scheduler progress

IPython/zmq/parallel/client.py

@@ -1,640 +1,654 b''
-#!/usr/bin/env python
 """A semi-synchronous Client for the ZMQ controller"""
+#-----------------------------------------------------------------------------
+#  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.
+#-----------------------------------------------------------------------------
 
-import time
+#-----------------------------------------------------------------------------
+# Imports
+#-----------------------------------------------------------------------------
 
+import time
 from pprint import pprint
 
+import zmq
+from zmq.eventloop import ioloop, zmqstream
+
 from IPython.external.decorator import decorator
 
 import streamsession as ss
-import zmq
-from zmq.eventloop import ioloop, zmqstream
 from remotenamespace import RemoteNamespace
 from view import DirectView
 from dependency import Dependency, depend, require
 
 def _push(ns):
     globals().update(ns)
 
 def _pull(keys):
     g = globals()
     if isinstance(keys, (list,tuple)):
         return map(g.get, keys)
     else:
         return g.get(keys)
 
 def _clear():
     globals().clear()
 
 def execute(code):
     exec code in globals()
 
 # decorators for methods:
 @decorator
 def spinfirst(f,self,*args,**kwargs):
     self.spin()
     return f(self, *args, **kwargs)
 
 @decorator
 def defaultblock(f, self, *args, **kwargs):
     block = kwargs.get('block',None)
     block = self.block if block is None else block
     saveblock = self.block
     self.block = block
     ret = f(self, *args, **kwargs)
     self.block = saveblock
     return ret
 
 class AbortedTask(object):
     def __init__(self, msg_id):
         self.msg_id = msg_id
 # @decorator
 # def checktargets(f):
 #     @wraps(f)
 #     def checked_method(self, *args, **kwargs):
 #         self._build_targets(kwargs['targets'])
 #         return f(self, *args, **kwargs)
 #     return checked_method
 
 
 # class _ZMQEventLoopThread(threading.Thread):
 #     
 #     def __init__(self, loop):
 #         self.loop = loop
 #         threading.Thread.__init__(self)
 #     
 #     def run(self):
 #         self.loop.start()
 # 
 class Client(object):
     """A semi-synchronous client to the IPython ZMQ controller
     
     Attributes
     ----------
     ids : set
         a set of engine IDs
         requesting the ids attribute always synchronizes
         the registration state. To request ids without synchronization,
         use _ids
     
     history : list of msg_ids
         a list of msg_ids, keeping track of all the execution
         messages you have submitted
     
     outstanding : set of msg_ids
         a set of msg_ids that have been submitted, but whose
         results have not 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
             
     barrier : wait on one or more msg_ids
     
     execution methods: apply/apply_bound/apply_to
         legacy: execute, run
     
     query methods: queue_status, get_result
     
     control methods: abort, kill
     
     
     
     """
     
     
     _connected=False
     _engines=None
     registration_socket=None
     query_socket=None
     control_socket=None
     notification_socket=None
     queue_socket=None
     task_socket=None
     block = False
     outstanding=None
     results = None
     history = None
     debug = False
     
     def __init__(self, addr, context=None, username=None, debug=False):
         if context is None:
             context = zmq.Context()
         self.context = context
         self.addr = addr
         if username is None:
             self.session = ss.StreamSession()
         else:
             self.session = ss.StreamSession(username)
         self.registration_socket = self.context.socket(zmq.PAIR)
         self.registration_socket.setsockopt(zmq.IDENTITY, self.session.session)
         self.registration_socket.connect(addr)
         self._engines = {}
         self._ids = set()
         self.outstanding=set()
         self.results = {}
         self.history = []
         self.debug = debug
         self.session.debug = debug
         
         self._notification_handlers = {'registration_notification' : self._register_engine,
                                     'unregistration_notification' : self._unregister_engine,
                                     }
         self._queue_handlers = {'execute_reply' : self._handle_execute_reply,
                                 'apply_reply' : self._handle_apply_reply}
         self._connect()
         
     
     @property
     def ids(self):
         self._flush_notifications()
         return self._ids
     
     def _update_engines(self, engines):
         for k,v in engines.iteritems():
             eid = int(k)
             self._engines[eid] = bytes(v) # force not unicode
             self._ids.add(eid)
     
     def _build_targets(self, targets):
         if targets is None:
             targets = self._ids
         elif isinstance(targets, str):
             if targets.lower() == 'all':
                 targets = self._ids
             else:
                 raise TypeError("%r not valid str target, must be 'all'"%(targets))
         elif isinstance(targets, int):
             targets = [targets]
         return [self._engines[t] for t in targets], list(targets)
     
     def _connect(self):
         """setup all our socket connections to the controller"""
         if self._connected:
             return
         self._connected=True
         self.session.send(self.registration_socket, 'connection_request')
         idents,msg = self.session.recv(self.registration_socket,mode=0)
         if self.debug:
             pprint(msg)
         msg = ss.Message(msg)
         content = msg.content
         if content.status == 'ok':
             if content.queue:
                 self.queue_socket = self.context.socket(zmq.PAIR)
                 self.queue_socket.setsockopt(zmq.IDENTITY, self.session.session)
                 self.queue_socket.connect(content.queue)
             if content.task:
                 self.task_socket = self.context.socket(zmq.PAIR)
                 self.task_socket.setsockopt(zmq.IDENTITY, self.session.session)
                 self.task_socket.connect(content.task)
             if content.notification:
                 self.notification_socket = self.context.socket(zmq.SUB)
                 self.notification_socket.connect(content.notification)
                 self.notification_socket.setsockopt(zmq.SUBSCRIBE, "")
             if content.query:
                 self.query_socket = self.context.socket(zmq.PAIR)
                 self.query_socket.setsockopt(zmq.IDENTITY, self.session.session)
                 self.query_socket.connect(content.query)
             if content.control:
                 self.control_socket = self.context.socket(zmq.PAIR)
                 self.control_socket.setsockopt(zmq.IDENTITY, self.session.session)
                 self.control_socket.connect(content.control)
             self._update_engines(dict(content.engines))
                 
         else:
             self._connected = False
             raise Exception("Failed to connect!")
     
     #### handlers and callbacks for incoming messages #######
     def _register_engine(self, msg):
         content = msg['content']
         eid = content['id']
         d = {eid : content['queue']}
         self._update_engines(d)
         self._ids.add(int(eid))
 
     def _unregister_engine(self, msg):
         # print 'unregister',msg
         content = msg['content']
         eid = int(content['id'])
         if eid in self._ids:
             self._ids.remove(eid)
             self._engines.pop(eid)
         
     def _handle_execute_reply(self, msg):
         # msg_id = msg['msg_id']
         parent = msg['parent_header']
         msg_id = parent['msg_id']
         if msg_id not in self.outstanding:
             print "got unknown result: %s"%msg_id
         else:
             self.outstanding.remove(msg_id)
         self.results[msg_id] = ss.unwrap_exception(msg['content'])
     
     def _handle_apply_reply(self, msg):
         # pprint(msg)
         # msg_id = msg['msg_id']
         parent = msg['parent_header']
         msg_id = parent['msg_id']
         if msg_id not in self.outstanding:
             print "got unknown result: %s"%msg_id
         else:
             self.outstanding.remove(msg_id)
         content = msg['content']
         if content['status'] == 'ok':
             self.results[msg_id] = ss.unserialize_object(msg['buffers'])
         elif content['status'] == 'aborted':
             self.results[msg_id] = AbortedTask(msg_id)
         elif content['status'] == 'resubmitted':
             pass # handle resubmission
         else:
             self.results[msg_id] = ss.unwrap_exception(content)
     
     def _flush_notifications(self):
         "flush incoming notifications of engine registrations"
         msg = self.session.recv(self.notification_socket, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
             msg = msg[-1]
             msg_type = msg['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)
             msg = self.session.recv(self.notification_socket, mode=zmq.NOBLOCK)
     
     def _flush_results(self, sock):
         "flush incoming task or queue results"
         msg = self.session.recv(sock, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
             msg = msg[-1]
             msg_type = msg['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)
             msg = self.session.recv(sock, mode=zmq.NOBLOCK)
     
     def _flush_control(self, sock):
         "flush incoming control replies"
         msg = self.session.recv(sock, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
             msg = self.session.recv(sock, mode=zmq.NOBLOCK)
     
