PyZMQ Documentation¶

Release:	14.6.0
Date:	June 17, 2015

PyZMQ is the Python bindings for ØMQ. This documentation currently contains notes on some important aspects of developing PyZMQ and an overview of what the ØMQ API looks like in Python. For information on how to use ØMQ in general, see the many examples in the excellent ØMQ Guide, all of which have a version in Python.

PyZMQ works with Python 3 (≥ 3.2), and Python 2 (≥ 2.6), with no transformations or 2to3, as well as PyPy (at least 2.0 beta), thanks to a new CFFI backend.

Please don’t hesitate to report pyzmq-specific issues to our tracker on GitHub. General questions about ØMQ are better sent to the ØMQ mailing list or IRC Channel.

Summary of Changes in PyZMQ

Supported LibZMQ¶

PyZMQ aims to support all stable ( ≥2.1.4, ≥ 3.2.2, ≥ 4.0.1 ) and active development ( ≥ 4.1.0 ) versions of libzmq. Building the same pyzmq against various versions of libzmq is supported, but only the functionality of the linked libzmq will be available.

Note

libzmq 3.0-3.1 are not, and will never be supported. There never was a stable release of either.

Binary distributions (wheels on PyPI or GitHub) of PyZMQ ship with the stable version of libzmq at the time of release, built with default configuration. For pyzmq-14.6.0, this is 4.0.5.

Using PyZMQ¶

The PyZMQ API¶

Release:	14.6.0
Date:	June 17, 2015

zmq¶

Python bindings for 0MQ.

Basic Classes¶

`Context`¶

class zmq.Context(io_threads=1, **kwargs)¶

Create a zmq Context

A zmq Context creates sockets via its ctx.socket method.

Attributes

sockopts

Methods

closed¶: boolean - whether the context has been terminated. If True, you can no longer use this Context.

destroy(linger=None)¶: Close all sockets associated with this context, and then terminate the context. If linger is specified, the LINGER sockopt of the sockets will be set prior to closing.

Warning

destroy involves calling zmq_close(), which is NOT threadsafe. If there are active sockets in other threads, this must not be called.

get(option)¶

Get the value of a context option.

See the 0MQ API documentation for zmq_ctx_get for details on specific options.

New in version libzmq-3.2.

New in version 13.0.

Parameters:

option : int

The option to get. Available values will depend on your version of libzmq. Examples include:
zmq.IO_THREADS, zmq.MAX_SOCKETS

Returns:

optval : int

The value of the option as an integer.

getsockopt(opt)¶: get default socket options for new sockets created by this Context

New in version 13.0.

classmethod instance(io_threads=1)¶

Returns a global Context instance.

Most single-threaded applications have a single, global Context. Use this method instead of passing around Context instances throughout your code.

A common pattern for classes that depend on Contexts is to use a default argument to enable programs with multiple Contexts but not require the argument for simpler applications:

class MyClass(object):

def __init__(self, context=None):

self.context = context or Context.instance()

set(option, optval)¶

Set a context option.

See the 0MQ API documentation for zmq_ctx_set for details on specific options.

New in version libzmq-3.2.

New in version 13.0.

Parameters:

option : int

The option to set. Available values will depend on your version of libzmq. Examples include:
zmq.IO_THREADS, zmq.MAX_SOCKETS

optval : int

The value of the option to set.

setsockopt(opt, value)¶: set default socket options for new sockets created by this Context

New in version 13.0.

classmethod shadow(address)¶

Shadow an existing libzmq context

address is the integer address of the libzmq context or an FFI pointer to it.

New in version 14.1.

classmethod shadow_pyczmq(ctx)¶

Shadow an existing pyczmq context

ctx is the FFI zctx_t * pointer

New in version 14.1.

socket(socket_type)¶

Create a Socket associated with this Context.

Parameters:

socket_type : int

The socket type, which can be any of the 0MQ socket types: REQ, REP, PUB, SUB, PAIR, DEALER, ROUTER, PULL, PUSH, etc.

term()¶

Close or terminate the context.

This can be called to close the context by hand. If this is not called, the context will automatically be closed when it is garbage collected.

underlying¶: The address of the underlying libzmq context

`Socket`¶

class zmq.Socket(*a, **kw)¶

The ZMQ socket object

To create a Socket, first create a Context:

ctx = zmq.Context.instance()

then call ctx.socket(socket_type):

s = ctx.socket(zmq.ROUTER)

Attributes

Methods

closed¶: boolean - whether the socket has been closed. If True, you can no longer use this Socket.

bind(addr)¶

Bind the socket to an address.

This causes the socket to listen on a network port. Sockets on the other side of this connection will use Socket.connect(addr) to connect to this socket.

Parameters:

addr : str

The address string. This has the form ‘protocol://interface:port’, for example ‘tcp://127.0.0.1:5555’. Protocols supported include tcp, udp, pgm, epgm, inproc and ipc. If the address is unicode, it is encoded to utf-8 first.

bind_to_random_port(addr, min_port=49152, max_port=65536, max_tries=100)¶

bind this socket to a random port in a range

If the port range is unspecified, the system will choose the port.

Parameters:

addr : str

The address string without the port to pass to Socket.bind().

min_port : int, optional

The minimum port in the range of ports to try (inclusive).

max_port : int, optional

The maximum port in the range of ports to try (exclusive).

max_tries : int, optional

The maximum number of bind attempts to make.

Returns:

port : int

The port the socket was bound to.

Raises:

ZMQBindError

if max_tries reached before successful bind

close(linger=None)¶

Close the socket.

If linger is specified, LINGER sockopt will be set prior to closing.

This can be called to close the socket by hand. If this is not called, the socket will automatically be closed when it is garbage collected.

connect(addr)¶

Connect to a remote 0MQ socket.

Parameters:

addr : str

The address string. This has the form ‘protocol://interface:port’, for example ‘tcp://127.0.0.1:5555’. Protocols supported are tcp, upd, pgm, inproc and ipc. If the address is unicode, it is encoded to utf-8 first.

disable_monitor()¶: Shutdown the PAIR socket (created using get_monitor_socket) that is serving socket events.

New in version 14.4.

disconnect(addr)¶

Disconnect from a remote 0MQ socket (undoes a call to connect).

New in version libzmq-3.2.

New in version 13.0.

Parameters:

addr : str

The address string. This has the form ‘protocol://interface:port’, for example ‘tcp://127.0.0.1:5555’. Protocols supported are tcp, upd, pgm, inproc and ipc. If the address is unicode, it is encoded to utf-8 first.

get(option)¶

Get the value of a socket option.

See the 0MQ API documentation for details on specific options.

