement for __import__, that optionally imports
             locally as well.
             """
             # don't override nested imports
             save_import = __builtin__.__import__
             __builtin__.__import__ = local_import
             
             if imp.lock_held():
                 # this is a side-effect import, don't do it remotely, or even
                 # ignore the local effects
                 return local_import(name, globals, locals, fromlist, level)
             
             imp.acquire_lock()
             if local:
                 mod = local_import(name, globals, locals, fromlist, level)
             else:
                 raise NotImplementedError("remote-only imports not yet implemented")
             imp.release_lock()
             
             key = name+':'+','.join(fromlist or [])
             if level == -1 and key not in modules:
                 modules.add(key)
                 if fromlist:
                     print "importing %s from %s on engine(s)"%(','.join(fromlist), name)
                 else:
                     print "importing %s on engine(s)"%name
                 results.append(self.apply_async(remote_import, name, fromlist, level))
             # restore override
             __builtin__.__import__ = save_import
                 
             return mod
         
         # override __import__
         __builtin__.__import__ = view_import
         try:
             # enter the block
             yield
         except ImportError:
             if not local:
                 # ignore import errors if not doing local imports
                 pass
         finally:
             # always restore __import__
             __builtin__.__import__ = local_import
         
         for r in results:
             # raise possible remote ImportErrors here
             r.get()
             
     
     @sync_results
     @save_ids
     def _really_apply(self, f, args=None, kwargs=None, targets=None, block=None, track=None):
         """calls f(*args, **kwargs) on remote engines, returning the result.
         
         This method sets all of `apply`'s flags via this View's attributes.
         
         Parameters
         ----------
         
         f : callable
         
         args : list [default: empty]
         
         kwargs : dict [default: empty]
         
         targets : target list [default: self.targets]
             where to run
         block : bool [default: self.block]
             whether to block 
         track : bool [default: self.track]
             whether to ask zmq to track the message, for safe non-copying sends
         
         Returns
         -------
         
         if self.block is False:
             returns AsyncResult
         else:
             returns actual result of f(*args, **kwargs) on the engine(s)
             This will be a list of self.targets is also a list (even length 1), or
             the single result if self.targets is an integer engine id
         """
         args = [] if args is None else args
         kwargs = {} if kwargs is None else kwargs
         block = self.block if block is None else block
         track = self.track if track is None else track
         targets = self.targets if targets is None else targets
         
         _idents = self.client._build_targets(targets)[0]
         msg_ids = []
         trackers = []
         for ident in _idents:
             msg = self.client.send_apply_message(self._socket, f, args, kwargs, track=track,
                                     ident=ident)
             if track:
                 trackers.append(msg['tracker'])
-            msg_ids.append(msg['msg_id'])
+            msg_ids.append(msg['header']['msg_id'])
         tracker = None if track is False else zmq.MessageTracker(*trackers)
         ar = AsyncResult(self.client, msg_ids, fname=f.__name__, targets=targets, tracker=tracker)
         if block:
             try:
                 return ar.get()
             except KeyboardInterrupt:
                 pass
         return ar
     
     @spin_after
     def map(self, f, *sequences, **kwargs):
         """view.map(f, *sequences, block=self.block) => list|AsyncMapResult
         
         Parallel version of builtin `map`, using this View's `targets`.
         
         There will be one task per target, so work will be chunked
         if the sequences are longer than `targets`.  
         
         Results can be iterated as they are ready, but will become available in chunks.
         
         Parameters
         ----------
         
         f : callable
             function to be mapped
         *sequences: one or more sequences of matching length
             the sequences to be distributed and passed to `f`
         block : bool
             whether to wait for the result or not [default self.block]
         
         Returns
         -------
         
         if block=False:
             AsyncMapResult
                 An object like AsyncResult, but which reassembles the sequence of results
                 into a single list. AsyncMapResults can be iterated through before all
                 results are complete.
         else:
             list
                 the result of map(f,*sequences)
         """
         
         block = kwargs.pop('block', self.block)
         for k in kwargs.keys():
             if k not in ['block', 'track']:
                 raise TypeError("invalid keyword arg, %r"%k)
         
         assert len(sequences) > 0, "must have some sequences to map onto!"
         pf = ParallelFunction(self, f, block=block, **kwargs)
         return pf.map(*sequences)
     
     def execute(self, code, targets=None, block=None):
         """Executes `code` on `targets` in blocking or nonblocking manner.
         
         ``execute`` is always `bound` (affects engine namespace)
         
         Parameters
         ----------
         
         code : str
                 the code string to be executed
         block : bool
                 whether or not to wait until done to return
                 default: self.block
         """
         return self._really_apply(util._execute, args=(code,), block=block, targets=targets)
     
     def run(self, filename, targets=None, block=None):
         """Execute contents of `filename` on my engine(s). 
         
         This simply reads the contents of the file and calls `execute`.
         
         Parameters
         ----------
         
         filename : str
                 The path to the file
         targets : int/str/list of ints/strs
                 the engines on which to execute
                 default : all
         block : bool
                 whether or not to wait until done
                 default: self.block
         
         """
         with open(filename, 'r') as f:
             # add newline in case of trailing indented whitespace
             # which will cause SyntaxError
             code = f.read()+'\n'
         return self.execute(code, block=block, targets=targets)
     
     def update(self, ns):
         """update remote namespace with dict `ns`
         
         See `push` for details.
         """
         return self.push(ns, block=self.block, track=self.track)
     
     def push(self, ns, targets=None, block=None, track=None):
         """update remote namespace with dict `ns`
         
         Parameters
         ----------
         
         ns : dict
             dict of keys with which to update engine namespace(s)
         block : bool [default : self.block]
             whether to wait to be notified of engine receipt
         
         """
         
         block = block if block is not None else self.block
         track = track if track is not None else self.track
         targets = targets if targets is not None else self.targets
         # applier = self.apply_sync if block else self.apply_async
         if not isinstance(ns, dict):
             raise TypeError("Must be a dict, not %s"%type(ns))
         return self._really_apply(util._push, (ns,), block=block, track=track, targets=targets)
 
     def get(self, key_s):
         """get object(s) by `key_s` from remote namespace
         
         see `pull` for details.
         """
         # block = block if block is not None else self.block
         return self.pull(key_s, block=True)
     
     def pull(self, names, targets=None, block=None):
         """get object(s) by `name` from remote namespace
         
         will return one object if it is a key.
         can also take a list of keys, in which case it will return a list of objects.
         """
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         applier = self.apply_sync if block else self.apply_async
         if isinstance(names, basestring):
             pass
         elif isinstance(names, (list,tuple,set)):
             for key in names:
                 if not isinstance(key, basestring):
                     raise TypeError("keys must be str, not type %r"%type(key))
         else:
             raise TypeError("names must be strs, not %r"%names)
         return self._really_apply(util._pull, (names,), block=block, targets=targets)
     
     def scatter(self, key, seq, dist='b', flatten=False, targets=None, block=None, track=None):
         """
         Partition a Python sequence and send the partitions to a set of engines.
         """
         block = block if block is not None else self.block
         track = track if track is not None else self.track
         targets = targets if targets is not None else self.targets
         
         mapObject = Map.dists[dist]()
         nparts = len(targets)
         msg_ids = []
         trackers = []
         for index, engineid in enumerate(targets):
             partition = mapObject.getPartition(seq, index, nparts)
             if flatten and len(partition) == 1:
                 ns = {key: partition[0]}
             else:
                 ns = {key: partition}
             r = self.push(ns, block=False, track=track, targets=engineid)
             msg_ids.extend(r.msg_ids)
             if track:
                 trackers.append(r._tracker)
         
         if track:
             tracker = zmq.MessageTracker(*trackers)
         else:
             tracker = None
         
         r = AsyncResult(self.client, msg_ids, fname='scatter', targets=targets, tracker=tracker)
         if block:
             r.wait()
         else:
             return r
     
     @sync_results
     @save_ids
     def gather(self, key, dist='b', targets=None, block=None):
         """
         Gather a partitioned sequence on a set of engines as a single local seq.
         """
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         mapObject = Map.dists[dist]()
         msg_ids = []
         
         for index, engineid in enumerate(targets):
             msg_ids.extend(self.pull(key, block=False, targets=engineid).msg_ids)
         
         r = AsyncMapResult(self.client, msg_ids, mapObject, fname='gather')
         
         if block:
             try:
                 return r.get()
             except KeyboardInterrupt:
                 pass
         return r
     
     def __getitem__(self, key):
         return self.get(key)
     
     def __setitem__(self,key, value):
         self.update({key:value})
     
     def clear(self, targets=None, block=False):
         """Clear the remote namespaces on my engines."""
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         return self.client.clear(targets=targets, block=block)
     
     def kill(self, targets=None, block=True):
         """Kill my engines."""
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         return self.client.kill(targets=targets, block=block)
     
     #----------------------------------------
     # activate for %px,%autopx magics
     #----------------------------------------
     def activate(self):
         """Make this `View` active for parallel magic commands.
         
         IPython has a magic command syntax to work with `MultiEngineClient` objects.
         In a given IPython session there is a single active one.  While
         there can be many `Views` created and used by the user, 
         there is only one active one.  The active `View` is used whenever 
         the magic commands %px and %autopx are used.
         
         The activate() method is called on a given `View` to make it 
         active.  Once this has been done, the magic commands can be used.
         """
         
         try:
             # This is injected into __builtins__.
             ip = get_ipython()
         except NameError:
             print "The IPython parallel magics (%result, %px, %autopx) only work within IPython."
         else:
             pmagic = ip.plugin_manager.get_plugin('parallelmagic')
             if pmagic is None:
                 ip.magic_load_ext('parallelmagic')
                 pmagic = ip.plugin_manager.get_plugin('parallelmagic')
 
             pmagic.active_view = self
 
 
 @skip_doctest
 class LoadBalancedView(View):
     """An load-balancing View that only executes via the Task scheduler.
     
     Load-balanced views can be created with the client's `view` method:
     
     >>> v = client.load_balanced_view()
     
     or targets can be specified, to restrict the potential destinations:
     
     >>> v = client.client.load_balanced_view(([1,3])
     
     which would restrict loadbalancing to between engines 1 and 3.
     
     """
     
     follow=Any()
     after=Any()
     timeout=CFloat()
     retries = CInt(0)
     
     _task_scheme = Any()
     _flag_names = List(['targets', 'block', 'track', 'follow', 'after', 'timeout', 'retries'])
     
     def __init__(self, client=None, socket=None, **flags):
         super(LoadBalancedView, self).__init__(client=client, socket=socket, **flags)
         self._task_scheme=client._task_scheme
     
     def _validate_dependency(self, dep):
         """validate a dependency.
         
         For use in `set_flags`.
         """
         if dep is None or isinstance(dep, (basestring, AsyncResult, Dependency)):
             return True
         elif isinstance(dep, (list,set, tuple)):
             for d in dep:
                 if not isinstance(d, (basestring, AsyncResult)):
                     return False
         elif isinstance(dep, dict):
             if set(dep.keys()) != set(Dependency().as_dict().keys()):
                 return False
             if not isinstance(dep['msg_ids'], list):
                 return False
             for d in dep['msg_ids']:
                 if not isinstance(d, basestring):
                     return False
         else:
             return False
         
         return True
                 
     def _render_dependency(self, dep):
         """helper for building jsonable dependencies from various input forms."""
         if isinstance(dep, Dependency):
             return dep.as_dict()
         elif isinstance(dep, AsyncResult):
             return dep.msg_ids
         elif dep is None:
             return []
         else:
             # pass to Dependency constructor
             return list(Dependency(dep))
 
     def set_flags(self, **kwargs):
         """set my attribute flags by keyword.
         
         A View is a wrapper for the Client's apply method, but with attributes
         that specify keyword arguments, those attributes can be set by keyword
         argument with this method.
         
         Parameters
         ----------
         
         block : bool
             whether to wait for results
         track : bool
             whether to create a MessageTracker to allow the user to 
             safely edit after arrays and buffers during non-copying
             sends.
 
         after : Dependency or collection of msg_ids
             Only for load-balanced execution (targets=None)
             Specify a list of msg_ids as a time-based dependency.
             This job will only be run *after* the dependencies
             have been met.
 
         follow : Dependency or collection of msg_ids
             Only for load-balanced execution (targets=None)
             Specify a list of msg_ids as a location-based dependency.
             This job will only be run on an engine where this dependency
             is met.
 
         timeout : float/int or None
             Only for load-balanced execution (targets=None)
             Specify an amount of time (in seconds) for the scheduler to
             wait for dependencies to be met before failing with a
             DependencyTimeout.
 
         retries : int
             Number of times a task will be retried on failure.
         """
         
         super(LoadBalancedView, self).set_flags(**kwargs)
         for name in ('follow', 'after'):
             if name in kwargs:
                 value = kwargs[name]
                 if self._validate_dependency(value):
                     setattr(self, name, value)
                 else:
                     raise ValueError("Invalid dependency: %r"%value)
         if 'timeout' in kwargs:
             t = kwargs['timeout']
             if not isinstance(t, (int, long, float, type(None))):
                 raise TypeError("Invalid type for timeout: %r"%type(t))
             if t is not None:
                 if t < 0:
                     raise ValueError("Invalid timeout: %s"%t)
             self.timeout = t
     
     @sync_results
     @save_ids
     def _really_apply(self, f, args=None, kwargs=None, block=None, track=None,
                                         after=None, follow=None, timeout=None,
                                         targets=None, retries=None):
         """calls f(*args, **kwargs) on a remote engine, returning the result.
         
         This method temporarily sets all of `apply`'s flags for a single call.
         
         Parameters
         ----------
         
         f : callable
         
         args : list [default: empty]
         
         kwargs : dict [default: empty]
         
         block : bool [default: self.block]
             whether to block 
         track : bool [default: self.track]
             whether to ask zmq to track the message, for safe non-copying sends
             
         !!!!!! TODO: THE REST HERE  !!!!
         
         Returns
         -------
         
         if self.block is False:
             returns AsyncResult
         else:
             returns actual result of f(*args, **kwargs) on the engine(s)
             This will be a list of self.targets is also a list (even length 1), or
             the single result if self.targets is an integer engine id
         """
 
         # validate whether we can run
         if self._socket.closed:
             msg = "Task farming is disabled"
             if self._task_scheme == 'pure':
                 msg += " because the pure ZMQ scheduler cannot handle"
                 msg += " disappearing engines."
             raise RuntimeError(msg)
         
         if self._task_scheme == 'pure':
             # pure zmq scheme doesn't support extra features
             msg = "Pure ZMQ scheduler doesn't support the following flags:"
             "follow, after, retries, targets, timeout"
             if (follow or after or retries or targets or timeout):
                 # hard fail on Scheduler flags
                 raise RuntimeError(msg)
             if isinstance(f, dependent):
                 # soft warn on functional dependencies
                 warnings.warn(msg, RuntimeWarning)
         
         # build args
         args = [] if args is None else args
         kwargs = {} if kwargs is None else kwargs
         block = self.block if block is None else block
         track = self.track if track is None else track
         after = self.after if after is None else after
         retries = self.retries if retries is None else retries
         follow = self.follow if follow is None else follow
         timeout = self.timeout if timeout is None else timeout
         targets = self.targets if targets is None else targets
         
         if not isinstance(retries, int):
             raise TypeError('retries must be int, not %r'%type(retries))
 
         if targets is None:
             idents = []
         else:
             idents = self.client._build_targets(targets)[0]
             # ensure *not* bytes
             idents = [ ident.decode() for ident in idents ]
         
         after = self._render_dependency(after)
         follow = self._render_dependency(follow)
         subheader = dict(after=after, follow=follow, timeout=timeout, targets=idents, retries=retries)
         
         msg = self.client.send_apply_message(self._socket, f, args, kwargs, track=track,
                                 subheader=subheader)
         tracker = None if track is False else msg['tracker']
         
-        ar = AsyncResult(self.client, msg['msg_id'], fname=f.__name__, targets=None, tracker=tracker)
+        ar = AsyncResult(self.client, msg['header']['msg_id'], fname=f.__name__, targets=None, tracker=tracker)
         
         if block:
             try:
                 return ar.get()
             except KeyboardInterrupt:
                 pass
         return ar
         
     @spin_after
     @save_ids
     def map(self, f, *sequences, **kwargs):
         """view.map(f, *sequences, block=self.block, chunksize=1) => list|AsyncMapResult
         
         Parallel version of builtin `map`, load-balanced by this View.
         
         `block`, and `chunksize` can be specified by keyword only.
         
         Each `chunksize` elements will be a separate task, and will be
         load-balanced. This lets individual elements be available for iteration
         as soon as they arrive.
         
         Parameters
         ----------
         
         f : callable
             function to be mapped
         *sequences: one or more sequences of matching length
             the sequences to be distributed and passed to `f`
         block : bool
             whether to wait for the result or not [default self.block]
         track : bool
             whether to create a MessageTracker to allow the user to 
             safely edit after arrays and buffers during non-copying
             sends.
         chunksize : int
             how many elements should be in each task [default 1]
         
         Returns
         -------
         
         if block=False:
             AsyncMapResult
                 An object like AsyncResult, but which reassembles the sequence of results
                 into a single list. AsyncMapResults can be iterated through before all
                 results are complete.
             else:
                 the result of map(f,*sequences)
         
         """
         
         # default
         block = kwargs.get('block', self.block)
         chunksize = kwargs.get('chunksize', 1)
         
         keyset = set(kwargs.keys())
         extra_keys = keyset.difference_update(set(['block', 'chunksize']))
         if extra_keys:
             raise TypeError("Invalid kwargs: %s"%list(extra_keys))
         
         assert len(sequences) > 0, "must have some sequences to map onto!"
         
         pf = ParallelFunction(self, f, block=block,  chunksize=chunksize)
         return pf.map(*sequences)
 
 __all__ = ['LoadBalancedView', 'DirectView']

IPython/parallel/controller/hub.py

@@ -1,1291 +1,1291 b''
 #!/usr/bin/env python
 """The IPython Controller Hub with 0MQ
 This is the master object that handles connections from engines and clients,
 and monitors traffic through the various queues.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 from __future__ import print_function
 
 import sys
 import time
 from datetime import datetime
 
 import zmq
 from zmq.eventloop import ioloop
 from zmq.eventloop.zmqstream import ZMQStream
 
 # internal:
 from IPython.utils.importstring import import_item
 from IPython.utils.traitlets import (
         HasTraits, Instance, Int, Unicode, Dict, Set, Tuple, CBytes, DottedObjectName
         )
 
 from IPython.parallel import error, util
 from IPython.parallel.factory import RegistrationFactory
 
 from IPython.zmq.session import SessionFactory
 
 from .heartmonitor import HeartMonitor
 
 #-----------------------------------------------------------------------------
 # Code
 #-----------------------------------------------------------------------------
 
 def _passer(*args, **kwargs):
     return
 
 def _printer(*args, **kwargs):
     print (args)
     print (kwargs)
 
 def empty_record():
     """Return an empty dict with all record keys."""
     return {
         'msg_id' : None,
         'header' : None,
         'content': None,
         'buffers': None,
         'submitted': None,
         'client_uuid' : None,
         'engine_uuid' : None,
         'started': None,
         'completed': None,
         'resubmitted': None,
         'result_header' : None,
         'result_content' : None,
         'result_buffers' : None,
         'queue' : None,
         'pyin' : None,
         'pyout': None,
         'pyerr': None,
         'stdout': '',
         'stderr': '',
     }
     
 def init_record(msg):
     """Initialize a TaskRecord based on a request."""
     header = msg['header']
     return {
         'msg_id' : header['msg_id'],
         'header' : header,
         'content': msg['content'],
         'buffers': msg['buffers'],
         'submitted': header['date'],
         'client_uuid' : None,
         'engine_uuid' : None,
         'started': None,
         'completed': None,
         'resubmitted': None,
         'result_header' : None,
         'result_content' : None,
         'result_buffers' : None,
         'queue' : None,
         'pyin' : None,
         'pyout': None,
         'pyerr': None,
         'stdout': '',
         'stderr': '',
     }
 
 
 class EngineConnector(HasTraits):
     """A simple object for accessing the various zmq connections of an object.
     Attributes are:
     id (int): engine ID
     uuid (str): uuid (unused?)
     queue (str): identity of queue's XREQ socket
     registration (str): identity of registration XREQ socket
     heartbeat (str): identity of heartbeat XREQ socket
     """
     id=Int(0)
     queue=CBytes()
     control=CBytes()
     registration=CBytes()
     heartbeat=CBytes()
     pending=Set()
 
 class HubFactory(RegistrationFactory):
     """The Configurable for setting up a Hub."""
     
     # port-pairs for monitoredqueues:
     hb = Tuple(Int,Int,config=True,
         help="""XREQ/SUB Port pair for Engine heartbeats""")
     def _hb_default(self):
         return tuple(util.select_random_ports(2))
 
     mux = Tuple(Int,Int,config=True,
         help="""Engine/Client Port pair for MUX queue""")
     
     def _mux_default(self):
         return tuple(util.select_random_ports(2))
     
     task = Tuple(Int,Int,config=True,
         help="""Engine/Client Port pair for Task queue""")
     def _task_default(self):
         return tuple(util.select_random_ports(2))
 
     control = Tuple(Int,Int,config=True,
         help="""Engine/Client Port pair for Control queue""")
     
     def _control_default(self):
         return tuple(util.select_random_ports(2))
 
     iopub = Tuple(Int,Int,config=True,
         help="""Engine/Client Port pair for IOPub relay""")
     
     def _iopub_default(self):
         return tuple(util.select_random_ports(2))
     
     # single ports:
     mon_port = Int(config=True,
         help="""Monitor (SUB) port for queue traffic""")
 
     def _mon_port_default(self):
         return util.select_random_ports(1)[0]
     
     notifier_port = Int(config=True,
         help="""PUB port for sending engine status notifications""")
 
     def _notifier_port_default(self):
         return util.select_random_ports(1)[0]
     
     engine_ip = Unicode('127.0.0.1', config=True,
         help="IP on which to listen for engine connections. [default: loopback]")
     engine_transport = Unicode('tcp', config=True,
         help="0MQ transport for engine connections. [default: tcp]")
     
     client_ip = Unicode('127.0.0.1', config=True,
         help="IP on which to listen for client connections. [default: loopback]")
     client_transport = Unicode('tcp', config=True,
         help="0MQ transport for client connections. [default : tcp]")
     
     monitor_ip = Unicode('127.0.0.1', config=True,
         help="IP on which to listen for monitor messages. [default: loopback]")
     monitor_transport = Unicode('tcp', config=True,
         help="0MQ transport for monitor messages. [default : tcp]")
     
     monitor_url = Unicode('')
     
     db_class = DottedObjectName('IPython.parallel.controller.dictdb.DictDB',
         config=True, help="""The class to use for the DB backend""")
     
     # not configurable
     db = Instance('IPython.parallel.controller.dictdb.BaseDB')
     heartmonitor = Instance('IPython.parallel.controller.heartmonitor.HeartMonitor')
     
     def _ip_changed(self, name, old, new):
         self.engine_ip = new
         self.client_ip = new
         self.monitor_ip = new
         self._update_monitor_url()
     
     def _update_monitor_url(self):
         self.monitor_url = "%s://%s:%i"%(self.monitor_transport, self.monitor_ip, self.mon_port)
     
     def _transport_changed(self, name, old, new):
         self.engine_transport = new
         self.client_transport = new
         self.monitor_transport = new
         self._update_monitor_url()
         
     def __init__(self, **kwargs):
         super(HubFactory, self).__init__(**kwargs)
         self._update_monitor_url()
     
     
     def construct(self):
         self.init_hub()
     
     def start(self):
         self.heartmonitor.start()
         self.log.info("Heartmonitor started")
     
     def init_hub(self):
         """construct"""
         client_iface = "%s://%s:"%(self.client_transport, self.client_ip) + "%i"
         engine_iface = "%s://%s:"%(self.engine_transport, self.engine_ip) + "%i"
         
         ctx = self.context
         loop = self.loop
         
         # Registrar socket
         q = ZMQStream(ctx.socket(zmq.XREP), loop)
         q.bind(client_iface % self.regport)
         self.log.info("Hub listening on %s for registration."%(client_iface%self.regport))
         if self.client_ip != self.engine_ip:
             q.bind(engine_iface % self.regport)
             self.log.info("Hub listening on %s for registration."%(engine_iface%self.regport))
         
         ### Engine connections ###
 
         # heartbeat
         hpub = ctx.socket(zmq.PUB)
         hpub.bind(engine_iface % self.hb[0])
         hrep = ctx.socket(zmq.XREP)
         hrep.bind(engine_iface % self.hb[1])
         self.heartmonitor = HeartMonitor(loop=loop, config=self.config, log=self.log,
                                 pingstream=ZMQStream(hpub,loop),
                                 pongstream=ZMQStream(hrep,loop)
                             )
 
         ### Client connections ###
         # Notifier socket
         n = ZMQStream(ctx.socket(zmq.PUB), loop)
         n.bind(client_iface%self.notifier_port)
 
         ### build and launch the queues ###
 
         # monitor socket
         sub = ctx.socket(zmq.SUB)
         sub.setsockopt(zmq.SUBSCRIBE, b"")
         sub.bind(self.monitor_url)
         sub.bind('inproc://monitor')
         sub = ZMQStream(sub, loop)
         
         # connect the db
         self.log.info('Hub using DB backend: %r'%(self.db_class.split()[-1]))
         # cdir = self.config.Global.cluster_dir
         self.db = import_item(str(self.db_class))(session=self.session.session, 
                                             config=self.config, log=self.log)
         time.sleep(.25)
         try:
             scheme = self.config.TaskScheduler.scheme_name
         except AttributeError:
             from .scheduler import TaskScheduler
             scheme = TaskScheduler.scheme_name.get_default_value()
         # build connection dicts
         self.engine_info = {
             'control' : engine_iface%self.control[1],
             'mux': engine_iface%self.mux[1],
             'heartbeat': (engine_iface%self.hb[0], engine_iface%self.hb[1]),
             'task' : engine_iface%self.task[1],
             'iopub' : engine_iface%self.iopub[1],
             # 'monitor' : engine_iface%self.mon_port,
             }
 
         self.client_info = {
             'control' : client_iface%self.control[0],
             'mux': client_iface%self.mux[0],
             'task' : (scheme, client_iface%self.task[0]),
             'iopub' : client_iface%self.iopub[0],
             'notification': client_iface%self.notifier_port
             }
         self.log.debug("Hub engine addrs: %s"%self.engine_info)
         self.log.debug("Hub client addrs: %s"%self.client_info)
 