     ###### get/setitem ########
     
     def __getitem__(self, key):
         if isinstance(key, int):
             if key not in self.ids:
                 raise IndexError("No such engine: %i"%key)
             return DirectView(self, key)
         
         if isinstance(key, slice):
             indices = range(len(self.ids))[key]
             ids = sorted(self._ids)
             key = [ ids[i] for i in indices ]
             # newkeys = sorted(self._ids)[thekeys[k]]
         
         if isinstance(key, (tuple, list, xrange)):
             _,targets = self._build_targets(list(key))
             return DirectView(self, targets)
         else:
             raise TypeError("key by int/iterable of ints only, not %s"%(type(key)))
     
     ############ begin real methods #############
     
     def spin(self):
         """flush incoming notifications and execution results."""
         if self.notification_socket:
             self._flush_notifications()
         if self.queue_socket:
             self._flush_results(self.queue_socket)
         if self.task_socket:
             self._flush_results(self.task_socket)
         if self.control_socket:
             self._flush_control(self.control_socket)
     
     @spinfirst
     def queue_status(self, targets=None, verbose=False):
         """fetch the status of engine queues
         
         Parameters
         ----------
         targets : int/str/list of ints/strs
                 the engines on which to execute
                 default : all
         verbose : bool
                 whether to return lengths only, or lists of ids for each element
                 
         """
         targets = self._build_targets(targets)[1]
         content = dict(targets=targets)
         self.session.send(self.query_socket, "queue_request", content=content)
         idents,msg = self.session.recv(self.query_socket, 0)
         if self.debug:
             pprint(msg)
         return msg['content']
         
     @spinfirst
     @defaultblock
     def clear(self, targets=None, block=None):
         """clear the namespace in target(s)"""
         targets = self._build_targets(targets)[0]
-        print targets
         for t in targets:
             self.session.send(self.control_socket, 'clear_request', content={},ident=t)
         error = False
         if self.block:
             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 = msg['content']
         if error:
             return error
         
     
     @spinfirst
     @defaultblock
     def abort(self, msg_ids = None, targets=None, block=None):
         """abort the Queues of target(s)"""
         targets = self._build_targets(targets)[0]
-        print targets
         if isinstance(msg_ids, basestring):
             msg_ids = [msg_ids]
         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 self.block:
             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 = msg['content']
         if error:
             return error
     
     @spinfirst
     @defaultblock
     def kill(self, targets=None, block=None):
         """Terminates one or more engine processes."""
         targets = self._build_targets(targets)[0]
-        print targets
         for t in targets:
             self.session.send(self.control_socket, 'kill_request', content={},ident=t)
         error = False
         if self.block:
             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 = msg['content']
         if error:
             return error
         
     @defaultblock
     def execute(self, code, targets='all', block=None):
         """executes `code` on `targets` in blocking or nonblocking manner.
         
         Parameters
         ----------
         code : str
                 the code string to be executed
         targets : int/str/list of ints/strs
                 the engines on which to execute
                 default : all
         block : bool
                 whether or not to wait until done
         """
         # block = self.block if block is None else block
         # saveblock = self.block
         # self.block = block
         result = self.apply(execute, (code,), targets=targets, block=block, bound=True)
         # self.block = saveblock
         return result
     
     def run(self, code, block=None):
         """runs `code` on an engine. 
         
         Calls to this are load-balanced.
         
         Parameters
         ----------
         code : str
                 the code string to be executed
         block : bool
                 whether or not to wait until done
         
         """
         result = self.apply(execute, (code,), targets=None, block=block, bound=False)
         return result
     
     def _apply_balanced(self, f, args, kwargs, bound=True, block=None,
                             after=None, follow=None):
         """the underlying method for applying functions in a load balanced
         manner."""
         block = block if block is not None else self.block
+        if isinstance(after, Dependency):
+            after = after.as_dict()
+        elif after is None:
+            after = []
+        if isinstance(follow, Dependency):
+            follow = follow.as_dict()
+        elif follow is None:
+            follow = []
+        subheader = dict(after=after, follow=follow)
         
         bufs = ss.pack_apply_message(f,args,kwargs)
         content = dict(bound=bound)
         msg = self.session.send(self.task_socket, "apply_request", 
-                content=content, buffers=bufs)
+                content=content, buffers=bufs, subheader=subheader)
         msg_id = msg['msg_id']
         self.outstanding.add(msg_id)
         self.history.append(msg_id)
         if block:
             self.barrier(msg_id)
             return self.results[msg_id]
         else:
             return msg_id
     
     def _apply_direct(self, f, args, kwargs, bound=True, block=None, targets=None,
                                 after=None, follow=None):
         """Then underlying method for applying functions to specific engines."""
         
         block = block if block is not None else self.block
         
         queues,targets = self._build_targets(targets)
-        print queues
         bufs = ss.pack_apply_message(f,args,kwargs)
         if isinstance(after, Dependency):
             after = after.as_dict()
         elif after is None:
             after = []
         if isinstance(follow, Dependency):
             follow = follow.as_dict()
         elif follow is None:
             follow = []
         subheader = dict(after=after, follow=follow)
         content = dict(bound=bound)
         msg_ids = []
         for queue in queues:
             msg = self.session.send(self.queue_socket, "apply_request", 
                     content=content, buffers=bufs,ident=queue, subheader=subheader)
             msg_id = msg['msg_id']
             self.outstanding.add(msg_id)
             self.history.append(msg_id)
             msg_ids.append(msg_id)
         if block:
             self.barrier(msg_ids)
         else:
             if len(msg_ids) == 1:
                 return msg_ids[0]
             else:
                 return msg_ids
         if len(msg_ids) == 1:
             return self.results[msg_ids[0]]
         else:
             result = {}
             for target,mid in zip(targets, msg_ids):
                     result[target] = self.results[mid]
             return result
     
     def apply(self, f, args=None, kwargs=None, bound=True, block=None, targets=None,
                         after=None, follow=None):
         """calls f(*args, **kwargs) on a remote engine(s), returning the result.
         
         if self.block is False:
             returns msg_id or list of msg_ids
         else:
             returns actual result of f(*args, **kwargs)
         """
         # enforce types of f,args,kwrags
         args = args if args is not None else []
         kwargs = kwargs if kwargs is not None else {}
         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))
             
         options  = dict(bound=bound, block=block, after=after, follow=follow)
             
         if targets is None:
             return self._apply_balanced(f, args, kwargs, **options)
         else:
             return self._apply_direct(f, args, kwargs, targets=targets, **options)
     
     def push(self, ns, targets=None, block=None):
         """push the contents of `ns` into the namespace on `target`"""
         if not isinstance(ns, dict):
             raise TypeError("Must be a dict, not %s"%type(ns))
         result = self.apply(_push, (ns,), targets=targets, block=block,bound=True)
         return result
     
     @spinfirst
     def pull(self, keys, targets=None, block=True):
         """pull objects from `target`'s namespace by `keys`"""
         
         result = self.apply(_pull, (keys,), targets=targets, block=block, bound=True)
         return result
     
     def barrier(self, msg_ids=None, timeout=-1):
         """waits on one or more `msg_ids`, for up to `timeout` seconds.
         