Parameters:

option : int

The option to get. Available values will depend on your version of libzmq. Examples include:
zmq.IDENTITY, HWM, LINGER, FD, EVENTS

Returns:

optval : int or bytes

The value of the option as a bytestring or int.

get_hwm()¶

get the High Water Mark

On libzmq ≥ 3, this gets SNDHWM if available, otherwise RCVHWM

get_monitor_socket(events=None, addr=None)¶

Return a connected PAIR socket ready to receive the event notifications.

New in version libzmq-4.0.

New in version 14.0.

Parameters:

events : bitfield (int) [default: ZMQ_EVENTS_ALL]

The bitmask defining which events are wanted.

addr : string [default: None]

The optional endpoint for the monitoring sockets.

Returns:

socket : (PAIR)

The socket is already connected and ready to receive messages.

get_string(option, encoding='utf-8')¶

get the value of a socket option

See the 0MQ documentation for details on specific options.

Parameters:

option : int

The option to retrieve.

Returns:

optval : unicode string (unicode on py2, str on py3)

The value of the option as a unicode string.

getsockopt(option)¶

Get the value of a socket option.

See the 0MQ API documentation for details on specific options.

Parameters:

option : int

The option to get. Available values will depend on your version of libzmq. Examples include:
zmq.IDENTITY, HWM, LINGER, FD, EVENTS

Returns:

optval : int or bytes

The value of the option as a bytestring or int.

getsockopt_string(option, encoding='utf-8')¶

get the value of a socket option

See the 0MQ documentation for details on specific options.

Parameters:

option : int

The option to retrieve.

Returns:

optval : unicode string (unicode on py2, str on py3)

The value of the option as a unicode string.

hwm¶

get the High Water Mark

On libzmq ≥ 3, this gets SNDHWM if available, otherwise RCVHWM

monitor(addr, flags)¶

Start publishing socket events on inproc. See libzmq docs for zmq_monitor for details.

While this function is available from libzmq 3.2, pyzmq cannot parse monitor messages from libzmq prior to 4.0.

Parameters:

addr : str

The inproc url used for monitoring. Passing None as the addr will cause an existing socket monitor to be deregistered.

events : int [default: zmq.EVENT_ALL]

The zmq event bitmask for which events will be sent to the monitor.

poll(timeout=None, flags=1)¶

poll the socket for events

The default is to poll forever for incoming events. Timeout is in milliseconds, if specified.

Parameters:

timeout : int [default: None]

The timeout (in milliseconds) to wait for an event. If unspecified (or specified None), will wait forever for an event.

flags : bitfield (int) [default: POLLIN]

The event flags to poll for (any combination of POLLIN|POLLOUT). The default is to check for incoming events (POLLIN).

Returns:

events : bitfield (int)

The events that are ready and waiting. Will be 0 if no events were ready by the time timeout was reached.

recv(flags=0, copy=True, track=False)¶

Receive a message.

Parameters:

flags : int

Any supported flag: NOBLOCK. If NOBLOCK is set, this method will raise a ZMQError with EAGAIN if a message is not ready. If NOBLOCK is not set, then this method will block until a message arrives.

copy : bool

Should the message be received in a copying or non-copying manner? If False a Frame object is returned, if True a string copy of message is returned.

track : bool

Should the message be tracked for notification that ZMQ has finished with it? (ignored if copy=True)

Returns:

msg : bytes, Frame

The received message frame. If copy is False, then it will be a Frame, otherwise it will be bytes.

Raises:

ZMQError

for any of the reasons zmq_msg_recv might fail.

recv_json(flags=0, **kwargs)¶

receive a Python object as a message using json to serialize

Keyword arguments are passed on to json.loads

Parameters:

flags : int

Any valid recv flag.

Returns:

obj : Python object

The Python object that arrives as a message.

recv_multipart(flags=0, copy=True, track=False)¶

receive a multipart message as a list of bytes or Frame objects

Parameters:

flags : int, optional

Any supported flag: NOBLOCK. If NOBLOCK is set, this method will raise a ZMQError with EAGAIN if a message is not ready. If NOBLOCK is not set, then this method will block until a message arrives.

copy : bool, optional

Should the message frame(s) be received in a copying or non-copying manner? If False a Frame object is returned for each part, if True a copy of the bytes is made for each frame.

track : bool, optional

Should the message frame(s) be tracked for notification that ZMQ has finished with it? (ignored if copy=True)

Returns:

msg_parts : list

A list of frames in the multipart message; either Frames or bytes, depending on copy.

recv_pyobj(flags=0)¶

receive a Python object as a message using pickle to serialize

Parameters:

flags : int

Any valid recv flag.

Returns:

obj : Python object

The Python object that arrives as a message.

recv_string(flags=0, encoding='utf-8')¶

receive a unicode string, as sent by send_string

Parameters:

flags : int

Any valid recv flag.

encoding : str [default: ‘utf-8’]

The encoding to be used

Returns:

s : unicode string (unicode on py2, str on py3)

The Python unicode string that arrives as encoded bytes.

send(data, flags=0, copy=True, track=False)¶

Send a message on this socket.

This queues the message to be sent by the IO thread at a later time.

Parameters:

data : object, str, Frame

The content of the message.

flags : int

Any supported flag: NOBLOCK, SNDMORE.

copy : bool

Should the message be sent in a copying or non-copying manner.

track : bool

Should the message be tracked for notification that ZMQ has finished with it? (ignored if copy=True)

Returns:

None : if copy or not track

None if message was sent, raises an exception otherwise.

MessageTracker : if track and not copy

a MessageTracker object, whose pending property will be True until the send is completed.

Raises:

TypeError

If a unicode object is passed

ValueError

If track=True, but an untracked Frame is passed.

ZMQError

If the send does not succeed for any reason.

send_json(obj, flags=0, **kwargs)¶

send a Python object as a message using json to serialize

Keyword arguments are passed on to json.dumps

Parameters:

obj : Python object

The Python object to send

flags : int

Any valid send flag

send_multipart(msg_parts, flags=0, copy=True, track=False)¶

send a sequence of buffers as a multipart message

The zmq.SNDMORE flag is added to all msg parts before the last.

Parameters:

msg_parts : iterable

A sequence of objects to send as a multipart message. Each element can be any sendable object (Frame, bytes, buffer-providers)

flags : int, optional

SNDMORE is handled automatically for frames before the last.

copy : bool, optional

Should the frame(s) be sent in a copying or non-copying manner.

track : bool, optional

Should the frame(s) be tracked for notification that ZMQ has finished with it (ignored if copy=True).

Returns:

None : if copy or not track

MessageTracker : if track and not copy

a MessageTracker object, whose pending property will be True until the last send is completed.

send_pyobj(obj, flags=0, protocol=2)¶

send a Python object as a message using pickle to serialize

Parameters:

obj : Python object

The Python object to send.

flags : int

Any valid send flag.

protocol : int

The pickle protocol number to use. The default is pickle.DEFAULT_PROTOCOL where defined, and pickle.HIGHEST_PROTOCOL elsewhere.

send_string(u, flags=0, copy=True, encoding='utf-8')¶

send a Python unicode string as a message with an encoding

0MQ communicates with raw bytes, so you must encode/decode text (unicode on py2, str on py3) around 0MQ.