         # resubmit stream
         r = ZMQStream(ctx.socket(zmq.XREQ), loop)
         url = util.disambiguate_url(self.client_info['task'][-1])
         r.setsockopt(zmq.IDENTITY, util.asbytes(self.session.session))
         r.connect(url)
 
         self.hub = Hub(loop=loop, session=self.session, monitor=sub, heartmonitor=self.heartmonitor,
                 query=q, notifier=n, resubmit=r, db=self.db,
                 engine_info=self.engine_info, client_info=self.client_info,
                 log=self.log)
     
 
 class Hub(SessionFactory):
     """The IPython Controller Hub with 0MQ connections
     
     Parameters
     ==========
     loop: zmq IOLoop instance
     session: Session object
     <removed> context: zmq context for creating new connections (?)
     queue: ZMQStream for monitoring the command queue (SUB)
     query: ZMQStream for engine registration and client queries requests (XREP)
     heartbeat: HeartMonitor object checking the pulse of the engines
     notifier: ZMQStream for broadcasting engine registration changes (PUB)
     db: connection to db for out of memory logging of commands
                 NotImplemented
     engine_info: dict of zmq connection information for engines to connect
                 to the queues.
     client_info: dict of zmq connection information for engines to connect
                 to the queues.
     """
     # internal data structures:
     ids=Set() # engine IDs
     keytable=Dict()
     by_ident=Dict()
     engines=Dict()
     clients=Dict()
     hearts=Dict()
     pending=Set()
     queues=Dict()  # pending msg_ids keyed by engine_id
     tasks=Dict() # pending msg_ids submitted as tasks, keyed by client_id
     completed=Dict() # completed msg_ids keyed by engine_id
     all_completed=Set() # completed msg_ids keyed by engine_id
     dead_engines=Set() # completed msg_ids keyed by engine_id
     unassigned=Set() # set of task msg_ds not yet assigned a destination
     incoming_registrations=Dict()
     registration_timeout=Int()
     _idcounter=Int(0)
     
     # objects from constructor:
     query=Instance(ZMQStream)
     monitor=Instance(ZMQStream)
     notifier=Instance(ZMQStream)
     resubmit=Instance(ZMQStream)
     heartmonitor=Instance(HeartMonitor)
     db=Instance(object)
     client_info=Dict()
     engine_info=Dict()
     
     
     def __init__(self, **kwargs):
         """
         # universal:
         loop: IOLoop for creating future connections
         session: streamsession for sending serialized data
         # engine:
         queue: ZMQStream for monitoring queue messages
         query: ZMQStream for engine+client registration and client requests
         heartbeat: HeartMonitor object for tracking engines
         # extra:
         db: ZMQStream for db connection (NotImplemented)
         engine_info: zmq address/protocol dict for engine connections
         client_info: zmq address/protocol dict for client connections
         """
         
         super(Hub, self).__init__(**kwargs)
         self.registration_timeout = max(5000, 2*self.heartmonitor.period)
         
         # validate connection dicts:
         for k,v in self.client_info.iteritems():
             if k == 'task':
                 util.validate_url_container(v[1])
             else:
                 util.validate_url_container(v)
         # util.validate_url_container(self.client_info)
         util.validate_url_container(self.engine_info)
         
         # register our callbacks
         self.query.on_recv(self.dispatch_query)
         self.monitor.on_recv(self.dispatch_monitor_traffic)
         
         self.heartmonitor.add_heart_failure_handler(self.handle_heart_failure)
         self.heartmonitor.add_new_heart_handler(self.handle_new_heart)
         
         self.monitor_handlers = {b'in' : self.save_queue_request,
                                 b'out': self.save_queue_result,
                                 b'intask': self.save_task_request,
                                 b'outtask': self.save_task_result,
                                 b'tracktask': self.save_task_destination,
                                 b'incontrol': _passer,
                                 b'outcontrol': _passer,
                                 b'iopub': self.save_iopub_message,
         }
         
         self.query_handlers = {'queue_request': self.queue_status,
                                 'result_request': self.get_results,
                                 'history_request': self.get_history,
                                 'db_request': self.db_query,
                                 'purge_request': self.purge_results,
                                 'load_request': self.check_load,
                                 'resubmit_request': self.resubmit_task,
                                 'shutdown_request': self.shutdown_request,
                                 'registration_request' : self.register_engine,
                                 'unregistration_request' : self.unregister_engine,
                                 'connection_request': self.connection_request,
         }
         
         # ignore resubmit replies
         self.resubmit.on_recv(lambda msg: None, copy=False)
 
         self.log.info("hub::created hub")
     
     @property
     def _next_id(self):
         """gemerate a new ID.
         
         No longer reuse old ids, just count from 0."""
         newid = self._idcounter
         self._idcounter += 1
         return newid
         # newid = 0
         # incoming = [id[0] for id in self.incoming_registrations.itervalues()]
         # # print newid, self.ids, self.incoming_registrations
         # while newid in self.ids or newid in incoming:
         #     newid += 1
         # return newid
     
     #-----------------------------------------------------------------------------
     # message validation
     #-----------------------------------------------------------------------------
     
     def _validate_targets(self, targets):
         """turn any valid targets argument into a list of integer ids"""
         if targets is None:
             # default to all
             targets = self.ids
             
         if isinstance(targets, (int,str,unicode)):
             # only one target specified
             targets = [targets]
         _targets = []
         for t in targets:
             # map raw identities to ids
             if isinstance(t, (str,unicode)):
                 t = self.by_ident.get(t, t)
             _targets.append(t)
         targets = _targets
         bad_targets = [ t for t in targets if t not in self.ids ]
         if bad_targets:
             raise IndexError("No Such Engine: %r"%bad_targets)
         if not targets:
             raise IndexError("No Engines Registered")
         return targets
     
     #-----------------------------------------------------------------------------
     # dispatch methods (1 per stream)
     #-----------------------------------------------------------------------------
     
     
     def dispatch_monitor_traffic(self, msg):
         """all ME and Task queue messages come through here, as well as
         IOPub traffic."""
         self.log.debug("monitor traffic: %r"%msg[:2])
         switch = msg[0]
         try:
             idents, msg = self.session.feed_identities(msg[1:])
         except ValueError:
             idents=[]
         if not idents:
             self.log.error("Bad Monitor Message: %r"%msg)
             return
         handler = self.monitor_handlers.get(switch, None)
         if handler is not None:
             handler(idents, msg)
         else:
             self.log.error("Invalid monitor topic: %r"%switch)
         
     
     def dispatch_query(self, msg):
         """Route registration requests and queries from clients."""
         try:
             idents, msg = self.session.feed_identities(msg)
         except ValueError:
             idents = []
         if not idents:
             self.log.error("Bad Query Message: %r"%msg)
             return
         client_id = idents[0]
         try:
-            msg = self.session.unpack_message(msg, content=True)
+            msg = self.session.unserialize(msg, content=True)
         except Exception:
             content = error.wrap_exception()
             self.log.error("Bad Query Message: %r"%msg, exc_info=True)
             self.session.send(self.query, "hub_error", ident=client_id, 
                     content=content)
             return
         # print client_id, header, parent, content
         #switch on message type:
-        msg_type = msg['msg_type']
+        msg_type = msg['header']['msg_type']
         self.log.info("client::client %r requested %r"%(client_id, msg_type))
         handler = self.query_handlers.get(msg_type, None)
         try:
             assert handler is not None, "Bad Message Type: %r"%msg_type
         except:
             content = error.wrap_exception()
             self.log.error("Bad Message Type: %r"%msg_type, exc_info=True)
             self.session.send(self.query, "hub_error", ident=client_id, 
                     content=content)
             return
 
         else:
             handler(idents, msg)
             
     def dispatch_db(self, msg):
         """"""
         raise NotImplementedError
     
     #---------------------------------------------------------------------------
     # handler methods (1 per event)
     #---------------------------------------------------------------------------
     
     #----------------------- Heartbeat --------------------------------------
     
     def handle_new_heart(self, heart):
         """handler to attach to heartbeater.
         Called when a new heart starts to beat.
         Triggers completion of registration."""
         self.log.debug("heartbeat::handle_new_heart(%r)"%heart)
         if heart not in self.incoming_registrations:
             self.log.info("heartbeat::ignoring new heart: %r"%heart)
         else:
             self.finish_registration(heart)
         
     
     def handle_heart_failure(self, heart):
         """handler to attach to heartbeater.
         called when a previously registered heart fails to respond to beat request.
         triggers unregistration"""
         self.log.debug("heartbeat::handle_heart_failure(%r)"%heart)
         eid = self.hearts.get(heart, None)
         queue = self.engines[eid].queue
         if eid is None:
             self.log.info("heartbeat::ignoring heart failure %r"%heart)
         else:
             self.unregister_engine(heart, dict(content=dict(id=eid, queue=queue)))
     
     #----------------------- MUX Queue Traffic ------------------------------
     
     def save_queue_request(self, idents, msg):
         if len(idents) < 2:
             self.log.error("invalid identity prefix: %r"%idents)
             return
         queue_id, client_id = idents[:2]
         try:
-            msg = self.session.unpack_message(msg)
+            msg = self.session.unserialize(msg)
         except Exception:
             self.log.error("queue::client %r sent invalid message to %r: %r"%(client_id, queue_id, msg), exc_info=True)
             return
         
         eid = self.by_ident.get(queue_id, None)
         if eid is None:
             self.log.error("queue::target %r not registered"%queue_id)
             self.log.debug("queue::    valid are: %r"%(self.by_ident.keys()))
             return
         record = init_record(msg)
         msg_id = record['msg_id']
         # Unicode in records
         record['engine_uuid'] = queue_id.decode('ascii')
         record['client_uuid'] = client_id.decode('ascii')
         record['queue'] = 'mux'
 
         try:
             # it's posible iopub arrived first:
             existing = self.db.get_record(msg_id)
             for key,evalue in existing.iteritems():
                 rvalue = record.get(key, None)
                 if evalue and rvalue and evalue != rvalue:
                     self.log.warn("conflicting initial state for record: %r:%r <%r> %r"%(msg_id, rvalue, key, evalue))
                 elif evalue and not rvalue:
                     record[key] = evalue
             try:
                 self.db.update_record(msg_id, record)
             except Exception:
                 self.log.error("DB Error updating record %r"%msg_id, exc_info=True)
         except KeyError:
             try:
                 self.db.add_record(msg_id, record)
             except Exception:
                 self.log.error("DB Error adding record %r"%msg_id, exc_info=True)
                 
         
         self.pending.add(msg_id)
         self.queues[eid].append(msg_id)
     
     def save_queue_result(self, idents, msg):
         if len(idents) < 2:
             self.log.error("invalid identity prefix: %r"%idents)
             return
             
         client_id, queue_id = idents[:2]
         try:
-            msg = self.session.unpack_message(msg)
+            msg = self.session.unserialize(msg)
         except Exception:
             self.log.error("queue::engine %r sent invalid message to %r: %r"%(
                     queue_id,client_id, msg), exc_info=True)
             return
         
         eid = self.by_ident.get(queue_id, None)
         if eid is None:
             self.log.error("queue::unknown engine %r is sending a reply: "%queue_id)
             return
         
         parent = msg['parent_header']
         if not parent:
             return
         msg_id = parent['msg_id']
         if msg_id in self.pending:
             self.pending.remove(msg_id)
             self.all_completed.add(msg_id)
             self.queues[eid].remove(msg_id)
             self.completed[eid].append(msg_id)
         elif msg_id not in self.all_completed:
             # it could be a result from a dead engine that died before delivering the
             # result
             self.log.warn("queue:: unknown msg finished %r"%msg_id)
             return
         # update record anyway, because the unregistration could have been premature
         rheader = msg['header']
         completed = rheader['date']
         started = rheader.get('started', None)
         result = {
             'result_header' : rheader,
             'result_content': msg['content'],
             'started' : started,
             'completed' : completed
         }
 
         result['result_buffers'] = msg['buffers']
         try:
             self.db.update_record(msg_id, result)
         except Exception:
             self.log.error("DB Error updating record %r"%msg_id, exc_info=True)
         
             
     #--------------------- Task Queue Traffic ------------------------------
     
     def save_task_request(self, idents, msg):
         """Save the submission of a task."""
         client_id = idents[0]
         
         try:
-            msg = self.session.unpack_message(msg)
+            msg = self.session.unserialize(msg)
         except Exception:
             self.log.error("task::client %r sent invalid task message: %r"%(
                     client_id, msg), exc_info=True)
             return
         record = init_record(msg)
 
         record['client_uuid'] = client_id
         record['queue'] = 'task'
         header = msg['header']
         msg_id = header['msg_id']
         self.pending.add(msg_id)
         self.unassigned.add(msg_id)
         try:
             # it's posible iopub arrived first:
             existing = self.db.get_record(msg_id)
             if existing['resubmitted']:
                 for key in ('submitted', 'client_uuid', 'buffers'):
                     # don't clobber these keys on resubmit
                     # submitted and client_uuid should be different
                     # and buffers might be big, and shouldn't have changed
                     record.pop(key)
                     # still check content,header which should not change
                     # but are not expensive to compare as buffers
 
             for key,evalue in existing.iteritems():
                 if key.endswith('buffers'):
                     # don't compare buffers
                     continue
                 rvalue = record.get(key, None)
                 if evalue and rvalue and evalue != rvalue:
                     self.log.warn("conflicting initial state for record: %r:%r <%r> %r"%(msg_id, rvalue, key, evalue))
                 elif evalue and not rvalue:
                     record[key] = evalue
             try:
                 self.db.update_record(msg_id, record)
             except Exception:
                 self.log.error("DB Error updating record %r"%msg_id, exc_info=True)
         except KeyError:
             try:
                 self.db.add_record(msg_id, record)
             except Exception:
                 self.log.error("DB Error adding record %r"%msg_id, exc_info=True)
         except Exception:
             self.log.error("DB Error saving task request %r"%msg_id, exc_info=True)
     
     def save_task_result(self, idents, msg):
         """save the result of a completed task."""
         client_id = idents[0]
         try:
-            msg = self.session.unpack_message(msg)
+            msg = self.session.unserialize(msg)
         except Exception:
             self.log.error("task::invalid task result message send to %r: %r"%(
                     client_id, msg), exc_info=True)
             return
         
         parent = msg['parent_header']
         if not parent:
             # print msg
             self.log.warn("Task %r had no parent!"%msg)
             return
         msg_id = parent['msg_id']
         if msg_id in self.unassigned:
             self.unassigned.remove(msg_id)
         
         header = msg['header']
         engine_uuid = header.get('engine', None)
         eid = self.by_ident.get(engine_uuid, None)
         
         if msg_id in self.pending:
             self.pending.remove(msg_id)
             self.all_completed.add(msg_id)
             if eid is not None:
                 self.completed[eid].append(msg_id)
                 if msg_id in self.tasks[eid]:
                     self.tasks[eid].remove(msg_id)
             completed = header['date']
             started = header.get('started', None)
             result = {
                 'result_header' : header,
                 'result_content': msg['content'],
                 'started' : started,
                 'completed' : completed,
                 'engine_uuid': engine_uuid
             }
 
             result['result_buffers'] = msg['buffers']
             try:
                 self.db.update_record(msg_id, result)
             except Exception:
                 self.log.error("DB Error saving task request %r"%msg_id, exc_info=True)
             
         else:
             self.log.debug("task::unknown task %r finished"%msg_id)
     
     def save_task_destination(self, idents, msg):
         try:
-            msg = self.session.unpack_message(msg, content=True)
+            msg = self.session.unserialize(msg, content=True)
         except Exception:
             self.log.error("task::invalid task tracking message", exc_info=True)
             return
         content = msg['content']
         # print (content)
         msg_id = content['msg_id']
         engine_uuid = content['engine_id']
         eid = self.by_ident[util.asbytes(engine_uuid)]
         
         self.log.info("task::task %r arrived on %r"%(msg_id, eid))
         if msg_id in self.unassigned:
             self.unassigned.remove(msg_id)
         # else:
         #     self.log.debug("task::task %r not listed as MIA?!"%(msg_id))
         
         self.tasks[eid].append(msg_id)
         # self.pending[msg_id][1].update(received=datetime.now(),engine=(eid,engine_uuid))
         try:
             self.db.update_record(msg_id, dict(engine_uuid=engine_uuid))
         except Exception:
             self.log.error("DB Error saving task destination %r"%msg_id, exc_info=True)
             
     
     def mia_task_request(self, idents, msg):
         raise NotImplementedError
         client_id = idents[0]
         # content = dict(mia=self.mia,status='ok')
         # self.session.send('mia_reply', content=content, idents=client_id)
         
         
     #--------------------- IOPub Traffic ------------------------------
     
     def save_iopub_message(self, topics, msg):
         """save an iopub message into the db"""
         # print (topics)
         try:
-            msg = self.session.unpack_message(msg, content=True)
+            msg = self.session.unserialize(msg, content=True)
         except Exception:
             self.log.error("iopub::invalid IOPub message", exc_info=True)
             return
         
         parent = msg['parent_header']
         if not parent:
             self.log.error("iopub::invalid IOPub message: %r"%msg)
             return
         msg_id = parent['msg_id']
-        msg_type = msg['msg_type']
+        msg_type = msg['header']['msg_type']
         content = msg['content']
         
         # ensure msg_id is in db
         try:
             rec = self.db.get_record(msg_id)
         except KeyError:
             rec = empty_record()
             rec['msg_id'] = msg_id
             self.db.add_record(msg_id, rec)
         # stream
         d = {}
         if msg_type == 'stream':
             name = content['name']
             s = rec[name] or ''
             d[name] = s + content['data']
             
         elif msg_type == 'pyerr':
             d['pyerr'] = content
         elif msg_type == 'pyin':
             d['pyin'] = content['code']
         else:
             d[msg_type] = content.get('data', '')
         
         try:
             self.db.update_record(msg_id, d)
         except Exception:
             self.log.error("DB Error saving iopub message %r"%msg_id, exc_info=True)
         
     
             
     #-------------------------------------------------------------------------
     # Registration requests
     #-------------------------------------------------------------------------
         
     def connection_request(self, client_id, msg):
         """Reply with connection addresses for clients."""
         self.log.info("client::client %r connected"%client_id)
         content = dict(status='ok')
         content.update(self.client_info)
         jsonable = {}
         for k,v in self.keytable.iteritems():
             if v not in self.dead_engines:
                 jsonable[str(k)] = v.decode('ascii')
         content['engines'] = jsonable
         self.session.send(self.query, 'connection_reply', content, parent=msg, ident=client_id)
     
     def register_engine(self, reg, msg):
         """Register a new engine."""
         content = msg['content']
         try:
             queue = util.asbytes(content['queue'])
         except KeyError:
             self.log.error("registration::queue not specified", exc_info=True)
             return
         heart = content.get('heartbeat', None)
         if heart:
             heart = util.asbytes(heart)
         """register a new engine, and create the socket(s) necessary"""
         eid = self._next_id
         # print (eid, queue, reg, heart)
         
         self.log.debug("registration::register_engine(%i, %r, %r, %r)"%(eid, queue, reg, heart))
         
         content = dict(id=eid,status='ok')
         content.update(self.engine_info)
         # check if requesting available IDs:
         if queue in self.by_ident:
             try:
                 raise KeyError("queue_id %r in use"%queue)
             except:
                 content = error.wrap_exception()
                 self.log.error("queue_id %r in use"%queue, exc_info=True)
         elif heart in self.hearts: # need to check unique hearts?
             try:
                 raise KeyError("heart_id %r in use"%heart)
             except:
                 self.log.error("heart_id %r in use"%heart, exc_info=True)
                 content = error.wrap_exception()
         else:
             for h, pack in self.incoming_registrations.iteritems():
                 if heart == h:
                     try:
                         raise KeyError("heart_id %r in use"%heart)
                     except:
                         self.log.error("heart_id %r in use"%heart, exc_info=True)
                         content = error.wrap_exception()
                     break
                 elif queue == pack[1]:
                     try:
                         raise KeyError("queue_id %r in use"%queue)
                     except:
                         self.log.error("queue_id %r in use"%queue, exc_info=True)
                         content = error.wrap_exception()
                     break
         
         msg = self.session.send(self.query, "registration_reply", 
                 content=content, 
                 ident=reg)
         
         if content['status'] == 'ok':
             if heart in self.heartmonitor.hearts:
                 # already beating
                 self.incoming_registrations[heart] = (eid,queue,reg[0],None)
                 self.finish_registration(heart)
             else:
                 purge = lambda : self._purge_stalled_registration(heart)
                 dc = ioloop.DelayedCallback(purge, self.registration_timeout, self.loop)
                 dc.start()
                 self.incoming_registrations[heart] = (eid,queue,reg[0],dc)
         else:
             self.log.error("registration::registration %i failed: %r"%(eid, content['evalue']))
         return eid
     
     def unregister_engine(self, ident, msg):
         """Unregister an engine that explicitly requested to leave."""
         try:
             eid = msg['content']['id']
         except:
             self.log.error("registration::bad engine id for unregistration: %r"%ident, exc_info=True)
             return
         self.log.info("registration::unregister_engine(%r)"%eid)
         # print (eid)
         uuid = self.keytable[eid]
         content=dict(id=eid, queue=uuid.decode('ascii'))
         self.dead_engines.add(uuid)
         # self.ids.remove(eid)
         # uuid = self.keytable.pop(eid)
         # 
         # ec = self.engines.pop(eid)
         # self.hearts.pop(ec.heartbeat)
         # self.by_ident.pop(ec.queue)
         # self.completed.pop(eid)
         handleit = lambda : self._handle_stranded_msgs(eid, uuid)
         dc = ioloop.DelayedCallback(handleit, self.registration_timeout, self.loop)
         dc.start()
         ############## TODO: HANDLE IT ################
         
         if self.notifier:
             self.session.send(self.notifier, "unregistration_notification", content=content)
     
     def _handle_stranded_msgs(self, eid, uuid):
         """Handle messages known to be on an engine when the engine unregisters.
         
