resubmitted tasks are now wholly separate (new msg_ids)

This removes much of the fragility due to trying to cram a resubmit into the original ID.  The `resubmit` record is now the resubmitted task ID, rather than the timestamp.

IPython/parallel/client/client.py

@@ -1,1570 +1,1566 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
 from threading import Thread, Event
 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.py3compat import cast_bytes
 from IPython.utils.traitlets import (HasTraits, Integer, Instance, Unicode,
                                     Dict, List, Bool, Set, Any)
 from IPython.external.decorator import decorator
 from IPython.external.ssh import tunnel
 
 from IPython.parallel import Reference
 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' : '',
               'outputs' : [],
             }
         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 ssh private 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)
     _spin_thread = Any()
     _stop_spinning = Any()
 
     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=Integer(0)
     _ignored_hub_replies=Integer(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._stop_spinning = Event()
 
         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."
 
         if not util.is_url(url_or_file):
             # it's not a url, try for a file
             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 = cast_bytes(exec_key)
                 extra_args['key'] = exec_key
         self.session = Session(**extra_args)
 
         self._query_socket = self._context.socket(zmq.DEALER)
         self._query_socket.setsockopt(zmq.IDENTITY, self.session.bsession)
         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 [cast_bytes(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 = self.session.bsession
             if content.mux:
                 self._mux_socket = self._context.socket(zmq.DEALER)
                 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.DEALER)
                 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.DEALER)
             #     self._query_socket.setsockopt(zmq.IDENTITY, self.session.bsession)
             #     connect_socket(self._query_socket, content.query)
             if content.control:
                 self._control_socket = self._context.socket(zmq.DEALER)
                 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)
 
         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] = content
         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 _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['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['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']
             # ignore IOPub messages with no parent.
             # Caused by print statements or warnings from before the first execution.
             if not parent:
                 continue
             msg_id = parent['msg_id']
             content = msg['content']
             header = msg['header']
             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']})
             elif msg_type == 'display_data':
                 md['outputs'].append(content.get('data'))
             elif msg_type == 'pyout':
                 md['pyout'] = content.get('data')
             else:
                 # unhandled msg_type (status, etc.)
                 pass
 
             # 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
         self.stop_spin_thread()
         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_every(self, interval=1):
         """target func for use in spin_thread"""
         while True:
             if self._stop_spinning.is_set():
                 return
             time.sleep(interval)
             self.spin()
 
     def spin_thread(self, interval=1):
         """call Client.spin() in a background thread on some regular interval
         
         This helps ensure that messages don't pile up too much in the zmq queue
         while you are working on other things, or just leaving an idle terminal.
         
         It also helps limit potential padding of the `received` timestamp
         on AsyncResult objects, used for timings.
         
         Parameters
         ----------
         
         interval : float, optional
             The interval on which to spin the client in the background thread
             (simply passed to time.sleep).
         
         Notes
         -----
         
         For precision timing, you may want to use this method to put a bound
         on the jitter (in seconds) in `received` timestamps used
         in AsyncResult.wall_time.
         
         """
         if self._spin_thread is not None:
             self.stop_spin_thread()
         self._stop_spinning.clear()
         self._spin_thread = Thread(target=self._spin_every, args=(interval,))
         self._spin_thread.daemon = True
         self._spin_thread.start()
     
     def stop_spin_thread(self):
         """stop background spin_thread, if any"""
         if self._spin_thread is not None:
             self._stop_spinning.set()
             self._spin_thread.join()
             self._spin_thread = None
 
     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
             
             If unspecified/None: abort all outstanding jobs.
 
         """
         block = self.block if block is None else block
         jobs = jobs if jobs is not None else list(self.outstanding)
         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_request(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) and not isinstance(f, Reference):
             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['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
 
     def send_execute_request(self, socket, code, silent=True, subheader=None, ident=None):
         """construct and send an execute request via a socket.
 
         """
 
         assert not self._closed, "cannot use me anymore, I'm closed!"
         # defaults:
         subheader = subheader if subheader is not None else {}
 
         # validate arguments
         if not isinstance(code, basestring):
             raise TypeError("code must be text, not %s" % type(code))
         if not isinstance(subheader, dict):
             raise TypeError("subheader must be dict, not %s" % type(subheader))
         
         content = dict(code=code, silent=bool(silent), user_variables=[], user_expressions={})
 
 
         msg = self.session.send(socket, "execute_request", content=content, ident=ident,
                             subheader=subheader)
 
         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.
         
         rc.direct_view('all') is distinguished from rc[:] in that 'all' will
         evaluate the target engines at each execution, whereas rc[:] will connect to
         all *current* engines, and that list will not change.
         
         That is, 'all' will always use all engines, whereas rc[:] will not use
         engines added after the DirectView is constructed.
 