Parameters:

u : Python unicode string (unicode on py2, str on py3)

The unicode string to send.

flags : int, optional

Any valid send flag.

encoding : str [default: ‘utf-8’]

The encoding to be used

set(option, optval)¶

Set socket options.

See the 0MQ API documentation for details on specific options.

Parameters:

option : int

The option to set. Available values will depend on your version of libzmq. Examples include:
zmq.SUBSCRIBE, UNSUBSCRIBE, IDENTITY, HWM, LINGER, FD

optval : int or bytes

The value of the option to set.

Notes

Warning

All options other than zmq.SUBSCRIBE, zmq.UNSUBSCRIBE and zmq.LINGER only take effect for subsequent socket bind/connects.

set_hwm(value)¶

set the High Water Mark

On libzmq ≥ 3, this sets both SNDHWM and RCVHWM

Warning

New values only take effect for subsequent socket bind/connects.

set_string(option, optval, encoding='utf-8')¶

set socket options with a unicode object

This is simply a wrapper for setsockopt to protect from encoding ambiguity.

See the 0MQ documentation for details on specific options.

Parameters:

option : int

The name of the option to set. Can be any of: SUBSCRIBE, UNSUBSCRIBE, IDENTITY

optval : unicode string (unicode on py2, str on py3)

The value of the option to set.

encoding : str

The encoding to be used, default is utf8

setsockopt(option, optval)¶

Set socket options.

See the 0MQ API documentation for details on specific options.

Parameters:

option : int

The option to set. Available values will depend on your version of libzmq. Examples include:
zmq.SUBSCRIBE, UNSUBSCRIBE, IDENTITY, HWM, LINGER, FD

optval : int or bytes

The value of the option to set.

Notes

Warning

All options other than zmq.SUBSCRIBE, zmq.UNSUBSCRIBE and zmq.LINGER only take effect for subsequent socket bind/connects.

setsockopt_string(option, optval, encoding='utf-8')¶

set socket options with a unicode object

This is simply a wrapper for setsockopt to protect from encoding ambiguity.

See the 0MQ documentation for details on specific options.

Parameters:

option : int

The name of the option to set. Can be any of: SUBSCRIBE, UNSUBSCRIBE, IDENTITY

optval : unicode string (unicode on py2, str on py3)

The value of the option to set.

encoding : str

The encoding to be used, default is utf8

classmethod shadow(address)¶

Shadow an existing libzmq socket

address is the integer address of the libzmq socket or an FFI pointer to it.

New in version 14.1.

socket_type¶

unbind(addr)¶

Unbind from an address (undoes a call to bind).

New in version libzmq-3.2.

New in version 13.0.

Parameters:

addr : str

The address string. This has the form ‘protocol://interface:port’, for example ‘tcp://127.0.0.1:5555’. Protocols supported are tcp, upd, pgm, inproc and ipc. If the address is unicode, it is encoded to utf-8 first.

underlying¶: The address of the underlying libzmq socket

`Frame`¶

class zmq.Frame¶

Attributes

Methods

buffer¶: A read-only buffer view of the message contents.

bytes¶

The message content as a Python bytes object.

The first time this property is accessed, a copy of the message contents is made. From then on that same copy of the message is returned.

get(option)¶

Get a Frame option or property.

See the 0MQ API documentation for zmq_msg_get and zmq_msg_gets for details on specific options.

New in version libzmq-3.2.

New in version 13.0.

Changed in version 14.3: add support for zmq_msg_gets (requires libzmq-4.1)

set(option, value)¶

Set a Frame option.

See the 0MQ API documentation for zmq_msg_set for details on specific options.

New in version libzmq-3.2.

New in version 13.0.

`MessageTracker`¶

class zmq.MessageTracker(*towatch)¶

A class for tracking if 0MQ is done using one or more messages.

When you send a 0MQ message, it is not sent immediately. The 0MQ IO thread sends the message at some later time. Often you want to know when 0MQ has actually sent the message though. This is complicated by the fact that a single 0MQ message can be sent multiple times using different sockets. This class allows you to track all of the 0MQ usages of a message.

Parameters:

*towatch : tuple of Event, MessageTracker, Message instances.

This list of objects to track. This class can track the low-level Events used by the Message class, other MessageTrackers or actual Messages.

Attributes

events
peers

Methods

done¶: Is 0MQ completely done with the message(s) being tracked?

wait(timeout=-1)¶

Wait for 0MQ to be done with the message or until timeout.

Parameters:

timeout : float [default: -1, wait forever]

Maximum time in (s) to wait before raising NotDone.

Returns:

None

if done before timeout

Raises:

NotDone

if timeout reached before I am done.

Polling¶

`Poller`¶

class zmq.Poller¶

A stateful poll interface that mirrors Python’s built-in poll.

Attributes

sockets

Methods

modify(socket, flags=3)¶: Modify the flags for an already registered 0MQ socket or native fd.

poll(timeout=None)¶

Poll the registered 0MQ or native fds for I/O.

Parameters:

timeout : float, int

The timeout in milliseconds. If None, no timeout (infinite). This is in milliseconds to be compatible with select.poll(). The underlying zmq_poll uses microseconds and we convert to that in this function.

Returns:

events : list of tuples

The list of events that are ready to be processed. This is a list of tuples of the form (socket, event), where the 0MQ Socket or integer fd is the first element, and the poll event mask (POLLIN, POLLOUT) is the second. It is common to call events = dict(poller.poll()), which turns the list of tuples into a mapping of socket : event.

register(socket, flags=POLLIN|POLLOUT)¶

Register a 0MQ socket or native fd for I/O monitoring.

register(s,0) is equivalent to unregister(s).

Parameters:

socket : zmq.Socket or native socket

A zmq.Socket or any Python object having a fileno() method that returns a valid file descriptor.

flags : int

The events to watch for. Can be POLLIN, POLLOUT or POLLIN|POLLOUT. If flags=0, socket will be unregistered.

unregister(socket)¶

Remove a 0MQ socket or native fd for I/O monitoring.

Parameters:

socket : Socket

The socket instance to stop polling.

zmq.select(rlist, wlist, xlist, timeout=None) -> (rlist, wlist, xlist)¶

Return the result of poll as a lists of sockets ready for r/w/exception.

This has the same interface as Python’s built-in select.select() function.