         It is possible that this will fire prematurely - that is, an engine will
         go down after completing a result, and the client will be notified
         that the result failed and later receive the actual result.
         """
         
         outstanding = self.queues[eid]
         
         for msg_id in outstanding:
             self.pending.remove(msg_id)
             self.all_completed.add(msg_id)
             try:
                 raise error.EngineError("Engine %r died while running task %r"%(eid, msg_id))
             except:
                 content = error.wrap_exception()
             # build a fake header:
             header = {}
             header['engine'] = uuid
             header['date'] = datetime.now()
             rec = dict(result_content=content, result_header=header, result_buffers=[])
             rec['completed'] = header['date']
             rec['engine_uuid'] = uuid
             try:
                 self.db.update_record(msg_id, rec)
             except Exception:
                 self.log.error("DB Error handling stranded msg %r"%msg_id, exc_info=True)
                 
     
     def finish_registration(self, heart):
         """Second half of engine registration, called after our HeartMonitor
         has received a beat from the Engine's Heart."""
         try: 
             (eid,queue,reg,purge) = self.incoming_registrations.pop(heart)
         except KeyError:
             self.log.error("registration::tried to finish nonexistant registration", exc_info=True)
             return
         self.log.info("registration::finished registering engine %i:%r"%(eid,queue))
         if purge is not None:
             purge.stop()
         control = queue
         self.ids.add(eid)
         self.keytable[eid] = queue
         self.engines[eid] = EngineConnector(id=eid, queue=queue, registration=reg, 
                                     control=control, heartbeat=heart)
         self.by_ident[queue] = eid
         self.queues[eid] = list()
         self.tasks[eid] = list()
         self.completed[eid] = list()
         self.hearts[heart] = eid
         content = dict(id=eid, queue=self.engines[eid].queue.decode('ascii'))
         if self.notifier:
             self.session.send(self.notifier, "registration_notification", content=content)
         self.log.info("engine::Engine Connected: %i"%eid)
     
     def _purge_stalled_registration(self, heart):
         if heart in self.incoming_registrations:
             eid = self.incoming_registrations.pop(heart)[0]
             self.log.info("registration::purging stalled registration: %i"%eid)
         else:
             pass
             
     #-------------------------------------------------------------------------
     # Client Requests
     #-------------------------------------------------------------------------
     
     def shutdown_request(self, client_id, msg):
         """handle shutdown request."""
         self.session.send(self.query, 'shutdown_reply', content={'status': 'ok'}, ident=client_id)
         # also notify other clients of shutdown
         self.session.send(self.notifier, 'shutdown_notice', content={'status': 'ok'})
         dc = ioloop.DelayedCallback(lambda : self._shutdown(), 1000, self.loop)
         dc.start()
     
     def _shutdown(self):
         self.log.info("hub::hub shutting down.")
         time.sleep(0.1)
         sys.exit(0)
         
     
     def check_load(self, client_id, msg):
         content = msg['content']
         try:
             targets = content['targets']
             targets = self._validate_targets(targets)
         except:
             content = error.wrap_exception()
             self.session.send(self.query, "hub_error", 
                     content=content, ident=client_id)
             return
         
         content = dict(status='ok')
         # loads = {}
         for t in targets:
             content[bytes(t)] = len(self.queues[t])+len(self.tasks[t])
         self.session.send(self.query, "load_reply", content=content, ident=client_id)
             
     
     def queue_status(self, client_id, msg):
         """Return the Queue status of one or more targets.
         if verbose: return the msg_ids
         else: return len of each type.
         keys: queue (pending MUX jobs)
             tasks (pending Task jobs)
             completed (finished jobs from both queues)"""
         content = msg['content']
         targets = content['targets']
         try:
             targets = self._validate_targets(targets)
         except:
             content = error.wrap_exception()
             self.session.send(self.query, "hub_error", 
                     content=content, ident=client_id)
             return
         verbose = content.get('verbose', False)
         content = dict(status='ok')
         for t in targets:
             queue = self.queues[t]
             completed = self.completed[t]
             tasks = self.tasks[t]
             if not verbose:
                 queue = len(queue)
                 completed = len(completed)
                 tasks = len(tasks)
             content[str(t)] = {'queue': queue, 'completed': completed , 'tasks': tasks}
         content['unassigned'] = list(self.unassigned) if verbose else len(self.unassigned)
         # print (content)
         self.session.send(self.query, "queue_reply", content=content, ident=client_id)
     
     def purge_results(self, client_id, msg):
         """Purge results from memory. This method is more valuable before we move
         to a DB based message storage mechanism."""
         content = msg['content']
         self.log.info("Dropping records with %s", content)
         msg_ids = content.get('msg_ids', [])
         reply = dict(status='ok')
         if msg_ids == 'all':
             try:
                 self.db.drop_matching_records(dict(completed={'$ne':None}))
             except Exception:
                 reply = error.wrap_exception()
         else:
             pending = filter(lambda m: m in self.pending, msg_ids)
             if pending:
                 try:
                     raise IndexError("msg pending: %r"%pending[0])
                 except:
                     reply = error.wrap_exception()
             else:
                 try:
                     self.db.drop_matching_records(dict(msg_id={'$in':msg_ids}))
                 except Exception:
                     reply = error.wrap_exception()
 
             if reply['status'] == 'ok':
                 eids = content.get('engine_ids', [])
                 for eid in eids:
                     if eid not in self.engines:
                         try:
                             raise IndexError("No such engine: %i"%eid)
                         except:
                             reply = error.wrap_exception()
                         break
                     uid = self.engines[eid].queue
                     try:
                         self.db.drop_matching_records(dict(engine_uuid=uid, completed={'$ne':None}))
                     except Exception:
                         reply = error.wrap_exception()
                         break
         
         self.session.send(self.query, 'purge_reply', content=reply, ident=client_id)
     
     def resubmit_task(self, client_id, msg):
         """Resubmit one or more tasks."""
         def finish(reply):
             self.session.send(self.query, 'resubmit_reply', content=reply, ident=client_id)
 
         content = msg['content']
         msg_ids = content['msg_ids']
         reply = dict(status='ok')
         try:
             records = self.db.find_records({'msg_id' : {'$in' : msg_ids}}, keys=[
                 'header', 'content', 'buffers'])
         except Exception:
             self.log.error('db::db error finding tasks to resubmit', exc_info=True)
             return finish(error.wrap_exception())
 
         # validate msg_ids
         found_ids = [ rec['msg_id'] for rec in records ]
         invalid_ids = filter(lambda m: m in self.pending, found_ids)
         if len(records) > len(msg_ids):
             try:
                 raise RuntimeError("DB appears to be in an inconsistent state."
                     "More matching records were found than should exist")
             except Exception:
                 return finish(error.wrap_exception())
         elif len(records) < len(msg_ids):
             missing = [ m for m in msg_ids if m not in found_ids ]
             try:
                 raise KeyError("No such msg(s): %r"%missing)
             except KeyError:
                 return finish(error.wrap_exception())
         elif invalid_ids:
             msg_id = invalid_ids[0]
             try:
                 raise ValueError("Task %r appears to be inflight"%(msg_id))
             except Exception:
                 return finish(error.wrap_exception())
 
         # clear the existing records
         now = datetime.now()
         rec = empty_record()
         map(rec.pop, ['msg_id', 'header', 'content', 'buffers', 'submitted'])
         rec['resubmitted'] = now
         rec['queue'] = 'task'
         rec['client_uuid'] = client_id[0]
         try:
             for msg_id in msg_ids:
                 self.all_completed.discard(msg_id)
                 self.db.update_record(msg_id, rec)
         except Exception:
             self.log.error('db::db error upating record', exc_info=True)
             reply = error.wrap_exception()
         else:
             # send the messages
             for rec in records:
                 header = rec['header']
                 # include resubmitted in header to prevent digest collision
                 header['resubmitted'] = now
                 msg = self.session.msg(header['msg_type'])
                 msg['content'] = rec['content']
                 msg['header'] = header
-                msg['msg_id'] = rec['msg_id']
+                msg['header']['msg_id'] = rec['msg_id']
                 self.session.send(self.resubmit, msg, buffers=rec['buffers'])
 
         finish(dict(status='ok'))
 
     
     def _extract_record(self, rec):
         """decompose a TaskRecord dict into subsection of reply for get_result"""
         io_dict = {}
         for key in 'pyin pyout pyerr stdout stderr'.split():
                 io_dict[key] = rec[key]
         content = { 'result_content': rec['result_content'],
                             'header': rec['header'],
                             'result_header' : rec['result_header'],
                             'io' : io_dict,
                           }
         if rec['result_buffers']:
             buffers = map(bytes, rec['result_buffers'])
         else:
             buffers = []
         
         return content, buffers
     
     def get_results(self, client_id, msg):
         """Get the result of 1 or more messages."""
         content = msg['content']
         msg_ids = sorted(set(content['msg_ids']))
         statusonly = content.get('status_only', False)
         pending = []
         completed = []
         content = dict(status='ok')
         content['pending'] = pending
         content['completed'] = completed
         buffers = []
         if not statusonly:
             try:
                 matches = self.db.find_records(dict(msg_id={'$in':msg_ids}))
                 # turn match list into dict, for faster lookup
                 records = {}
                 for rec in matches:
                     records[rec['msg_id']] = rec
             except Exception:
                 content = error.wrap_exception()
                 self.session.send(self.query, "result_reply", content=content, 
                                                     parent=msg, ident=client_id)
                 return
         else:
             records = {}
         for msg_id in msg_ids:
             if msg_id in self.pending:
                 pending.append(msg_id)
             elif msg_id in self.all_completed:
                 completed.append(msg_id)
                 if not statusonly:
                     c,bufs = self._extract_record(records[msg_id])
                     content[msg_id] = c
                     buffers.extend(bufs)
             elif msg_id in records:
                 if rec['completed']:
                     completed.append(msg_id)
                     c,bufs = self._extract_record(records[msg_id])
                     content[msg_id] = c
                     buffers.extend(bufs)
                 else:
                     pending.append(msg_id)
             else:
                 try:
                     raise KeyError('No such message: '+msg_id)
                 except:
                     content = error.wrap_exception()
                 break
         self.session.send(self.query, "result_reply", content=content, 
                                             parent=msg, ident=client_id,
                                             buffers=buffers)
 
     def get_history(self, client_id, msg):
         """Get a list of all msg_ids in our DB records"""
         try:
             msg_ids = self.db.get_history()
         except Exception as e:
             content = error.wrap_exception()
         else:
             content = dict(status='ok', history=msg_ids)
         
         self.session.send(self.query, "history_reply", content=content, 
                                             parent=msg, ident=client_id)
 
     def db_query(self, client_id, msg):
         """Perform a raw query on the task record database."""
         content = msg['content']
         query = content.get('query', {})
         keys = content.get('keys', None)
         buffers = []
         empty = list()
         try:
             records = self.db.find_records(query, keys)
         except Exception as e:
             content = error.wrap_exception()
         else:
             # extract buffers from reply content:
             if keys is not None:
                 buffer_lens = [] if 'buffers' in keys else None
                 result_buffer_lens = [] if 'result_buffers' in keys else None
             else:
                 buffer_lens = []
                 result_buffer_lens = []
             
             for rec in records:
                 # buffers may be None, so double check
                 if buffer_lens is not None:
                     b = rec.pop('buffers', empty) or empty
                     buffer_lens.append(len(b))
                     buffers.extend(b)
                 if result_buffer_lens is not None:
                     rb = rec.pop('result_buffers', empty) or empty
                     result_buffer_lens.append(len(rb))
                     buffers.extend(rb)
             content = dict(status='ok', records=records, buffer_lens=buffer_lens,
                                     result_buffer_lens=result_buffer_lens)
         # self.log.debug (content)
         self.session.send(self.query, "db_reply", content=content, 
                                             parent=msg, ident=client_id,
                                             buffers=buffers)
 

IPython/parallel/controller/scheduler.py

@@ -1,714 +1,714 b''
 """The Python scheduler for rich scheduling.
 
 The Pure ZMQ scheduler does not allow routing schemes other than LRU,
 nor does it check msg_id DAG dependencies. For those, a slightly slower
 Python Scheduler exists.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #----------------------------------------------------------------------
 # Imports
 #----------------------------------------------------------------------
 
 from __future__ import print_function
 
 import logging
 import sys
 
 from datetime import datetime, timedelta
 from random import randint, random
 from types import FunctionType
 
 try:
     import numpy
 except ImportError:
     numpy = None
 
 import zmq
 from zmq.eventloop import ioloop, zmqstream
 
 # local imports
 from IPython.external.decorator import decorator
 from IPython.config.application import Application
 from IPython.config.loader import Config
 from IPython.utils.traitlets import Instance, Dict, List, Set, Int, Enum, CBytes
 
 from IPython.parallel import error
 from IPython.parallel.factory import SessionFactory
 from IPython.parallel.util import connect_logger, local_logger, asbytes
 
 from .dependency import Dependency
 
 @decorator
 def logged(f,self,*args,**kwargs):
     # print ("#--------------------")
     self.log.debug("scheduler::%s(*%s,**%s)", f.func_name, args, kwargs)
     # print ("#--")
     return f(self,*args, **kwargs)
 
 #----------------------------------------------------------------------
 # Chooser functions
 #----------------------------------------------------------------------
 
 def plainrandom(loads):
     """Plain random pick."""
     n = len(loads)
     return randint(0,n-1)
 
 def lru(loads):
     """Always pick the front of the line.
     
     The content of `loads` is ignored.
     
     Assumes LRU ordering of loads, with oldest first.
     """
     return 0
 
 def twobin(loads):
     """Pick two at random, use the LRU of the two.
     
     The content of loads is ignored.
     
     Assumes LRU ordering of loads, with oldest first.
     """
     n = len(loads)
     a = randint(0,n-1)
     b = randint(0,n-1)
     return min(a,b)
 
 def weighted(loads):
     """Pick two at random using inverse load as weight.
     
     Return the less loaded of the two.
     """
     # weight 0 a million times more than 1:
     weights = 1./(1e-6+numpy.array(loads))
     sums = weights.cumsum()
     t = sums[-1]
     x = random()*t
     y = random()*t
     idx = 0
     idy = 0
     while sums[idx] < x:
         idx += 1
     while sums[idy] < y:
         idy += 1
     if weights[idy] > weights[idx]:
         return idy
     else:
         return idx
 
 def leastload(loads):
     """Always choose the lowest load.
     
     If the lowest load occurs more than once, the first
     occurance will be used.  If loads has LRU ordering, this means
     the LRU of those with the lowest load is chosen.
     """
     return loads.index(min(loads))
 
 #---------------------------------------------------------------------
 # Classes
 #---------------------------------------------------------------------
 # store empty default dependency:
 MET = Dependency([])
 
 class TaskScheduler(SessionFactory):
     """Python TaskScheduler object.
     
     This is the simplest object that supports msg_id based
     DAG dependencies. *Only* task msg_ids are checked, not
     msg_ids of jobs submitted via the MUX queue.
     
     """
     
     hwm = Int(0, config=True, shortname='hwm',
         help="""specify the High Water Mark (HWM) for the downstream
         socket in the Task scheduler. This is the maximum number
         of allowed outstanding tasks on each engine."""
     )
     scheme_name = Enum(('leastload', 'pure', 'lru', 'plainrandom', 'weighted', 'twobin'),
         'leastload', config=True, shortname='scheme', allow_none=False,
         help="""select the task scheduler scheme  [default: Python LRU]
         Options are: 'pure', 'lru', 'plainrandom', 'weighted', 'twobin','leastload'"""
     )
     def _scheme_name_changed(self, old, new):
         self.log.debug("Using scheme %r"%new)
         self.scheme = globals()[new]
     
     # input arguments:
     scheme = Instance(FunctionType) # function for determining the destination
     def _scheme_default(self):
         return leastload
     client_stream = Instance(zmqstream.ZMQStream) # client-facing stream
     engine_stream = Instance(zmqstream.ZMQStream) # engine-facing stream
     notifier_stream = Instance(zmqstream.ZMQStream) # hub-facing sub stream
     mon_stream = Instance(zmqstream.ZMQStream) # hub-facing pub stream
     
     # internals:
     graph = Dict() # dict by msg_id of [ msg_ids that depend on key ]
     retries = Dict() # dict by msg_id of retries remaining (non-neg ints)
     # waiting = List() # list of msg_ids ready to run, but haven't due to HWM
     depending = Dict() # dict by msg_id of (msg_id, raw_msg, after, follow)
     pending = Dict() # dict by engine_uuid of submitted tasks
     completed = Dict() # dict by engine_uuid of completed tasks
     failed = Dict() # dict by engine_uuid of failed tasks
     destinations = Dict() # dict by msg_id of engine_uuids where jobs ran (reverse of completed+failed)
     clients = Dict() # dict by msg_id for who submitted the task
     targets = List() # list of target IDENTs
     loads = List() # list of engine loads
     # full = Set() # set of IDENTs that have HWM outstanding tasks
     all_completed = Set() # set of all completed tasks
     all_failed = Set() # set of all failed tasks
     all_done = Set() # set of all finished tasks=union(completed,failed)
     all_ids = Set() # set of all submitted task IDs
     blacklist = Dict() # dict by msg_id of locations where a job has encountered UnmetDependency
     auditor = Instance('zmq.eventloop.ioloop.PeriodicCallback')
     
     ident = CBytes() # ZMQ identity. This should just be self.session.session
                      # but ensure Bytes
     def _ident_default(self):
         return asbytes(self.session.session)
     
     def start(self):
         self.engine_stream.on_recv(self.dispatch_result, copy=False)
         self._notification_handlers = dict(
             registration_notification = self._register_engine,
             unregistration_notification = self._unregister_engine
         )
         self.notifier_stream.on_recv(self.dispatch_notification)
         self.auditor = ioloop.PeriodicCallback(self.audit_timeouts, 2e3, self.loop) # 1 Hz
         self.auditor.start()
         self.log.info("Scheduler started [%s]"%self.scheme_name)
     
     def resume_receiving(self):
         """Resume accepting jobs."""
         self.client_stream.on_recv(self.dispatch_submission, copy=False)
     
     def stop_receiving(self):
         """Stop accepting jobs while there are no engines.
         Leave them in the ZMQ queue."""
         self.client_stream.on_recv(None)
     
     #-----------------------------------------------------------------------
     # [Un]Registration Handling
     #-----------------------------------------------------------------------
     
     def dispatch_notification(self, msg):
         """dispatch register/unregister events."""
         try:
             idents,msg = self.session.feed_identities(msg)
         except ValueError:
             self.log.warn("task::Invalid Message: %r",msg)
             return
         try:
-            msg = self.session.unpack_message(msg)
+            msg = self.session.unserialize(msg)
         except ValueError:
             self.log.warn("task::Unauthorized message from: %r"%idents)
             return
         
-        msg_type = msg['msg_type']
+        msg_type = msg['header']['msg_type']
         
         handler = self._notification_handlers.get(msg_type, None)
         if handler is None:
             self.log.error("Unhandled message type: %r"%msg_type)
         else:
             try:
                 handler(asbytes(msg['content']['queue']))
             except Exception:
                 self.log.error("task::Invalid notification msg: %r",msg)
     
     def _register_engine(self, uid):
         """New engine with ident `uid` became available."""
         # head of the line:
         self.targets.insert(0,uid)
         self.loads.insert(0,0)
 
         # initialize sets
         self.completed[uid] = set()
         self.failed[uid] = set()
         self.pending[uid] = {}
         if len(self.targets) == 1:
             self.resume_receiving()
         # rescan the graph:
         self.update_graph(None)
 
     def _unregister_engine(self, uid):
         """Existing engine with ident `uid` became unavailable."""
         if len(self.targets) == 1:
             # this was our only engine
             self.stop_receiving()
         
         # handle any potentially finished tasks:
         self.engine_stream.flush()
         
         # don't pop destinations, because they might be used later
         # map(self.destinations.pop, self.completed.pop(uid))
         # map(self.destinations.pop, self.failed.pop(uid))
 
         # prevent this engine from receiving work
         idx = self.targets.index(uid)
         self.targets.pop(idx)
         self.loads.pop(idx)
         
         # wait 5 seconds before cleaning up pending jobs, since the results might
         # still be incoming
         if self.pending[uid]:
             dc = ioloop.DelayedCallback(lambda : self.handle_stranded_tasks(uid), 5000, self.loop)
             dc.start()
         else:
             self.completed.pop(uid)
             self.failed.pop(uid)
 
     
     def handle_stranded_tasks(self, engine):
         """Deal with jobs resident in an engine that died."""
         lost = self.pending[engine]
         for msg_id in lost.keys():
             if msg_id not in self.pending[engine]:
                 # prevent double-handling of messages
                 continue
 
             raw_msg = lost[msg_id][0]
             idents,msg = self.session.feed_identities(raw_msg, copy=False)
             parent = self.session.unpack(msg[1].bytes)
             idents = [engine, idents[0]]
 
             # build fake error reply
             try:
                 raise error.EngineError("Engine %r died while running task %r"%(engine, msg_id))
             except:
                 content = error.wrap_exception()
             msg = self.session.msg('apply_reply', content, parent=parent, subheader={'status':'error'})
             raw_reply = map(zmq.Message, self.session.serialize(msg, ident=idents))
             # and dispatch it
             self.dispatch_result(raw_reply)
 
         # finally scrub completed/failed lists
         self.completed.pop(engine)
         self.failed.pop(engine)
     
     
     #-----------------------------------------------------------------------
     # Job Submission
     #-----------------------------------------------------------------------
     def dispatch_submission(self, raw_msg):
         """Dispatch job submission to appropriate handlers."""
         # ensure targets up to date:
         self.notifier_stream.flush()
         try:
             idents, msg = self.session.feed_identities(raw_msg, copy=False)
-            msg = self.session.unpack_message(msg, content=False, copy=False)
+            msg = self.session.unserialize(msg, content=False, copy=False)
         except Exception:
             self.log.error("task::Invaid task msg: %r"%raw_msg, exc_info=True)
             return
         
         
         # send to monitor
         self.mon_stream.send_multipart([b'intask']+raw_msg, copy=False)
         
         header = msg['header']
         msg_id = header['msg_id']
         self.all_ids.add(msg_id)
         
         # get targets as a set of bytes objects
         # from a list of unicode objects
         targets = header.get('targets', [])
         targets = map(asbytes, targets)
         targets = set(targets)
             
         retries = header.get('retries', 0)
         self.retries[msg_id] = retries
         
         # time dependencies
         after = header.get('after', None)
         if after:
             after = Dependency(after)
             if after.all:
                 if after.success:
                     after = Dependency(after.difference(self.all_completed),
                                 success=after.success,
                                 failure=after.failure,
                                 all=after.all,
                     )
                 if after.failure:
                     after = Dependency(after.difference(self.all_failed),
                                 success=after.success,
                                 failure=after.failure,
                                 all=after.all,
                     )
             if after.check(self.all_completed, self.all_failed):
                 # recast as empty set, if `after` already met,
                 # to prevent unnecessary set comparisons
                 after = MET
         else:
             after = MET
         
         # location dependencies
         follow = Dependency(header.get('follow', []))
         
         # turn timeouts into datetime objects:
         timeout = header.get('timeout', None)
         if timeout:
             timeout = datetime.now() + timedelta(0,timeout,0)
         
         args = [raw_msg, targets, after, follow, timeout]
         
         # validate and reduce dependencies:
         for dep in after,follow:
             if not dep: # empty dependency
                 continue
             # check valid:
             if msg_id in dep or dep.difference(self.all_ids):
                 self.depending[msg_id] = args
                 return self.fail_unreachable(msg_id, error.InvalidDependency)
             # check if unreachable:
             if dep.unreachable(self.all_completed, self.all_failed):
                 self.depending[msg_id] = args
                 return self.fail_unreachable(msg_id)
         
         if after.check(self.all_completed, self.all_failed):
             # time deps already met, try to run
             if not self.maybe_run(msg_id, *args):
                 # can't run yet
                 if msg_id not in self.all_failed:
                     # could have failed as unreachable
                     self.save_unmet(msg_id, *args)
         else:
             self.save_unmet(msg_id, *args)
     
     def audit_timeouts(self):
         """Audit all waiting tasks for expired timeouts."""
         now = datetime.now()
         for msg_id in self.depending.keys():
             # must recheck, in case one failure cascaded to another:
             if msg_id in self.depending:
                 raw,after,targets,follow,timeout = self.depending[msg_id]
                 if timeout and timeout < now:
                     self.fail_unreachable(msg_id, error.TaskTimeout)
                 
     def fail_unreachable(self, msg_id, why=error.ImpossibleDependency):
         """a task has become unreachable, send a reply with an ImpossibleDependency
         error."""
         if msg_id not in self.depending:
             self.log.error("msg %r already failed!", msg_id)
             return
         raw_msg,targets,after,follow,timeout = self.depending.pop(msg_id)
         for mid in follow.union(after):
             if mid in self.graph:
                 self.graph[mid].remove(msg_id)
         
         # FIXME: unpacking a message I've already unpacked, but didn't save:
         idents,msg = self.session.feed_identities(raw_msg, copy=False)
         header = self.session.unpack(msg[1].bytes)
         
         try:
             raise why()
         except:
             content = error.wrap_exception()
         
         self.all_done.add(msg_id)
         self.all_failed.add(msg_id)
         
         msg = self.session.send(self.client_stream, 'apply_reply', content, 
                                                 parent=header, ident=idents)
         self.session.send(self.mon_stream, msg, ident=[b'outtask']+idents)
         
         self.update_graph(msg_id, success=False)
     
     def maybe_run(self, msg_id, raw_msg, targets, after, follow, timeout):
         """check location dependencies, and run if they are met."""
         blacklist = self.blacklist.setdefault(msg_id, set())
         if follow or targets or blacklist or self.hwm:
             # we need a can_run filter
             def can_run(idx):
                 # check hwm
                 if self.hwm and self.loads[idx] == self.hwm:
                     return False
                 target = self.targets[idx]
                 # check blacklist
                 if target in blacklist:
                     return False
                 # check targets
                 if targets and target not in targets:
                     return False
                 # check follow
                 return follow.check(self.completed[target], self.failed[target])
             
             indices = filter(can_run, range(len(self.targets)))
 
             if not indices:
                 # couldn't run
                 if follow.all:
                     # check follow for impossibility
                     dests = set()
                     relevant = set()
                     if follow.success:
                         relevant = self.all_completed
                     if follow.failure:
                         relevant = relevant.union(self.all_failed)
                     for m in follow.intersection(relevant):
                         dests.add(self.destinations[m])
                     if len(dests) > 1:
                         self.depending[msg_id] = (raw_msg, targets, after, follow, timeout)
                         self.fail_unreachable(msg_id)
                         return False
                 if targets:
                     # check blacklist+targets for impossibility
                     targets.difference_update(blacklist)
                     if not targets or not targets.intersection(self.targets):
                         self.depending[msg_id] = (raw_msg, targets, after, follow, timeout)
                         self.fail_unreachable(msg_id)
                         return False
                 return False
         else:
             indices = None
             
         self.submit_task(msg_id, raw_msg, targets, follow, timeout, indices)
         return True
             
     def save_unmet(self, msg_id, raw_msg, targets, after, follow, timeout):
         """Save a message for later submission when its dependencies are met."""
         self.depending[msg_id] = [raw_msg,targets,after,follow,timeout]
         # track the ids in follow or after, but not those already finished
         for dep_id in after.union(follow).difference(self.all_done):
             if dep_id not in self.graph:
                 self.graph[dep_id] = set()
             self.graph[dep_id].add(msg_id)
     
     def submit_task(self, msg_id, raw_msg, targets, follow, timeout, indices=None):
         """Submit a task to any of a subset of our targets."""
         if indices:
             loads = [self.loads[i] for i in indices]
         else:
             loads = self.loads
         idx = self.scheme(loads)
         if indices:
             idx = indices[idx]
         target = self.targets[idx]
         # print (target, map(str, msg[:3]))
         # send job to the engine
         self.engine_stream.send(target, flags=zmq.SNDMORE, copy=False)
         self.engine_stream.send_multipart(raw_msg, copy=False)
         # update load
         self.add_job(idx)
         self.pending[target][msg_id] = (raw_msg, targets, MET, follow, timeout)
         # notify Hub
         content = dict(msg_id=msg_id, engine_id=target.decode('ascii'))
         self.session.send(self.mon_stream, 'task_destination', content=content, 
                         ident=[b'tracktask',self.ident])
         
     
     #-----------------------------------------------------------------------
     # Result Handling
     #-----------------------------------------------------------------------
     def dispatch_result(self, raw_msg):
         """dispatch method for result replies"""
         try:
             idents,msg = self.session.feed_identities(raw_msg, copy=False)
-            msg = self.session.unpack_message(msg, content=False, copy=False)
+            msg = self.session.unserialize(msg, content=False, copy=False)
             engine = idents[0]
             try:
                 idx = self.targets.index(engine)
             except ValueError:
                 pass # skip load-update for dead engines
             else:
                 self.finish_job(idx)
         except Exception:
             self.log.error("task::Invaid result: %r", raw_msg, exc_info=True)
             return
 
         header = msg['header']
         parent = msg['parent_header']
         if header.get('dependencies_met', True):
             success = (header['status'] == 'ok')
             msg_id = parent['msg_id']
             retries = self.retries[msg_id]
             if not success and retries > 0:
                 # failed
                 self.retries[msg_id] = retries - 1
                 self.handle_unmet_dependency(idents, parent)
             else:
                 del self.retries[msg_id]
                 # relay to client and update graph
                 self.handle_result(idents, parent, raw_msg, success)
                 # send to Hub monitor
                 self.mon_stream.send_multipart([b'outtask']+raw_msg, copy=False)
         else:
             self.handle_unmet_dependency(idents, parent)
         
     def handle_result(self, idents, parent, raw_msg, success=True):
         """handle a real task result, either success or failure"""
         # first, relay result to client
         engine = idents[0]
         client = idents[1]
         # swap_ids for XREP-XREP mirror
         raw_msg[:2] = [client,engine]
         # print (map(str, raw_msg[:4]))
         self.client_stream.send_multipart(raw_msg, copy=False)
         # now, update our data structures
         msg_id = parent['msg_id']
         self.blacklist.pop(msg_id, None)
         self.pending[engine].pop(msg_id)
         if success:
             self.completed[engine].add(msg_id)
             self.all_completed.add(msg_id)
         else:
             self.failed[engine].add(msg_id)
             self.all_failed.add(msg_id)
         self.all_done.add(msg_id)
         self.destinations[msg_id] = engine
         
         self.update_graph(msg_id, success)
         
     def handle_unmet_dependency(self, idents, parent):
         """handle an unmet dependency"""
         engine = idents[0]
         msg_id = parent['msg_id']
         
         if msg_id not in self.blacklist:
             self.blacklist[msg_id] = set()
         self.blacklist[msg_id].add(engine)
         
         args = self.pending[engine].pop(msg_id)
         raw,targets,after,follow,timeout = args
         
         if self.blacklist[msg_id] == targets:
             self.depending[msg_id] = args
             self.fail_unreachable(msg_id)
         elif not self.maybe_run(msg_id, *args):
             # resubmit failed
             if msg_id not in self.all_failed:
                 # put it back in our dependency tree
                 self.save_unmet(msg_id, *args)
         
         if self.hwm:
             try:
                 idx = self.targets.index(engine)
             except ValueError:
                 pass # skip load-update for dead engines
             else:
                 if self.loads[idx] == self.hwm-1:
                     self.update_graph(None)
         
         
     
     def update_graph(self, dep_id=None, success=True):
         """dep_id just finished. Update our dependency
         graph and submit any jobs that just became runable.
         