         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)
+        mapping = content['resubmitted']
+        new_ids = [ mapping[msg_id] for msg_id in theids ]
 
-        ar = AsyncHubResult(self, msg_ids=theids)
+        ar = AsyncHubResult(self, msg_ids=new_ids)
 
         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))
                 if rec.get('received'):
                     md['received'] = rec['received']
                 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
         """
         if targets == 'all':
             # allow 'all' to be evaluated on the engine
             engine_ids = None
         else:
             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/controller/hub.py

@@ -1,1304 +1,1310 b''
 """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) 2010-2011  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 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.py3compat import cast_bytes
 from IPython.utils.traitlets import (
         HasTraits, Instance, Integer, 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,
         'received': 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,
         'received': 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=Integer(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(Integer,Integer,config=True,
         help="""XREQ/SUB Port pair for Engine heartbeats""")
     def _hb_default(self):
         return tuple(util.select_random_ports(2))
 
     mux = Tuple(Integer,Integer,config=True,
         help="""Engine/Client Port pair for MUX queue""")
 
     def _mux_default(self):
         return tuple(util.select_random_ports(2))
 
     task = Tuple(Integer,Integer,config=True,
         help="""Engine/Client Port pair for Task queue""")
     def _task_default(self):
         return tuple(util.select_random_ports(2))
 
     control = Tuple(Integer,Integer,config=True,
         help="""Engine/Client Port pair for Control queue""")
 
     def _control_default(self):
         return tuple(util.select_random_ports(2))
 
     iopub = Tuple(Integer,Integer,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 = Integer(config=True,
         help="""Monitor (SUB) port for queue traffic""")
 
     def _mon_port_default(self):
         return util.select_random_ports(1)[0]
 
     notifier_port = Integer(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.ROUTER), 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.ROUTER)
         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.DEALER), loop)
         url = util.disambiguate_url(self.client_info['task'][-1])
         r.setsockopt(zmq.IDENTITY, self.session.bsession)
         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=Integer()
     _idcounter=Integer(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
             return 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(cast_bytes(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)
     #-----------------------------------------------------------------------------
 
 
     @util.log_errors
     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[0])
         switch = msg[0]
         try:
             idents, msg = self.session.feed_identities(msg[1:])
         except ValueError:
             idents=[]
         if not idents:
             self.log.error("Monitor message without topic: %r", msg)
             return
         handler = self.monitor_handlers.get(switch, None)
         if handler is not None:
             handler(idents, msg)
         else:
             self.log.error("Unrecognized monitor topic: %r", switch)
 
 
     @util.log_errors
     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.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['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 or self.keytable[eid] in self.dead_engines:
             self.log.info("heartbeat::ignoring heart failure %r (not an engine or already dead)", 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.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']
         self.log.info("queue::client %r submitted request %r to %s", client_id, msg_id, eid)
         # 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.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)
             self.log.info("queue::request %r completed on %s", msg_id, eid)
         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'],
             'received': datetime.now(),
             '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.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.decode('ascii')
         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.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', u'')
         eid = self.by_ident.get(cast_bytes(engine_uuid), None)
         
         status = header.get('status', None)
 
         if msg_id in self.pending:
             self.log.info("task::task %r finished on %s", msg_id, eid)
             self.pending.remove(msg_id)
             self.all_completed.add(msg_id)
             if eid is not None:
                 if status != 'aborted':
                     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,
                 'received' : datetime.now(),
                 '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.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[cast_bytes(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.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['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 = cast_bytes(content['queue'])
         except KeyError:
             self.log.error("registration::queue not specified", exc_info=True)
             return
         heart = content.get('heartbeat', None)
         if heart:
             heart = cast_bytes(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)
+        pending_ids = [ msg_id for msg_id in found_ids if msg_id in self.pending ]
         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]
+        elif pending_ids:
+            pass
+            # no need to raise on resubmit of pending task, now that we
+            # resubmit under new ID, but do we want to raise anyway?
+            # msg_id = invalid_ids[0]
+            # try:
+            #     raise ValueError("Task(s) %r appears to be inflight" % )
+            # except Exception:
+            #     return finish(error.wrap_exception())
+
+        # mapping of original IDs to resubmitted IDs
+        resubmitted = {}
+
+        # send the messages
+        for rec in records:
+            header = rec['header']
+            msg = self.session.msg(header['msg_type'])
+            msg_id = msg['msg_id']
+            msg['content'] = rec['content']
+            header.update(msg['header'])
+            msg['header'] = header
+
+            self.session.send(self.resubmit, msg, buffers=rec['buffers'])
+
+            resubmitted[rec['msg_id']] = msg_id
+            self.pending.add(msg_id)
+            msg['buffers'] = []
             try:
-                raise ValueError("Task %r appears to be inflight" % msg_id)
+                self.db.add_record(msg_id, init_record(msg))
             except Exception:
-                return finish(error.wrap_exception())
+                self.log.error("db::DB Error updating record: %s", msg_id, exc_info=True)
 
-        # 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['header']['msg_id'] = rec['msg_id']
-                self.session.send(self.resubmit, msg, buffers=rec['buffers'])
-
-        finish(dict(status='ok'))
+        finish(dict(status='ok', resubmitted=resubmitted))
+        
+        # store the new IDs in the Task DB
+        for msg_id, resubmit_id in resubmitted.iteritems():
+            try:
+                self.db.update_record(msg_id, {'resubmitted' : resubmit_id})
+            except Exception:
+                self.log.error("db::DB Error updating record: %s", msg_id, exc_info=True)
 