Parameters:

timeout : float, int, optional

The timeout in seconds. If None, no timeout (infinite). This is in seconds to be compatible with select.select(). The underlying zmq_poll uses microseconds and we convert to that in this function.

rlist : list of sockets/FDs

sockets/FDs to be polled for read events

wlist : list of sockets/FDs

sockets/FDs to be polled for write events

xlist : list of sockets/FDs

sockets/FDs to be polled for error events

Returns:

(rlist, wlist, xlist) : tuple of lists of sockets (length 3)

Lists correspond to sockets available for read/write/error events respectively.

Exceptions¶

`ZMQError`¶

class zmq.ZMQError(errno=None, msg=None)¶

Wrap an errno style error.

Parameters:

errno : int

The ZMQ errno or None. If None, then zmq_errno() is called and used.

msg : string

Description of the error or None.

Attributes

errno

`ZMQVersionError`¶

class zmq.ZMQVersionError(min_version, msg='Feature')¶

Raised when a feature is not provided by the linked version of libzmq.

New in version 14.2.

Attributes

min_version

`Again`¶

class zmq.Again(errno=None, msg=None)¶

Wrapper for zmq.EAGAIN

New in version 13.0.

Attributes

errno

`ContextTerminated`¶

class zmq.ContextTerminated(errno=None, msg=None)¶

Wrapper for zmq.ETERM

New in version 13.0.

Attributes

errno

`NotDone`¶

class zmq.NotDone¶

Raised when timeout is reached while waiting for 0MQ to finish with a Message

See also

MessageTracker.wait: object for tracking when ZeroMQ is done

Attributes

`ZMQBindError`¶

class zmq.ZMQBindError¶

An error for Socket.bind_to_random_port().

See also

Socket.bind_to_random_port

Attributes

Functions¶

zmq.zmq_version()¶: return the version of libzmq as a string

zmq.pyzmq_version()¶: return the version of pyzmq as a string

zmq.zmq_version_info()¶: Return the version of ZeroMQ itself as a 3-tuple of ints.

zmq.pyzmq_version_info()¶

return the pyzmq version as a tuple of at least three numbers

If pyzmq is a development version, inf will be appended after the third integer.

zmq.has()¶: Check for zmq capability by name (e.g. ‘ipc’, ‘curve’)

New in version libzmq-4.1.

New in version 14.1.

zmq.device(device_type, frontend, backend)¶

Start a zeromq device.

Deprecated since version libzmq-3.2: Use zmq.proxy

Parameters:

device_type : (QUEUE, FORWARDER, STREAMER)

The type of device to start.

frontend : Socket

The Socket instance for the incoming traffic.

backend : Socket

The Socket instance for the outbound traffic.

zmq.proxy(frontend, backend, capture)¶

Start a zeromq proxy (replacement for device).

New in version libzmq-3.2.

New in version 13.0.

Parameters:

frontend : Socket

The Socket instance for the incoming traffic.

backend : Socket

The Socket instance for the outbound traffic.

capture : Socket (optional)

The Socket instance for capturing traffic.

zmq.curve_keypair()¶

generate a Z85 keypair for use with zmq.CURVE security

Requires libzmq (≥ 4.0) to have been linked with libsodium.

New in version libzmq-4.0.

New in version 14.0.

Returns:

(public, secret) : two bytestrings

The public and private keypair as 40 byte z85-encoded bytestrings.

zmq.get_includes()¶: Return a list of directories to include for linking against pyzmq with cython.

devices¶

Functions¶

zmq.device(device_type, frontend, backend)¶

Start a zeromq device.

Deprecated since version libzmq-3.2: Use zmq.proxy

Parameters:

device_type : (QUEUE, FORWARDER, STREAMER)

The type of device to start.

frontend : Socket

The Socket instance for the incoming traffic.

backend : Socket

The Socket instance for the outbound traffic.

zmq.proxy(frontend, backend, capture)¶

Start a zeromq proxy (replacement for device).

New in version libzmq-3.2.

New in version 13.0.

Parameters:

frontend : Socket

The Socket instance for the incoming traffic.

backend : Socket

The Socket instance for the outbound traffic.

capture : Socket (optional)

The Socket instance for capturing traffic.

Module: `zmq.devices`¶

0MQ Device classes for running in background threads or processes.

Base Devices¶

`Device`¶

class zmq.devices.Device(device_type=3, in_type=None, out_type=None)¶

A 0MQ Device to be run in the background.

You do not pass Socket instances to this, but rather Socket types:

Device(device_type, in_socket_type, out_socket_type)

For instance:

dev = Device(zmq.QUEUE, zmq.DEALER, zmq.ROUTER)

Similar to zmq.device, but socket types instead of sockets themselves are passed, and the sockets are created in the work thread, to avoid issues with thread safety. As a result, additional bind_{in|out} and connect_{in|out} methods and setsockopt_{in|out} allow users to specify connections for the sockets.

Parameters:

device_type : int

The 0MQ Device type

{in|out}_type : int

zmq socket types, to be passed later to context.socket(). e.g. zmq.PUB, zmq.SUB, zmq.REQ. If out_type is < 0, then in_socket is used for both in_socket and out_socket.

Attributes

daemon

(int) sets whether the thread should be run as a daemon Default is true, because if it is false, the thread will not exit unless it is killed

Methods

bind_{in_out}(iface)	passthrough for `{in\|out}_socket.bind(iface)`, to be called in the thread
connect_{in_out}(iface)	passthrough for `{in\|out}_socket.connect(iface)`, to be called in the thread
setsockopt_{in_out}(opt,value)	passthrough for `{in\|out}_socket.setsockopt(opt, value)`, to be called in the thread

bind_in(addr)¶

Enqueue ZMQ address for binding on in_socket.

See zmq.Socket.bind for details.

bind_out(addr)¶

Enqueue ZMQ address for binding on out_socket.

See zmq.Socket.bind for details.

connect_in(addr)¶

Enqueue ZMQ address for connecting on in_socket.

See zmq.Socket.connect for details.

connect_out(addr)¶

Enqueue ZMQ address for connecting on out_socket.

See zmq.Socket.connect for details.

join(timeout=None)¶

wait for me to finish, like Thread.join.

Reimplemented appropriately by subclasses.

setsockopt_in(opt, value)¶

Enqueue setsockopt(opt, value) for in_socket

See zmq.Socket.setsockopt for details.

setsockopt_out(opt, value)¶

Enqueue setsockopt(opt, value) for out_socket

See zmq.Socket.setsockopt for details.

start()¶: Start the device. Override me in subclass for other launchers.

`ThreadDevice`¶

class zmq.devices.ThreadDevice(device_type=3, in_type=None, out_type=None)¶

A Device that will be run in a background Thread.

See Device for details.

Attributes

launcher

Methods

`ProcessDevice`¶

class zmq.devices.ProcessDevice(device_type=3, in_type=None, out_type=None)¶

A Device that will be run in a background Process.

See Device for details.

Attributes

launcher

Methods

context_factory(io_threads=1, **kwargs)¶

Callable that returns a context. Typically either Context.instance or Context, depending on whether the device should share the global instance or not.