         Parameters
         ----------
         msg_ids : int, str, or list of ints and/or strs
                 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, msg_ids still outstanding
         """
         tic = time.time()
         if msg_ids is None:
             theids = self.outstanding
         else:
             if isinstance(msg_ids, (int, str)):
                 msg_ids = [msg_ids]
             theids = set()
             for msg_id in msg_ids:
                 if isinstance(msg_id, int):
                     msg_id = self.history[msg_id]
                 theids.add(msg_id)
         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
     
     @spinfirst
     def get_results(self, msg_ids,status_only=False):
         """returns the result of the execute or task request with `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]
             theids.append(msg_id)
         
         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)
         
         # while True:
         #     try:
         #     except zmq.ZMQError:
         #         time.sleep(1e-3)
         #         continue
         #     else:
         #         break
         return msg['content']
 
 class AsynClient(Client):
     """An Asynchronous client, using the Tornado Event Loop"""
     io_loop = None
     queue_stream = None
     notifier_stream = None
     
     def __init__(self, addr, context=None, username=None, debug=False, io_loop=None):
         Client.__init__(self, addr, context, username, debug)
         if io_loop is None:
             io_loop = ioloop.IOLoop.instance()
         self.io_loop = io_loop
         
         self.queue_stream = zmqstream.ZMQStream(self.queue_socket, io_loop)
         self.control_stream = zmqstream.ZMQStream(self.control_socket, io_loop)
         self.task_stream = zmqstream.ZMQStream(self.task_socket, io_loop)
         self.notification_stream = zmqstream.ZMQStream(self.notification_socket, io_loop)
     
     def spin(self):
         for stream in (self.queue_stream, self.notifier_stream, 
                         self.task_stream, self.control_stream):
             stream.flush()
     
  No newline at end of file

IPython/zmq/parallel/controller.py

@@ -1,921 +1,919 b''
 #!/usr/bin/env python
-# encoding: utf-8
-
 """The IPython Controller with 0MQ
 This is the master object that handles connections from engines, clients, and 
 """
 #-----------------------------------------------------------------------------
-#  Copyright (C) 2008-2009  The IPython Development Team
+#  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 datetime import datetime
 import logging
 
 import zmq
 from zmq.eventloop import zmqstream, ioloop
 import uuid
 
 # internal:
 from IPython.zmq.log import logger # a Logger object
 from IPython.zmq.entry_point import bind_port
 
 from streamsession import Message, wrap_exception
 from entry_point import (make_argument_parser, select_random_ports, split_ports,
                         connect_logger)
 # from messages import json # use the same import switches
 
 #-----------------------------------------------------------------------------
 # Code
 #-----------------------------------------------------------------------------
 
 class ReverseDict(dict):
     """simple double-keyed subset of dict methods."""
     
     def __init__(self, *args, **kwargs):
         dict.__init__(self, *args, **kwargs)
         self.reverse = dict()
         for key, value in self.iteritems():
             self.reverse[value] = key
     
     def __getitem__(self, key):
         try:
             return dict.__getitem__(self, key)
         except KeyError:
             return self.reverse[key]
     
     def __setitem__(self, key, value):
         if key in self.reverse:
             raise KeyError("Can't have key %r on both sides!"%key)
         dict.__setitem__(self, key, value)
         self.reverse[value] = key
     
     def pop(self, key):
         value = dict.pop(self, key)
         self.d1.pop(value)
         return value
     
 
 class EngineConnector(object):
     """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=0
     queue=None
     control=None
     registration=None
     heartbeat=None
     pending=None
     
     def __init__(self, id, queue, registration, control, heartbeat=None):
         logger.info("engine::Engine Connected: %i"%id)
         self.id = id
         self.queue = queue
         self.registration = registration
         self.control = control
         self.heartbeat = heartbeat
         
 class Controller(object):
     """The IPython Controller with 0MQ connections
     
     Parameters
     ==========
     loop: zmq IOLoop instance
     session: StreamSession object
     <removed> context: zmq context for creating new connections (?)
     registrar: ZMQStream for engine registration requests (XREP)
     clientele: ZMQStream for client connections (XREP)
                 not used for jobs, only query/control commands
     queue: ZMQStream for monitoring the command queue (SUB)
     heartbeat: HeartMonitor object checking the pulse of the engines
     db_stream: connection to db for out of memory logging of commands
                 NotImplemented
     queue_addr: zmq connection address of the XREP socket for the queue
     hb_addr: zmq connection address of the PUB socket for heartbeats
     task_addr: zmq connection address of the XREQ socket for task queue
     """
     # internal data structures:
     ids=None # engine IDs
     keytable=None
     engines=None
     clients=None
     hearts=None
     pending=None
     results=None
     tasks=None
     completed=None
     mia=None
     incoming_registrations=None
     registration_timeout=None
     
     #objects from constructor:
     loop=None
     registrar=None
     clientelle=None
     queue=None
     heartbeat=None
     notifier=None
     db=None
     client_addr=None
     engine_addrs=None
     
     
     def __init__(self, loop, session, queue, registrar, heartbeat, clientele, notifier, db, engine_addrs, client_addrs):
         """
         # universal:
         loop: IOLoop for creating future connections
         session: streamsession for sending serialized data
         # engine:
         queue: ZMQStream for monitoring queue messages
         registrar: ZMQStream for engine registration
         heartbeat: HeartMonitor object for tracking engines
         # client:
         clientele: ZMQStream for client connections
         # extra:
         db: ZMQStream for db connection (NotImplemented)
         engine_addrs: zmq address/protocol dict for engine connections
         client_addrs: zmq address/protocol dict for client connections
         """
         self.ids = set()
         self.keytable={}
         self.incoming_registrations={}
         self.engines = {}
         self.by_ident = {}
         self.clients = {}
         self.hearts = {}
         self.mia = set()
         
         # self.sockets = {}
         self.loop = loop
         self.session = session
         self.registrar = registrar
         self.clientele = clientele
         self.queue = queue
         self.heartbeat = heartbeat
         self.notifier = notifier
         self.db = db
         
         self.client_addrs = client_addrs
         assert isinstance(client_addrs['queue'], str)
         # self.hb_addrs = hb_addrs
         self.engine_addrs = engine_addrs
         assert isinstance(engine_addrs['queue'], str)
         assert len(engine_addrs['heartbeat']) == 2
         
         
         # register our callbacks
         self.registrar.on_recv(self.dispatch_register_request)
         self.clientele.on_recv(self.dispatch_client_msg)
         self.queue.on_recv(self.dispatch_queue_traffic)
         
         if heartbeat is not None:
             heartbeat.add_heart_failure_handler(self.handle_heart_failure)
             heartbeat.add_new_heart_handler(self.handle_new_heart)
         
         if self.db is not None:
             self.db.on_recv(self.dispatch_db)
             
         self.client_handlers = {'queue_request': self.queue_status,
                                 'result_request': self.get_results,
                                 'purge_request': self.purge_results,
                                 'resubmit_request': self.resubmit_task,
                                 }
         
         self.registrar_handlers = {'registration_request' : self.register_engine,
                                 'unregistration_request' : self.unregister_engine,
                                 'connection_request': self.connection_request,
         
         }
         # 
         # this is the stuff that will move to DB:
         self.results = {} # completed results
         self.pending = {} # pending messages, keyed by msg_id
         self.queues = {} # pending msg_ids keyed by engine_id
         self.tasks = {} # pending msg_ids submitted as tasks, keyed by client_id
         self.completed = {} # completed msg_ids keyed by engine_id
         self.registration_timeout = max(5000, 2*self.heartbeat.period)
         
         logger.info("controller::created controller")
     
     def _new_id(self):
         """gemerate a new ID"""
         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
     
     def _validate_client_msg(self, msg):
         """validates and unpacks headers of a message. Returns False if invalid,
         (ident, header, parent, content)"""
         client_id = msg[0]
         try:
             msg = self.session.unpack_message(msg[1:], content=True)
         except:
             logger.error("client::Invalid Message %s"%msg)
             return False
         
         msg_type = msg.get('msg_type', None)
         if msg_type is None:
             return False
         header = msg.get('header')
         # session doesn't handle split content for now:
         return client_id, msg
         
     
     #-----------------------------------------------------------------------------
-    # dispatch methods (1 per socket)
+    # dispatch methods (1 per stream)
     #-----------------------------------------------------------------------------
     
     def dispatch_register_request(self, msg):
         """"""
         logger.debug("registration::dispatch_register_request(%s)"%msg)
         idents,msg = self.session.feed_identities(msg)
         print idents,msg, len(msg)
         try:
             msg = self.session.unpack_message(msg,content=True)
         except Exception, e:
             logger.error("registration::got bad registration message: %s"%msg)
             raise e
             return
         
         msg_type = msg['msg_type']
         content = msg['content']
         
         handler = self.registrar_handlers.get(msg_type, None)
         if handler is None:
             logger.error("registration::got bad registration message: %s"%msg)
         else:
             handler(idents, msg)
     
     def dispatch_queue_traffic(self, msg):
         """all ME and Task queue messages come through here"""
         logger.debug("queue traffic: %s"%msg[:2])
         switch = msg[0]
         idents, msg = self.session.feed_identities(msg[1:])
         if switch == 'in':
             self.save_queue_request(idents, msg)
         elif switch == 'out':
             self.save_queue_result(idents, msg)
         elif switch == 'intask':
             self.save_task_request(idents, msg)
         elif switch == 'outtask':
             self.save_task_result(idents, msg)
         elif switch == 'tracktask':
             self.save_task_destination(idents, msg)
         elif switch in ('incontrol', 'outcontrol'):
             pass
         else:
             logger.error("Invalid message topic: %s"%switch)
         
     
     def dispatch_client_msg(self, msg):
         """Route messages from clients"""
         idents, msg = self.session.feed_identities(msg)
         client_id = idents[0]
         try:
             msg = self.session.unpack_message(msg, content=True)
         except:
             content = wrap_exception()
             logger.error("Bad Client Message: %s"%msg)
             self.session.send(self.clientele, "controller_error", ident=client_id, 
                     content=content)
             return
         