 
     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'):
                 io_dict[key] = rec[key]
         content = { 'result_content': rec['result_content'],
                             'header': rec['header'],
                             'result_header' : rec['result_header'],
                             'received' : rec['received'],
                             '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 = None
                 result_buffer_lens = None
 
             for rec in records:
                 # buffers may be None, so double check
                 b = rec.pop('buffers', empty) or empty
                 if buffer_lens is not None:
                     buffer_lens.append(len(b))
                     buffers.extend(b)
                 rb = rec.pop('result_buffers', empty) or empty
                 if result_buffer_lens is not None:
                     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/sqlitedb.py

@@ -1,412 +1,412 b''
 """A TaskRecord backend using sqlite3
 
 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.
 #-----------------------------------------------------------------------------
 
 import json
 import os
 import cPickle as pickle
 from datetime import datetime
 
 try:
     import sqlite3
 except ImportError:
     sqlite3 = None
 
 from zmq.eventloop import ioloop
 
 from IPython.utils.traitlets import Unicode, Instance, List, Dict
 from .dictdb import BaseDB
 from IPython.utils.jsonutil import date_default, extract_dates, squash_dates
 
 #-----------------------------------------------------------------------------
 # SQLite operators, adapters, and converters
 #-----------------------------------------------------------------------------
 
 try:
     buffer
 except NameError:
     # py3k
     buffer = memoryview
 
 operators = {
  '$lt' : "<",
  '$gt' : ">",
  # null is handled weird with ==,!=
  '$eq' : "=",
  '$ne' : "!=",
  '$lte': "<=",
  '$gte': ">=",
  '$in' : ('=', ' OR '),
  '$nin': ('!=', ' AND '),
  # '$all': None,
  # '$mod': None,
  # '$exists' : None
 }
 null_operators = {
 '=' : "IS NULL",
 '!=' : "IS NOT NULL",
 }
 
 def _adapt_dict(d):
     return json.dumps(d, default=date_default)
 
 def _convert_dict(ds):
     if ds is None:
         return ds
     else:
         if isinstance(ds, bytes):
             # If I understand the sqlite doc correctly, this will always be utf8
             ds = ds.decode('utf8')
         return extract_dates(json.loads(ds))
 
 def _adapt_bufs(bufs):
     # this is *horrible*
     # copy buffers into single list and pickle it:
     if bufs and isinstance(bufs[0], (bytes, buffer)):
         return sqlite3.Binary(pickle.dumps(map(bytes, bufs),-1))
     elif bufs:
         return bufs
     else:
         return None
 
 def _convert_bufs(bs):
     if bs is None:
         return []
     else:
         return pickle.loads(bytes(bs))
 
 #-----------------------------------------------------------------------------
 # SQLiteDB class
 #-----------------------------------------------------------------------------
 
 class SQLiteDB(BaseDB):
     """SQLite3 TaskRecord backend."""
 
     filename = Unicode('tasks.db', config=True,
         help="""The filename of the sqlite task database. [default: 'tasks.db']""")
     location = Unicode('', config=True,
         help="""The directory containing the sqlite task database.  The default
         is to use the cluster_dir location.""")
     table = Unicode("", config=True,
         help="""The SQLite Table to use for storing tasks for this session. If unspecified,
         a new table will be created with the Hub's IDENT.  Specifying the table will result
         in tasks from previous sessions being available via Clients' db_query and
         get_result methods.""")
 
     if sqlite3 is not None:
         _db = Instance('sqlite3.Connection')
     else:
         _db = None
     # the ordered list of column names
     _keys = List(['msg_id' ,
             'header' ,
             'content',
             'buffers',
             'submitted',
             'client_uuid' ,
             'engine_uuid' ,
             'started',
             'completed',
             'resubmitted',
             'received',
             'result_header' ,
             'result_content' ,
             'result_buffers' ,
             'queue' ,
             'pyin' ,
             'pyout',
             'pyerr',
             'stdout',
             'stderr',
         ])
     # sqlite datatypes for checking that db is current format
     _types = Dict({'msg_id' : 'text' ,
             'header' : 'dict text',
             'content' : 'dict text',
             'buffers' : 'bufs blob',
             'submitted' : 'timestamp',
             'client_uuid' : 'text',
             'engine_uuid' : 'text',
             'started' : 'timestamp',
             'completed' : 'timestamp',
-            'resubmitted' : 'timestamp',
+            'resubmitted' : 'text',
             'received' : 'timestamp',
             'result_header' : 'dict text',
             'result_content' : 'dict text',
             'result_buffers' : 'bufs blob',
             'queue' : 'text',
             'pyin' : 'text',
             'pyout' : 'text',
             'pyerr' : 'text',
             'stdout' : 'text',
             'stderr' : 'text',
         })
 
     def __init__(self, **kwargs):
         super(SQLiteDB, self).__init__(**kwargs)
         if sqlite3 is None:
             raise ImportError("SQLiteDB requires sqlite3")
         if not self.table:
             # use session, and prefix _, since starting with # is illegal
             self.table = '_'+self.session.replace('-','_')
         if not self.location:
             # get current profile
             from IPython.core.application import BaseIPythonApplication
             if BaseIPythonApplication.initialized():
                 app = BaseIPythonApplication.instance()
                 if app.profile_dir is not None:
                     self.location = app.profile_dir.location
                 else:
                     self.location = u'.'
             else:
                 self.location = u'.'
         self._init_db()
 