         Called with dep_id=None to update entire graph for hwm, but without finishing
         a task.
         """
         # print ("\n\n***********")
         # pprint (dep_id)
         # pprint (self.graph)
         # pprint (self.depending)
         # pprint (self.all_completed)
         # pprint (self.all_failed)
         # print ("\n\n***********\n\n")
         # update any jobs that depended on the dependency
         jobs = self.graph.pop(dep_id, [])
 
         # recheck *all* jobs if
         # a) we have HWM and an engine just become no longer full
         # or b) dep_id was given as None
         if dep_id is None or self.hwm and any( [ load==self.hwm-1 for load in self.loads ]):
             jobs = self.depending.keys()
         
         for msg_id in jobs:
             raw_msg, targets, after, follow, timeout = self.depending[msg_id]
             
             if after.unreachable(self.all_completed, self.all_failed)\
                     or follow.unreachable(self.all_completed, self.all_failed):
                 self.fail_unreachable(msg_id)
             
             elif after.check(self.all_completed, self.all_failed): # time deps met, maybe run
                 if self.maybe_run(msg_id, raw_msg, targets, MET, follow, timeout):
                     
                     self.depending.pop(msg_id)
                     for mid in follow.union(after):
                         if mid in self.graph:
                             self.graph[mid].remove(msg_id)
     
     #----------------------------------------------------------------------
     # methods to be overridden by subclasses
     #----------------------------------------------------------------------
     
     def add_job(self, idx):
         """Called after self.targets[idx] just got the job with header.
         Override with subclasses.  The default ordering is simple LRU.
         The default loads are the number of outstanding jobs."""
         self.loads[idx] += 1
         for lis in (self.targets, self.loads):
             lis.append(lis.pop(idx))
             
     
     def finish_job(self, idx):
         """Called after self.targets[idx] just finished a job.
         Override with subclasses."""
         self.loads[idx] -= 1
     
 
 
 def launch_scheduler(in_addr, out_addr, mon_addr, not_addr, config=None,
                         logname='root', log_url=None, loglevel=logging.DEBUG,
                         identity=b'task', in_thread=False):
     
     ZMQStream = zmqstream.ZMQStream
     
     if config:
         # unwrap dict back into Config
         config = Config(config)
 
     if in_thread:
         # use instance() to get the same Context/Loop as our parent
         ctx = zmq.Context.instance()
         loop = ioloop.IOLoop.instance()
     else:
         # in a process, don't use instance() 
         # for safety with multiprocessing
         ctx = zmq.Context()
         loop = ioloop.IOLoop()
     ins = ZMQStream(ctx.socket(zmq.XREP),loop)
     ins.setsockopt(zmq.IDENTITY, identity)
     ins.bind(in_addr)
     
     outs = ZMQStream(ctx.socket(zmq.XREP),loop)
     outs.setsockopt(zmq.IDENTITY, identity)
     outs.bind(out_addr)
     mons = zmqstream.ZMQStream(ctx.socket(zmq.PUB),loop)
     mons.connect(mon_addr)
     nots = zmqstream.ZMQStream(ctx.socket(zmq.SUB),loop)
     nots.setsockopt(zmq.SUBSCRIBE, b'')
     nots.connect(not_addr)
     
     # setup logging.
     if in_thread:
         log = Application.instance().log
     else:
         if log_url:
             log = connect_logger(logname, ctx, log_url, root="scheduler", loglevel=loglevel)
         else:
             log = local_logger(logname, loglevel)
     
     scheduler = TaskScheduler(client_stream=ins, engine_stream=outs,
                             mon_stream=mons, notifier_stream=nots,
                             loop=loop, log=log,
                             config=config)
     scheduler.start()
     if not in_thread:
         try:
             loop.start()
         except KeyboardInterrupt:
             print ("interrupted, exiting...", file=sys.__stderr__)
 

IPython/parallel/engine/engine.py

@@ -1,174 +1,174 b''
 #!/usr/bin/env python
 """A simple engine that talks to a controller over 0MQ.
 it handles registration, etc. and launches a kernel
 connected to the Controller's Schedulers.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 from __future__ import print_function
 
 import sys
 import time
 
 import zmq
 from zmq.eventloop import ioloop, zmqstream
 
 # internal
 from IPython.utils.traitlets import Instance, Dict, Int, Type, CFloat, Unicode, CBytes
 # from IPython.utils.localinterfaces import LOCALHOST 
 
 from IPython.parallel.controller.heartmonitor import Heart
 from IPython.parallel.factory import RegistrationFactory
 from IPython.parallel.util import disambiguate_url, asbytes
 
 from IPython.zmq.session import Message
 
 from .streamkernel import Kernel
 
 class EngineFactory(RegistrationFactory):
     """IPython engine"""
     
     # configurables:
     out_stream_factory=Type('IPython.zmq.iostream.OutStream', config=True,
         help="""The OutStream for handling stdout/err.
         Typically 'IPython.zmq.iostream.OutStream'""")
     display_hook_factory=Type('IPython.zmq.displayhook.ZMQDisplayHook', config=True,
         help="""The class for handling displayhook.
         Typically 'IPython.zmq.displayhook.ZMQDisplayHook'""")
     location=Unicode(config=True,
         help="""The location (an IP address) of the controller.  This is
         used for disambiguating URLs, to determine whether
         loopback should be used to connect or the public address.""")
     timeout=CFloat(2,config=True,
         help="""The time (in seconds) to wait for the Controller to respond
         to registration requests before giving up.""")
     
     # not configurable:
     user_ns=Dict()
     id=Int(allow_none=True)
     registrar=Instance('zmq.eventloop.zmqstream.ZMQStream')
     kernel=Instance(Kernel)
     
     bident = CBytes()
     ident = Unicode()
     def _ident_changed(self, name, old, new):
         self.bident = asbytes(new)
     
     
     def __init__(self, **kwargs):
         super(EngineFactory, self).__init__(**kwargs)
         self.ident = self.session.session
         ctx = self.context
         
         reg = ctx.socket(zmq.XREQ)
         reg.setsockopt(zmq.IDENTITY, self.bident)
         reg.connect(self.url)
         self.registrar = zmqstream.ZMQStream(reg, self.loop)
         
     def register(self):
         """send the registration_request"""
         
         self.log.info("Registering with controller at %s"%self.url)
         content = dict(queue=self.ident, heartbeat=self.ident, control=self.ident)
         self.registrar.on_recv(self.complete_registration)
         # print (self.session.key)
         self.session.send(self.registrar, "registration_request",content=content)
     
     def complete_registration(self, msg):
         # print msg
         self._abort_dc.stop()
         ctx = self.context
         loop = self.loop
         identity = self.bident
         idents,msg = self.session.feed_identities(msg)
-        msg = Message(self.session.unpack_message(msg))
+        msg = Message(self.session.unserialize(msg))
         
         if msg.content.status == 'ok':
             self.id = int(msg.content.id)
             
             # create Shell Streams (MUX, Task, etc.):
             queue_addr = msg.content.mux
             shell_addrs = [ str(queue_addr) ]
             task_addr = msg.content.task
             if task_addr:
                 shell_addrs.append(str(task_addr))
             
             # Uncomment this to go back to two-socket model
             # shell_streams = []
             # for addr in shell_addrs:
             #     stream = zmqstream.ZMQStream(ctx.socket(zmq.XREP), loop)
             #     stream.setsockopt(zmq.IDENTITY, identity)
             #     stream.connect(disambiguate_url(addr, self.location))
             #     shell_streams.append(stream)
             
             # Now use only one shell stream for mux and tasks
             stream = zmqstream.ZMQStream(ctx.socket(zmq.XREP), loop)
             stream.setsockopt(zmq.IDENTITY, identity)
             shell_streams = [stream]
             for addr in shell_addrs:
                 stream.connect(disambiguate_url(addr, self.location))
             # end single stream-socket
             
             # control stream:
             control_addr = str(msg.content.control)
             control_stream = zmqstream.ZMQStream(ctx.socket(zmq.XREP), loop)
             control_stream.setsockopt(zmq.IDENTITY, identity)
             control_stream.connect(disambiguate_url(control_addr, self.location))
             
             # create iopub stream:
             iopub_addr = msg.content.iopub
             iopub_stream = zmqstream.ZMQStream(ctx.socket(zmq.PUB), loop)
             iopub_stream.setsockopt(zmq.IDENTITY, identity)
             iopub_stream.connect(disambiguate_url(iopub_addr, self.location))
             
             # launch heartbeat
             hb_addrs = msg.content.heartbeat
             # print (hb_addrs)
             
             # # Redirect input streams and set a display hook.
             if self.out_stream_factory:
                 sys.stdout = self.out_stream_factory(self.session, iopub_stream, u'stdout')
                 sys.stdout.topic = 'engine.%i.stdout'%self.id
                 sys.stderr = self.out_stream_factory(self.session, iopub_stream, u'stderr')
                 sys.stderr.topic = 'engine.%i.stderr'%self.id
             if self.display_hook_factory:
                 sys.displayhook = self.display_hook_factory(self.session, iopub_stream)
                 sys.displayhook.topic = 'engine.%i.pyout'%self.id
 
             self.kernel = Kernel(config=self.config, int_id=self.id, ident=self.ident, session=self.session, 
                     control_stream=control_stream, shell_streams=shell_streams, iopub_stream=iopub_stream, 
                     loop=loop, user_ns = self.user_ns, log=self.log)
             self.kernel.start()
             hb_addrs = [ disambiguate_url(addr, self.location) for addr in hb_addrs ]
             heart = Heart(*map(str, hb_addrs), heart_id=identity)
             heart.start()
             
             
         else:
             self.log.fatal("Registration Failed: %s"%msg)
             raise Exception("Registration Failed: %s"%msg)
         
         self.log.info("Completed registration with id %i"%self.id)
     
     
     def abort(self):
         self.log.fatal("Registration timed out after %.1f seconds"%self.timeout)
         self.session.send(self.registrar, "unregistration_request", content=dict(id=self.id))
         time.sleep(1)
         sys.exit(255)
     
     def start(self):
         dc = ioloop.DelayedCallback(self.register, 0, self.loop)
         dc.start()
         self._abort_dc = ioloop.DelayedCallback(self.abort, self.timeout*1000, self.loop)
         self._abort_dc.start()
 

IPython/parallel/engine/kernelstarter.py

@@ -1,230 +1,230 b''
 """KernelStarter class that intercepts Control Queue messages, and handles process management.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 from zmq.eventloop import ioloop
 
 from IPython.zmq.session import Session
 
 class KernelStarter(object):
     """Object for resetting/killing the Kernel."""
     
     
     def __init__(self, session, upstream, downstream, *kernel_args, **kernel_kwargs):
         self.session = session
         self.upstream = upstream
         self.downstream = downstream
         self.kernel_args = kernel_args
         self.kernel_kwargs = kernel_kwargs
         self.handlers = {}
         for method in 'shutdown_request shutdown_reply'.split():
             self.handlers[method] = getattr(self, method)
     
     def start(self):
         self.upstream.on_recv(self.dispatch_request)
         self.downstream.on_recv(self.dispatch_reply)
         
     #--------------------------------------------------------------------------
     # Dispatch methods
     #--------------------------------------------------------------------------
     
     def dispatch_request(self, raw_msg):
         idents, msg = self.session.feed_identities()
         try:
-            msg = self.session.unpack_message(msg, content=False)
+            msg = self.session.unserialize(msg, content=False)
         except:
             print ("bad msg: %s"%msg)
         
-        msgtype = msg['msg_type']
+        msgtype = msg['header']['msg_type']
         handler = self.handlers.get(msgtype, None)
         if handler is None:
             self.downstream.send_multipart(raw_msg, copy=False)
         else:
             handler(msg)
         
     def dispatch_reply(self, raw_msg):
         idents, msg = self.session.feed_identities()
         try:
-            msg = self.session.unpack_message(msg, content=False)
+            msg = self.session.unserialize(msg, content=False)
         except:
             print ("bad msg: %s"%msg)
         
-        msgtype = msg['msg_type']
+        msgtype = msg['header']['msg_type']
         handler = self.handlers.get(msgtype, None)
         if handler is None:
             self.upstream.send_multipart(raw_msg, copy=False)
         else:
             handler(msg)
     
     #--------------------------------------------------------------------------
     # Handlers
     #--------------------------------------------------------------------------
     
     def shutdown_request(self, msg):
         """"""
         self.downstream.send_multipart(msg)
     
     #--------------------------------------------------------------------------
     # Kernel process management methods, from KernelManager:
     #--------------------------------------------------------------------------
     
     def _check_local(addr):
         if isinstance(addr, tuple):
             addr = addr[0]
         return addr in LOCAL_IPS
     
     def start_kernel(self, **kw):
         """Starts a kernel process and configures the manager to use it.
 
         If random ports (port=0) are being used, this method must be called
         before the channels are created.
 
         Parameters:
         -----------
         ipython : bool, optional (default True)
              Whether to use an IPython kernel instead of a plain Python kernel.
         """
         self.kernel = Process(target=make_kernel, args=self.kernel_args,
                                             kwargs=self.kernel_kwargs)
 
     def shutdown_kernel(self, restart=False):
         """ Attempts to the stop the kernel process cleanly. If the kernel
         cannot be stopped, it is killed, if possible.
         """
         # FIXME: Shutdown does not work on Windows due to ZMQ errors!
         if sys.platform == 'win32':
             self.kill_kernel()
             return
 
         # Don't send any additional kernel kill messages immediately, to give
         # the kernel a chance to properly execute shutdown actions. Wait for at
         # most 1s, checking every 0.1s.
         self.xreq_channel.shutdown(restart=restart)
         for i in range(10):
             if self.is_alive:
                 time.sleep(0.1)
             else:
                 break
         else:
             # OK, we've waited long enough.
             if self.has_kernel:
                 self.kill_kernel()
     
     def restart_kernel(self, now=False):
         """Restarts a kernel with the same arguments that were used to launch
         it. If the old kernel was launched with random ports, the same ports
         will be used for the new kernel.
 
         Parameters
         ----------
         now : bool, optional
           If True, the kernel is forcefully restarted *immediately*, without
           having a chance to do any cleanup action.  Otherwise the kernel is
           given 1s to clean up before a forceful restart is issued.
 
           In all cases the kernel is restarted, the only difference is whether
           it is given a chance to perform a clean shutdown or not.
         """
         if self._launch_args is None:
             raise RuntimeError("Cannot restart the kernel. "
                                "No previous call to 'start_kernel'.")
         else:
             if self.has_kernel:
                 if now:
                     self.kill_kernel()
                 else:
                     self.shutdown_kernel(restart=True)
             self.start_kernel(**self._launch_args)
 
             # FIXME: Messages get dropped in Windows due to probable ZMQ bug
             # unless there is some delay here.
             if sys.platform == 'win32':
                 time.sleep(0.2)
 
     @property
     def has_kernel(self):
         """Returns whether a kernel process has been specified for the kernel
         manager.
         """
         return self.kernel is not None
 
     def kill_kernel(self):
         """ Kill the running kernel. """
         if self.has_kernel:
             # Pause the heart beat channel if it exists.
             if self._hb_channel is not None:
                 self._hb_channel.pause()
 
             # Attempt to kill the kernel.
             try:
                 self.kernel.kill()
             except OSError, e:
                 # In Windows, we will get an Access Denied error if the process
                 # has already terminated. Ignore it.
                 if not (sys.platform == 'win32' and e.winerror == 5):
                     raise
             self.kernel = None
         else:
             raise RuntimeError("Cannot kill kernel. No kernel is running!")
 
     def interrupt_kernel(self):
         """ Interrupts the kernel. Unlike ``signal_kernel``, this operation is
         well supported on all platforms.
         """
         if self.has_kernel:
             if sys.platform == 'win32':
                 from parentpoller import ParentPollerWindows as Poller
                 Poller.send_interrupt(self.kernel.win32_interrupt_event)
             else:
                 self.kernel.send_signal(signal.SIGINT)
         else:
             raise RuntimeError("Cannot interrupt kernel. No kernel is running!")
 
     def signal_kernel(self, signum):
         """ Sends a signal to the kernel. Note that since only SIGTERM is
         supported on Windows, this function is only useful on Unix systems.
         """
         if self.has_kernel:
             self.kernel.send_signal(signum)
         else:
             raise RuntimeError("Cannot signal kernel. No kernel is running!")
 
     @property
     def is_alive(self):
         """Is the kernel process still running?"""
         # FIXME: not using a heartbeat means this method is broken for any
         # remote kernel, it's only capable of handling local kernels.
         if self.has_kernel:
             if self.kernel.poll() is None:
                 return True
             else:
                 return False
         else:
             # We didn't start the kernel with this KernelManager so we don't
             # know if it is running. We should use a heartbeat for this case.
             return True
 
 
 def make_starter(up_addr, down_addr, *args, **kwargs):
     """entry point function for launching a kernelstarter in a subprocess"""
     loop = ioloop.IOLoop.instance()
     ctx = zmq.Context()
     session = Session()
     upstream = zmqstream.ZMQStream(ctx.socket(zmq.XREQ),loop)
     upstream.connect(up_addr)
     downstream = zmqstream.ZMQStream(ctx.socket(zmq.XREQ),loop)
     downstream.connect(down_addr)
     
     starter = KernelStarter(session, upstream, downstream, *args, **kwargs)
     starter.start()
     loop.start()
-    
  No newline at end of file
+    

IPython/parallel/engine/streamkernel.py

@@ -1,438 +1,440 b''
 #!/usr/bin/env python
 """
 Kernel adapted from kernel.py to use ZMQ Streams
 
 Authors:
 
 * Min RK
 * Brian Granger
 * Fernando Perez
 * Evan Patterson
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 # Standard library imports.
 from __future__ import print_function
 
 import sys
 import time
 
 from code import CommandCompiler
 from datetime import datetime
 from pprint import pprint
 
 # System library imports.
 import zmq
 from zmq.eventloop import ioloop, zmqstream
 
 # Local imports.
 from IPython.utils.traitlets import Instance, List, Int, Dict, Set, Unicode, CBytes
 from IPython.zmq.completer import KernelCompleter
 
 from IPython.parallel.error import wrap_exception
 from IPython.parallel.factory import SessionFactory
 from IPython.parallel.util import serialize_object, unpack_apply_message, asbytes
 
 def printer(*args):
     pprint(args, stream=sys.__stdout__)
 
 
 class _Passer(zmqstream.ZMQStream):
     """Empty class that implements `send()` that does nothing.
     
     Subclass ZMQStream for Session typechecking
     
     """
     def __init__(self, *args, **kwargs):
         pass
     
     def send(self, *args, **kwargs):
         pass
     send_multipart = send
     
 
 #-----------------------------------------------------------------------------
 # Main kernel class
 #-----------------------------------------------------------------------------
 
 class Kernel(SessionFactory):
 
     #---------------------------------------------------------------------------
     # Kernel interface
     #---------------------------------------------------------------------------
     
     # kwargs:
     exec_lines = List(Unicode, config=True,
         help="List of lines to execute")
     
     # identities:
     int_id = Int(-1)
     bident = CBytes()
     ident = Unicode()
     def _ident_changed(self, name, old, new):
         self.bident = asbytes(new)
     
     user_ns = Dict(config=True,  help="""Set the user's namespace of the Kernel""")
     
     control_stream = Instance(zmqstream.ZMQStream)
     task_stream = Instance(zmqstream.ZMQStream)
     iopub_stream = Instance(zmqstream.ZMQStream)
     client = Instance('IPython.parallel.Client')
     
     # internals
     shell_streams = List()
     compiler = Instance(CommandCompiler, (), {})
     completer = Instance(KernelCompleter)
     
     aborted = Set()
     shell_handlers = Dict()
     control_handlers = Dict()
     
     def _set_prefix(self):
         self.prefix = "engine.%s"%self.int_id
     
     def _connect_completer(self):
         self.completer = KernelCompleter(self.user_ns)
     
     def __init__(self, **kwargs):
         super(Kernel, self).__init__(**kwargs)
         self._set_prefix()
         self._connect_completer()
         
         self.on_trait_change(self._set_prefix, 'id')
         self.on_trait_change(self._connect_completer, 'user_ns')
         
         # Build dict of handlers for message types
         for msg_type in ['execute_request', 'complete_request', 'apply_request', 
                 'clear_request']:
             self.shell_handlers[msg_type] = getattr(self, msg_type)
         
         for msg_type in ['shutdown_request', 'abort_request']+self.shell_handlers.keys():
             self.control_handlers[msg_type] = getattr(self, msg_type)
         
         self._initial_exec_lines()
     
     def _wrap_exception(self, method=None):
         e_info = dict(engine_uuid=self.ident, engine_id=self.int_id, method=method)
         content=wrap_exception(e_info)
         return content
     
     def _initial_exec_lines(self):
         s = _Passer()
         content = dict(silent=True, user_variable=[],user_expressions=[])
         for line in self.exec_lines:
             self.log.debug("executing initialization: %s"%line)
             content.update({'code':line})
             msg = self.session.msg('execute_request', content)
             self.execute_request(s, [], msg)
         
         
     #-------------------- control handlers -----------------------------
     def abort_queues(self):
         for stream in self.shell_streams:
             if stream:
                 self.abort_queue(stream)
     
     def abort_queue(self, stream):
         while True:
             idents,msg = self.session.recv(stream, zmq.NOBLOCK, content=True)
             if msg is None:
                 return
                 
             self.log.info("Aborting:")
             self.log.info(str(msg))
-            msg_type = msg['msg_type']
+            msg_type = msg['header']['msg_type']
             reply_type = msg_type.split('_')[0] + '_reply'
             # reply_msg = self.session.msg(reply_type, {'status' : 'aborted'}, msg)
             # self.reply_socket.send(ident,zmq.SNDMORE)
             # self.reply_socket.send_json(reply_msg)
             reply_msg = self.session.send(stream, reply_type, 
                         content={'status' : 'aborted'}, parent=msg, ident=idents)
             self.log.debug(str(reply_msg))
             # We need to wait a bit for requests to come in. This can probably
             # be set shorter for true asynchronous clients.
             time.sleep(0.05)
     
     def abort_request(self, stream, ident, parent):
         """abort a specifig msg by id"""
         msg_ids = parent['content'].get('msg_ids', None)
         if isinstance(msg_ids, basestring):
             msg_ids = [msg_ids]
         if not msg_ids:
             self.abort_queues()
         for mid in msg_ids:
             self.aborted.add(str(mid))
         
         content = dict(status='ok')
         reply_msg = self.session.send(stream, 'abort_reply', content=content, 
                 parent=parent, ident=ident)
         self.log.debug(str(reply_msg))
     
     def shutdown_request(self, stream, ident, parent):
         """kill ourself.  This should really be handled in an external process"""
         try:
             self.abort_queues()
         except:
             content = self._wrap_exception('shutdown')
         else:
             content = dict(parent['content'])
             content['status'] = 'ok'
         msg = self.session.send(stream, 'shutdown_reply',
                                 content=content, parent=parent, ident=ident)
         self.log.debug(str(msg))
         dc = ioloop.DelayedCallback(lambda : sys.exit(0), 1000, self.loop)
         dc.start()
     
     def dispatch_control(self, msg):
         idents,msg = self.session.feed_identities(msg, copy=False)
         try:
-            msg = self.session.unpack_message(msg, content=True, copy=False)
+            msg = self.session.unserialize(msg, content=True, copy=False)
         except:
             self.log.error("Invalid Message", exc_info=True)
             return
         else:
             self.log.debug("Control received, %s", msg)
         
         header = msg['header']
         msg_id = header['msg_id']
-        
-        handler = self.control_handlers.get(msg['msg_type'], None)
+        msg_type = header['msg_type']
+
+        handler = self.control_handlers.get(msg_type, None)
         if handler is None:
-            self.log.error("UNKNOWN CONTROL MESSAGE TYPE: %r"%msg['msg_type'])
+            self.log.error("UNKNOWN CONTROL MESSAGE TYPE: %r"%msg_type)
         else:
             handler(self.control_stream, idents, msg)
     