         # print client_id, header, parent, content
         #switch on message type:
         msg_type = msg['msg_type']
         logger.info("client:: client %s requested %s"%(client_id, msg_type))
         handler = self.client_handlers.get(msg_type, None)
         try:
             assert handler is not None, "Bad Message Type: %s"%msg_type
         except:
             content = wrap_exception()
             logger.error("Bad Message Type: %s"%msg_type)
             self.session.send(self.clientele, "controller_error", ident=client_id, 
                     content=content)
             return
         else:
             handler(client_id, 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."""
         logger.debug("heartbeat::handle_new_heart(%r)"%heart)
         if heart not in self.incoming_registrations:
             logger.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"""
         logger.debug("heartbeat::handle_heart_failure(%r)"%heart)
         eid = self.hearts.get(heart, None)
         queue = self.engines[eid].queue
         if eid is None:
             logger.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):
         queue_id, client_id = idents[:2]
             
         try:
             msg = self.session.unpack_message(msg, content=False)
         except:
             logger.error("queue::client %r sent invalid message to %r: %s"%(client_id, queue_id, msg))
             return
         
         eid = self.by_ident.get(queue_id, None)
         if eid is None:
             logger.error("queue::target %r not registered"%queue_id)
             logger.debug("queue::    valid are: %s"%(self.by_ident.keys()))
             return
             
         header = msg['header']
         msg_id = header['msg_id']
         info = dict(submit=datetime.now(),
                     received=None,
                     engine=(eid, queue_id))
         self.pending[msg_id] = ( msg, info )
         self.queues[eid][0].append(msg_id)
     
     def save_queue_result(self, idents, msg):
         client_id, queue_id = idents[:2]
         
         try:
             msg = self.session.unpack_message(msg, content=False)
         except:
             logger.error("queue::engine %r sent invalid message to %r: %s"%(
                     queue_id,client_id, msg))
             return
         
         eid = self.by_ident.get(queue_id, None)
         if eid is None:
             logger.error("queue::unknown engine %r is sending a reply: "%queue_id)
             logger.debug("queue::       %s"%msg[2:])
             return
         
         parent = msg['parent_header']
         if not parent:
             return
         msg_id = parent['msg_id']
         self.results[msg_id] = msg
         if msg_id in self.pending:
             self.pending.pop(msg_id)
             self.queues[eid][0].remove(msg_id)
             self.completed[eid].append(msg_id)
         else:
             logger.debug("queue:: unknown msg finished %s"%msg_id)
             
     #--------------------- Task Queue Traffic ------------------------------
     
     def save_task_request(self, idents, msg):
         client_id = idents[0]
         
         try:
             msg = self.session.unpack_message(msg, content=False)
         except:
             logger.error("task::client %r sent invalid task message: %s"%(
                     client_id, msg))
             return
         
         header = msg['header']
         msg_id = header['msg_id']
         self.mia.add(msg_id)
         self.pending[msg_id] = msg
         if not self.tasks.has_key(client_id):
             self.tasks[client_id] = []
         self.tasks[client_id].append(msg_id)
     
     def save_task_result(self, idents, msg):
         client_id = idents[0]
         try:
             msg = self.session.unpack_message(msg, content=False)
         except:
             logger.error("task::invalid task result message send to %r: %s"%(
                     client_id, msg))
             return
         
         parent = msg['parent_header']
         if not parent:
             # print msg
             # logger.warn("")
             return
         msg_id = parent['msg_id']
         self.results[msg_id] = msg
         if msg_id in self.pending:
             self.pending.pop(msg_id)
             if msg_id in self.mia:
                 self.mia.remove(msg_id)
         else:
-            logger.debug("task:: unknown task %s finished"%msg_id)
+            logger.debug("task::unknown task %s finished"%msg_id)
     
     def save_task_destination(self, idents, msg):
         try:
             msg = self.session.unpack_message(msg, content=True)
         except:
             logger.error("task::invalid task tracking message")
             return
         content = msg['content']
         print content
         msg_id = content['msg_id']
         engine_uuid = content['engine_id']
         for eid,queue_id in self.keytable.iteritems():
             if queue_id == engine_uuid:
                 break
         
-        logger.info("task:: task %s arrived on %s"%(msg_id, eid))
+        logger.info("task::task %s arrived on %s"%(msg_id, eid))
         if msg_id in self.mia:
             self.mia.remove(msg_id)
         else:
             logger.debug("task::task %s not listed as MIA?!"%(msg_id))
         self.tasks[engine_uuid].append(msg_id)
     
     def mia_task_request(self, idents, msg):
         client_id = idents[0]
         content = dict(mia=self.mia,status='ok')
         self.session.send('mia_reply', content=content, idents=client_id)
         
         
             
     #-------------------- Registration -----------------------------
         
     def connection_request(self, client_id, msg):
         """reply with connection addresses for clients"""
         logger.info("client::client %s connected"%client_id)
         content = dict(status='ok')
         content.update(self.client_addrs)
         jsonable = {}
         for k,v in self.keytable.iteritems():
             jsonable[str(k)] = v
         content['engines'] = jsonable
         self.session.send(self.registrar, 'connection_reply', content, parent=msg, ident=client_id)
     
     def register_engine(self, reg, msg):
         """register an engine"""
         content = msg['content']
         try:
             queue = content['queue']
         except KeyError:
             logger.error("registration::queue not specified")
             return
         heart = content.get('heartbeat', None)
         """register a new engine, and create the socket(s) necessary"""
         eid = self._new_id()
         # print (eid, queue, reg, heart)
         
         logger.debug("registration::register_engine(%i, %r, %r, %r)"%(eid, queue, reg, heart))
         
         content = dict(id=eid,status='ok')
         content.update(self.engine_addrs)
         # check if requesting available IDs:
         if queue in self.by_ident:
             content = {'status': 'error', 'reason': "queue_id %r in use"%queue}
         elif heart in self.hearts: # need to check unique hearts?
             content = {'status': 'error', 'reason': "heart_id %r in use"%heart}
         else:
             for h, pack in self.incoming_registrations.iteritems():
                 if heart == h:
                     content = {'status': 'error', 'reason': "heart_id %r in use"%heart}
                     break
                 elif queue == pack[1]:
                     content = {'status': 'error', 'reason': "queue_id %r in use"%queue}
                     break
         
         msg = self.session.send(self.registrar, "registration_reply", 
                 content=content, 
                 ident=reg)
         
         if content['status'] == 'ok':
             if heart in self.heartbeat.hearts:
                 # already beating
                 self.incoming_registrations[heart] = (eid,queue,reg,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,dc)
         else:
             logger.error("registration::registration %i failed: %s"%(eid, content['reason']))
         return eid
     
     def unregister_engine(self, ident, msg):
         try:
             eid = msg['content']['id']
         except:
             logger.error("registration::bad engine id for unregistration: %s"%ident)
             return
         logger.info("registration::unregister_engine(%s)"%eid)
         content=dict(id=eid, queue=self.engines[eid].queue)
         self.ids.remove(eid)
         self.keytable.pop(eid)
         ec = self.engines.pop(eid)
         self.hearts.pop(ec.heartbeat)
         self.by_ident.pop(ec.queue)
         self.completed.pop(eid)
         for msg_id in self.queues.pop(eid)[0]:
             msg = self.pending.pop(msg_id)
             ############## TODO: HANDLE IT ################
         
         if self.notifier:
             self.session.send(self.notifier, "unregistration_notification", content=content)
     
     def finish_registration(self, heart):
         try: 
             (eid,queue,reg,purge) = self.incoming_registrations.pop(heart)
         except KeyError:
             logger.error("registration::tried to finish nonexistant registration")
             return
         logger.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(eid, queue, reg, control, heart)
         self.by_ident[queue] = eid
         self.queues[eid] = ([],[])
         self.completed[eid] = list()
         self.hearts[heart] = eid
         content = dict(id=eid, queue=self.engines[eid].queue)
         if self.notifier:
             self.session.send(self.notifier, "registration_notification", content=content)
     
     def _purge_stalled_registration(self, heart):
         if heart in self.incoming_registrations:
             eid = self.incoming_registrations.pop(heart)[0]
             logger.info("registration::purging stalled registration: %i"%eid)
         else:
             pass
             