         # register db commit as 2s periodic callback
         # to prevent clogging pipes
         # assumes we are being run in a zmq ioloop app
         loop = ioloop.IOLoop.instance()
         pc = ioloop.PeriodicCallback(self._db.commit, 2000, loop)
         pc.start()
 
     def _defaults(self, keys=None):
         """create an empty record"""
         d = {}
         keys = self._keys if keys is None else keys
         for key in keys:
             d[key] = None
         return d
 
     def _check_table(self):
         """Ensure that an incorrect table doesn't exist
 
         If a bad (old) table does exist, return False
         """
         cursor = self._db.execute("PRAGMA table_info(%s)"%self.table)
         lines = cursor.fetchall()
         if not lines:
             # table does not exist
             return True
         types = {}
         keys = []
         for line in lines:
             keys.append(line[1])
             types[line[1]] = line[2]
         if self._keys != keys:
             # key mismatch
             self.log.warn('keys mismatch')
             return False
         for key in self._keys:
             if types[key] != self._types[key]:
                 self.log.warn(
                     'type mismatch: %s: %s != %s'%(key,types[key],self._types[key])
                 )
                 return False
         return True
 
     def _init_db(self):
         """Connect to the database and get new session number."""
         # register adapters
         sqlite3.register_adapter(dict, _adapt_dict)
         sqlite3.register_converter('dict', _convert_dict)
         sqlite3.register_adapter(list, _adapt_bufs)
         sqlite3.register_converter('bufs', _convert_bufs)
         # connect to the db
         dbfile = os.path.join(self.location, self.filename)
         self._db = sqlite3.connect(dbfile, detect_types=sqlite3.PARSE_DECLTYPES,
             # isolation_level = None)#,
              cached_statements=64)
         # print dir(self._db)
         first_table = previous_table = self.table
         i=0
         while not self._check_table():
             i+=1
             self.table = first_table+'_%i'%i
             self.log.warn(
                 "Table %s exists and doesn't match db format, trying %s"%
                 (previous_table, self.table)
             )
             previous_table = self.table
 
         self._db.execute("""CREATE TABLE IF NOT EXISTS %s
                 (msg_id text PRIMARY KEY,
                 header dict text,
                 content dict text,
                 buffers bufs blob,
                 submitted timestamp,
                 client_uuid text,
                 engine_uuid text,
                 started timestamp,
                 completed timestamp,
-                resubmitted timestamp,
+                resubmitted text,
                 received timestamp,
                 result_header dict text,
                 result_content dict text,
                 result_buffers bufs blob,
                 queue text,
                 pyin text,
                 pyout text,
                 pyerr text,
                 stdout text,
                 stderr text)
                 """%self.table)
         self._db.commit()
 
     def _dict_to_list(self, d):
         """turn a mongodb-style record dict into a list."""
 
         return [ d[key] for key in self._keys ]
 
     def _list_to_dict(self, line, keys=None):
         """Inverse of dict_to_list"""
         keys = self._keys if keys is None else keys
         d = self._defaults(keys)
         for key,value in zip(keys, line):
             d[key] = value
 
         return d
 
     def _render_expression(self, check):
         """Turn a mongodb-style search dict into an SQL query."""
         expressions = []
         args = []
 
         skeys = set(check.keys())
         skeys.difference_update(set(self._keys))
         skeys.difference_update(set(['buffers', 'result_buffers']))
         if skeys:
             raise KeyError("Illegal testing key(s): %s"%skeys)
 
         for name,sub_check in check.iteritems():
             if isinstance(sub_check, dict):
                 for test,value in sub_check.iteritems():
                     try:
                         op = operators[test]
                     except KeyError:
                         raise KeyError("Unsupported operator: %r"%test)
                     if isinstance(op, tuple):
                         op, join = op
 
                     if value is None and op in null_operators:
                         expr = "%s %s" % (name, null_operators[op])
                     else:
                         expr = "%s %s ?"%(name, op)
                         if isinstance(value, (tuple,list)):
                             if op in null_operators and any([v is None for v in value]):
                                 # equality tests don't work with NULL
                                 raise ValueError("Cannot use %r test with NULL values on SQLite backend"%test)
                             expr = '( %s )'%( join.join([expr]*len(value)) )
                             args.extend(value)
                         else:
                             args.append(value)
                     expressions.append(expr)
             else:
                 # it's an equality check
                 if sub_check is None:
                     expressions.append("%s IS NULL" % name)
                 else:
                     expressions.append("%s = ?"%name)
                     args.append(sub_check)
 
         expr = " AND ".join(expressions)
         return expr, args
 
     def add_record(self, msg_id, rec):
         """Add a new Task Record, by msg_id."""
         d = self._defaults()
         d.update(rec)
         d['msg_id'] = msg_id
         line = self._dict_to_list(d)
         tups = '(%s)'%(','.join(['?']*len(line)))
         self._db.execute("INSERT INTO %s VALUES %s"%(self.table, tups), line)
         # self._db.commit()
 