 
     #-------------------- queue helpers ------------------------------
     
     def check_dependencies(self, dependencies):
         if not dependencies:
             return True
         if len(dependencies) == 2 and dependencies[0] in 'any all'.split():
             anyorall = dependencies[0]
             dependencies = dependencies[1]
         else:
             anyorall = 'all'
         results = self.client.get_results(dependencies,status_only=True)
         if results['status'] != 'ok':
             return False
         
         if anyorall == 'any':
             if not results['completed']:
                 return False
         else:
             if results['pending']:
                 return False
         
         return True
     
     def check_aborted(self, msg_id):
         return msg_id in self.aborted
     
     #-------------------- queue handlers -----------------------------
     
     def clear_request(self, stream, idents, parent):
         """Clear our namespace."""
         self.user_ns = {}
         msg = self.session.send(stream, 'clear_reply', ident=idents, parent=parent, 
                 content = dict(status='ok'))
         self._initial_exec_lines()
     
     def execute_request(self, stream, ident, parent):
         self.log.debug('execute request %s'%parent)
         try:
             code = parent[u'content'][u'code']
         except:
             self.log.error("Got bad msg: %s"%parent, exc_info=True)
             return
         self.session.send(self.iopub_stream, u'pyin', {u'code':code},parent=parent,
                             ident=asbytes('%s.pyin'%self.prefix))
         started = datetime.now()
         try:
             comp_code = self.compiler(code, '<zmq-kernel>')
             # allow for not overriding displayhook
             if hasattr(sys.displayhook, 'set_parent'):
                 sys.displayhook.set_parent(parent)
                 sys.stdout.set_parent(parent)
                 sys.stderr.set_parent(parent)
             exec comp_code in self.user_ns, self.user_ns
         except:
             exc_content = self._wrap_exception('execute')
             # exc_msg = self.session.msg(u'pyerr', exc_content, parent)
             self.session.send(self.iopub_stream, u'pyerr', exc_content, parent=parent,
                             ident=asbytes('%s.pyerr'%self.prefix))
             reply_content = exc_content
         else:
             reply_content = {'status' : 'ok'}
         
         reply_msg = self.session.send(stream, u'execute_reply', reply_content, parent=parent, 
                     ident=ident, subheader = dict(started=started))
         self.log.debug(str(reply_msg))
         if reply_msg['content']['status'] == u'error':
             self.abort_queues()
 
     def complete_request(self, stream, ident, parent):
         matches = {'matches' : self.complete(parent),
                    'status' : 'ok'}
         completion_msg = self.session.send(stream, 'complete_reply',
                                            matches, parent, ident)
         # print >> sys.__stdout__, completion_msg
 
     def complete(self, msg):
         return self.completer.complete(msg.content.line, msg.content.text)
     
     def apply_request(self, stream, ident, parent):
         # flush previous reply, so this request won't block it
         stream.flush(zmq.POLLOUT)
         try:
             content = parent[u'content']
             bufs = parent[u'buffers']
             msg_id = parent['header']['msg_id']
             # bound = parent['header'].get('bound', False)
         except:
             self.log.error("Got bad msg: %s"%parent, exc_info=True)
             return
         # pyin_msg = self.session.msg(u'pyin',{u'code':code}, parent=parent)
         # self.iopub_stream.send(pyin_msg)
         # self.session.send(self.iopub_stream, u'pyin', {u'code':code},parent=parent)
         sub = {'dependencies_met' : True, 'engine' : self.ident,
                 'started': datetime.now()}
         try:
             # allow for not overriding displayhook
             if hasattr(sys.displayhook, 'set_parent'):
                 sys.displayhook.set_parent(parent)
                 sys.stdout.set_parent(parent)
                 sys.stderr.set_parent(parent)
             # exec "f(*args,**kwargs)" in self.user_ns, self.user_ns
             working = self.user_ns
             # suffix = 
             prefix = "_"+str(msg_id).replace("-","")+"_"
             
             f,args,kwargs = unpack_apply_message(bufs, working, copy=False)
             # if bound:
             #     bound_ns = Namespace(working)
             #     args = [bound_ns]+list(args)
 
             fname = getattr(f, '__name__', 'f')
             
             fname = prefix+"f"
             argname = prefix+"args"
             kwargname = prefix+"kwargs"
             resultname = prefix+"result"
             
             ns = { fname : f, argname : args, kwargname : kwargs , resultname : None }
             # print ns
             working.update(ns)
             code = "%s=%s(*%s,**%s)"%(resultname, fname, argname, kwargname)
             try:
                 exec code in working,working
                 result = working.get(resultname)
             finally:
                 for key in ns.iterkeys():
                     working.pop(key)
             # if bound:
             #     working.update(bound_ns)
             
             packed_result,buf = serialize_object(result)
             result_buf = [packed_result]+buf
         except:
             exc_content = self._wrap_exception('apply')
             # exc_msg = self.session.msg(u'pyerr', exc_content, parent)
             self.session.send(self.iopub_stream, u'pyerr', exc_content, parent=parent,
                                 ident=asbytes('%s.pyerr'%self.prefix))
             reply_content = exc_content
             result_buf = []
             
             if exc_content['ename'] == 'UnmetDependency':
                 sub['dependencies_met'] = False
         else:
             reply_content = {'status' : 'ok'}
         
         # put 'ok'/'error' status in header, for scheduler introspection:
         sub['status'] = reply_content['status']
         
         reply_msg = self.session.send(stream, u'apply_reply', reply_content, 
                     parent=parent, ident=ident,buffers=result_buf, subheader=sub)
         
         # flush i/o
         # should this be before reply_msg is sent, like in the single-kernel code, 
         # or should nothing get in the way of real results?
         sys.stdout.flush()
         sys.stderr.flush()
     
     def dispatch_queue(self, stream, msg):
         self.control_stream.flush()
         idents,msg = self.session.feed_identities(msg, copy=False)
         try:
-            msg = self.session.unpack_message(msg, content=True, copy=False)
+            msg = self.session.unserialize(msg, content=True, copy=False)
         except:
             self.log.error("Invalid Message", exc_info=True)
             return
         else:
             self.log.debug("Message received, %s", msg)
             
         
         header = msg['header']
         msg_id = header['msg_id']
+        msg_type = msg['header']['msg_type']
         if self.check_aborted(msg_id):
             self.aborted.remove(msg_id)
             # is it safe to assume a msg_id will not be resubmitted?
-            reply_type = msg['msg_type'].split('_')[0] + '_reply'
+            reply_type = msg_type.split('_')[0] + '_reply'
             status = {'status' : 'aborted'}
             reply_msg = self.session.send(stream, reply_type, subheader=status,
                         content=status, parent=msg, ident=idents)
             return
-        handler = self.shell_handlers.get(msg['msg_type'], None)
+        handler = self.shell_handlers.get(msg_type, None)
         if handler is None:
-            self.log.error("UNKNOWN MESSAGE TYPE: %r"%msg['msg_type'])
+            self.log.error("UNKNOWN MESSAGE TYPE: %r"%msg_type)
         else:
             handler(stream, idents, msg)
     
     def start(self):
         #### stream mode:
         if self.control_stream:
             self.control_stream.on_recv(self.dispatch_control, copy=False)
             self.control_stream.on_err(printer)
         
         def make_dispatcher(stream):
             def dispatcher(msg):
                 return self.dispatch_queue(stream, msg)
             return dispatcher
         
         for s in self.shell_streams:
             s.on_recv(make_dispatcher(s), copy=False)
             s.on_err(printer)
         
         if self.iopub_stream:
             self.iopub_stream.on_err(printer)
         
         #### while True mode:
         # while True:
         #     idle = True
         #     try:
         #         msg = self.shell_stream.socket.recv_multipart(
         #                     zmq.NOBLOCK, copy=False)
         #     except zmq.ZMQError, e:
         #         if e.errno != zmq.EAGAIN:
         #             raise e
         #     else:
         #         idle=False
         #         self.dispatch_queue(self.shell_stream, msg)
         #             
         #     if not self.task_stream.empty():
         #         idle=False
         #         msg = self.task_stream.recv_multipart()
         #         self.dispatch_queue(self.task_stream, msg)
         #     if idle:
         #         # don't busywait
         #         time.sleep(1e-3)
 

IPython/parallel/tests/test_db.py

@@ -1,179 +1,180 b''
 """Tests for db backends
 
 Authors:
 
 * Min RK
 """
 
 #-------------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-------------------------------------------------------------------------------
 
 #-------------------------------------------------------------------------------
 # Imports
 #-------------------------------------------------------------------------------
 
 from __future__ import division
 
 import tempfile
 import time
 
 from datetime import datetime, timedelta
 from unittest import TestCase
 
 from nose import SkipTest
 
 from IPython.parallel import error
 from IPython.parallel.controller.dictdb import DictDB
 from IPython.parallel.controller.sqlitedb import SQLiteDB
 from IPython.parallel.controller.hub import init_record, empty_record
 
 from IPython.zmq.session import Session
 
 
 #-------------------------------------------------------------------------------
 # TestCases
 #-------------------------------------------------------------------------------
 
 class TestDictBackend(TestCase):
     def setUp(self):
         self.session = Session()
         self.db = self.create_db()
         self.load_records(16)
     
     def create_db(self):
         return DictDB()
     
     def load_records(self, n=1):
         """load n records for testing"""
         #sleep 1/10 s, to ensure timestamp is different to previous calls
         time.sleep(0.1)
         msg_ids = []
         for i in range(n):
             msg = self.session.msg('apply_request', content=dict(a=5))
             msg['buffers'] = []
             rec = init_record(msg)
-            msg_ids.append(msg['msg_id'])
-            self.db.add_record(msg['msg_id'], rec)
+            msg_id = msg['header']['msg_id']
+            msg_ids.append(msg_id)
+            self.db.add_record(msg_id, rec)
         return msg_ids
     
     def test_add_record(self):
         before = self.db.get_history()
         self.load_records(5)
         after = self.db.get_history()
         self.assertEquals(len(after), len(before)+5)
         self.assertEquals(after[:-5],before)
         
     def test_drop_record(self):
         msg_id = self.load_records()[-1]
         rec = self.db.get_record(msg_id)
         self.db.drop_record(msg_id)
         self.assertRaises(KeyError,self.db.get_record, msg_id)
     
     def _round_to_millisecond(self, dt):
         """necessary because mongodb rounds microseconds"""
         micro = dt.microsecond
         extra = int(str(micro)[-3:])
         return dt - timedelta(microseconds=extra)
     
     def test_update_record(self):
         now = self._round_to_millisecond(datetime.now())
         # 
         msg_id = self.db.get_history()[-1]
         rec1 = self.db.get_record(msg_id)
         data = {'stdout': 'hello there', 'completed' : now}
         self.db.update_record(msg_id, data)
         rec2 = self.db.get_record(msg_id)
         self.assertEquals(rec2['stdout'], 'hello there')
         self.assertEquals(rec2['completed'], now)
         rec1.update(data)
         self.assertEquals(rec1, rec2)
     
     # def test_update_record_bad(self):
     #     """test updating nonexistant records"""
     #     msg_id = str(uuid.uuid4())
     #     data = {'stdout': 'hello there'}
     #     self.assertRaises(KeyError, self.db.update_record, msg_id, data)
 
     def test_find_records_dt(self):
         """test finding records by date"""
         hist = self.db.get_history()
         middle = self.db.get_record(hist[len(hist)//2])
         tic = middle['submitted']
         before = self.db.find_records({'submitted' : {'$lt' : tic}})
         after = self.db.find_records({'submitted' : {'$gte' : tic}})
         self.assertEquals(len(before)+len(after),len(hist))
         for b in before:
             self.assertTrue(b['submitted'] < tic)
         for a in after:
             self.assertTrue(a['submitted'] >= tic)
         same = self.db.find_records({'submitted' : tic})
         for s in same:
             self.assertTrue(s['submitted'] == tic)
     
     def test_find_records_keys(self):
         """test extracting subset of record keys"""
         found = self.db.find_records({'msg_id': {'$ne' : ''}},keys=['submitted', 'completed'])
         for rec in found:
             self.assertEquals(set(rec.keys()), set(['msg_id', 'submitted', 'completed']))
     
     def test_find_records_msg_id(self):
         """ensure msg_id is always in found records"""
         found = self.db.find_records({'msg_id': {'$ne' : ''}},keys=['submitted', 'completed'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
         found = self.db.find_records({'msg_id': {'$ne' : ''}},keys=['submitted'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
         found = self.db.find_records({'msg_id': {'$ne' : ''}},keys=['msg_id'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
     
     def test_find_records_in(self):
         """test finding records with '$in','$nin' operators"""
         hist = self.db.get_history()
         even = hist[::2]
         odd = hist[1::2]
         recs = self.db.find_records({ 'msg_id' : {'$in' : even}})
         found = [ r['msg_id'] for r in recs ]
         self.assertEquals(set(even), set(found))
         recs = self.db.find_records({ 'msg_id' : {'$nin' : even}})
         found = [ r['msg_id'] for r in recs ]
         self.assertEquals(set(odd), set(found))
     
     def test_get_history(self):
         msg_ids = self.db.get_history()
         latest = datetime(1984,1,1)
         for msg_id in msg_ids:
             rec = self.db.get_record(msg_id)
             newt = rec['submitted']
             self.assertTrue(newt >= latest)
             latest = newt
         msg_id = self.load_records(1)[-1]
         self.assertEquals(self.db.get_history()[-1],msg_id)
     
     def test_datetime(self):
         """get/set timestamps with datetime objects"""
         msg_id = self.db.get_history()[-1]
         rec = self.db.get_record(msg_id)
         self.assertTrue(isinstance(rec['submitted'], datetime))
         self.db.update_record(msg_id, dict(completed=datetime.now()))
         rec = self.db.get_record(msg_id)
         self.assertTrue(isinstance(rec['completed'], datetime))
 
     def test_drop_matching(self):
         msg_ids = self.load_records(10)
         query = {'msg_id' : {'$in':msg_ids}}
         self.db.drop_matching_records(query)
         recs = self.db.find_records(query)
         self.assertEquals(len(recs), 0)
             
 class TestSQLiteBackend(TestDictBackend):
     def create_db(self):
         return SQLiteDB(location=tempfile.gettempdir())
     
     def tearDown(self):
         self.db._db.close()

IPython/zmq/ipkernel.py

@@ -1,678 +1,680 b''
 #!/usr/bin/env python
 """A simple interactive kernel that talks to a frontend over 0MQ.
 
 Things to do:
 
 * Implement `set_parent` logic. Right before doing exec, the Kernel should
   call set_parent on all the PUB objects with the message about to be executed.
 * Implement random port and security key logic.
 * Implement control messages.
 * Implement event loop and poll version.
 """
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 from __future__ import print_function
 
 # Standard library imports.
 import __builtin__
 import atexit
 import sys
 import time
 import traceback
 import logging
 # System library imports.
 import zmq
 
 # Local imports.
 from IPython.config.configurable import Configurable
 from IPython.config.application import boolean_flag
 from IPython.core.application import ProfileDir
 from IPython.core.shellapp import (
     InteractiveShellApp, shell_flags, shell_aliases
 )
 from IPython.utils import io
 from IPython.utils.jsonutil import json_clean
 from IPython.lib import pylabtools
 from IPython.utils.traitlets import (
     List, Instance, Float, Dict, Bool, Int, Unicode, CaselessStrEnum
 )
 
 from entry_point import base_launch_kernel
 from kernelapp import KernelApp, kernel_flags, kernel_aliases
 from iostream import OutStream
 from session import Session, Message
 from zmqshell import ZMQInteractiveShell
 
 
 
 #-----------------------------------------------------------------------------
 # Main kernel class
 #-----------------------------------------------------------------------------
 
 class Kernel(Configurable):
 
     #---------------------------------------------------------------------------
     # Kernel interface
     #---------------------------------------------------------------------------
 
     shell = Instance('IPython.core.interactiveshell.InteractiveShellABC')
     session = Instance(Session)
     shell_socket = Instance('zmq.Socket')
     iopub_socket = Instance('zmq.Socket')
     stdin_socket = Instance('zmq.Socket')
     log = Instance(logging.Logger)
 
     # Private interface
 
     # Time to sleep after flushing the stdout/err buffers in each execute
     # cycle.  While this introduces a hard limit on the minimal latency of the
     # execute cycle, it helps prevent output synchronization problems for
     # clients.
     # Units are in seconds.  The minimum zmq latency on local host is probably
     # ~150 microseconds, set this to 500us for now.  We may need to increase it
     # a little if it's not enough after more interactive testing.
     _execute_sleep = Float(0.0005, config=True)
 
     # Frequency of the kernel's event loop.
     # Units are in seconds, kernel subclasses for GUI toolkits may need to
     # adapt to milliseconds.
     _poll_interval = Float(0.05, config=True)
 
     # If the shutdown was requested over the network, we leave here the
     # necessary reply message so it can be sent by our registered atexit
     # handler.  This ensures that the reply is only sent to clients truly at
     # the end of our shutdown process (which happens after the underlying
     # IPython shell's own shutdown).
     _shutdown_message = None
 
     # This is a dict of port number that the kernel is listening on. It is set
     # by record_ports and used by connect_request.
     _recorded_ports = Dict()
 
 
 
     def __init__(self, **kwargs):
         super(Kernel, self).__init__(**kwargs)
 
         # Before we even start up the shell, register *first* our exit handlers
         # so they come before the shell's
         atexit.register(self._at_shutdown)
 
         # Initialize the InteractiveShell subclass
         self.shell = ZMQInteractiveShell.instance(config=self.config)
         self.shell.displayhook.session = self.session
         self.shell.displayhook.pub_socket = self.iopub_socket
         self.shell.display_pub.session = self.session
         self.shell.display_pub.pub_socket = self.iopub_socket
 
         # TMP - hack while developing
         self.shell._reply_content = None
 
         # Build dict of handlers for message types
         msg_types = [ 'execute_request', 'complete_request', 
                       'object_info_request', 'history_request',
                       'connect_request', 'shutdown_request']
         self.handlers = {}
         for msg_type in msg_types:
             self.handlers[msg_type] = getattr(self, msg_type)
 
     def do_one_iteration(self):
         """Do one iteration of the kernel's evaluation loop.
         """
         ident,msg = self.session.recv(self.shell_socket, zmq.NOBLOCK)
         if msg is None:
             return
-        
+
+        msg_type = msg['header']['msg_type']
+
         # This assert will raise in versions of zeromq 2.0.7 and lesser.
         # We now require 2.0.8 or above, so we can uncomment for safety.
         # print(ident,msg, file=sys.__stdout__)
         assert ident is not None, "Missing message part."
         
         # Print some info about this message and leave a '--->' marker, so it's
         # easier to trace visually the message chain when debugging.  Each
         # handler prints its message at the end.
-        self.log.debug('\n*** MESSAGE TYPE:'+str(msg['msg_type'])+'***')
+        self.log.debug('\n*** MESSAGE TYPE:'+str(msg_type)+'***')
         self.log.debug('   Content: '+str(msg['content'])+'\n   --->\n   ')
 
         # Find and call actual handler for message
-        handler = self.handlers.get(msg['msg_type'], None)
+        handler = self.handlers.get(msg_type, None)
         if handler is None:
             self.log.error("UNKNOWN MESSAGE TYPE:" +str(msg))
         else:
             handler(ident, msg)
             
         # Check whether we should exit, in case the incoming message set the
         # exit flag on
         if self.shell.exit_now:
             self.log.debug('\nExiting IPython kernel...')
             # We do a normal, clean exit, which allows any actions registered
             # via atexit (such as history saving) to take place.
             sys.exit(0)
 
 
     def start(self):
         """ Start the kernel main loop.
         """
         poller = zmq.Poller()
         poller.register(self.shell_socket, zmq.POLLIN)
         while True:
             try:
                 # scale by extra factor of 10, because there is no
                 # reason for this to be anything less than ~ 0.1s
                 # since it is a real poller and will respond
                 # to events immediately
                 poller.poll(10*1000*self._poll_interval)
                 self.do_one_iteration()
             except KeyboardInterrupt:
                 # Ctrl-C shouldn't crash the kernel
                 io.raw_print("KeyboardInterrupt caught in kernel")
 
     def record_ports(self, ports):
         """Record the ports that this kernel is using.
 
         The creator of the Kernel instance must call this methods if they
         want the :meth:`connect_request` method to return the port numbers.
         """
         self._recorded_ports = ports
 
     #---------------------------------------------------------------------------
     # Kernel request handlers
     #---------------------------------------------------------------------------
 
     def _publish_pyin(self, code, parent):
         """Publish the code request on the pyin stream."""
 
         pyin_msg = self.session.send(self.iopub_socket, u'pyin',{u'code':code}, parent=parent)
 
     def execute_request(self, ident, parent):
         
         status_msg = self.session.send(self.iopub_socket,
             u'status',
             {u'execution_state':u'busy'},
             parent=parent
         )
         
         try:
             content = parent[u'content']
             code = content[u'code']
             silent = content[u'silent'] 
         except:
             self.log.error("Got bad msg: ")
             self.log.error(str(Message(parent)))
             return
 
         shell = self.shell # we'll need this a lot here
 
         # Replace raw_input. Note that is not sufficient to replace 
         # raw_input in the user namespace.
         raw_input = lambda prompt='': self._raw_input(prompt, ident, parent)
         __builtin__.raw_input = raw_input
 
         # Set the parent message of the display hook and out streams.
         shell.displayhook.set_parent(parent)
         shell.display_pub.set_parent(parent)
         sys.stdout.set_parent(parent)
         sys.stderr.set_parent(parent)
 
         # Re-broadcast our input for the benefit of listening clients, and
         # start computing output
         if not silent:
             self._publish_pyin(code, parent)
 
         reply_content = {}
         try:
             if silent:
                 # run_code uses 'exec' mode, so no displayhook will fire, and it
                 # doesn't call logging or history manipulations.  Print
                 # statements in that code will obviously still execute.
                 shell.run_code(code)
             else:
                 # FIXME: the shell calls the exception handler itself.
                 shell.run_cell(code)
         except:
             status = u'error'
             # FIXME: this code right now isn't being used yet by default,
             # because the run_cell() call above directly fires off exception
             # reporting.  This code, therefore, is only active in the scenario
             # where runlines itself has an unhandled exception.  We need to
             # uniformize this, for all exception construction to come from a
             # single location in the codbase.
             etype, evalue, tb = sys.exc_info()
             tb_list = traceback.format_exception(etype, evalue, tb)
             reply_content.update(shell._showtraceback(etype, evalue, tb_list))
         else:
             status = u'ok'
 
         reply_content[u'status'] = status
         
         # Return the execution counter so clients can display prompts
         reply_content['execution_count'] = shell.execution_count -1
 
         # FIXME - fish exception info out of shell, possibly left there by
         # runlines.  We'll need to clean up this logic later.
         if shell._reply_content is not None:
             reply_content.update(shell._reply_content)
             # reset after use
             shell._reply_content = None
 
         # At this point, we can tell whether the main code execution succeeded
         # or not.  If it did, we proceed to evaluate user_variables/expressions
         if reply_content['status'] == 'ok':
             reply_content[u'user_variables'] = \
                          shell.user_variables(content[u'user_variables'])
             reply_content[u'user_expressions'] = \
                          shell.user_expressions(content[u'user_expressions'])
         else:
             # If there was an error, don't even try to compute variables or
             # expressions
             reply_content[u'user_variables'] = {}
             reply_content[u'user_expressions'] = {}
 
         # Payloads should be retrieved regardless of outcome, so we can both
         # recover partial output (that could have been generated early in a
         # block, before an error) and clear the payload system always.
         reply_content[u'payload'] = shell.payload_manager.read_payload()
         # Be agressive about clearing the payload because we don't want
         # it to sit in memory until the next execute_request comes in.
         shell.payload_manager.clear_payload()
 
         # Flush output before sending the reply.
         sys.stdout.flush()
         sys.stderr.flush()
         # FIXME: on rare occasions, the flush doesn't seem to make it to the
         # clients... This seems to mitigate the problem, but we definitely need
         # to better understand what's going on.
         if self._execute_sleep:
             time.sleep(self._execute_sleep)
         
         # Send the reply.
         reply_msg = self.session.send(self.shell_socket, u'execute_reply',
                                       reply_content, parent, ident=ident)
         self.log.debug(str(reply_msg))
 
         if reply_msg['content']['status'] == u'error':
             self._abort_queue()
 
         status_msg = self.session.send(self.iopub_socket,
             u'status',
             {u'execution_state':u'idle'},
             parent=parent
         )
 
     def complete_request(self, ident, parent):
         txt, matches = self._complete(parent)
         matches = {'matches' : matches,
                    'matched_text' : txt,
                    'status' : 'ok'}
         completion_msg = self.session.send(self.shell_socket, 'complete_reply',
                                            matches, parent, ident)
         self.log.debug(str(completion_msg))
 
     def object_info_request(self, ident, parent):
         object_info = self.shell.object_inspect(parent['content']['oname'])
         # Before we send this object over, we scrub it for JSON usage
         oinfo = json_clean(object_info)
         msg = self.session.send(self.shell_socket, 'object_info_reply',
                                 oinfo, parent, ident)
         self.log.debug(msg)
 
     def history_request(self, ident, parent):
         # We need to pull these out, as passing **kwargs doesn't work with
         # unicode keys before Python 2.6.5.
         hist_access_type = parent['content']['hist_access_type']
         raw = parent['content']['raw']
         output = parent['content']['output']
         if hist_access_type == 'tail':
             n = parent['content']['n']
             hist = self.shell.history_manager.get_tail(n, raw=raw, output=output,
                                                             include_latest=True)
         
         elif hist_access_type == 'range':
             session = parent['content']['session']
             start = parent['content']['start']
             stop = parent['content']['stop']
             hist = self.shell.history_manager.get_range(session, start, stop,
                                                         raw=raw, output=output)
         
         elif hist_access_type == 'search':
             pattern = parent['content']['pattern']
             hist = self.shell.history_manager.search(pattern, raw=raw, output=output)
         
         else:
             hist = []
         content = {'history' : list(hist)}
         msg = self.session.send(self.shell_socket, 'history_reply',
                                 content, parent, ident)
         self.log.debug(str(msg))
 
     def connect_request(self, ident, parent):
         if self._recorded_ports is not None:
             content = self._recorded_ports.copy()
         else:
             content = {}
         msg = self.session.send(self.shell_socket, 'connect_reply',
                                 content, parent, ident)
         self.log.debug(msg)
 
     def shutdown_request(self, ident, parent):
         self.shell.exit_now = True
         self._shutdown_message = self.session.msg(u'shutdown_reply', parent['content'], parent)
         sys.exit(0)
 