     #------------------- Client Requests -------------------------------
     
     def check_load(self, client_id, msg):
         content = msg['content']
         try:
             targets = content['targets']
             targets = self._validate_targets(targets)
         except:
             content = wrap_exception()
             self.session.send(self.clientele, "controller_error", 
                     content=content, ident=client_id)
             return
         
         content = dict(status='ok')
         # loads = {}
         for t in targets:
             content[str(t)] = len(self.queues[t])
         self.session.send(self.clientele, "load_reply", content=content, ident=client_id)
             
     
     def queue_status(self, client_id, msg):
         """handle queue_status request"""
         content = msg['content']
         targets = content['targets']
         try:
             targets = self._validate_targets(targets)
         except:
             content = wrap_exception()
             self.session.send(self.clientele, "controller_error", 
                     content=content, ident=client_id)
             return
         verbose = msg.get('verbose', False)
         content = dict()
         for t in targets:
             queue = self.queues[t]
             completed = self.completed[t]
             if not verbose:
                 queue = len(queue)
                 completed = len(completed)
             content[str(t)] = {'queue': queue, 'completed': completed }
             # pending
         self.session.send(self.clientele, "queue_reply", content=content, ident=client_id)
     
     def purge_results(self, client_id, msg):
         content = msg['content']
         msg_ids = content.get('msg_ids', [])
         reply = dict(status='ok')
         if msg_ids == 'all':
             self.results = {}
         else:
             for msg_id in msg_ids:
                 if msg_id in self.results:
                     self.results.pop(msg_id)
                 else:
                     if msg_id in self.pending:
                         reply = dict(status='error', reason="msg pending: %r"%msg_id)
                     else:
                         reply = dict(status='error', reason="No such msg: %r"%msg_id)
                     break
             eids = content.get('engine_ids', [])
             for eid in eids:
                 if eid not in self.engines:
                     reply = dict(status='error', reason="No such engine: %i"%eid)
                     break
                 msg_ids = self.completed.pop(eid)
                 for msg_id in msg_ids:
                     self.results.pop(msg_id)
         
         self.sesison.send(self.clientele, 'purge_reply', content=reply, ident=client_id)
     
     def resubmit_task(self, client_id, msg, buffers):
         content = msg['content']
         header = msg['header']
         
         
         msg_ids = content.get('msg_ids', [])
         reply = dict(status='ok')
         if msg_ids == 'all':
             self.results = {}
         else:
             for msg_id in msg_ids:
                 if msg_id in self.results:
                     self.results.pop(msg_id)
                 else:
                     if msg_id in self.pending:
                         reply = dict(status='error', reason="msg pending: %r"%msg_id)
                     else:
                         reply = dict(status='error', reason="No such msg: %r"%msg_id)
                     break
             eids = content.get('engine_ids', [])
             for eid in eids:
                 if eid not in self.engines:
                     reply = dict(status='error', reason="No such engine: %i"%eid)
                     break
                 msg_ids = self.completed.pop(eid)
                 for msg_id in msg_ids:
                     self.results.pop(msg_id)
         
         self.sesison.send(self.clientele, 'purge_reply', content=reply, ident=client_id)
     
     def get_results(self, client_id, msg):
         """get the result of 1 or more messages"""
         content = msg['content']
         msg_ids = set(content['msg_ids'])
         statusonly = content.get('status_only', False)
         pending = []
         completed = []
         content = dict(status='ok')
         content['pending'] = pending
         content['completed'] = completed
         for msg_id in msg_ids:
             if msg_id in self.pending:
                 pending.append(msg_id)
             elif msg_id in self.results:
                 completed.append(msg_id)
                 if not statusonly:
                     content[msg_id] = self.results[msg_id]['content']
             else:
                 content = dict(status='error')
                 content['reason'] = 'no such message: '+msg_id
                 break
         self.session.send(self.clientele, "result_reply", content=content, 
                                             parent=msg, ident=client_id)
     
 
 
 ############ OLD METHODS for Python Relay Controller ###################
     def _validate_engine_msg(self, msg):
         """validates and unpacks headers of a message. Returns False if invalid,
         (ident, message)"""
         ident = msg[0]
         try:
             msg = self.session.unpack_message(msg[1:], content=False)
         except:
             logger.error("engine.%s::Invalid Message %s"%(ident, msg))
             return False
         
         try:
             eid = msg.header.username
             assert self.engines.has_key(eid)
         except:
             logger.error("engine::Invalid Engine ID %s"%(ident))
             return False
         
         return eid, msg
     
 
 #--------------------
 # Entry Point
 #--------------------
 
 def main():
     import time
     from multiprocessing import Process
     
     from zmq.eventloop.zmqstream import ZMQStream
     from zmq.devices import ProcessMonitoredQueue
     from zmq.log import handlers
     
     import streamsession as session
     import heartmonitor
     from scheduler import launch_scheduler
     
     parser = make_argument_parser()
     
     parser.add_argument('--client', type=int, metavar='PORT', default=0,
                         help='set the XREP port for clients [default: random]')
     parser.add_argument('--notice', type=int, metavar='PORT', default=0,
                         help='set the PUB socket for registration notification [default: random]')
     parser.add_argument('--hb', type=str, metavar='PORTS',
                         help='set the 2 ports for heartbeats [default: random]')
     parser.add_argument('--ping', type=int, default=3000,
                         help='set the heartbeat period in ms [default: 3000]')
     parser.add_argument('--monitor', type=int, metavar='PORT', default=0,
                         help='set the SUB port for queue monitoring [default: random]')
     parser.add_argument('--mux', type=str, metavar='PORTS',
                         help='set the XREP ports for the MUX queue [default: random]')
     parser.add_argument('--task', type=str, metavar='PORTS',
                         help='set the XREP/XREQ ports for the task queue [default: random]')
     parser.add_argument('--control', type=str, metavar='PORTS',
                         help='set the XREP ports for the control queue [default: random]')
     parser.add_argument('--scheduler', type=str, default='pure',
                         choices = ['pure', 'lru', 'plainrandom', 'weighted', 'twobin','leastload'],
                         help='select the task scheduler  [default: pure ZMQ]')
     
     args = parser.parse_args()
     
     if args.url:
         args.transport,iface = args.url.split('://')
         iface = iface.split(':')
         args.ip = iface[0]
         if iface[1]:
             args.regport = iface[1]
     
     iface="%s://%s"%(args.transport,args.ip)+':%i'
     
     random_ports = 0
     if args.hb:
         hb = split_ports(args.hb, 2)
     else:
         hb = select_random_ports(2)
     if args.mux:
         mux = split_ports(args.mux, 2)
     else:
         mux = None
         random_ports += 2
     if args.task:
         task = split_ports(args.task, 2)
     else:
         task = None
         random_ports += 2
     if args.control:
         control = split_ports(args.control, 2)
     else:
         control = None
         random_ports += 2
     
     ctx = zmq.Context()
     loop = ioloop.IOLoop.instance()
     
     # setup logging
     connect_logger(ctx, iface%args.logport, root="controller", loglevel=args.loglevel)
     
     # Registrar socket
     reg = ZMQStream(ctx.socket(zmq.XREP), loop)
     regport = bind_port(reg, args.ip, args.regport)
     
     ### Engine connections ###
     
     # heartbeat
     hpub = ctx.socket(zmq.PUB)
     bind_port(hpub, args.ip, hb[0])
     hrep = ctx.socket(zmq.XREP)
     bind_port(hrep, args.ip, hb[1])
     
     hmon = heartmonitor.HeartMonitor(loop, ZMQStream(hpub,loop), ZMQStream(hrep,loop),args.ping)
     hmon.start()
     