     def get_record(self, msg_id):
         """Get a specific Task Record, by msg_id."""
         cursor = self._db.execute("""SELECT * FROM %s WHERE msg_id==?"""%self.table, (msg_id,))
         line = cursor.fetchone()
         if line is None:
             raise KeyError("No such msg: %r"%msg_id)
         return self._list_to_dict(line)
 
     def update_record(self, msg_id, rec):
         """Update the data in an existing record."""
         query = "UPDATE %s SET "%self.table
         sets = []
         keys = sorted(rec.keys())
         values = []
         for key in keys:
             sets.append('%s = ?'%key)
             values.append(rec[key])
         query += ', '.join(sets)
         query += ' WHERE msg_id == ?'
         values.append(msg_id)
         self._db.execute(query, values)
         # self._db.commit()
 
     def drop_record(self, msg_id):
         """Remove a record from the DB."""
         self._db.execute("""DELETE FROM %s WHERE msg_id==?"""%self.table, (msg_id,))
         # self._db.commit()
 
     def drop_matching_records(self, check):
         """Remove a record from the DB."""
         expr,args = self._render_expression(check)
         query = "DELETE FROM %s WHERE %s"%(self.table, expr)
         self._db.execute(query,args)
         # self._db.commit()
 
     def find_records(self, check, keys=None):
         """Find records matching a query dict, optionally extracting subset of keys.
 
         Returns list of matching records.
 
         Parameters
         ----------
 
         check: dict
             mongodb-style query argument
         keys: list of strs [optional]
             if specified, the subset of keys to extract.  msg_id will *always* be
             included.
         """
         if keys:
             bad_keys = [ key for key in keys if key not in self._keys ]
             if bad_keys:
                 raise KeyError("Bad record key(s): %s"%bad_keys)
 
         if keys:
             # ensure msg_id is present and first:
             if 'msg_id' in keys:
                 keys.remove('msg_id')
             keys.insert(0, 'msg_id')
             req = ', '.join(keys)
         else:
             req = '*'
         expr,args = self._render_expression(check)
         query = """SELECT %s FROM %s WHERE %s"""%(req, self.table, expr)
         cursor = self._db.execute(query, args)
         matches = cursor.fetchall()
         records = []
         for line in matches:
             rec = self._list_to_dict(line, keys)
             records.append(rec)
         return records
 
     def get_history(self):
         """get all msg_ids, ordered by time submitted."""
         query = """SELECT msg_id FROM %s ORDER by submitted ASC"""%self.table
         cursor = self._db.execute(query)
         # will be a list of length 1 tuples
         return [ tup[0] for tup in cursor.fetchall()]
 
 __all__ = ['SQLiteDB']
  No newline at end of file

IPython/parallel/tests/test_client.py

@@ -1,386 +1,387 b''
 """Tests for parallel client.py
 
 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 time
 from datetime import datetime
 from tempfile import mktemp
 
 import zmq
 
 from IPython.parallel.client import client as clientmod
 from IPython.parallel import error
 from IPython.parallel import AsyncResult, AsyncHubResult
 from IPython.parallel import LoadBalancedView, DirectView
 
 from clienttest import ClusterTestCase, segfault, wait, add_engines
 
 def setup():
     add_engines(4, total=True)
 
 class TestClient(ClusterTestCase):
     
     def test_ids(self):
         n = len(self.client.ids)
         self.add_engines(2)
         self.assertEquals(len(self.client.ids), n+2)
     
     def test_view_indexing(self):
         """test index access for views"""
         self.minimum_engines(4)
         targets = self.client._build_targets('all')[-1]
         v = self.client[:]
         self.assertEquals(v.targets, targets)
         t = self.client.ids[2]
         v = self.client[t]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, t)
         t = self.client.ids[2:4]
         v = self.client[t]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, t)
         v = self.client[::2]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, targets[::2])
         v = self.client[1::3]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, targets[1::3])
         v = self.client[:-3]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, targets[:-3])
         v = self.client[-1]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, targets[-1])
         self.assertRaises(TypeError, lambda : self.client[None])
     
     def test_lbview_targets(self):
         """test load_balanced_view targets"""
         v = self.client.load_balanced_view()
         self.assertEquals(v.targets, None)
         v = self.client.load_balanced_view(-1)
         self.assertEquals(v.targets, [self.client.ids[-1]])
         v = self.client.load_balanced_view('all')
         self.assertEquals(v.targets, None)
     
     def test_dview_targets(self):
         """test direct_view targets"""
         v = self.client.direct_view()
         self.assertEquals(v.targets, 'all')
         v = self.client.direct_view('all')
         self.assertEquals(v.targets, 'all')
         v = self.client.direct_view(-1)
         self.assertEquals(v.targets, self.client.ids[-1])
     
     def test_lazy_all_targets(self):
         """test lazy evaluation of rc.direct_view('all')"""
         v = self.client.direct_view()
         self.assertEquals(v.targets, 'all')
         
         def double(x):
             return x*2
         seq = range(100)
         ref = [ double(x) for x in seq ]
         