Attributes

sockopts

Methods

alias of Context

Proxy Devices¶

`Proxy`¶

class zmq.devices.Proxy(in_type, out_type, mon_type=1)¶

Threadsafe Proxy object.

See zmq.devices.Device for most of the spec. This subclass adds a <method>_mon version of each <method>_{in|out} method, for configuring the monitor socket.

A Proxy is a 3-socket ZMQ Device that functions just like a QUEUE, except each message is also sent out on the monitor socket.

A PUB socket is the most logical choice for the mon_socket, but it is not required.

Methods

bind_mon(addr)¶

Enqueue ZMQ address for binding on mon_socket.

See zmq.Socket.bind for details.

connect_mon(addr)¶

Enqueue ZMQ address for connecting on mon_socket.

See zmq.Socket.bind for details.

setsockopt_mon(opt, value)¶

Enqueue setsockopt(opt, value) for mon_socket

See zmq.Socket.setsockopt for details.

`ThreadProxy`¶

class zmq.devices.ThreadProxy(in_type, out_type, mon_type=1)¶

Proxy in a Thread. See Proxy for more.

Attributes

launcher

Methods

`ProcessProxy`¶

class zmq.devices.ProcessProxy(in_type, out_type, mon_type=1)¶

Proxy in a Process. See Proxy for more.

Attributes

launcher

Methods

MonitoredQueue Devices¶

zmq.devices.monitored_queue(in_socket, out_socket, mon_socket, in_prefix=b'in', out_prefix=b'out')¶

Start a monitored queue device.

A monitored queue is very similar to the zmq.proxy device (monitored queue came first).

Differences from zmq.proxy:

monitored_queue supports both in and out being ROUTER sockets (via swapping IDENTITY prefixes).
monitor messages are prefixed, making in and out messages distinguishable.

Parameters:

in_socket : Socket

One of the sockets to the Queue. Its messages will be prefixed with ‘in’.

out_socket : Socket

One of the sockets to the Queue. Its messages will be prefixed with ‘out’. The only difference between in/out socket is this prefix.

mon_socket : Socket

This socket sends out every message received by each of the others with an in/out prefix specifying which one it was.

in_prefix : str

Prefix added to broadcast messages from in_socket.

out_prefix : str

Prefix added to broadcast messages from out_socket.

`MonitoredQueue`¶

class zmq.devices.MonitoredQueue(in_type, out_type, mon_type=1, in_prefix='in', out_prefix='out')¶

Class for running monitored_queue in the background.

See zmq.devices.Device for most of the spec. MonitoredQueue differs from Proxy, only in that it adds a prefix to messages sent on the monitor socket, with a different prefix for each direction.

MQ also supports ROUTER on both sides, which zmq.proxy does not.

If a message arrives on in_sock, it will be prefixed with in_prefix on the monitor socket. If it arrives on out_sock, it will be prefixed with out_prefix.

A PUB socket is the most logical choice for the mon_socket, but it is not required.

Methods

`ThreadMonitoredQueue`¶

class zmq.devices.ThreadMonitoredQueue(in_type, out_type, mon_type=1, in_prefix='in', out_prefix='out')¶

Run zmq.monitored_queue in a background thread.

See MonitoredQueue and Proxy for details.

Attributes

launcher

Methods

`ProcessMonitoredQueue`¶

class zmq.devices.ProcessMonitoredQueue(in_type, out_type, mon_type=1, in_prefix='in', out_prefix='out')¶

Run zmq.monitored_queue in a background thread.

See MonitoredQueue and Proxy for details.

Attributes

launcher

Methods

green¶

Module: `green`¶

zmq.green - gevent compatibility with zeromq.

Usage¶

Instead of importing zmq directly, do so in the following manner:

import zmq.green as zmq

Any calls that would have blocked the current thread will now only block the current green thread.

This compatibility is accomplished by ensuring the nonblocking flag is set before any blocking operation and the ØMQ file descriptor is polled internally to trigger needed events.

auth¶

Module: `auth`¶

Utilities for ZAP authentication.

To run authentication in a background thread, see zmq.auth.thread. For integration with the tornado eventloop, see zmq.auth.ioloop.

New in version 14.1.

`Authenticator`¶

class zmq.auth.Authenticator(context=None, encoding='utf-8', log=None)¶

Implementation of ZAP authentication for zmq connections.

Note: - libzmq provides four levels of security: default NULL (which the Authenticator does

not see), and authenticated NULL, PLAIN, and CURVE, which the Authenticator can see.

until you add policies, all incoming NULL connections are allowed

(classic ZeroMQ behavior), and all PLAIN and CURVE connections are denied.

Methods

allow(*addresses)¶

Allow (whitelist) IP address(es).

Connections from addresses not in the whitelist will be rejected.

For NULL, all clients from this address will be accepted.
For PLAIN and CURVE, they will be allowed to continue with authentication.

whitelist is mutually exclusive with blacklist.

configure_curve(domain='*', location=None)¶

Configure CURVE authentication for a given domain.

CURVE authentication uses a directory that holds all public client certificates, i.e. their public keys.

To cover all domains, use “*”.

You can add and remove certificates in that directory at any time.

To allow all client keys without checking, specify CURVE_ALLOW_ANY for the location.

configure_plain(domain='*', passwords=None)¶

Configure PLAIN authentication for a given domain.

PLAIN authentication uses a plain-text password file. To cover all domains, use “*”. You can modify the password file at any time; it is reloaded automatically.

deny(*addresses)¶

Deny (blacklist) IP address(es).

Addresses not in the blacklist will be allowed to continue with authentication.

Blacklist is mutually exclusive with whitelist.

handle_zap_message(msg)¶: Perform ZAP authentication

start()¶: Create and bind the ZAP socket

stop()¶: Close the ZAP socket

Functions¶

zmq.auth.create_certificates(key_dir, name, metadata=None)¶

Create zmq certificates.

Returns the file paths to the public and secret certificate files.

zmq.auth.load_certificate(filename)¶

Load public and secret key from a zmq certificate.

Returns (public_key, secret_key)

If the certificate file only contains the public key, secret_key will be None.

If there is no public key found in the file, ValueError will be raised.

zmq.auth.load_certificates(directory='.')¶: Load public keys from all certificates in a directory

auth.thread¶

Module: `auth.thread`¶

ZAP Authenticator in a Python Thread.

New in version 14.1.

Classes¶

`ThreadAuthenticator`¶

class zmq.auth.thread.ThreadAuthenticator(context=None, encoding='utf-8', log=None)¶

Run ZAP authentication in a background thread

Methods

allow(*addresses)¶

Allow (whitelist) IP address(es).

Connections from addresses not in the whitelist will be rejected.