     #---------------------------------------------------------------------------
     # Protected interface
     #---------------------------------------------------------------------------
 
     def _abort_queue(self):
         while True:
             ident,msg = self.session.recv(self.shell_socket, zmq.NOBLOCK)
             if msg is None:
                 break
             else:
                 assert ident is not None, \
                        "Unexpected missing message part."
 
             self.log.debug("Aborting:\n"+str(Message(msg)))
-            msg_type = msg['msg_type']
+            msg_type = msg['header']['msg_type']
             reply_type = msg_type.split('_')[0] + '_reply'
             reply_msg = self.session.send(self.shell_socket, reply_type,
                     {'status' : 'aborted'}, msg, ident=ident)
             self.log.debug(reply_msg)
             # We need to wait a bit for requests to come in. This can probably
             # be set shorter for true asynchronous clients.
             time.sleep(0.1)
 
     def _raw_input(self, prompt, ident, parent):
         # Flush output before making the request.
         sys.stderr.flush()
         sys.stdout.flush()
 
         # Send the input request.
         content = dict(prompt=prompt)
         msg = self.session.send(self.stdin_socket, u'input_request', content, parent)
 
         # Await a response.
         ident, reply = self.session.recv(self.stdin_socket, 0)
         try:
             value = reply['content']['value']
         except:
             self.log.error("Got bad raw_input reply: ")
             self.log.error(str(Message(parent)))
             value = ''
         return value
     
     def _complete(self, msg):
         c = msg['content']
         try:
             cpos = int(c['cursor_pos'])
         except:
             # If we don't get something that we can convert to an integer, at
             # least attempt the completion guessing the cursor is at the end of
             # the text, if there's any, and otherwise of the line
             cpos = len(c['text'])
             if cpos==0:
                 cpos = len(c['line'])
         return self.shell.complete(c['text'], c['line'], cpos)
 
     def _object_info(self, context):
         symbol, leftover = self._symbol_from_context(context)
         if symbol is not None and not leftover:
             doc = getattr(symbol, '__doc__', '')
         else:
             doc = ''
         object_info = dict(docstring = doc)
         return object_info
 
     def _symbol_from_context(self, context):
         if not context:
             return None, context
 
         base_symbol_string = context[0]
         symbol = self.shell.user_ns.get(base_symbol_string, None)
         if symbol is None:
             symbol = __builtin__.__dict__.get(base_symbol_string, None)
         if symbol is None:
             return None, context
 
         context = context[1:]
         for i, name in enumerate(context):
             new_symbol = getattr(symbol, name, None)
             if new_symbol is None:
                 return symbol, context[i:]
             else:
                 symbol = new_symbol
 
         return symbol, []
 
     def _at_shutdown(self):
         """Actions taken at shutdown by the kernel, called by python's atexit.
         """
         # io.rprint("Kernel at_shutdown") # dbg
         if self._shutdown_message is not None:
             self.session.send(self.shell_socket, self._shutdown_message)
             self.session.send(self.iopub_socket, self._shutdown_message)
             self.log.debug(str(self._shutdown_message))
             # A very short sleep to give zmq time to flush its message buffers
             # before Python truly shuts down.
             time.sleep(0.01)
 
 
 class QtKernel(Kernel):
     """A Kernel subclass with Qt support."""
 
     def start(self):
         """Start a kernel with QtPy4 event loop integration."""
 
         from IPython.external.qt_for_kernel import QtCore
         from IPython.lib.guisupport import get_app_qt4, start_event_loop_qt4
 
         self.app = get_app_qt4([" "])
         self.app.setQuitOnLastWindowClosed(False)
         self.timer = QtCore.QTimer()
         self.timer.timeout.connect(self.do_one_iteration)
         # Units for the timer are in milliseconds
         self.timer.start(1000*self._poll_interval)
         start_event_loop_qt4(self.app)
 
 
 class WxKernel(Kernel):
     """A Kernel subclass with Wx support."""
 
     def start(self):
         """Start a kernel with wx event loop support."""
 
         import wx
         from IPython.lib.guisupport import start_event_loop_wx
 
         doi = self.do_one_iteration
          # Wx uses milliseconds
         poll_interval = int(1000*self._poll_interval)
 
         # We have to put the wx.Timer in a wx.Frame for it to fire properly.
         # We make the Frame hidden when we create it in the main app below.
         class TimerFrame(wx.Frame):
             def __init__(self, func):
                 wx.Frame.__init__(self, None, -1)
                 self.timer = wx.Timer(self)
                 # Units for the timer are in milliseconds
                 self.timer.Start(poll_interval)
                 self.Bind(wx.EVT_TIMER, self.on_timer)
                 self.func = func
 
             def on_timer(self, event):
                 self.func()
 
         # We need a custom wx.App to create our Frame subclass that has the
         # wx.Timer to drive the ZMQ event loop.
         class IPWxApp(wx.App):
             def OnInit(self):
                 self.frame = TimerFrame(doi)
                 self.frame.Show(False)
                 return True
 
         # The redirect=False here makes sure that wx doesn't replace
         # sys.stdout/stderr with its own classes.
         self.app = IPWxApp(redirect=False)
         start_event_loop_wx(self.app)
 
 
 class TkKernel(Kernel):
     """A Kernel subclass with Tk support."""
 
     def start(self):
         """Start a Tk enabled event loop."""
 
         import Tkinter
         doi = self.do_one_iteration
         # Tk uses milliseconds
         poll_interval = int(1000*self._poll_interval)
         # For Tkinter, we create a Tk object and call its withdraw method.
         class Timer(object):
             def __init__(self, func):
                 self.app = Tkinter.Tk()
                 self.app.withdraw()
                 self.func = func
                 
             def on_timer(self):
                 self.func()
                 self.app.after(poll_interval, self.on_timer)
                 
             def start(self):
                 self.on_timer()  # Call it once to get things going.
                 self.app.mainloop()
 
         self.timer = Timer(doi)
         self.timer.start()
 
 
 class GTKKernel(Kernel):
     """A Kernel subclass with GTK support."""
     
     def start(self):
         """Start the kernel, coordinating with the GTK event loop"""
         from .gui.gtkembed import GTKEmbed
         
         gtk_kernel = GTKEmbed(self)
         gtk_kernel.start()
 
 
 #-----------------------------------------------------------------------------
 # Aliases and Flags for the IPKernelApp
 #-----------------------------------------------------------------------------
 
 flags = dict(kernel_flags)
 flags.update(shell_flags)
 
 addflag = lambda *args: flags.update(boolean_flag(*args))
 
 flags['pylab'] = (
     {'IPKernelApp' : {'pylab' : 'auto'}},
     """Pre-load matplotlib and numpy for interactive use with
     the default matplotlib backend."""
 )
 
 aliases = dict(kernel_aliases)
 aliases.update(shell_aliases)
 
 # it's possible we don't want short aliases for *all* of these:
 aliases.update(dict(
     pylab='IPKernelApp.pylab',
 ))
 
 #-----------------------------------------------------------------------------
 # The IPKernelApp class
 #-----------------------------------------------------------------------------
 
 class IPKernelApp(KernelApp, InteractiveShellApp):
     name = 'ipkernel'
 
     aliases = Dict(aliases)
     flags = Dict(flags)
     classes = [Kernel, ZMQInteractiveShell, ProfileDir, Session]
     # configurables
     pylab = CaselessStrEnum(['tk', 'qt', 'wx', 'gtk', 'osx', 'inline', 'auto'],
         config=True,
         help="""Pre-load matplotlib and numpy for interactive use,
         selecting a particular matplotlib backend and loop integration.
         """
     )
     pylab_import_all = Bool(True, config=True,
         help="""If true, an 'import *' is done from numpy and pylab,
         when using pylab"""
     )
     def initialize(self, argv=None):
         super(IPKernelApp, self).initialize(argv)
         self.init_shell()
         self.init_extensions()
         self.init_code()
 
     def init_kernel(self):
         kernel_factory = Kernel
 
         kernel_map = {
             'qt' : QtKernel,
             'qt4': QtKernel,
             'inline': Kernel,
             'osx': TkKernel,
             'wx' : WxKernel,
             'tk' : TkKernel,
             'gtk': GTKKernel,
         }
 
         if self.pylab:
             key = None if self.pylab == 'auto' else self.pylab
             gui, backend = pylabtools.find_gui_and_backend(key)
             kernel_factory = kernel_map.get(gui)
             if kernel_factory is None:
                 raise ValueError('GUI is not supported: %r' % gui)
             pylabtools.activate_matplotlib(backend)
 
         kernel = kernel_factory(config=self.config, session=self.session,
                                 shell_socket=self.shell_socket,
                                 iopub_socket=self.iopub_socket,
                                 stdin_socket=self.stdin_socket,
                                 log=self.log
         )
         self.kernel = kernel
         kernel.record_ports(self.ports)
 
         if self.pylab:
             import_all = self.pylab_import_all
             pylabtools.import_pylab(kernel.shell.user_ns, backend, import_all,
                                     shell=kernel.shell)
     
     def init_shell(self):
         self.shell = self.kernel.shell
 
 
 #-----------------------------------------------------------------------------
 # Kernel main and launch functions
 #-----------------------------------------------------------------------------
 
 def launch_kernel(*args, **kwargs):
     """Launches a localhost IPython kernel, binding to the specified ports.
 
     This function simply calls entry_point.base_launch_kernel with the right first
     command to start an ipkernel.  See base_launch_kernel for arguments.
 
     Returns
     -------
     A tuple of form:
         (kernel_process, shell_port, iopub_port, stdin_port, hb_port)
     where kernel_process is a Popen object and the ports are integers.
     """
     return base_launch_kernel('from IPython.zmq.ipkernel import main; main()',
                               *args, **kwargs)
 
 
 def main():
     """Run an IPKernel as an application"""
     app = IPKernelApp.instance()
     app.initialize()
     app.start()
 
 
 if __name__ == '__main__':
     main()

IPython/zmq/pykernel.py

@@ -1,278 +1,278 b''
 #!/usr/bin/env python
 """A simple interactive kernel that talks to a frontend over 0MQ.
 
 Things to do:
 
 * Implement `set_parent` logic. Right before doing exec, the Kernel should
   call set_parent on all the PUB objects with the message about to be executed.
 * Implement random port and security key logic.
 * Implement control messages.
 * Implement event loop and poll version.
 """
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 # Standard library imports.
 import __builtin__
 from code import CommandCompiler
 import sys
 import time
 import traceback
 
 # System library imports.
 import zmq
 
 # Local imports.
 from IPython.utils.traitlets import HasTraits, Instance, Dict, Float
 from completer import KernelCompleter
 from entry_point import base_launch_kernel
 from session import Session, Message
 from kernelapp import KernelApp
 
 #-----------------------------------------------------------------------------
 # Main kernel class
 #-----------------------------------------------------------------------------
 
 class Kernel(HasTraits):
 
     # Private interface
 
     # Time to sleep after flushing the stdout/err buffers in each execute
     # cycle.  While this introduces a hard limit on the minimal latency of the
     # execute cycle, it helps prevent output synchronization problems for
     # clients.
     # Units are in seconds.  The minimum zmq latency on local host is probably
     # ~150 microseconds, set this to 500us for now.  We may need to increase it
     # a little if it's not enough after more interactive testing.
     _execute_sleep = Float(0.0005, config=True)
 
     # This is a dict of port number that the kernel is listening on. It is set
     # by record_ports and used by connect_request.
     _recorded_ports = Dict()
 
     #---------------------------------------------------------------------------
     # Kernel interface
     #---------------------------------------------------------------------------
 
     session = Instance(Session)
     shell_socket = Instance('zmq.Socket')
     iopub_socket = Instance('zmq.Socket')
     stdin_socket = Instance('zmq.Socket')
     log = Instance('logging.Logger')
 
     def __init__(self, **kwargs):
         super(Kernel, self).__init__(**kwargs)
         self.user_ns = {}
         self.history = []
         self.compiler = CommandCompiler()
         self.completer = KernelCompleter(self.user_ns)
 
         # Build dict of handlers for message types
         msg_types = [ 'execute_request', 'complete_request', 
                       'object_info_request', 'shutdown_request' ]
         self.handlers = {}
         for msg_type in msg_types:
             self.handlers[msg_type] = getattr(self, msg_type)
 
     def start(self):
         """ Start the kernel main loop.
         """
         while True:
             ident,msg = self.session.recv(self.shell_socket,0)
             assert ident is not None, "Missing message part."
             omsg = Message(msg)
             self.log.debug(str(omsg))
             handler = self.handlers.get(omsg.msg_type, None)
             if handler is None:
                 self.log.error("UNKNOWN MESSAGE TYPE: %s"%omsg)
             else:
                 handler(ident, omsg)
 
     def record_ports(self, ports):
         """Record the ports that this kernel is using.
 
         The creator of the Kernel instance must call this methods if they
         want the :meth:`connect_request` method to return the port numbers.
         """
         self._recorded_ports = ports
 
     #---------------------------------------------------------------------------
     # Kernel request handlers
     #---------------------------------------------------------------------------
 
     def execute_request(self, ident, parent):
         try:
             code = parent[u'content'][u'code']
         except:
             self.log.error("Got bad msg: %s"%Message(parent))
             return
         pyin_msg = self.session.send(self.iopub_socket, u'pyin',{u'code':code}, parent=parent)
 
         try:
             comp_code = self.compiler(code, '<zmq-kernel>')
 
             # Replace raw_input. Note that is not sufficient to replace 
             # raw_input in the user namespace.
             raw_input = lambda prompt='': self._raw_input(prompt, ident, parent)
             __builtin__.raw_input = raw_input
 
             # Set the parent message of the display hook and out streams.
             sys.displayhook.set_parent(parent)
             sys.stdout.set_parent(parent)
             sys.stderr.set_parent(parent)
 
             exec comp_code in self.user_ns, self.user_ns
         except:
             etype, evalue, tb = sys.exc_info()
             tb = traceback.format_exception(etype, evalue, tb)
             exc_content = {
                 u'status' : u'error',
                 u'traceback' : tb,
                 u'ename' : unicode(etype.__name__),
                 u'evalue' : unicode(evalue)
             }
             exc_msg = self.session.send(self.iopub_socket, u'pyerr', exc_content, parent)
             reply_content = exc_content
         else:
             reply_content = { 'status' : 'ok', 'payload' : {} }
             
         # Flush output before sending the reply.
         sys.stderr.flush()
         sys.stdout.flush()
         # FIXME: on rare occasions, the flush doesn't seem to make it to the
         # clients... This seems to mitigate the problem, but we definitely need
         # to better understand what's going on.
         if self._execute_sleep:
             time.sleep(self._execute_sleep)
 
         # Send the reply.
         reply_msg = self.session.send(self.shell_socket, u'execute_reply', reply_content, parent, ident=ident)
         self.log.debug(Message(reply_msg))
         if reply_msg['content']['status'] == u'error':
             self._abort_queue()
 
     def complete_request(self, ident, parent):
         matches = {'matches' : self._complete(parent),
                    'status' : 'ok'}
         completion_msg = self.session.send(self.shell_socket, 'complete_reply',
                                            matches, parent, ident)
         self.log.debug(completion_msg)
 
     def object_info_request(self, ident, parent):
         context = parent['content']['oname'].split('.')
         object_info = self._object_info(context)
         msg = self.session.send(self.shell_socket, 'object_info_reply',
                                 object_info, parent, ident)
         self.log.debug(msg)
 
     def shutdown_request(self, ident, parent):
         content = dict(parent['content'])
         msg = self.session.send(self.shell_socket, 'shutdown_reply',
                                 content, parent, ident)
         msg = self.session.send(self.iopub_socket, 'shutdown_reply',
                                 content, parent, ident)
         self.log.debug(msg)
         time.sleep(0.1)
         sys.exit(0)
 
     #---------------------------------------------------------------------------
     # Protected interface
     #---------------------------------------------------------------------------
 
     def _abort_queue(self):
         while True:
             ident,msg = self.session.recv(self.shell_socket, zmq.NOBLOCK)
             if msg is None:
                 # msg=None on EAGAIN
                 break
             else:
                 assert ident is not None, "Missing message part."
             self.log.debug("Aborting: %s"%Message(msg))
-            msg_type = msg['msg_type']
+            msg_type = msg['header']['msg_type']
             reply_type = msg_type.split('_')[0] + '_reply'
             reply_msg = self.session.send(self.shell_socket, reply_type, {'status':'aborted'}, msg, ident=ident)
             self.log.debug(Message(reply_msg))
             # We need to wait a bit for requests to come in. This can probably
             # be set shorter for true asynchronous clients.
             time.sleep(0.1)
 
     def _raw_input(self, prompt, ident, parent):
         # Flush output before making the request.
         sys.stderr.flush()
         sys.stdout.flush()
 
         # Send the input request.
         content = dict(prompt=prompt)
         msg = self.session.send(self.stdin_socket, u'input_request', content, parent)
 
         # Await a response.
         ident,reply = self.session.recv(self.stdin_socket, 0)
         try:
             value = reply['content']['value']
         except:
             self.log.error("Got bad raw_input reply: %s"%Message(parent))
             value = ''
         return value
 
     def _complete(self, msg):
         return self.completer.complete(msg.content.line, msg.content.text)
 
     def _object_info(self, context):
         symbol, leftover = self._symbol_from_context(context)
         if symbol is not None and not leftover:
             doc = getattr(symbol, '__doc__', '')
         else:
             doc = ''
         object_info = dict(docstring = doc)
         return object_info
 
     def _symbol_from_context(self, context):
         if not context:
             return None, context
 
         base_symbol_string = context[0]
         symbol = self.user_ns.get(base_symbol_string, None)
         if symbol is None:
             symbol = __builtin__.__dict__.get(base_symbol_string, None)
         if symbol is None:
             return None, context
 
         context = context[1:]
         for i, name in enumerate(context):
             new_symbol = getattr(symbol, name, None)
             if new_symbol is None:
                 return symbol, context[i:]
             else:
                 symbol = new_symbol
 
         return symbol, []
 
 #-----------------------------------------------------------------------------
 # Kernel main and launch functions
 #-----------------------------------------------------------------------------
 
 def launch_kernel(*args, **kwargs):
     """ Launches a simple Python kernel, binding to the specified ports.
     
     This function simply calls entry_point.base_launch_kernel with the right first
     command to start a pykernel.  See base_launch_kernel for arguments.
 
     Returns
     -------
     A tuple of form:
         (kernel_process, xrep_port, pub_port, req_port, hb_port)
     where kernel_process is a Popen object and the ports are integers.
     """
     return base_launch_kernel('from IPython.zmq.pykernel import main; main()',
                                 *args, **kwargs)
 
 def main():
     """Run a PyKernel as an application"""
     app = KernelApp.instance()
     app.initialize()
     app.start()
 
 if __name__ == '__main__':
     main()

IPython/zmq/session.py

@@ -1,679 +1,697 b''
 #!/usr/bin/env python
 """Session object for building, serializing, sending, and receiving messages in
 IPython. The Session object supports serialization, HMAC signatures, and
 metadata on messages.
 
 Also defined here are utilities for working with Sessions:
 * A SessionFactory to be used as a base class for configurables that work with
 Sessions.
 * A Message object for convenience that allows attribute-access to the msg dict.
 
 Authors:
 
 * Min RK
 * Brian Granger
 * Fernando Perez
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import hmac
 import logging
 import os
 import pprint
 import uuid
 from datetime import datetime
 
 try:
     import cPickle
     pickle = cPickle
 except:
     cPickle = None
     import pickle
 
 import zmq
 from zmq.utils import jsonapi
 from zmq.eventloop.ioloop import IOLoop
 from zmq.eventloop.zmqstream import ZMQStream
 
 from IPython.config.configurable import Configurable, LoggingConfigurable
 from IPython.utils.importstring import import_item
 from IPython.utils.jsonutil import extract_dates, squash_dates, date_default
 from IPython.utils.traitlets import (CBytes, Unicode, Bool, Any, Instance, Set,
                                         DottedObjectName)
 
 #-----------------------------------------------------------------------------
 # utility functions
 #-----------------------------------------------------------------------------
 
 def squash_unicode(obj):
     """coerce unicode back to bytestrings."""
     if isinstance(obj,dict):
         for key in obj.keys():
             obj[key] = squash_unicode(obj[key])
             if isinstance(key, unicode):
                 obj[squash_unicode(key)] = obj.pop(key)
     elif isinstance(obj, list):
         for i,v in enumerate(obj):
             obj[i] = squash_unicode(v)
     elif isinstance(obj, unicode):
         obj = obj.encode('utf8')
     return obj
 
 #-----------------------------------------------------------------------------
 # globals and defaults
 #-----------------------------------------------------------------------------
 key = 'on_unknown' if jsonapi.jsonmod.__name__ == 'jsonlib' else 'default'
 json_packer = lambda obj: jsonapi.dumps(obj, **{key:date_default})
 json_unpacker = lambda s: extract_dates(jsonapi.loads(s))
 
 pickle_packer = lambda o: pickle.dumps(o,-1)
 pickle_unpacker = pickle.loads
 
 default_packer = json_packer
 default_unpacker = json_unpacker
 
 
 DELIM=b"<IDS|MSG>"
 
 #-----------------------------------------------------------------------------
 # Classes
 #-----------------------------------------------------------------------------
 
 class SessionFactory(LoggingConfigurable):
     """The Base class for configurables that have a Session, Context, logger,
     and IOLoop.
     """
     
     logname = Unicode('')
     def _logname_changed(self, name, old, new):
         self.log = logging.getLogger(new)
     
     # not configurable:
     context = Instance('zmq.Context')
     def _context_default(self):
         return zmq.Context.instance()
     
     session = Instance('IPython.zmq.session.Session')
     
     loop = Instance('zmq.eventloop.ioloop.IOLoop', allow_none=False)
     def _loop_default(self):
         return IOLoop.instance()
     
     def __init__(self, **kwargs):
         super(SessionFactory, self).__init__(**kwargs)
         
         if self.session is None:
             # construct the session
             self.session = Session(**kwargs)
     
 
 class Message(object):
     """A simple message object that maps dict keys to attributes.
 
     A Message can be created from a dict and a dict from a Message instance
     simply by calling dict(msg_obj)."""
     
     def __init__(self, msg_dict):
         dct = self.__dict__
         for k, v in dict(msg_dict).iteritems():
             if isinstance(v, dict):
                 v = Message(v)
             dct[k] = v
 
     # Having this iterator lets dict(msg_obj) work out of the box.
     def __iter__(self):
         return iter(self.__dict__.iteritems())
     
     def __repr__(self):
         return repr(self.__dict__)
 
     def __str__(self):
         return pprint.pformat(self.__dict__)
 
     def __contains__(self, k):
         return k in self.__dict__
 
     def __getitem__(self, k):
         return self.__dict__[k]
 
 
 def msg_header(msg_id, msg_type, username, session):
     date = datetime.now()
     return locals()
 
 def extract_header(msg_or_header):
     """Given a message or header, return the header."""
     if not msg_or_header:
         return {}
     try:
         # See if msg_or_header is the entire message.
         h = msg_or_header['header']
     except KeyError:
         try:
             # See if msg_or_header is just the header
             h = msg_or_header['msg_id']
         except KeyError:
             raise
         else:
             h = msg_or_header
     if not isinstance(h, dict):
         h = dict(h)
     return h
 
 class Session(Configurable):
     """Object for handling serialization and sending of messages.
     
     The Session object handles building messages and sending them
     with ZMQ sockets or ZMQStream objects.  Objects can communicate with each
     other over the network via Session objects, and only need to work with the
     dict-based IPython message spec. The Session will handle 
     serialization/deserialization, security, and metadata.
     
     Sessions support configurable serialiization via packer/unpacker traits,
     and signing with HMAC digests via the key/keyfile traits.
     
     Parameters
     ----------
     
     debug : bool
         whether to trigger extra debugging statements
     packer/unpacker : str : 'json', 'pickle' or import_string
         importstrings for methods to serialize message parts.  If just
         'json' or 'pickle', predefined JSON and pickle packers will be used.
         Otherwise, the entire importstring must be used.
         
         The functions must accept at least valid JSON input, and output *bytes*.
         
         For example, to use msgpack:
         packer = 'msgpack.packb', unpacker='msgpack.unpackb'
     pack/unpack : callables
         You can also set the pack/unpack callables for serialization directly.
     session : bytes
         the ID of this Session object.  The default is to generate a new UUID.
     username : unicode
         username added to message headers.  The default is to ask the OS.
     key : bytes
         The key used to initialize an HMAC signature.  If unset, messages
         will not be signed or checked.
     keyfile : filepath
         The file containing a key.  If this is set, `key` will be initialized
         to the contents of the file.
     
     """
     
     debug=Bool(False, config=True, help="""Debug output in the Session""")
     
     packer = DottedObjectName('json',config=True,
             help="""The name of the packer for serializing messages.
             Should be one of 'json', 'pickle', or an import name
             for a custom callable serializer.""")
     def _packer_changed(self, name, old, new):
         if new.lower() == 'json':
             self.pack = json_packer
             self.unpack = json_unpacker
         elif new.lower() == 'pickle':
             self.pack = pickle_packer
             self.unpack = pickle_unpacker
         else:
             self.pack = import_item(str(new))
 
     unpacker = DottedObjectName('json', config=True,
         help="""The name of the unpacker for unserializing messages.
         Only used with custom functions for `packer`.""")
     def _unpacker_changed(self, name, old, new):
         if new.lower() == 'json':
             self.pack = json_packer
             self.unpack = json_unpacker
         elif new.lower() == 'pickle':
             self.pack = pickle_packer
             self.unpack = pickle_unpacker
         else:
             self.unpack = import_item(str(new))
         
     session = CBytes(b'', config=True,
         help="""The UUID identifying this session.""")
     def _session_default(self):
         return bytes(uuid.uuid4())
     
-    username = Unicode(os.environ.get('USER','username'), config=True,
+    username = Unicode(os.environ.get('USER',u'username'), config=True,
         help="""Username for the Session. Default is your system username.""")
     