         # add some engines, which should be used
         self.add_engines(1)
         n1 = len(self.client.ids)
         
         # simple apply
         r = v.apply_sync(lambda : 1)
         self.assertEquals(r, [1] * n1)
         
         # map goes through remotefunction
         r = v.map_sync(double, seq)
         self.assertEquals(r, ref)
 
         # add a couple more engines, and try again
         self.add_engines(2)
         n2 = len(self.client.ids)
         self.assertNotEquals(n2, n1)
         
         # apply
         r = v.apply_sync(lambda : 1)
         self.assertEquals(r, [1] * n2)
         
         # map
         r = v.map_sync(double, seq)
         self.assertEquals(r, ref)
     
     def test_targets(self):
         """test various valid targets arguments"""
         build = self.client._build_targets
         ids = self.client.ids
         idents,targets = build(None)
         self.assertEquals(ids, targets)
     
     def test_clear(self):
         """test clear behavior"""
         self.minimum_engines(2)
         v = self.client[:]
         v.block=True
         v.push(dict(a=5))
         v.pull('a')
         id0 = self.client.ids[-1]
         self.client.clear(targets=id0, block=True)
         a = self.client[:-1].get('a')
         self.assertRaisesRemote(NameError, self.client[id0].get, 'a')
         self.client.clear(block=True)
         for i in self.client.ids:
             self.assertRaisesRemote(NameError, self.client[i].get, 'a')
     
     def test_get_result(self):
         """test getting results from the Hub."""
         c = clientmod.Client(profile='iptest')
         t = c.ids[-1]
         ar = c[t].apply_async(wait, 1)
         # give the monitor time to notice the message
         time.sleep(.25)
         ahr = self.client.get_result(ar.msg_ids)
         self.assertTrue(isinstance(ahr, AsyncHubResult))
         self.assertEquals(ahr.get(), ar.get())
         ar2 = self.client.get_result(ar.msg_ids)
         self.assertFalse(isinstance(ar2, AsyncHubResult))
         c.close()
     
     def test_ids_list(self):
         """test client.ids"""
         ids = self.client.ids
         self.assertEquals(ids, self.client._ids)
         self.assertFalse(ids is self.client._ids)
         ids.remove(ids[-1])
         self.assertNotEquals(ids, self.client._ids)
     
     def test_queue_status(self):
         ids = self.client.ids
         id0 = ids[0]
         qs = self.client.queue_status(targets=id0)
         self.assertTrue(isinstance(qs, dict))
         self.assertEquals(sorted(qs.keys()), ['completed', 'queue', 'tasks'])
         allqs = self.client.queue_status()
         self.assertTrue(isinstance(allqs, dict))
         intkeys = list(allqs.keys())
         intkeys.remove('unassigned')
         self.assertEquals(sorted(intkeys), sorted(self.client.ids))
         unassigned = allqs.pop('unassigned')
         for eid,qs in allqs.items():
             self.assertTrue(isinstance(qs, dict))
             self.assertEquals(sorted(qs.keys()), ['completed', 'queue', 'tasks'])
 
     def test_shutdown(self):
         ids = self.client.ids
         id0 = ids[0]
         self.client.shutdown(id0, block=True)
         while id0 in self.client.ids:
             time.sleep(0.1)
             self.client.spin()
         
         self.assertRaises(IndexError, lambda : self.client[id0])
         
     def test_result_status(self):
         pass
         # to be written
     
     def test_db_query_dt(self):
         """test db query by date"""
         hist = self.client.hub_history()
         middle = self.client.db_query({'msg_id' : hist[len(hist)//2]})[0]
         tic = middle['submitted']
         before = self.client.db_query({'submitted' : {'$lt' : tic}})
         after = self.client.db_query({'submitted' : {'$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.client.db_query({'submitted' : tic})
         for s in same:
             self.assertTrue(s['submitted'] == tic)
     
     def test_db_query_keys(self):
         """test extracting subset of record keys"""
         found = self.client.db_query({'msg_id': {'$ne' : ''}},keys=['submitted', 'completed'])
         for rec in found:
             self.assertEquals(set(rec.keys()), set(['msg_id', 'submitted', 'completed']))
     
     def test_db_query_default_keys(self):
         """default db_query excludes buffers"""
         found = self.client.db_query({'msg_id': {'$ne' : ''}})
         for rec in found:
             keys = set(rec.keys())
             self.assertFalse('buffers' in keys, "'buffers' should not be in: %s" % keys)
             self.assertFalse('result_buffers' in keys, "'result_buffers' should not be in: %s" % keys)
     
     def test_db_query_msg_id(self):
         """ensure msg_id is always in db queries"""
         found = self.client.db_query({'msg_id': {'$ne' : ''}},keys=['submitted', 'completed'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
         found = self.client.db_query({'msg_id': {'$ne' : ''}},keys=['submitted'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
         found = self.client.db_query({'msg_id': {'$ne' : ''}},keys=['msg_id'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
     
     def test_db_query_get_result(self):
         """pop in db_query shouldn't pop from result itself"""
         self.client[:].apply_sync(lambda : 1)
         found = self.client.db_query({'msg_id': {'$ne' : ''}})
         rc2 = clientmod.Client(profile='iptest')
         # If this bug is not fixed, this call will hang:
         ar = rc2.get_result(self.client.history[-1])
         ar.wait(2)
         self.assertTrue(ar.ready())
         ar.get()
         rc2.close()
     