For NULL, all clients from this address will be accepted.
For PLAIN and CURVE, they will be allowed to continue with authentication.

whitelist is mutually exclusive with blacklist.

configure_curve(domain='*', location='')¶

Configure CURVE authentication for a given domain.

CURVE authentication uses a directory that holds all public client certificates, i.e. their public keys.

To cover all domains, use “*”.

You can add and remove certificates in that directory at any time.

To allow all client keys without checking, specify CURVE_ALLOW_ANY for the location.

configure_plain(domain='*', passwords=None)¶

Configure PLAIN authentication for a given domain.

PLAIN authentication uses a plain-text password file. To cover all domains, use “*”. You can modify the password file at any time; it is reloaded automatically.

deny(*addresses)¶

Deny (blacklist) IP address(es).

Addresses not in the blacklist will be allowed to continue with authentication.

Blacklist is mutually exclusive with whitelist.

is_alive()¶: Is the ZAP thread currently running?

start()¶: Start the authentication thread

stop()¶: Stop the authentication thread

auth.ioloop¶

Module: `auth.ioloop`¶

ZAP Authenticator integrated with the tornado IOLoop.

New in version 14.1.

`IOLoopAuthenticator`¶

class zmq.auth.ioloop.IOLoopAuthenticator(context=None, encoding='utf-8', log=None, io_loop=None)¶

ZAP authentication for use in the tornado IOLoop

Methods

allow(*addresses)¶

Allow (whitelist) IP address(es).

Connections from addresses not in the whitelist will be rejected.

For NULL, all clients from this address will be accepted.
For PLAIN and CURVE, they will be allowed to continue with authentication.

whitelist is mutually exclusive with blacklist.

configure_curve(domain='*', location=None)¶

Configure CURVE authentication for a given domain.

CURVE authentication uses a directory that holds all public client certificates, i.e. their public keys.

To cover all domains, use “*”.

You can add and remove certificates in that directory at any time.

To allow all client keys without checking, specify CURVE_ALLOW_ANY for the location.

configure_plain(domain='*', passwords=None)¶

Configure PLAIN authentication for a given domain.

PLAIN authentication uses a plain-text password file. To cover all domains, use “*”. You can modify the password file at any time; it is reloaded automatically.

deny(*addresses)¶

Deny (blacklist) IP address(es).

Addresses not in the blacklist will be allowed to continue with authentication.

Blacklist is mutually exclusive with whitelist.

handle_zap_message(msg)¶: Perform ZAP authentication

start()¶: Start ZAP authentication

stop()¶: Stop ZAP authentication

utils.win32¶

Module: `zmq.utils.win32`¶

Win32 compatibility utilities.

`allow_interrupt`¶

class zmq.utils.win32.allow_interrupt(action=None)¶

Utility for fixing CTRL-C events on Windows.

On Windows, the Python interpreter intercepts CTRL-C events in order to translate them into KeyboardInterrupt exceptions. It (presumably) does this by setting a flag in its “control control handler” and checking it later at a convenient location in the interpreter.

However, when the Python interpreter is blocked waiting for the ZMQ poll operation to complete, it must wait for ZMQ’s select() operation to complete before translating the CTRL-C event into the KeyboardInterrupt exception.

The only way to fix this seems to be to add our own “console control handler” and perform some application-defined operation that will unblock the ZMQ polling operation in order to force ZMQ to pass control back to the Python interpreter.

This context manager performs all that Windows-y stuff, providing you with a hook that is called when a CTRL-C event is intercepted. This hook allows you to unblock your ZMQ poll operation immediately, which will then result in the expected KeyboardInterrupt exception.

Without this context manager, your ZMQ-based application will not respond normally to CTRL-C events on Windows. If a CTRL-C event occurs while blocked on ZMQ socket polling, the translation to a KeyboardInterrupt exception will be delayed until the I/O completes and control returns to the Python interpreter (this may never happen if you use an infinite timeout).

A no-op implementation is provided on non-Win32 systems to avoid the application from having to conditionally use it.

Example usage:

def stop_my_application():
    # ...

with allow_interrupt(stop_my_application):
    # main polling loop.

In a typical ZMQ application, you would use the “self pipe trick” to send message to a PAIR socket in order to interrupt your blocking socket polling operation.

In a Tornado event loop, you can use the IOLoop.stop method to unblock your I/O loop.

More Than Just Bindings¶

PyZMQ is ostensibly the Python bindings for ØMQ, but the project, following Python’s ‘batteries included’ philosophy, provides more than just Python methods and objects for calling into the ØMQ C++ library.

The Core as Bindings¶

PyZMQ is currently broken up into four subpackages. First, is the Core. zmq.core contains the actual bindings for ZeroMQ, and no extended functionality beyond the very basic. The core modules are split, such that each basic ZeroMQ object (or function, if no object is associated) is a separate module, e.g. zmq.core.context contains the Context object, zmq.core.poll contains a Poller object, as well as the select() function, etc. ZMQ constants are, for convenience, all kept together in zmq.core.constants.

There are two reasons for breaking the core into submodules: recompilation and derivative projects. The monolithic PyZMQ became quite tedious to have to recompile everything for a small change to a single object. With separate files, that’s no longer necessary. The second reason has to do with Cython. PyZMQ is written in Cython, a tool for efficiently writing C-extensions for Python. By separating out our objects into individual pyx files, each with their declarations in a pxd header, other projects can write extensions in Cython and call directly to ZeroMQ at the C-level without the penalty of going through our Python objects.

Thread Safety¶

In ØMQ, Contexts are threadsafe objects, but Sockets are not. It is safe to use a single Context (e.g. via zmq.Context.instance()) in your entire multithreaded application, but you should create sockets on a per-thread basis. If you share sockets across threads, you are likely to encounter uncatchable c-level crashes of your application unless you use judicious application of threading.Lock, but this approach is not recommended.

See also

ZeroMQ API note on threadsafety on 2.2 or 3.2

Socket Options as Attributes¶

New in version 2.1.9.

In 0MQ, socket options are set/retrieved with the set/getsockopt() methods. With the class-based approach in pyzmq, it would be logical to perform these operations with simple attribute access, and this has been added in pyzmq 2.1.9. Simply assign to or request a Socket attribute with the (case-insensitive) name of a sockopt, and it should behave just as you would expect:

s = ctx.socket(zmq.DEALER)
s.identity = b'dealer'
s.hwm = 10
s.events
# 0
s.fd
# 16

Default Options on the Context¶

New in version 2.1.11.

Just like setting socket options as attributes on Sockets, you can do the same on Contexts. This affects the default options of any new sockets created after the assignment.

ctx = zmq.Context()
ctx.linger = 0
rep = ctx.socket(zmq.REP)
req = ctx.socket(zmq.REQ)

Socket options that do not apply to a socket (e.g. SUBSCRIBE on non-SUB sockets) will simply be ignored.