     # message signature related traits:
     key = CBytes(b'', config=True,
         help="""execution key, for extra authentication.""")
     def _key_changed(self, name, old, new):
         if new:
             self.auth = hmac.HMAC(new)
         else:
             self.auth = None
     auth = Instance(hmac.HMAC)
     digest_history = Set()
     
     keyfile = Unicode('', config=True,
         help="""path to file containing execution key.""")
     def _keyfile_changed(self, name, old, new):
         with open(new, 'rb') as f:
             self.key = f.read().strip()
 
     pack = Any(default_packer) # the actual packer function
     def _pack_changed(self, name, old, new):
         if not callable(new):
             raise TypeError("packer must be callable, not %s"%type(new))
     
     unpack = Any(default_unpacker) # the actual packer function
     def _unpack_changed(self, name, old, new):
         # unpacker is not checked - it is assumed to be 
         if not callable(new):
             raise TypeError("unpacker must be callable, not %s"%type(new))
 
     def __init__(self, **kwargs):
         """create a Session object
         
         Parameters
         ----------
 
         debug : bool
             whether to trigger extra debugging statements
         packer/unpacker : str : 'json', 'pickle' or import_string
             importstrings for methods to serialize message parts.  If just
             'json' or 'pickle', predefined JSON and pickle packers will be used.
             Otherwise, the entire importstring must be used.
 
             The functions must accept at least valid JSON input, and output
             *bytes*.
 
             For example, to use msgpack:
             packer = 'msgpack.packb', unpacker='msgpack.unpackb'
         pack/unpack : callables
             You can also set the pack/unpack callables for serialization
             directly.
         session : bytes
             the ID of this Session object.  The default is to generate a new 
             UUID.
         username : unicode
             username added to message headers.  The default is to ask the OS.
         key : bytes
             The key used to initialize an HMAC signature.  If unset, messages
             will not be signed or checked.
         keyfile : filepath
             The file containing a key.  If this is set, `key` will be 
             initialized to the contents of the file.
         """
         super(Session, self).__init__(**kwargs)
         self._check_packers()
         self.none = self.pack({})
 
     @property
     def msg_id(self):
         """always return new uuid"""
         return str(uuid.uuid4())
 
     def _check_packers(self):
         """check packers for binary data and datetime support."""
         pack = self.pack
         unpack = self.unpack
         
         # check simple serialization
         msg = dict(a=[1,'hi'])
         try:
             packed = pack(msg)
         except Exception:
             raise ValueError("packer could not serialize a simple message")
         
         # ensure packed message is bytes
         if not isinstance(packed, bytes):
             raise ValueError("message packed to %r, but bytes are required"%type(packed))
         
         # check that unpack is pack's inverse
         try:
             unpacked = unpack(packed)
         except Exception:
             raise ValueError("unpacker could not handle the packer's output")
         
         # check datetime support
         msg = dict(t=datetime.now())
         try:
             unpacked = unpack(pack(msg))
         except Exception:
             self.pack = lambda o: pack(squash_dates(o))
             self.unpack = lambda s: extract_dates(unpack(s))
     
     def msg_header(self, msg_type):
         return msg_header(self.msg_id, msg_type, self.username, self.session)
 
-    def msg(self, msg_type, content=None, parent=None, subheader=None):
+    def msg(self, msg_type, content=None, parent=None, subheader=None, header=None):
         """Return the nested message dict.
 
         This format is different from what is sent over the wire. The
-        self.serialize method converts this nested message dict to the wire
-        format, which uses a message list.
+        serialize/unserialize methods converts this nested message dict to the wire
+        format, which is a list of message parts.
         """
         msg = {}
-        msg['header'] = self.msg_header(msg_type)
-        msg['msg_id'] = msg['header']['msg_id']
+        msg['header'] = self.msg_header(msg_type) if header is None else header
         msg['parent_header'] = {} if parent is None else extract_header(parent)
-        msg['msg_type'] = msg_type
         msg['content'] = {} if content is None else content
         sub = {} if subheader is None else subheader
         msg['header'].update(sub)
         return msg
 
     def sign(self, msg_list):
         """Sign a message with HMAC digest. If no auth, return b''.
 
         Parameters
         ----------
         msg_list : list
             The [p_header,p_parent,p_content] part of the message list. 
         """
         if self.auth is None:
             return b''
         h = self.auth.copy()
         for m in msg_list:
             h.update(m)
         return h.hexdigest()
     
     def serialize(self, msg, ident=None):
         """Serialize the message components to bytes.
 
+        This is roughly the inverse of unserialize. The serialize/unserialize
+        methods work with full message lists, whereas pack/unpack work with
+        the individual message parts in the message list.
+
         Parameters
         ----------
         msg : dict or Message
             The nexted message dict as returned by the self.msg method.
 
         Returns
         -------
         msg_list : list
             The list of bytes objects to be sent with the format:
             [ident1,ident2,...,DELIM,HMAC,p_header,p_parent,p_content,
              buffer1,buffer2,...]. In this list, the p_* entities are
             the packed or serialized versions, so if JSON is used, these
             are uft8 encoded JSON strings.
         """
         content = msg.get('content', {})
         if content is None:
             content = self.none
         elif isinstance(content, dict):
             content = self.pack(content)
         elif isinstance(content, bytes):
             # content is already packed, as in a relayed message
             pass
         elif isinstance(content, unicode):
             # should be bytes, but JSON often spits out unicode
             content = content.encode('utf8')
         else:
             raise TypeError("Content incorrect type: %s"%type(content))
         
         real_message = [self.pack(msg['header']), 
                         self.pack(msg['parent_header']), 
                         content
         ]
         
         to_send = []
 
         if isinstance(ident, list):
             # accept list of idents
             to_send.extend(ident)
         elif ident is not None:
             to_send.append(ident)
         to_send.append(DELIM)
         
         signature = self.sign(real_message)
         to_send.append(signature)
         
         to_send.extend(real_message)
 
         return to_send
         
-    def send(self, stream, msg_or_type, content=None, parent=None, ident=None, 
-                                    buffers=None, subheader=None, track=False):
+    def send(self, stream, msg_or_type, content=None, parent=None, ident=None,
+             buffers=None, subheader=None, track=False, header=None):
         """Build and send a message via stream or socket.
 
         The message format used by this function internally is as follows:
 
         [ident1,ident2,...,DELIM,HMAC,p_header,p_parent,p_content,
          buffer1,buffer2,...] 
 
-        The self.serialize method converts the nested message dict into this
+        The serialize/unserialize methods convert the nested message dict into this
         format.
 
         Parameters
         ----------
         
         stream : zmq.Socket or ZMQStream
-            the socket-like object used to send the data
+            The socket-like object used to send the data.
         msg_or_type : str or Message/dict
             Normally, msg_or_type will be a msg_type unless a message is being 
-            sent more than once.
+            sent more than once. If a header is supplied, this can be set to
+            None and the msg_type will be pulled from the header.
         
         content : dict or None
-            the content of the message (ignored if msg_or_type is a message)
+            The content of the message (ignored if msg_or_type is a message).
+        header : dict or None
+            The header dict for the message (ignores if msg_to_type is a message).
         parent : Message or dict or None
-            the parent or parent header describing the parent of this message
+            The parent or parent header describing the parent of this message
+            (ignored if msg_or_type is a message).
         ident : bytes or list of bytes
-            the zmq.IDENTITY routing path
+            The zmq.IDENTITY routing path.
         subheader : dict or None
-            extra header keys for this message's header
+            Extra header keys for this message's header (ignored if msg_or_type
+            is a message).
         buffers : list or None
-            the already-serialized buffers to be appended to the message
+            The already-serialized buffers to be appended to the message.
         track : bool
-            whether to track.  Only for use with Sockets, 
-            because ZMQStream objects cannot track messages.
+            Whether to track.  Only for use with Sockets, because ZMQStream
+            objects cannot track messages.
         
         Returns
         -------
-        msg : message dict
-            the constructed message
-        (msg,tracker) : (message dict, MessageTracker)
+        msg : dict
+            The constructed message.
+        (msg,tracker) : (dict, MessageTracker)
             if track=True, then a 2-tuple will be returned, 
             the first element being the constructed
             message, and the second being the MessageTracker
             
         """
 
         if not isinstance(stream, (zmq.Socket, ZMQStream)):
             raise TypeError("stream must be Socket or ZMQStream, not %r"%type(stream))
         elif track and isinstance(stream, ZMQStream):
             raise TypeError("ZMQStream cannot track messages")
         
         if isinstance(msg_or_type, (Message, dict)):
-            # we got a Message, not a msg_type
-            # don't build a new Message
+            # We got a Message or message dict, not a msg_type so don't
+            # build a new Message.
             msg = msg_or_type
         else:
-            msg = self.msg(msg_or_type, content, parent, subheader)
-        
+            msg = self.msg(msg_or_type, content=content, parent=parent, 
+                           subheader=subheader, header=header)
+
         buffers = [] if buffers is None else buffers
         to_send = self.serialize(msg, ident)
         flag = 0
         if buffers:
             flag = zmq.SNDMORE
             _track = False
         else:
             _track=track
         if track:
             tracker = stream.send_multipart(to_send, flag, copy=False, track=_track)
         else:
             tracker = stream.send_multipart(to_send, flag, copy=False)
         for b in buffers[:-1]:
             stream.send(b, flag, copy=False)
         if buffers:
             if track:
                 tracker = stream.send(buffers[-1], copy=False, track=track)
             else:
                 tracker = stream.send(buffers[-1], copy=False)
 
         # omsg = Message(msg)
         if self.debug:
             pprint.pprint(msg)
             pprint.pprint(to_send)
             pprint.pprint(buffers)
         
         msg['tracker'] = tracker
         
         return msg
-    
+
     def send_raw(self, stream, msg_list, flags=0, copy=True, ident=None):
         """Send a raw message via ident path.
 
         This method is used to send a already serialized message.
 
         Parameters
         ----------
         stream : ZMQStream or Socket
             The ZMQ stream or socket to use for sending the message.
         msg_list : list
             The serialized list of messages to send. This only includes the
             [p_header,p_parent,p_content,buffer1,buffer2,...] portion of
             the message.
         ident : ident or list
             A single ident or a list of idents to use in sending.
         """
         to_send = []
         if isinstance(ident, bytes):
             ident = [ident]
         if ident is not None:
             to_send.extend(ident)
-            
+
         to_send.append(DELIM)
         to_send.append(self.sign(msg_list))
         to_send.extend(msg_list)
         stream.send_multipart(msg_list, flags, copy=copy)
     
     def recv(self, socket, mode=zmq.NOBLOCK, content=True, copy=True):
         """Receive and unpack a message.
 
         Parameters
         ----------
         socket : ZMQStream or Socket
             The socket or stream to use in receiving.
 
         Returns
         -------
         [idents], msg
             [idents] is a list of idents and msg is a nested message dict of
             same format as self.msg returns.
         """
         if isinstance(socket, ZMQStream):
             socket = socket.socket
         try:
             msg_list = socket.recv_multipart(mode)
         except zmq.ZMQError as e:
             if e.errno == zmq.EAGAIN:
                 # We can convert EAGAIN to None as we know in this case
                 # recv_multipart won't return None.
                 return None,None
             else:
                 raise
         # split multipart message into identity list and message dict
         # invalid large messages can cause very expensive string comparisons
         idents, msg_list = self.feed_identities(msg_list, copy)
         try:
-            return idents, self.unpack_message(msg_list, content=content, copy=copy)
+            return idents, self.unserialize(msg_list, content=content, copy=copy)
         except Exception as e:
             print (idents, msg_list)
             # TODO: handle it
             raise e
     
     def feed_identities(self, msg_list, copy=True):
         """Split the identities from the rest of the message.
 
         Feed until DELIM is reached, then return the prefix as idents and
         remainder as msg_list. This is easily broken by setting an IDENT to DELIM,
         but that would be silly.
         
         Parameters
         ----------
         msg_list : a list of Message or bytes objects
             The message to be split.
         copy : bool
             flag determining whether the arguments are bytes or Messages
         
         Returns
         -------
-        (idents,msg_list) : two lists
-            idents will always be a list of bytes - the indentity prefix
-            msg_list will be a list of bytes or Messages, unchanged from input
-            msg_list should be unpackable via self.unpack_message at this point.
+        (idents, msg_list) : two lists
+            idents will always be a list of bytes, each of which is a ZMQ
+            identity. msg_list will be a list of bytes or zmq.Messages of the
+            form [HMAC,p_header,p_parent,p_content,buffer1,buffer2,...] and
+            should be unpackable/unserializable via self.unserialize at this
+            point.
         """
         if copy:
             idx = msg_list.index(DELIM)
             return msg_list[:idx], msg_list[idx+1:]
         else:
             failed = True
             for idx,m in enumerate(msg_list):
                 if m.bytes == DELIM:
                     failed = False
                     break
             if failed:
                 raise ValueError("DELIM not in msg_list")
             idents, msg_list = msg_list[:idx], msg_list[idx+1:]
             return [m.bytes for m in idents], msg_list
     
-    def unpack_message(self, msg_list, content=True, copy=True):
-        """Return a message object from the format
-        sent by self.send.
-        
+    def unserialize(self, msg_list, content=True, copy=True):
+        """Unserialize a msg_list to a nested message dict.
+
+        This is roughly the inverse of serialize. The serialize/unserialize
+        methods work with full message lists, whereas pack/unpack work with
+        the individual message parts in the message list.
+
         Parameters:
         -----------
-        
+        msg_list : list of bytes or Message objects
+            The list of message parts of the form [HMAC,p_header,p_parent,
+            p_content,buffer1,buffer2,...].
         content : bool (True)
-            whether to unpack the content dict (True), 
-            or leave it serialized (False)
-        
+            Whether to unpack the content dict (True), or leave it packed
+            (False).
         copy : bool (True)
-            whether to return the bytes (True), 
-            or the non-copying Message object in each place (False)
-        
+            Whether to return the bytes (True), or the non-copying Message
+            object in each place (False).
+
+        Returns
+        -------
+        msg : dict
+            The nested message dict with top-level keys [header, parent_header,
+            content, buffers].
         """
         minlen = 4
         message = {}
         if not copy:
             for i in range(minlen):
                 msg_list[i] = msg_list[i].bytes
         if self.auth is not None:
             signature = msg_list[0]
             if signature in self.digest_history:
                 raise ValueError("Duplicate Signature: %r"%signature)
             self.digest_history.add(signature)
             check = self.sign(msg_list[1:4])
             if not signature == check:
                 raise ValueError("Invalid Signature: %r"%signature)
         if not len(msg_list) >= minlen:
             raise TypeError("malformed message, must have at least %i elements"%minlen)
         message['header'] = self.unpack(msg_list[1])
-        message['msg_type'] = message['header']['msg_type']
         message['parent_header'] = self.unpack(msg_list[2])
         if content:
             message['content'] = self.unpack(msg_list[3])
         else:
             message['content'] = msg_list[3]
         
         message['buffers'] = msg_list[4:]
         return message
 
 def test_msg2obj():
     am = dict(x=1)
     ao = Message(am)
     assert ao.x == am['x']
 
     am['y'] = dict(z=1)
     ao = Message(am)
     assert ao.y.z == am['y']['z']
     
     k1, k2 = 'y', 'z'
     assert ao[k1][k2] == am[k1][k2]
     
     am2 = dict(ao)
     assert am['x'] == am2['x']
     assert am['y']['z'] == am2['y']['z']
 

IPython/zmq/tests/test_session.py

@@ -1,111 +1,177 b''
 """test building messages with streamsession"""
 
 #-------------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-------------------------------------------------------------------------------
 
 #-------------------------------------------------------------------------------
 # Imports
 #-------------------------------------------------------------------------------
 
 import os
 import uuid
 import zmq
 
 from zmq.tests import BaseZMQTestCase
 from zmq.eventloop.zmqstream import ZMQStream
 
 from IPython.zmq import session as ss
 
 class SessionTestCase(BaseZMQTestCase):
     
     def setUp(self):
         BaseZMQTestCase.setUp(self)
         self.session = ss.Session()
 
+
+class MockSocket(zmq.Socket):
+
+    def __init__(self, *args, **kwargs):
+        super(MockSocket,self).__init__(*args,**kwargs)
+        self.data = []
+
+    def send_multipart(self, msgparts, *args, **kwargs):
+        self.data.extend(msgparts)
+
+    def send(self, part, *args, **kwargs):
+        self.data.append(part)
+
+    def recv_multipart(self, *args, **kwargs):
+        return self.data
+
 class TestSession(SessionTestCase):
     
     def test_msg(self):
         """message format"""
         msg = self.session.msg('execute')
-        thekeys = set('header msg_id parent_header msg_type content'.split())
+        thekeys = set('header parent_header content'.split())
         s = set(msg.keys())
         self.assertEquals(s, thekeys)
         self.assertTrue(isinstance(msg['content'],dict))
         self.assertTrue(isinstance(msg['header'],dict))
         self.assertTrue(isinstance(msg['parent_header'],dict))
-        self.assertEquals(msg['msg_type'], 'execute')
-        
-            
-    
+        self.assertEquals(msg['header']['msg_type'], 'execute')
+
+    def test_serialize(self):
+        msg = self.session.msg('execute',content=dict(a=10))
+        msg_list = self.session.serialize(msg, ident=b'foo')
+        ident, msg_list = self.session.feed_identities(msg_list)
+        new_msg = self.session.unserialize(msg_list)
+        self.assertEquals(ident[0], b'foo')
+        self.assertEquals(new_msg['header'],msg['header'])
+        self.assertEquals(new_msg['content'],msg['content'])
+        self.assertEquals(new_msg['parent_header'],msg['parent_header'])
+
+    def test_send(self):
+        socket = MockSocket(zmq.Context.instance(),zmq.PAIR)
+
+        msg = self.session.msg('execute', content=dict(a=10))
+        self.session.send(socket, msg, ident=b'foo', buffers=[b'bar'])
+        ident, msg_list = self.session.feed_identities(socket.data)
+        new_msg = self.session.unserialize(msg_list)
+        self.assertEquals(ident[0], b'foo')
+        self.assertEquals(new_msg['header'],msg['header'])
+        self.assertEquals(new_msg['content'],msg['content'])
+        self.assertEquals(new_msg['parent_header'],msg['parent_header'])        
+        self.assertEquals(new_msg['buffers'],[b'bar'])
+
+        socket.data = []
+
+        content = msg['content']
+        header = msg['header']
+        parent = msg['parent_header']
+        msg_type = header['msg_type']
+        self.session.send(socket, None, content=content, parent=parent,
+            header=header, ident=b'foo', buffers=[b'bar'])
+        ident, msg_list = self.session.feed_identities(socket.data)
+        new_msg = self.session.unserialize(msg_list)
+        self.assertEquals(ident[0], b'foo')
+        self.assertEquals(new_msg['header'],msg['header'])
+        self.assertEquals(new_msg['content'],msg['content'])
+        self.assertEquals(new_msg['parent_header'],msg['parent_header'])        
+        self.assertEquals(new_msg['buffers'],[b'bar'])
+
+        socket.data = []
+
+        self.session.send(socket, msg, ident=b'foo', buffers=[b'bar'])
+        ident, new_msg = self.session.recv(socket)
+        self.assertEquals(ident[0], b'foo')
+        self.assertEquals(new_msg['header'],msg['header'])
+        self.assertEquals(new_msg['content'],msg['content'])
+        self.assertEquals(new_msg['parent_header'],msg['parent_header'])        
+        self.assertEquals(new_msg['buffers'],[b'bar'])
+
+        socket.close()
+
     def test_args(self):
         """initialization arguments for Session"""
         s = self.session
         self.assertTrue(s.pack is ss.default_packer)
         self.assertTrue(s.unpack is ss.default_unpacker)
-        self.assertEquals(s.username, os.environ.get('USER', 'username'))
+        self.assertEquals(s.username, os.environ.get('USER', u'username'))
         
         s = ss.Session()
-        self.assertEquals(s.username, os.environ.get('USER', 'username'))
+        self.assertEquals(s.username, os.environ.get('USER', u'username'))
         
         self.assertRaises(TypeError, ss.Session, pack='hi')
         self.assertRaises(TypeError, ss.Session, unpack='hi')
         u = str(uuid.uuid4())
-        s = ss.Session(username='carrot', session=u)
+        s = ss.Session(username=u'carrot', session=u)
         self.assertEquals(s.session, u)
-        self.assertEquals(s.username, 'carrot')
+        self.assertEquals(s.username, u'carrot')
         
     def test_tracking(self):
         """test tracking messages"""
         a,b = self.create_bound_pair(zmq.PAIR, zmq.PAIR)
         s = self.session
         stream = ZMQStream(a)
         msg = s.send(a, 'hello', track=False)
         self.assertTrue(msg['tracker'] is None)
         msg = s.send(a, 'hello', track=True)
         self.assertTrue(isinstance(msg['tracker'], zmq.MessageTracker))
         M = zmq.Message(b'hi there', track=True)
         msg = s.send(a, 'hello', buffers=[M], track=True)
         t = msg['tracker']
         self.assertTrue(isinstance(t, zmq.MessageTracker))
         self.assertRaises(zmq.NotDone, t.wait, .1)
         del M
         t.wait(1) # this will raise
         
         
     # def test_rekey(self):
     #     """rekeying dict around json str keys"""
     #     d = {'0': uuid.uuid4(), 0:uuid.uuid4()}
     #     self.assertRaises(KeyError, ss.rekey, d)
     #     
     #     d = {'0': uuid.uuid4(), 1:uuid.uuid4(), 'asdf':uuid.uuid4()}
     #     d2 = {0:d['0'],1:d[1],'asdf':d['asdf']}
     #     rd = ss.rekey(d)
     #     self.assertEquals(d2,rd)
     #     
     #     d = {'1.5':uuid.uuid4(),'1':uuid.uuid4()}
     #     d2 = {1.5:d['1.5'],1:d['1']}
     #     rd = ss.rekey(d)
     #     self.assertEquals(d2,rd)
     #     
     #     d = {'1.0':uuid.uuid4(),'1':uuid.uuid4()}
     #     self.assertRaises(KeyError, ss.rekey, d)
     # 
     def test_unique_msg_ids(self):
         """test that messages receive unique ids"""
         ids = set()
         for i in range(2**12):
             h = self.session.msg_header('test')
             msg_id = h['msg_id']
             self.assertTrue(msg_id not in ids)
             ids.add(msg_id)
     
     def test_feed_identities(self):
         """scrub the front for zmq IDENTITIES"""
         theids = "engine client other".split()
         content = dict(code='whoda',stuff=object())
         themsg = self.session.msg('execute',content=content)
         pmsg = theids
+

docs/source/development/messaging.txt

@@ -1,937 +1,937 b''
 .. _messaging:
 
 ======================
  Messaging in IPython
 ======================
 
 
 Introduction
 ============
 
 This document explains the basic communications design and messaging
 specification for how the various IPython objects interact over a network
 transport.  The current implementation uses the ZeroMQ_ library for messaging
 within and between hosts.
 
 .. Note::
 
    This document should be considered the authoritative description of the
    IPython messaging protocol, and all developers are strongly encouraged to
    keep it updated as the implementation evolves, so that we have a single
    common reference for all protocol details.
    
 The basic design is explained in the following diagram:
 
 .. image:: figs/frontend-kernel.png
    :width: 450px
    :alt: IPython kernel/frontend messaging architecture.
    :align: center
    :target: ../_images/frontend-kernel.png
 
 A single kernel can be simultaneously connected to one or more frontends.  The
 kernel has three sockets that serve the following functions:
 
 1. REQ: this socket is connected to a *single* frontend at a time, and it allows
    the kernel to request input from a frontend when :func:`raw_input` is called.
    The frontend holding the matching REP socket acts as a 'virtual keyboard'
    for the kernel while this communication is happening (illustrated in the
    figure by the black outline around the central keyboard).  In practice,
    frontends may display such kernel requests using a special input widget or
    otherwise indicating that the user is to type input for the kernel instead
    of normal commands in the frontend.
 
 2. XREP: this single sockets allows multiple incoming connections from
    frontends, and this is the socket where requests for code execution, object
    information, prompts, etc. are made to the kernel by any frontend.  The
    communication on this socket is a sequence of request/reply actions from
    each frontend and the kernel.
 
 3. PUB: this socket is the 'broadcast channel' where the kernel publishes all
    side effects (stdout, stderr, etc.) as well as the requests coming from any
    client over the XREP socket and its own requests on the REP socket.  There
    are a number of actions in Python which generate side effects: :func:`print`
    writes to ``sys.stdout``, errors generate tracebacks, etc.  Additionally, in
    a multi-client scenario, we want all frontends to be able to know what each
    other has sent to the kernel (this can be useful in collaborative scenarios,
    for example).  This socket allows both side effects and the information
    about communications taking place with one client over the XREQ/XREP channel
    to be made available to all clients in a uniform manner.
 
    All messages are tagged with enough information (details below) for clients
    to know which messages come from their own interaction with the kernel and
    which ones are from other clients, so they can display each type
    appropriately.
 
 The actual format of the messages allowed on each of these channels is
 specified below.  Messages are dicts of dicts with string keys and values that
 are reasonably representable in JSON.  Our current implementation uses JSON
 explicitly as its message format, but this shouldn't be considered a permanent
 feature.  As we've discovered that JSON has non-trivial performance issues due
 to excessive copying, we may in the future move to a pure pickle-based raw
 message format.  However, it should be possible to easily convert from the raw
 objects to JSON, since we may have non-python clients (e.g. a web frontend).
 As long as it's easy to make a JSON version of the objects that is a faithful
 representation of all the data, we can communicate with such clients.
 
 .. Note::
 
    Not all of these have yet been fully fleshed out, but the key ones are, see
    kernel and frontend files for actual implementation details.
 