     def test_db_query_in(self):
         """test db query with '$in','$nin' operators"""
         hist = self.client.hub_history()
         even = hist[::2]
         odd = hist[1::2]
         recs = self.client.db_query({ 'msg_id' : {'$in' : even}})
         found = [ r['msg_id'] for r in recs ]
         self.assertEquals(set(even), set(found))
         recs = self.client.db_query({ 'msg_id' : {'$nin' : even}})
         found = [ r['msg_id'] for r in recs ]
         self.assertEquals(set(odd), set(found))
     
     def test_hub_history(self):
         hist = self.client.hub_history()
         recs = self.client.db_query({ 'msg_id' : {"$ne":''}})
         recdict = {}
         for rec in recs:
             recdict[rec['msg_id']] = rec
         
         latest = datetime(1984,1,1)
         for msg_id in hist:
             rec = recdict[msg_id]
             newt = rec['submitted']
             self.assertTrue(newt >= latest)
             latest = newt
         ar = self.client[-1].apply_async(lambda : 1)
         ar.get()
         time.sleep(0.25)
         self.assertEquals(self.client.hub_history()[-1:],ar.msg_ids)
     
     def _wait_for_idle(self):
         """wait for an engine to become idle, according to the Hub"""
         rc = self.client
         
         # timeout 2s, polling every 100ms
         for i in range(20):
             qs = rc.queue_status()
             if qs['unassigned'] or any(qs[eid]['tasks'] for eid in rc.ids):
                 time.sleep(0.1)
             else:
                 break
         
         # ensure Hub up to date:
         qs = rc.queue_status()
         self.assertEquals(qs['unassigned'], 0)
         for eid in rc.ids:
             self.assertEquals(qs[eid]['tasks'], 0)
     
     
     def test_resubmit(self):
         def f():
             import random
             return random.random()
         v = self.client.load_balanced_view()
         ar = v.apply_async(f)
         r1 = ar.get(1)
         # give the Hub a chance to notice:
         self._wait_for_idle()
         ahr = self.client.resubmit(ar.msg_ids)
         r2 = ahr.get(1)
         self.assertFalse(r1 == r2)
 
     def test_resubmit_aborted(self):
         def f():
             import random
             return random.random()
         v = self.client.load_balanced_view()
         # restrict to one engine, so we can put a sleep
         # ahead of the task, so it will get aborted
         eid = self.client.ids[-1]
         v.targets = [eid]
         sleep = v.apply_async(time.sleep, 0.5)
         ar = v.apply_async(f)
         ar.abort()
         self.assertRaises(error.TaskAborted, ar.get)
         # Give the Hub a chance to get up to date:
         self._wait_for_idle()
         ahr = self.client.resubmit(ar.msg_ids)
         r2 = ahr.get(1)
 
     def test_resubmit_inflight(self):
-        """ensure ValueError on resubmit of inflight task"""
+        """resubmit of inflight task"""
         v = self.client.load_balanced_view()
         ar = v.apply_async(time.sleep,1)
         # give the message a chance to arrive
         time.sleep(0.2)
-        self.assertRaisesRemote(ValueError, self.client.resubmit, ar.msg_ids)
+        ahr = self.client.resubmit(ar.msg_ids)
         ar.get(2)
+        ahr.get(2)
 
     def test_resubmit_badkey(self):
         """ensure KeyError on resubmit of nonexistant task"""
         self.assertRaisesRemote(KeyError, self.client.resubmit, ['invalid'])
 
     def test_purge_results(self):
         # ensure there are some tasks
         for i in range(5):
             self.client[:].apply_sync(lambda : 1)
         # Wait for the Hub to realise the result is done:
         # This prevents a race condition, where we
         # might purge a result the Hub still thinks is pending.
         time.sleep(0.1)
         rc2 = clientmod.Client(profile='iptest')
         hist = self.client.hub_history()
         ahr = rc2.get_result([hist[-1]])
         ahr.wait(10)
         self.client.purge_results(hist[-1])
         newhist = self.client.hub_history()
         self.assertEquals(len(newhist)+1,len(hist))
         rc2.spin()
         rc2.close()
         
     def test_purge_all_results(self):
         self.client.purge_results('all')
         hist = self.client.hub_history()
         self.assertEquals(len(hist), 0)
     
     def test_spin_thread(self):
         self.client.spin_thread(0.01)
         ar = self.client[-1].apply_async(lambda : 1)
         time.sleep(0.1)
         self.assertTrue(ar.wall_time < 0.1,
             "spin should have kept wall_time < 0.1, but got %f" % ar.wall_time
         )
     
     def test_stop_spin_thread(self):
         self.client.spin_thread(0.01)
         self.client.stop_spin_thread()
         ar = self.client[-1].apply_async(lambda : 1)
         time.sleep(0.15)
         self.assertTrue(ar.wall_time > 0.1,
             "Shouldn't be spinning, but got wall_time=%f" % ar.wall_time
         )
     