     ### Client connections ###
     # Clientele socket
     c = ZMQStream(ctx.socket(zmq.XREP), loop)
     cport = bind_port(c, args.ip, args.client)
     # Notifier socket
     n = ZMQStream(ctx.socket(zmq.PUB), loop)
     nport = bind_port(n, args.ip, args.notice)
     
     thesession = session.StreamSession(username=args.ident or "controller")
     
     ### build and launch the queues ###
     
     # monitor socket
     sub = ctx.socket(zmq.SUB)
     sub.setsockopt(zmq.SUBSCRIBE, "")
     monport = bind_port(sub, args.ip, args.monitor)
     sub = ZMQStream(sub, loop)
     
     ports = select_random_ports(random_ports)
     # Multiplexer Queue (in a Process)
     if not mux:
         mux = (ports.pop(),ports.pop())
     q = ProcessMonitoredQueue(zmq.XREP, zmq.XREP, zmq.PUB, 'in', 'out')
     q.bind_in(iface%mux[0])
     q.bind_out(iface%mux[1])
     q.connect_mon(iface%monport)
     q.daemon=True
     q.start()
     
     # Control Queue (in a Process)
     if not control:
         control = (ports.pop(),ports.pop())
     q = ProcessMonitoredQueue(zmq.XREP, zmq.XREP, zmq.PUB, 'incontrol', 'outcontrol')
     q.bind_in(iface%control[0])
     q.bind_out(iface%control[1])
     q.connect_mon(iface%monport)
     q.daemon=True
     q.start()
     
     # Task Queue (in a Process)
     if not task:
         task = (ports.pop(),ports.pop())
     if args.scheduler == 'pure':
         q = ProcessMonitoredQueue(zmq.XREP, zmq.XREQ, zmq.PUB, 'intask', 'outtask')
         q.bind_in(iface%task[0])
         q.bind_out(iface%task[1])
         q.connect_mon(iface%monport)
         q.daemon=True
         q.start()
     else:
         sargs = (iface%task[0],iface%task[1],iface%monport,iface%nport,args.scheduler)
         print sargs
         p = Process(target=launch_scheduler, args=sargs)
         p.daemon=True
         p.start()
     
     time.sleep(.25)
     
     # build connection dicts
     engine_addrs = {
         'control' : iface%control[1],
         'queue': iface%mux[1],
         'heartbeat': (iface%hb[0], iface%hb[1]),
         'task' : iface%task[1],
         'monitor' : iface%monport,
         }
     
     client_addrs = {
         'control' : iface%control[0],
         'query': iface%cport,
         'queue': iface%mux[0],
         'task' : iface%task[0],
         'notification': iface%nport
         }
     con = Controller(loop, thesession, sub, reg, hmon, c, n, None, engine_addrs, client_addrs)
     loop.start()
     

IPython/zmq/parallel/scheduler.py

@@ -1,401 +1,404 b''
 #----------------------------------------------------------------------
 # Imports
 #----------------------------------------------------------------------
 
 from random import randint,random
 
 try:
     import numpy
 except ImportError:
     numpy = None
 
 import zmq
 from zmq.eventloop import ioloop, zmqstream
 
 # local imports
 from IPython.zmq.log import logger # a Logger object
 from client import Client
 from dependency import Dependency
 import streamsession as ss
 
 from IPython.external.decorator import decorator
 
 @decorator
 def logged(f,self,*args,**kwargs):
     print ("#--------------------")
     print ("%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
 #---------------------------------------------------------------------
 class TaskScheduler(object):
-    """Simple Python TaskScheduler object.
+    """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.
     
     """
     
     scheme = leastload # function for determining the destination
     client_stream = None # client-facing stream
     engine_stream = None # engine-facing stream
     mon_stream = None # controller-facing stream
     dependencies = None # dict by msg_id of [ msg_ids that depend on key ]
     depending = None # dict by msg_id of (msg_id, raw_msg, after, follow)
     pending = None # dict by engine_uuid of submitted tasks
     completed = None # dict by engine_uuid of completed tasks
     clients = None # dict by msg_id for who submitted the task
     targets = None # list of target IDENTs
     loads = None # list of engine loads
     all_done = None # set of all completed tasks
     blacklist = None # dict by msg_id of locations where a job has encountered UnmetDependency
     
     
     def __init__(self, client_stream, engine_stream, mon_stream, 
                 notifier_stream, scheme=None, io_loop=None):
         if io_loop is None:
             io_loop = ioloop.IOLoop.instance()
         self.io_loop = io_loop
         self.client_stream = client_stream
         self.engine_stream = engine_stream
         self.mon_stream = mon_stream
         self.notifier_stream = notifier_stream
         
         if scheme is not None:
             self.scheme = scheme
         else:
             self.scheme = TaskScheduler.scheme
         
         self.session = ss.StreamSession(username="TaskScheduler")
         
         self.dependencies = {}
         self.depending = {}
         self.completed = {}
         self.pending = {}
         self.all_done = set()
+        self.blacklist = {}
         
         self.targets = []
         self.loads = []
         
         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)
     
     def resume_receiving(self):
         """resume accepting jobs"""
         self.client_stream.on_recv(self.dispatch_submission, copy=False)
     
     def stop_receiving(self):
         self.client_stream.on_recv(None)
     
     #-----------------------------------------------------------------------
     # [Un]Registration Handling
     #-----------------------------------------------------------------------
     
     def dispatch_notification(self, msg):
         """dispatch register/unregister events."""
         idents,msg = self.session.feed_identities(msg)
         msg = self.session.unpack_message(msg)
         msg_type = msg['msg_type']
         handler = self._notification_handlers.get(msg_type, None)
         if handler is None:
             raise Exception("Unhandled message type: %s"%msg_type)
         else:
             try:
                 handler(str(msg['content']['queue']))
             except KeyError:
                 logger.error("task::Invalid notification msg: %s"%msg)
     @logged
     def _register_engine(self, uid):
         """new engine became available"""
         # head of the line:
         self.targets.insert(0,uid)
         self.loads.insert(0,0)
         # initialize sets
         self.completed[uid] = set()
         self.pending[uid] = {}
         if len(self.targets) == 1:
             self.resume_receiving()
 
     def _unregister_engine(self, uid):
         """existing engine became unavailable"""
         # handle any potentially finished tasks:
         if len(self.targets) == 1:
             self.stop_receiving()
         self.engine_stream.flush()
         
         self.completed.pop(uid)
         lost = self.pending.pop(uid)
         
         idx = self.targets.index(uid)
         self.targets.pop(idx)
         self.loads.pop(idx)
         
         self.handle_stranded_tasks(lost)
     
     def handle_stranded_tasks(self, lost):
         """deal with jobs resident in an engine that died."""
         # TODO: resubmit the tasks?
         for msg_id in lost:
             pass
     
     
     #-----------------------------------------------------------------------
     # Job Submission
     #-----------------------------------------------------------------------
     @logged
     def dispatch_submission(self, raw_msg):
         """dispatch job submission"""
         # ensure targets up to date:
         self.notifier_stream.flush()
         try:
             idents, msg = self.session.feed_identities(raw_msg, copy=False)
         except Exception, e:
             logger.error("task::Invaid msg: %s"%msg)
             return
         
         msg = self.session.unpack_message(msg, content=False, copy=False)
-        print idents,msg
         header = msg['header']
         msg_id = header['msg_id']
         after = Dependency(header.get('after', []))
         if after.mode == 'all':
             after.difference_update(self.all_done)
         if after.check(self.all_done):
             # recast as empty set, if we are already met, 
             # to prevent 
             after = Dependency([])
         
         follow = Dependency(header.get('follow', []))
-        print raw_msg
         if len(after) == 0:
             # time deps already met, try to run
             if not self.maybe_run(msg_id, raw_msg, follow):
                 # can't run yet
                 self.save_unmet(msg_id, raw_msg, after, follow)
         else:
             self.save_unmet(msg_id, raw_msg, after, follow)
         # send to monitor
         self.mon_stream.send_multipart(['intask']+raw_msg, copy=False)
     @logged
     def maybe_run(self, msg_id, raw_msg, follow=None):
         """check location dependencies, and run if they are met."""
             