Core Extensions¶

We have extended the core functionality in two ways that appear inside the core bindings, and are not general ØMQ features.

Builtin Serialization¶

First, we added common serialization with the builtin json and pickle as first-class methods to the Socket class. A socket has the methods send_json() and send_pyobj(), which correspond to sending an object over the wire after serializing with json and pickle respectively, and any object sent via those methods can be reconstructed with the recv_json() and recv_pyobj() methods. Unicode strings are other objects that are not unambiguously sendable over the wire, so we include send_string() and recv_string() that simply send bytes after encoding the message (‘utf-8’ is the default).

See also

Further information on serialization in pyzmq.
Our Unicode discussion for more information on the trials and tribulations of working with Unicode in a C extension while supporting Python 2 and 3.

MessageTracker¶

The second extension of basic ØMQ functionality is the MessageTracker. The MessageTracker is an object used to track when the underlying ZeroMQ is done with a message buffer. One of the main use cases for ØMQ in Python is the ability to perform non-copying sends. Thanks to Python’s buffer interface, many objects (including NumPy arrays) provide the buffer interface, and are thus directly sendable. However, as with any asynchronous non-copying messaging system like ØMQ or MPI, it can be important to know when the message has actually been sent, so it is safe again to edit the buffer without worry of corrupting the message. This is what the MessageTracker is for.

The MessageTracker is a simple object, but there is a penalty to its use. Since by its very nature, the MessageTracker must involve threadsafe communication (specifically a builtin Queue object), instantiating a MessageTracker takes a modest amount of time (10s of µs), so in situations instantiating many small messages, this can actually dominate performance. As a result, tracking is optional, via the track flag, which is optionally passed, always defaulting to False, in each of the three places where a Frame object (the pyzmq object for wrapping a segment of a message) is instantiated: The Frame constructor, and non-copying sends and receives.

A MessageTracker is very simple, and has just one method and one attribute. The property MessageTracker.done will be True when the Frame(s) being tracked are no longer in use by ØMQ, and MessageTracker.wait() will block, waiting for the Frame(s) to be released.

Note

A Frame cannot be tracked after it has been instantiated without tracking. If a Frame is to even have the option of tracking, it must be constructed with track=True.

Extensions¶

So far, PyZMQ includes four extensions to core ØMQ that we found basic enough to be included in PyZMQ itself:

zmq.log : Logging handlers for hooking Python logging up to the network
zmq.devices : Custom devices and objects for running devices in the background
zmq.eventloop : The Tornado event loop, adapted for use with ØMQ sockets.
zmq.ssh : Simple tools for tunneling zeromq connections via ssh.

Serializing messages with PyZMQ¶

When sending messages over a network, you often need to marshall your data into bytes.

Builtin serialization¶

PyZMQ is primarily bindings for libzmq, but we do provide three builtin serialization methods for convenience, to help Python developers learn libzmq. Python has two primary packages for serializing objects: json and pickle, so we provide simple convenience methods for sending and receiving objects serialized with these modules. A socket has the methods send_json() and send_pyobj(), which correspond to sending an object over the wire after serializing with json and pickle respectively, and any object sent via those methods can be reconstructed with the recv_json() and recv_pyobj() methods.

These methods designed for convenience, not for performance, so developers who do want to emphasize performance should use their own serialized send/recv methods.

Using your own serialization¶

In general, you will want to provide your own serialization that is optimized for your application or library availability. This may include using your own preferred serialization ([msgpack], [protobuf]), or adding compression via [zlib] in the standard library, or the super fast [blosc] library.

There are two simple models for implementing your own serialization: write a function that takes the socket as an argument, or subclass Socket for use in your own apps.

For instance, pickles can often be reduced substantially in size by compressing the data. The following will send compressed pickles over the wire:

import zlib, cPickle as pickle

def send_zipped_pickle(socket, obj, flags=0, protocol=-1):
    """pickle an object, and zip the pickle before sending it"""
    p = pickle.dumps(obj, protocol)
    z = zlib.compress(p)
    return socket.send(z, flags=flags)

def recv_zipped_pickle(socket, flags=0, protocol=-1):
    """inverse of send_zipped_pickle"""
    z = socket.recv(flags)
    p = zlib.decompress(z)
    return pickle.loads(p)

A common data structure in Python is the numpy array. PyZMQ supports sending numpy arrays without copying any data, since they provide the Python buffer interface. However just the buffer is not enough information to reconstruct the array on the receiving side. Here is an example of a send/recv that allow non-copying sends/recvs of numpy arrays including the dtype/shape data necessary for reconstructing the array.

import numpy

def send_array(socket, A, flags=0, copy=True, track=False):
    """send a numpy array with metadata"""
    md = dict(
        dtype = str(A.dtype),
        shape = A.shape,
    )
    socket.send_json(md, flags|zmq.SNDMORE)
    return socket.send(A, flags, copy=copy, track=track)

def recv_array(socket, flags=0, copy=True, track=False):
    """recv a numpy array"""
    md = socket.recv_json(flags=flags)
    msg = socket.recv(flags=flags, copy=copy, track=track)
    buf = buffer(msg)
    A = numpy.frombuffer(buf, dtype=md['dtype'])
    return A.reshape(md['shape'])

[msgpack]

Message Pack serialization library http://msgpack.org

[protobuf]

Google Protocol Buffers http://code.google.com/p/protobuf

[zlib] Python stdlib module for zip compression: zlib

[blosc]

Blosc: A blocking, shuffling and loss-less (and crazy-fast) compression library http://blosc.pytables.org/trac

Devices in PyZMQ¶

See also

ØMQ Guide Device coverage.

ØMQ has a notion of Devices - simple programs that manage a send-recv pattern for connecting two or more sockets. Being full programs, devices include a while(True) loop and thus block execution permanently once invoked. We have provided in the devices subpackage some facilities for running these devices in the background, as well as a custom three-socket MonitoredQueue device.

BackgroundDevices¶

It seems fairly rare that in a Python program one would actually want to create a zmq device via device() in the main thread, since such a call would block execution forever. The most likely model for launching devices is in background threads or processes. We have provided classes for launching devices in a background thread with ThreadDevice and via multiprocessing with ProcessDevice. For threadsafety and running across processes, these methods do not take Socket objects as arguments, but rather socket types, and then the socket creation and configuration happens via the BackgroundDevice’s foo_in() proxy methods. For each configuration method (bind/connect/setsockopt), there are proxy methods for calling those methods on the Socket objects created in the background thread or process, prefixed with ‘in_’ or ‘out_’, corresponding to the in_socket and out_socket:

from zmq.devices import ProcessDevice

pd = ProcessDevice(zmq.QUEUE, zmq.ROUTER, zmq.DEALER)
pd.bind_in('tcp://*:12345')
pd.connect_out('tcp://127.0.0.1:12543')
pd.setsockopt_in(zmq.IDENTITY, 'ROUTER')
pd.setsockopt_out(zmq.IDENTITY, 'DEALER')
pd.start()
# it will now be running in a background process

MonitoredQueue¶

One of ØMQ’s builtin devices is the QUEUE. This is a symmetric two-socket device that fully supports passing messages in either direction via any pattern. We saw a logical extension of the QUEUE as one that behaves in the same way with respect to the in/out sockets, but also sends every message in either direction also on a third monitor socket. For performance reasons, this monitored_queue() function is written in Cython, so the loop does not involve Python, and should have the same performance as the basic QUEUE device.