 

docs/source/parallel/parallel_db.txt

@@ -1,137 +1,137 b''
 .. _parallel_db:
 
 =======================
 IPython's Task Database
 =======================
 
 The IPython Hub stores all task requests and results in a database. Currently supported backends
 are: MongoDB, SQLite (the default), and an in-memory DictDB.  The most common use case for
 this is clients requesting results for tasks they did not submit, via:
 
 .. sourcecode:: ipython
 
     In [1]: rc.get_result(task_id)
 
 However, since we have this DB backend, we provide a direct query method in the :class:`client`
 for users who want deeper introspection into their task history. The :meth:`db_query` method of
 the Client is modeled after MongoDB queries, so if you have used MongoDB it should look
 familiar.  In fact, when the MongoDB backend is in use, the query is relayed directly.  However,
 when using other backends, the interface is emulated and only a subset of queries is possible.
 
 .. seealso::
 
     MongoDB query docs: http://www.mongodb.org/display/DOCS/Querying
 
 :meth:`Client.db_query` takes a dictionary query object, with keys from the TaskRecord key list,
 and values of either exact values to test, or MongoDB queries, which are dicts of The form:
 ``{'operator' : 'argument(s)'}``. There is also an optional `keys` argument, that specifies
 which subset of keys should be retrieved. The default is to retrieve all keys excluding the
 request and result buffers. :meth:`db_query` returns a list of TaskRecord dicts. Also like
 MongoDB, the `msg_id` key will always be included, whether requested or not.
 
 TaskRecord keys:
 
 =============== =============== =============
 Key             Type            Description
 =============== =============== =============
-msg_id          uuid(bytes)     The msg ID
+msg_id          uuid(ascii)     The msg ID
 header          dict            The request header
 content         dict            The request content (likely empty)
 buffers         list(bytes)     buffers containing serialized request objects
 submitted       datetime        timestamp for time of submission (set by client)
 client_uuid     uuid(bytes)     IDENT of client's socket
 engine_uuid     uuid(bytes)     IDENT of engine's socket
 started         datetime        time task began execution on engine
 completed       datetime        time task finished execution (success or failure) on engine
-resubmitted     datetime        time of resubmission (if applicable)
+resubmitted     uuid(ascii)     msg_id of resubmitted task (if applicable)
 result_header   dict            header for result
 result_content  dict            content for result
 result_buffers  list(bytes)     buffers containing serialized request objects
 queue           bytes           The name of the queue for the task ('mux' or 'task')
 pyin            <unused>        Python input (unused)
 pyout           <unused>        Python output (unused)
 pyerr           <unused>        Python traceback (unused)
 stdout          str             Stream of stdout data
 stderr          str             Stream of stderr data
 
 =============== =============== =============
 
 MongoDB operators we emulate on all backends:
 
 ==========  =================
 Operator    Python equivalent
 ==========  =================
   '$in'       in
   '$nin'      not in
   '$eq'       ==
   '$ne'       !=
   '$ge'       >
   '$gte'      >=
   '$le'       <
   '$lte'      <=
 ==========  =================
 
 
 The DB Query is useful for two primary cases:
 
 1. deep polling of task status or metadata
 2. selecting a subset of tasks, on which to perform a later operation (e.g. wait on result, purge records, resubmit,...)
 
 Example Queries
 ===============
 
 
 To get all msg_ids that are not completed, only retrieving their ID and start time:
 
 .. sourcecode:: ipython
 
     In [1]: incomplete = rc.db_query({'complete' : None}, keys=['msg_id', 'started'])
 
 All jobs started in the last hour by me:
 
 .. sourcecode:: ipython
 
     In [1]: from datetime import datetime, timedelta
 
     In [2]: hourago = datetime.now() - timedelta(1./24)
 
     In [3]: recent = rc.db_query({'started' : {'$gte' : hourago },
                                     'client_uuid' : rc.session.session})
 
 All jobs started more than an hour ago, by clients *other than me*:
 
 .. sourcecode:: ipython
 
     In [3]: recent = rc.db_query({'started' : {'$le' : hourago },
                                     'client_uuid' : {'$ne' : rc.session.session}})
 
 Result headers for all jobs on engine 3 or 4:
 
 .. sourcecode:: ipython
 
     In [1]: uuids = map(rc._engines.get, (3,4))
 
     In [2]: hist34 = rc.db_query({'engine_uuid' : {'$in' : uuids }, keys='result_header')
 
 
 Cost
 ====
 
 The advantage of the database backends is, of course, that large amounts of
 data can be stored that won't fit in memory.  The default 'backend' is actually
 to just store all of this information in a Python dictionary.  This is very fast,
 but will run out of memory quickly if you move a lot of data around, or your
 cluster is to run for a long time.
 
 Unfortunately, the DB backends (SQLite and MongoDB) right now are rather slow,
 and can still consume large amounts of resources, particularly if large tasks
 or results are being created at a high frequency.
 
 For this reason, we have added :class:`~.NoDB`,a dummy backend that doesn't
 actually store any information. When you use this database, nothing is stored,
 and any request for results will result in a KeyError.  This obviously prevents
 later requests for results and task resubmission from functioning, but
 sometimes those nice features are not as useful as keeping Hub memory under
 control.