         if follow:
             def can_run(idx):
                 target = self.targets[idx]
                 return target not in self.blacklist.get(msg_id, []) and\
                         follow.check(self.completed[target])
             
             indices = filter(can_run, range(len(self.targets)))
             if not indices:
                 return False
         else:
             indices = None
             
         self.submit_task(msg_id, raw_msg, indices)
         return True
             
     @logged
     def save_unmet(self, msg_id, msg, after, follow):
         """Save a message for later submission when its dependencies are met."""
         self.depending[msg_id] = (msg_id,msg,after,follow)
         # track the ids in both follow/after, but not those already completed
         for dep_id in after.union(follow).difference(self.all_done):
+            print dep_id
             if dep_id not in self.dependencies:
                 self.dependencies[dep_id] = set()
             self.dependencies[dep_id].add(msg_id)
+    
     @logged
     def submit_task(self, msg_id, msg, follow=None, 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])
-        self.engine_stream.socket.send(target, flags=zmq.SNDMORE, copy=False)
-        self.engine_stream.socket.send_multipart(msg, copy=False)
+        self.engine_stream.send(target, flags=zmq.SNDMORE, copy=False)
+        self.engine_stream.send_multipart(msg, copy=False)
         self.add_job(idx)
         self.pending[target][msg_id] = (msg, follow)
     
     #-----------------------------------------------------------------------
     # Result Handling
     #-----------------------------------------------------------------------
     @logged
     def dispatch_result(self, raw_msg):
         try:
             idents,msg = self.session.feed_identities(raw_msg, copy=False)
         except Exception, e:
             logger.error("task::Invaid result: %s"%msg)
             return
         msg = self.session.unpack_message(msg, content=False, copy=False)
         header = msg['header']
         if header.get('dependencies_met', True):
             self.handle_result_success(idents, msg['parent_header'], raw_msg)
             # send to monitor
             self.mon_stream.send_multipart(['outtask']+raw_msg, copy=False)
         else:
-            self.handle_unmet_dependency(self, idents, msg['parent_header'])
+            self.handle_unmet_dependency(idents, msg['parent_header'])
         
     @logged
     def handle_result_success(self, idents, parent, raw_msg):
         # 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.pending[engine].pop(msg_id)
         self.completed[engine].add(msg_id)
         
         self.update_dependencies(msg_id)
         
     @logged
     def handle_unmet_dependency(self, idents, parent):
         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)
         raw_msg,follow = self.pending[engine].pop(msg_id)
-        if not self.maybe_run(raw_msg, follow):
+        if not self.maybe_run(msg_id, raw_msg, follow):
             # resubmit failed, put it back in our dependency tree
             self.save_unmet(msg_id, raw_msg, Dependency(), follow)
         pass
     @logged
     def update_dependencies(self, dep_id):
         """dep_id just finished. Update our dependency
         table and submit any jobs that just became runable."""
         if dep_id not in self.dependencies:
             return
         jobs = self.dependencies.pop(dep_id)
         for job in jobs:
             msg_id, raw_msg, after, follow = self.depending[job]
             if msg_id in after:
                 after.remove(msg_id)
             if not after: # time deps met
                 if self.maybe_run(msg_id, raw_msg, follow):
                     self.depending.pop(job)
                     for mid in follow:
-                        self.dependencies[mid].remove(msg_id)
+                        if mid in self.dependencies:
+                            self.dependencies[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, scheme='weighted'):
     from zmq.eventloop import ioloop
     from zmq.eventloop.zmqstream import ZMQStream
     
     ctx = zmq.Context()
     loop = ioloop.IOLoop()
     
     scheme = globals().get(scheme)
     
     ins = ZMQStream(ctx.socket(zmq.XREP),loop)
     ins.bind(in_addr)
     outs = ZMQStream(ctx.socket(zmq.XREP),loop)
     outs.bind(out_addr)
     mons = ZMQStream(ctx.socket(zmq.PUB),loop)
     mons.connect(mon_addr)
     nots = ZMQStream(ctx.socket(zmq.SUB),loop)
     nots.setsockopt(zmq.SUBSCRIBE, '')
     nots.connect(not_addr)
     
     scheduler = TaskScheduler(ins,outs,mons,nots,scheme,loop)
     
     loop.start()
 
 
 if __name__ == '__main__':
     iface = 'tcp://127.0.0.1:%i'
     launch_scheduler(iface%12345,iface%1236,iface%12347,iface%12348)

IPython/zmq/parallel/streamsession.py

@@ -1,498 +1,499 b''
 #!/usr/bin/env python
 """edited session.py to work with streams, and move msg_type to the header
 """
 
 
 import os
 import sys
 import traceback
 import pprint
 import uuid
 
 import zmq
 from zmq.utils import jsonapi
 from zmq.eventloop.zmqstream import ZMQStream
 
 from IPython.zmq.pickleutil import can, uncan, canSequence, uncanSequence
 from IPython.zmq.newserialized import serialize, unserialize
 
 try:
     import cPickle
     pickle = cPickle
 except:
     cPickle = None
     import pickle
 
 # packer priority: jsonlib[2], cPickle, simplejson/json, pickle
 json_name = '' if not jsonapi.jsonmod else jsonapi.jsonmod.__name__
 if json_name in ('jsonlib', 'jsonlib2'):
     use_json = True
 elif json_name:
     if cPickle is None:
         use_json = True
     else:
         use_json = False
 else:
     use_json = False
 
 def squash_unicode(obj):
     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
 
 if use_json:
     default_packer = jsonapi.dumps
     default_unpacker = lambda s: squash_unicode(jsonapi.loads(s))
 else:
     default_packer = lambda o: pickle.dumps(o,-1)
     default_unpacker = pickle.loads
 
 
 DELIM="<IDS|MSG>"
 
 def wrap_exception():
     etype, evalue, tb = sys.exc_info()
     tb = traceback.format_exception(etype, evalue, tb)
     exc_content = {
         u'status' : u'error',
         u'traceback' : tb,
         u'etype' : unicode(etype),
         u'evalue' : unicode(evalue)
     }
     return exc_content
 
 class KernelError(Exception):
     pass
 
 def unwrap_exception(content):
     err = KernelError(content['etype'], content['evalue'])
     err.evalue = content['evalue']
     err.etype = content['etype']
     err.traceback = ''.join(content['traceback'])
     return err
     
 
 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):
     return locals()
     # return {
     #     'msg_id' : msg_id,
     #     'msg_type': msg_type,
     #     'username' : username,
     #     'session' : session
     # }
 
 
 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
 
 def rekey(dikt):
     """rekey a dict that has been forced to use str keys where there should be
     ints by json.  This belongs in the jsonutil added by fperez."""
     for k in dikt.iterkeys():
         if isinstance(k, str):
             ik=fk=None
             try:
                 ik = int(k)
             except ValueError:
                 try:
                     fk = float(k)
                 except ValueError:
                     continue
             if ik is not None:
                 nk = ik
             else:
                 nk = fk
             if nk in dikt:
                 raise KeyError("already have key %r"%nk)
             dikt[nk] = dikt.pop(k)
     return dikt
 
 def serialize_object(obj, threshold=64e-6):
     """serialize an object into a list of sendable buffers.
     
     Returns: (pmd, bufs)
         where pmd is the pickled metadata wrapper, and bufs
         is a list of data buffers"""
     # threshold is 100 B
     databuffers = []
     if isinstance(obj, (list, tuple)):
         clist = canSequence(obj)
         slist = map(serialize, clist)
         for s in slist:
             if s.typeDescriptor in ('buffer', 'ndarray') or s.getDataSize() > threshold:
                 databuffers.append(s.getData())
                 s.data = None
         return pickle.dumps(slist,-1), databuffers
     elif isinstance(obj, dict):
         sobj = {}
         for k in sorted(obj.iterkeys()):
             s = serialize(can(obj[k]))
             if s.getDataSize() > threshold:
                 databuffers.append(s.getData())
                 s.data = None
             sobj[k] = s
         return pickle.dumps(sobj,-1),databuffers
     else:
         s = serialize(can(obj))
         if s.getDataSize() > threshold:
             databuffers.append(s.getData())
             s.data = None
         return pickle.dumps(s,-1),databuffers
             