    
 Python functional API
 =====================
 
 As messages are dicts, they map naturally to a ``func(**kw)`` call form.  We
 should develop, at a few key points, functional forms of all the requests that
 take arguments in this manner and automatically construct the necessary dict
 for sending.
 
 
 General Message Format
 ======================
 
 All messages send or received by any IPython process should have the following
 generic structure::
     
     {
       # The message header contains a pair of unique identifiers for the
       # originating session and the actual message id, in addition to the
       # username for the process that generated the message.  This is useful in
       # collaborative settings where multiple users may be interacting with the
       # same kernel simultaneously, so that frontends can label the various
       # messages in a meaningful way.
-      'header' : { 'msg_id' : uuid,
-                   'username' : str,
-           'session' : uuid
+      'header' : {
+                    'msg_id' : uuid,
+                    'username' : str,
+                    'session' : uuid
+                    # All recognized message type strings are listed below.
+                    'msg_type' : str,
          },
 
       # In a chain of messages, the header from the parent is copied so that
       # clients can track where messages come from.
       'parent_header' : dict,
 
-      # All recognized message type strings are listed below.
-      'msg_type' : str,
-
       # The actual content of the message must be a dict, whose structure
       # depends on the message type.x
       'content' : dict,
     }
 
 For each message type, the actual content will differ and all existing message
 types are specified in what follows of this document.
 
 
 Messages on the XREP/XREQ socket
 ================================
 
 .. _execute:
 
 Execute
 -------
 
 This message type is used by frontends to ask the kernel to execute code on
 behalf of the user, in a namespace reserved to the user's variables (and thus
 separate from the kernel's own internal code and variables).
 
 Message type: ``execute_request``::
 
     content = {
         # Source code to be executed by the kernel, one or more lines.
     'code' : str,
 
     # A boolean flag which, if True, signals the kernel to execute this
     # code as quietly as possible.  This means that the kernel will compile
     # the code witIPython/core/tests/h 'exec' instead of 'single' (so
     # sys.displayhook will not fire), and will *not*:
     #   - broadcast exceptions on the PUB socket
     #   - do any logging
     #   - populate any history
     #
     # The default is False.
     'silent' : bool,
 
     # A list of variable names from the user's namespace to be retrieved.  What
     # returns is a JSON string of the variable's repr(), not a python object.
     'user_variables' : list,
 
     # Similarly, a dict mapping names to expressions to be evaluated in the
     # user's dict.
     'user_expressions' : dict,
     }
 
 The ``code`` field contains a single string (possibly multiline).  The kernel
 is responsible for splitting this into one or more independent execution blocks
 and deciding whether to compile these in 'single' or 'exec' mode (see below for
 detailed execution semantics).
 
 The ``user_`` fields deserve a detailed explanation.  In the past, IPython had
 the notion of a prompt string that allowed arbitrary code to be evaluated, and
 this was put to good use by many in creating prompts that displayed system
 status, path information, and even more esoteric uses like remote instrument
 status aqcuired over the network.  But now that IPython has a clean separation
 between the kernel and the clients, the kernel has no prompt knowledge; prompts
 are a frontend-side feature, and it should be even possible for different
 frontends to display different prompts while interacting with the same kernel.
 
 The kernel now provides the ability to retrieve data from the user's namespace
 after the execution of the main ``code``, thanks to two fields in the
 ``execute_request`` message:
 
 - ``user_variables``: If only variables from the user's namespace are needed, a
   list of variable names can be passed and a dict with these names as keys and
   their :func:`repr()` as values will be returned.
 
 - ``user_expressions``: For more complex expressions that require function
   evaluations, a dict can be provided with string keys and arbitrary python
   expressions as values.  The return message will contain also a dict with the
   same keys and the :func:`repr()` of the evaluated expressions as value.
 
 With this information, frontends can display any status information they wish
 in the form that best suits each frontend (a status line, a popup, inline for a
 terminal, etc).
 
 .. Note::
 
    In order to obtain the current execution counter for the purposes of
    displaying input prompts, frontends simply make an execution request with an
    empty code string and ``silent=True``.
 
 Execution semantics
 ~~~~~~~~~~~~~~~~~~~
 
 When the silent flag is false, the execution of use code consists of the
 following phases (in silent mode, only the ``code`` field is executed):
 
 1. Run the ``pre_runcode_hook``.
 
 2. Execute the ``code`` field, see below for details.
 
 3. If #2 succeeds, compute ``user_variables`` and ``user_expressions`` are
    computed.  This ensures that any error in the latter don't harm the main
    code execution.
 
 4. Call any method registered with :meth:`register_post_execute`.
 
 .. warning::
 
    The API for running code before/after the main code block is likely to
    change soon.  Both the ``pre_runcode_hook`` and the
    :meth:`register_post_execute` are susceptible to modification, as we find a
    consistent model for both.
 
 To understand how the ``code`` field is executed, one must know that Python
 code can be compiled in one of three modes (controlled by the ``mode`` argument
 to the :func:`compile` builtin):
 
 *single*
   Valid for a single interactive statement (though the source can contain
   multiple lines, such as a for loop).  When compiled in this mode, the
   generated bytecode contains special instructions that trigger the calling of
   :func:`sys.displayhook` for any expression in the block that returns a value.
   This means that a single statement can actually produce multiple calls to
   :func:`sys.displayhook`, if for example it contains a loop where each
   iteration computes an unassigned expression would generate 10 calls::
 
       for i in range(10):
           i**2
 
 *exec*
   An arbitrary amount of source code, this is how modules are compiled.
   :func:`sys.displayhook` is *never* implicitly called.
 
 *eval*
   A single expression that returns a value.  :func:`sys.displayhook` is *never*
   implicitly called.
 
 
 The ``code`` field is split into individual blocks each of which is valid for
 execution in 'single' mode, and then:
 
 - If there is only a single block: it is executed in 'single' mode.
 
 - If there is more than one block:
 
   * if the last one is a single line long, run all but the last in 'exec' mode
     and the very last one in 'single' mode.  This makes it easy to type simple
     expressions at the end to see computed values.
 
   * if the last one is no more than two lines long, run all but the last in
     'exec' mode and the very last one in 'single' mode.  This makes it easy to
     type simple expressions at the end to see computed values.  - otherwise
     (last one is also multiline), run all in 'exec' mode
 
   * otherwise (last one is also multiline), run all in 'exec' mode as a single
     unit.
 
 Any error in retrieving the ``user_variables`` or evaluating the
 ``user_expressions`` will result in a simple error message in the return fields
 of the form::
 
    [ERROR] ExceptionType: Exception message
 
 The user can simply send the same variable name or expression for evaluation to
 see a regular traceback.
 
 Errors in any registered post_execute functions are also reported similarly,
 and the failing function is removed from the post_execution set so that it does
 not continue triggering failures.
 
 Upon completion of the execution request, the kernel *always* sends a reply,
 with a status code indicating what happened and additional data depending on
 the outcome.  See :ref:`below <execution_results>` for the possible return
 codes and associated data.
 
 
 Execution counter (old prompt number)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 The kernel has a single, monotonically increasing counter of all execution
 requests that are made with ``silent=False``.  This counter is used to populate
 the ``In[n]``, ``Out[n]`` and ``_n`` variables, so clients will likely want to
 display it in some form to the user, which will typically (but not necessarily)
 be done in the prompts.  The value of this counter will be returned as the
 ``execution_count`` field of all ``execute_reply`` messages.
 
 .. _execution_results:
 
 Execution results
 ~~~~~~~~~~~~~~~~~
     
 Message type: ``execute_reply``::
 
     content = {
       # One of: 'ok' OR 'error' OR 'abort'
       'status' : str,
 
       # The global kernel counter that increases by one with each non-silent
       # executed request.  This will typically be used by clients to display
       # prompt numbers to the user.  If the request was a silent one, this will
       # be the current value of the counter in the kernel.
       'execution_count' : int,
     }
 
 When status is 'ok', the following extra fields are present::
 
     {
       # The execution payload is a dict with string keys that may have been
       # produced by the code being executed.  It is retrieved by the kernel at
       # the end of the execution and sent back to the front end, which can take
       # action on it as needed.  See main text for further details.
       'payload' : dict,
 
       # Results for the user_variables and user_expressions.
       'user_variables' : dict,
       'user_expressions' : dict,
 
       # The kernel will often transform the input provided to it.  If the
       # '---->' transform had been applied, this is filled, otherwise it's the
       # empty string.  So transformations like magics don't appear here, only
       # autocall ones.
       'transformed_code' : str,
       }
 
 .. admonition:: Execution payloads
     
    The notion of an 'execution payload' is different from a return value of a
    given set of code, which normally is just displayed on the pyout stream
    through the PUB socket.  The idea of a payload is to allow special types of
    code, typically magics, to populate a data container in the IPython kernel
    that will be shipped back to the caller via this channel.  The kernel will
    have an API for this, probably something along the lines of::
 
        ip.exec_payload_add(key, value)
 
    though this API is still in the design stages.  The data returned in this
    payload will allow frontends to present special views of what just happened.
 
    
 When status is 'error', the following extra fields are present::
 
     {
       'exc_name' : str,   # Exception name, as a string
       'exc_value' : str,  # Exception value, as a string
 
       # The traceback will contain a list of frames, represented each as a
       # string.  For now we'll stick to the existing design of ultraTB, which
       # controls exception level of detail statefully.  But eventually we'll
       # want to grow into a model where more information is collected and
       # packed into the traceback object, with clients deciding how little or
       # how much of it to unpack.  But for now, let's start with a simple list
       # of strings, since that requires only minimal changes to ultratb as
       # written.
       'traceback' : list,
     }
 
 
 When status is 'abort', there are for now no additional data fields.  This
 happens when the kernel was interrupted by a signal.
 
 Kernel attribute access
 -----------------------
 
 .. warning::
 
    This part of the messaging spec is not actually implemented in the kernel
    yet.
  
 While this protocol does not specify full RPC access to arbitrary methods of
 the kernel object, the kernel does allow read (and in some cases write) access
 to certain attributes.
 
 The policy for which attributes can be read is: any attribute of the kernel, or
 its sub-objects, that belongs to a :class:`Configurable` object and has been
 declared at the class-level with Traits validation, is in principle accessible
 as long as its name does not begin with a leading underscore.  The attribute
 itself will have metadata indicating whether it allows remote read and/or write
 access.  The message spec follows for attribute read and write requests.
 
 Message type: ``getattr_request``::
 
     content = {
         # The (possibly dotted) name of the attribute
 	'name' : str,
     }
 
 When a ``getattr_request`` fails, there are two possible error types:
 
 - AttributeError: this type of error was raised when trying to access the
   given name by the kernel itself.  This means that the attribute likely
   doesn't exist.
 
 - AccessError: the attribute exists but its value is not readable remotely.
 
 
 Message type: ``getattr_reply``::
 
     content = {
         # One of ['ok', 'AttributeError', 'AccessError'].
         'status' : str,
 	# If status is 'ok', a JSON object.
 	'value' : object,
     }
 
 Message type: ``setattr_request``::
 
     content = {
         # The (possibly dotted) name of the attribute
 	'name' : str,
 	
 	# A JSON-encoded object, that will be validated by the Traits
 	# information in the kernel
 	'value' : object,
     }
 
 When a ``setattr_request`` fails, there are also two possible error types with
 similar meanings  as those of the ``getattr_request`` case, but for writing.
     
 Message type: ``setattr_reply``::
 
     content = {
         # One of ['ok', 'AttributeError', 'AccessError'].
         'status' : str,
     }
 
 
     
 Object information
 ------------------
 
 One of IPython's most used capabilities is the introspection of Python objects
 in the user's namespace, typically invoked via the ``?`` and ``??`` characters
 (which in reality are shorthands for the ``%pinfo`` magic).  This is used often
 enough that it warrants an explicit message type, especially because frontends
 may want to get object information in response to user keystrokes (like Tab or
 F1) besides from the user explicitly typing code like ``x??``.
 
 Message type: ``object_info_request``::
 
     content = {
         # The (possibly dotted) name of the object to be searched in all
 	# relevant namespaces
         'name' : str,
 
     	# The level of detail desired.  The default (0) is equivalent to typing
 	# 'x?' at the prompt, 1 is equivalent to 'x??'.
 	'detail_level' : int,
     }
 
 The returned information will be a dictionary with keys very similar to the
 field names that IPython prints at the terminal.
     
 Message type: ``object_info_reply``::
 
     content = {
     # The name the object was requested under
     'name' : str,
     
     # Boolean flag indicating whether the named object was found or not.  If
     # it's false, all other fields will be empty.
     'found' : bool,
     
     # Flags for magics and system aliases
     'ismagic' : bool,
     'isalias' : bool,
 
     # The name of the namespace where the object was found ('builtin',
     # 'magics', 'alias', 'interactive', etc.)
     'namespace' : str,
 
     # The type name will be type.__name__ for normal Python objects, but it
     # can also be a string like 'Magic function' or 'System alias'
     'type_name' : str,
 
     # The string form of the object, possibly truncated for length if 
     # detail_level is 0
     'string_form' : str,
 
     # For objects with a __class__ attribute this will be set
     'base_class' : str,
 
     # For objects with a __len__ attribute this will be set
     'length' : int,
 
     # If the object is a function, class or method whose file we can find,
     # we give its full path
     'file' : str,
 
     # For pure Python callable objects, we can reconstruct the object
     # definition line which provides its call signature.  For convenience this
     # is returned as a single 'definition' field, but below the raw parts that
     # compose it are also returned as the argspec field.
     'definition' : str,
 
     # The individual parts that together form the definition string.  Clients
     # with rich display capabilities may use this to provide a richer and more
     # precise representation of the definition line (e.g. by highlighting
     # arguments based on the user's cursor position).  For non-callable
     # objects, this field is empty.
     'argspec' : { # The names of all the arguments
                   args : list,
 		  # The name of the varargs (*args), if any
                   varargs : str,
 		  # The name of the varkw (**kw), if any
 		  varkw : str,
 		  # The values (as strings) of all default arguments.  Note
 		  # that these must be matched *in reverse* with the 'args'
 		  # list above, since the first positional args have no default
 		  # value at all.
 		  defaults : list,
 		  },
 
     # For instances, provide the constructor signature (the definition of
     # the __init__ method):
     'init_definition' : str,
     
     # Docstrings: for any object (function, method, module, package) with a
     # docstring, we show it.  But in addition, we may provide additional
     # docstrings.  For example, for instances we will show the constructor
     # and class docstrings as well, if available.
     'docstring' : str,
 
     # For instances, provide the constructor and class docstrings
     'init_docstring' : str,
     'class_docstring' : str,
     
     # If it's a callable object whose call method has a separate docstring and
     # definition line:
     'call_def' : str,
     'call_docstring' : str,
     
     # If detail_level was 1, we also try to find the source code that
     # defines the object, if possible.  The string 'None' will indicate
     # that no source was found.
     'source' : str,
     }
 '
     
 Complete
 --------
 
 Message type: ``complete_request``::
 
     content = {
         # The text to be completed, such as 'a.is'
     'text' : str,
 
     # The full line, such as 'print a.is'.  This allows completers to
     # make decisions that may require information about more than just the
     # current word.
     'line' : str,
 
     # The entire block of text where the line is.  This may be useful in the
     # case of multiline completions where more context may be needed.  Note: if
     # in practice this field proves unnecessary, remove it to lighten the
     # messages.
     
     'block' : str,
 
     # The position of the cursor where the user hit 'TAB' on the line.
     'cursor_pos' : int,
     }
 
 Message type: ``complete_reply``::
 
     content = {
         # The list of all matches to the completion request, such as
     # ['a.isalnum', 'a.isalpha'] for the above example.
     'matches' : list
     }
 
     
 History
 -------
 
 For clients to explicitly request history from a kernel.  The kernel has all
 the actual execution history stored in a single location, so clients can
 request it from the kernel when needed.
 
 Message type: ``history_request``::
 
     content = {
     
       # If True, also return output history in the resulting dict.
       'output' : bool,
 
       # If True, return the raw input history, else the transformed input.
       'raw' : bool,
 
       # So far, this can be 'range', 'tail' or 'search'.
       'hist_access_type' : str,
       
       # If hist_access_type is 'range', get a range of input cells. session can
       # be a positive session number, or a negative number to count back from
       # the current session.
       'session' : int,
       # start and stop are line numbers within that session.
       'start' : int,
       'stop' : int,
       
       # If hist_access_type is 'tail', get the last n cells.
       'n' : int,
       
       # If hist_access_type is 'search', get cells matching the specified glob
       # pattern (with * and ? as wildcards).
       'pattern' : str,
       
     }
 
 Message type: ``history_reply``::
 
     content = {
       # A list of 3 tuples, either:
       # (session, line_number, input) or
       # (session, line_number, (input, output)),
       # depending on whether output was False or True, respectively.
       'history' : list,
     }
 
 
 Connect
 -------
 
 When a client connects to the request/reply socket of the kernel, it can issue
 a connect request to get basic information about the kernel, such as the ports
 the other ZeroMQ sockets are listening on. This allows clients to only have
 to know about a single port (the XREQ/XREP channel) to connect to a kernel.
 
 Message type: ``connect_request``::
 
     content = {
     }
 
 Message type: ``connect_reply``::
 
     content = {
         'xrep_port' : int  # The port the XREP socket is listening on.
         'pub_port' : int   # The port the PUB socket is listening on.
         'req_port' : int   # The port the REQ socket is listening on.
         'hb_port' : int    # The port the heartbeat socket is listening on.
     }
 
 
 
 Kernel shutdown
 ---------------
 
 The clients can request the kernel to shut itself down; this is used in
 multiple cases:
 
 - when the user chooses to close the client application via a menu or window
   control.
 - when the user types 'exit' or 'quit' (or their uppercase magic equivalents).
 - when the user chooses a GUI method (like the 'Ctrl-C' shortcut in the
   IPythonQt client) to force a kernel restart to get a clean kernel without
   losing client-side state like history or inlined figures.
 
 The client sends a shutdown request to the kernel, and once it receives the
 reply message (which is otherwise empty), it can assume that the kernel has
 completed shutdown safely.
 
 Upon their own shutdown, client applications will typically execute a last
 minute sanity check and forcefully terminate any kernel that is still alive, to
 avoid leaving stray processes in the user's machine.
 
 For both shutdown request and reply, there is no actual content that needs to
 be sent, so the content dict is empty.
 
 Message type: ``shutdown_request``::
 
     content = {
         'restart' : bool # whether the shutdown is final, or precedes a restart
     }
 
 Message type: ``shutdown_reply``::
 
     content = {
         'restart' : bool # whether the shutdown is final, or precedes a restart
     }
 
 .. Note::
 
    When the clients detect a dead kernel thanks to inactivity on the heartbeat
    socket, they simply send a forceful process termination signal, since a dead
    process is unlikely to respond in any useful way to messages.
     
 
 Messages on the PUB/SUB socket
 ==============================
 
 Streams (stdout,  stderr, etc)
 ------------------------------
 
 Message type: ``stream``::
 
     content = {
         # The name of the stream is one of 'stdin', 'stdout', 'stderr'
         'name' : str,
     
         # The data is an arbitrary string to be written to that stream
         'data' : str,
     }
 
 When a kernel receives a raw_input call, it should also broadcast it on the pub
 socket with the names 'stdin' and 'stdin_reply'.  This will allow other clients
 to monitor/display kernel interactions and possibly replay them to their user
 or otherwise expose them.
 
 Display Data
 ------------
 
 This type of message is used to bring back data that should be diplayed (text,
 html, svg, etc.) in the frontends. This data is published to all frontends.
 Each message can have multiple representations of the data; it is up to the
 frontend to decide which to use and how. A single message should contain all
 possible representations of the same information. Each representation should
 be a JSON'able data structure, and should be a valid MIME type.
 
 Some questions remain about this design:
 
 * Do we use this message type for pyout/displayhook? Probably not, because
   the displayhook also has to handle the Out prompt display. On the other hand
   we could put that information into the metadata secion.
 
 Message type: ``display_data``::
 
     content = {
 
         # Who create the data
         'source' : str,
 
         # The data dict contains key/value pairs, where the kids are MIME
         # types and the values are the raw data of the representation in that
         # format. The data dict must minimally contain the ``text/plain``
         # MIME type which is used as a backup representation.        
         'data' : dict,
 
         # Any metadata that describes the data
         'metadata' : dict
     }
 
 Python inputs
 -------------
 
 These messages are the re-broadcast of the ``execute_request``.
 
 Message type: ``pyin``::
 
     content = {
         'code' : str  # Source code to be executed, one or more lines
     }
 
 Python outputs
 --------------
 
 When Python produces output from code that has been compiled in with the
 'single' flag to :func:`compile`, any expression that produces a value (such as
 ``1+1``) is passed to ``sys.displayhook``, which is a callable that can do with
 this value whatever it wants.  The default behavior of ``sys.displayhook`` in
 the Python interactive prompt is to print to ``sys.stdout`` the :func:`repr` of
 the value as long as it is not ``None`` (which isn't printed at all).  In our
 case, the kernel instantiates as ``sys.displayhook`` an object which has
 similar behavior, but which instead of printing to stdout, broadcasts these
 values as ``pyout`` messages for clients to display appropriately.
 
 IPython's displayhook can handle multiple simultaneous formats depending on its
 configuration. The default pretty-printed repr text is always given with the
 ``data`` entry in this message. Any other formats are provided in the
 ``extra_formats`` list. Frontends are free to display any or all of these
 according to its capabilities. ``extra_formats`` list contains 3-tuples of an ID
 string, a type string, and the data. The ID is unique to the formatter
 implementation that created the data. Frontends will typically ignore the ID
 unless if it has requested a particular formatter. The type string tells the
 frontend how to interpret the data. It is often, but not always a MIME type.
 Frontends should ignore types that it does not understand. The data itself is
 any JSON object and depends on the format. It is often, but not always a string.
 
 Message type: ``pyout``::
 
     content = {
 
         # The counter for this execution is also provided so that clients can
         # display it, since IPython automatically creates variables called _N
         # (for prompt N).
         'execution_count' : int,
         
         # The data dict contains key/value pairs, where the kids are MIME
         # types and the values are the raw data of the representation in that
         # format. The data dict must minimally contain the ``text/plain``
         # MIME type which is used as a backup representation.
         'data' : dict,
 
     }
     
 Python errors
 -------------
 
 When an error occurs during code execution
 
 Message type: ``pyerr``::
 
     content = {
        # Similar content to the execute_reply messages for the 'error' case,
        # except the 'status' field is omitted.
     }
 
 Kernel status
 -------------
 
 This message type is used by frontends to monitor the status of the kernel.
 
 Message type: ``status``::
 
     content = {
         # When the kernel starts to execute code, it will enter the 'busy'
         # state and when it finishes, it will enter the 'idle' state.
         execution_state : ('busy', 'idle')
     }
 
 Kernel crashes
 --------------
 
 When the kernel has an unexpected exception, caught by the last-resort
 sys.excepthook, we should broadcast the crash handler's output before exiting.
 This will allow clients to notice that a kernel died, inform the user and
 propose further actions.
 
 Message type: ``crash``::
 
     content = {
        # Similarly to the 'error' case for execute_reply messages, this will
        # contain exc_name, exc_type and traceback fields.
 
        # An additional field with supplementary information such as where to
        # send the crash message
        'info' : str,
     }
 
 
 Future ideas
 ------------
     
 Other potential message types, currently unimplemented, listed below as ideas.
     
 Message type: ``file``::
 
     content = {
     'path' : 'cool.jpg',
     'mimetype' : str,
     'data' : str,
     }
 
     
 Messages on the REQ/REP socket
 ==============================
 
 This is a socket that goes in the opposite direction: from the kernel to a
 *single* frontend, and its purpose is to allow ``raw_input`` and similar
 operations that read from ``sys.stdin`` on the kernel to be fulfilled by the
 client.  For now we will keep these messages as simple as possible, since they
 basically only mean to convey the ``raw_input(prompt)`` call.
 
 Message type: ``input_request``::
 
     content = { 'prompt' : str }
 
 Message type: ``input_reply``::
 
     content = { 'value' : str }
     
 .. Note::
 
    We do not explicitly try to forward the raw ``sys.stdin`` object, because in
    practice the kernel should behave like an interactive program.  When a
    program is opened on the console, the keyboard effectively takes over the
    ``stdin`` file descriptor, and it can't be used for raw reading anymore.
    Since the IPython kernel effectively behaves like a console program (albeit
    one whose "keyboard" is actually living in a separate process and
    transported over the zmq connection), raw ``stdin`` isn't expected to be
    available.
 
    
 Heartbeat for kernels
 =====================
 
 Initially we had considered using messages like those above over ZMQ for a
 kernel 'heartbeat' (a way to detect quickly and reliably whether a kernel is
 alive at all, even if it may be busy executing user code).  But this has the
 problem that if the kernel is locked inside extension code, it wouldn't execute
 the python heartbeat code.  But it turns out that we can implement a basic
 heartbeat with pure ZMQ, without using any Python messaging at all.
 
 The monitor sends out a single zmq message (right now, it is a str of the
 monitor's lifetime in seconds), and gets the same message right back, prefixed
 with the zmq identity of the XREQ socket in the heartbeat process. This can be
 a uuid, or even a full message, but there doesn't seem to be a need for packing
 up a message when the sender and receiver are the exact same Python object.
 
 The model is this::
 
     monitor.send(str(self.lifetime)) # '1.2345678910'
 
 and the monitor receives some number of messages of the form::
 
     ['uuid-abcd-dead-beef', '1.2345678910']
 
 where the first part is the zmq.IDENTITY of the heart's XREQ on the engine, and
 the rest is the message sent by the monitor.  No Python code ever has any
 access to the message between the monitor's send, and the monitor's recv.
 
 
 ToDo
 ====
 
 Missing things include:
 
 * Important: finish thinking through the payload concept and API.
 
 * Important: ensure that we have a good solution for magics like %edit.  It's
   likely that with the payload concept we can build a full solution, but not
   100% clear yet.
 
 * Finishing the details of the heartbeat protocol.
 
 * Signal handling: specify what kind of information kernel should broadcast (or
   not) when it receives signals.
 
 .. include:: ../links.rst