         
 def unserialize_object(bufs):
     """reconstruct an object serialized by serialize_object from data buffers"""
     bufs = list(bufs)
     sobj = pickle.loads(bufs.pop(0))
     if isinstance(sobj, (list, tuple)):
         for s in sobj:
             if s.data is None:
                 s.data = bufs.pop(0)
         return uncanSequence(map(unserialize, sobj))
     elif isinstance(sobj, dict):
         newobj = {}
         for k in sorted(sobj.iterkeys()):
             s = sobj[k]
             if s.data is None:
                 s.data = bufs.pop(0)
             newobj[k] = uncan(unserialize(s))
         return newobj
     else:
         if sobj.data is None:
             sobj.data = bufs.pop(0)
         return uncan(unserialize(sobj))
 
 def pack_apply_message(f, args, kwargs, threshold=64e-6):
     """pack up a function, args, and kwargs to be sent over the wire
     as a series of buffers. Any object whose data is larger than `threshold`
     will not have their data copied (currently only numpy arrays support zero-copy)"""
     msg = [pickle.dumps(can(f),-1)]
     databuffers = [] # for large objects
     sargs, bufs = serialize_object(args,threshold)
     msg.append(sargs)
     databuffers.extend(bufs)
     skwargs, bufs = serialize_object(kwargs,threshold)
     msg.append(skwargs)
     databuffers.extend(bufs)
     msg.extend(databuffers)
     return msg
 
 def unpack_apply_message(bufs, g=None, copy=True):
     """unpack f,args,kwargs from buffers packed by pack_apply_message()
     Returns: original f,args,kwargs"""
     bufs = list(bufs) # allow us to pop
     assert len(bufs) >= 3, "not enough buffers!"
     if not copy:
         for i in range(3):
             bufs[i] = bufs[i].bytes
     cf = pickle.loads(bufs.pop(0))
     sargs = list(pickle.loads(bufs.pop(0)))
     skwargs = dict(pickle.loads(bufs.pop(0)))
     # print sargs, skwargs
     f = uncan(cf, g)
     for sa in sargs:
         if sa.data is None:
             m = bufs.pop(0)
             if sa.getTypeDescriptor() in ('buffer', 'ndarray'):
                 if copy:
                     sa.data = buffer(m)
                 else:
                     sa.data = m.buffer
             else:
                 if copy:
                     sa.data = m
                 else:
                     sa.data = m.bytes
     
     args = uncanSequence(map(unserialize, sargs), g)
     kwargs = {}
     for k in sorted(skwargs.iterkeys()):
         sa = skwargs[k]
         if sa.data is None:
             sa.data = bufs.pop(0)
         kwargs[k] = uncan(unserialize(sa), g)
     
     return f,args,kwargs
 
 class StreamSession(object):
     """tweaked version of IPython.zmq.session.Session, for development in Parallel"""
     debug=False
     def __init__(self, username=None, session=None, packer=None, unpacker=None):
         if username is None:
             username = os.environ.get('USER','username')
         self.username = username
         if session is None:
             self.session = str(uuid.uuid4())
         else:
             self.session = session
         self.msg_id = str(uuid.uuid4())
         if packer is None:
             self.pack = default_packer
         else:
             if not callable(packer):
                 raise TypeError("packer must be callable, not %s"%type(packer))
             self.pack = packer
         
         if unpacker is None:
             self.unpack = default_unpacker
         else:
             if not callable(unpacker):
                 raise TypeError("unpacker must be callable, not %s"%type(unpacker))
             self.unpack = unpacker
         
         self.none = self.pack({})
             
     def msg_header(self, msg_type):
         h = msg_header(self.msg_id, msg_type, self.username, self.session)
         self.msg_id = str(uuid.uuid4())
         return h
 
     def msg(self, msg_type, content=None, parent=None, subheader=None):
         msg = {}
         msg['header'] = self.msg_header(msg_type)
         msg['msg_id'] = msg['header']['msg_id']
         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 send(self, stream, msg_type, content=None, buffers=None, parent=None, subheader=None, ident=None):
         """Build and send a message via stream or socket.
         
         Parameters
         ----------
         
         msg_type : str or Message/dict
             Normally, msg_type will be 
             
             
         
         Returns
         -------
         (msg,sent) : tuple
             msg : Message
                 the nice wrapped dict-like object containing the headers
             
         """
         if isinstance(msg_type, (Message, dict)):
             # we got a Message, not a msg_type
             # don't build a new Message
             msg = msg_type
             content = msg['content']
         else:
             msg = self.msg(msg_type, content, parent, subheader)
         buffers = [] if buffers is None else buffers
         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)
         to_send.append(self.pack(msg['header']))
         to_send.append(self.pack(msg['parent_header']))
         # if parent is None:
         #     to_send.append(self.none)
         # else:
         #     to_send.append(self.pack(dict(parent)))
         if content is None:
             content = self.none
         elif isinstance(content, dict):
             content = self.pack(content)
         elif isinstance(content, str):
             # content is already packed, as in a relayed message
             pass
         else:
             raise TypeError("Content incorrect type: %s"%type(content))
         to_send.append(content)
         flag = 0
         if buffers:
             flag = zmq.SNDMORE
         stream.send_multipart(to_send, flag, copy=False)
         for b in buffers[:-1]:
             stream.send(b, flag, copy=False)
         if buffers:
             stream.send(buffers[-1], copy=False)
         omsg = Message(msg)
         if self.debug:
             pprint.pprint(omsg)
             pprint.pprint(to_send)
             pprint.pprint(buffers)
         # return both the msg object and the buffers
         return omsg
     
     def send_raw(self, stream, msg, flags=0, copy=True, idents=None):
         """send a raw message via idents.
         
         Parameters
         ----------
         msg : list of sendable buffers"""
         to_send = []
         if isinstance(ident, str):
             ident = [ident]
         if ident is not None:
             to_send.extend(ident)
         to_send.append(DELIM)
         to_send.extend(msg)
         stream.send_multipart(msg, flags, copy=copy)
     
     def recv(self, socket, mode=zmq.NOBLOCK, content=True, copy=True):
         """receives and unpacks a message
         returns [idents], msg"""
         if isinstance(socket, ZMQStream):
             socket = socket.socket
         try:
             msg = socket.recv_multipart(mode)
         except zmq.ZMQError, e:
             if e.errno == zmq.EAGAIN:
                 # We can convert EAGAIN to None as we know in this case
                 # recv_json won't return None.
                 return None
             else:
                 raise
         # return an actual Message object
         # determine the number of idents by trying to unpack them.
         # this is terrible:
         idents, msg = self.feed_identities(msg, copy)
         try:
             return idents, self.unpack_message(msg, content=content, copy=copy)
         except Exception, e:
             print idents, msg
             # TODO: handle it
             raise e
     
     def feed_identities(self, msg, copy=True):
         """This is a completely horrible thing, but it strips the zmq
         ident prefixes off of a message. It will break if any identities
         are unpackable by self.unpack."""
         msg = list(msg)
         idents = []
         while len(msg) > 3:
             if copy:
                 s = msg[0]
             else:
                 s = msg[0].bytes
             if s == DELIM:
                 msg.pop(0)
                 break
             else:
                 idents.append(s)
                 msg.pop(0)
                 
         return idents, msg
     
     def unpack_message(self, msg, content=True, copy=True):
-        """return a message object from the format
+        """Return a message object from the format
         sent by self.send.
         
-        parameters:
+        Parameters:
+        -----------
         
         content : bool (True)
             whether to unpack the content dict (True), 
             or leave it serialized (False)
         
         copy : bool (True)
             whether to return the bytes (True), 
             or the non-copying Message object in each place (False)
         
         """
         if not len(msg) >= 3:
             raise TypeError("malformed message, must have at least 3 elements")
         message = {}
         if not copy:
             for i in range(3):
                 msg[i] = msg[i].bytes
         message['header'] = self.unpack(msg[0])
         message['msg_type'] = message['header']['msg_type']
         message['parent_header'] = self.unpack(msg[1])
         if content:
             message['content'] = self.unpack(msg[2])
         else:
             message['content'] = msg[2]
     
         # message['buffers'] = msg[3:]
         # else:
         #     message['header'] = self.unpack(msg[0].bytes)
         #     message['msg_type'] = message['header']['msg_type']
         #     message['parent_header'] = self.unpack(msg[1].bytes)
         #     if content:
         #         message['content'] = self.unpack(msg[2].bytes)
         #     else:
         #         message['content'] = msg[2].bytes
         
         message['buffers'] = msg[3:]# [ m.buffer for m in msg[3:] ]
         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']