One shortcoming of the QUEUE device is that it does not support having ROUTER sockets as both input and output. This is because ROUTER sockets, when they receive a message, prepend the IDENTITY of the socket that sent the message (for use in routing the reply). The result is that the output socket will always try to route the incoming message back to the original sender, which is presumably not the intended pattern. In order for the queue to support a ROUTER-ROUTER connection, it must swap the first two parts of the message in order to get the right message out the other side.

To invoke a monitored queue is similar to invoking a regular ØMQ device:

from zmq.devices import monitored_queue
ins = ctx.socket(zmq.ROUTER)
outs = ctx.socket(zmq.DEALER)
mons = ctx.socket(zmq.PUB)
configure_sockets(ins,outs,mons)
monitored_queue(ins, outs, mons, in_prefix='in', out_prefix='out')

The in_prefix and out_prefix default to ‘in’ and ‘out’ respectively, and a PUB socket is most logical for the monitor socket, since it will never receive messages, and the in/out prefix is well suited to the PUB/SUB topic subscription model. All messages sent on mons will be multipart, the first part being the prefix corresponding to the socket that received the message.

Or for launching an MQ in the background, there are ThreadMonitoredQueue and ProcessMonitoredQueue, which function just like the base BackgroundDevice objects, but add foo_mon() methods for configuring the monitor socket.

Eventloops and PyZMQ¶

Tornado IOLoop¶

Facebook’s Tornado includes an eventloop for handing poll events on filedescriptors and native sockets. We have included a small part of Tornado (specifically its ioloop), and adapted its IOStream class into ZMQStream for handling poll events on ØMQ sockets. A ZMQStream object works much like a Socket object, but instead of calling recv() directly, you register a callback with on_recv(). Callbacks can also be registered for send events with on_send().

`install()`¶

With PyZMQ’s ioloop, you can use zmq sockets in any tornado application. You must first install PyZMQ’s IOLoop, with the ioloop.install() function:

from zmq.eventloop import ioloop
ioloop.install()

This sets the global instance of tornado.ioloop.IOLoop with the global instance of our IOLoop class. The reason this must happen is that tornado objects avoid having to pass the active IOLoop instance around by having a staticmethod IOLoop.instance(), which always returns the active instance. If PyZMQ’s IOLoop is installed after the first call to IOLoop.instance() (called in almost every tornado object constructor), then it will raise an AssertionError, because the global IOLoop instance has already been created, and proceeding would result in not all objects being associated with the right IOLoop.

It is possible to use PyZMQ sockets with tornado without calling ioloop.install(), but it is less convenient. First, you must instruct the tornado IOLoop to use the zmq poller:

from zmq.eventloop.ioloop import ZMQIOLoop

loop = ZMQIOLoop()

Then, when you instantiate tornado and ZMQStream objects, you must pass the io_loop argument to ensure that they use this loop, instead of the global instance.

This is especially useful for writing tests, such as this:

from tornado.testing import AsyncTestCase
from zmq.eventloop.ioloop import ZMQIOLoop
from zmq.eventloop.zmqstream import ZMQStream

class TestZMQBridge(AsyncTestCase):

     # Use a ZMQ-compatible I/O loop so that we can use `ZMQStream`.
     def get_new_ioloop(self):
         return ZMQIOLoop()

You can also manually install this IOLoop as the global tornado instance, with:

from zmq.eventloop.ioloop import ZMQIOLoop
loop = ZMQIOLoop()
loop.install()

but it will NOT be the global pyzmq IOLoop instance, so it must still be passed to your ZMQStream constructors.

`send()`¶

ZMQStream objects do have send() and send_multipart() methods, which behaves the same way as Socket.send(), but instead of sending right away, the IOLoop will wait until socket is able to send (for instance if HWM is met, or a REQ/REP pattern prohibits sending at a certain point). Messages sent via send will also be passed to the callback registered with on_send() after sending.

`on_recv()`¶

ZMQStream.on_recv() is the primary method for using a ZMQStream. It registers a callback to fire with messages as they are received, which will always be multipart, even if its length is 1. You can easily use this to build things like an echo socket:

s = ctx.socket(zmq.REP)
s.bind('tcp://localhost:12345')
stream = ZMQStream(s)
def echo(msg):
    stream.send_multipart(msg)
stream.on_recv(echo)
ioloop.IOLoop.instance().start()

on_recv can also take a copy flag, just like Socket.recv(). If copy=False, then callbacks registered with on_recv will receive tracked Frame objects instead of bytes.

Note

A callback must be registered using either ZMQStream.on_recv() or ZMQStream.on_recv_stream() before any data will be received on the underlying socket. This allows you to temporarily pause processing on a socket by setting both callbacks to None. Processing can later be resumed by restoring either callback.

`on_recv_stream()`¶

ZMQStream.on_recv_stream() is just like on_recv above, but the callback will be passed both the message and the stream, rather than just the message. This is meant to make it easier to use a single callback with multiple streams.

s1 = ctx.socket(zmq.REP)
s1.bind('tcp://localhost:12345')
stream1 = ZMQStream(s1)

s2 = ctx.socket(zmq.REP)
s2.bind('tcp://localhost:54321')
stream2 = ZMQStream(s2)

def echo(stream, msg):
    stream.send_multipart(msg)

stream1.on_recv_stream(echo)
stream2.on_recv_stream(echo)

ioloop.IOLoop.instance().start()

`flush()`¶

Sometimes with an eventloop, there can be multiple events ready on a single iteration of the loop. The flush() method allows developers to pull messages off of the queue to enforce some priority over the event loop ordering. flush pulls any pending events off of the queue. You can specify to flush only recv events, only send events, or any events, and you can specify a limit for how many events to flush in order to prevent starvation.

PyZMQ and gevent¶

PyZMQ ≥ 2.2.0.1 ships with a gevent compatible API as zmq.green. To use it, simply:

import zmq.green as zmq

Then write your code as normal.

Socket.send/recv and zmq.Poller are gevent-aware.

In PyZMQ ≥ 2.2.0.2, green.device and green.eventloop should be gevent-friendly as well.

Note

The green device does not release the GIL, unlike the true device in zmq.core.

zmq.green.eventloop includes minimally patched IOLoop/ZMQStream in order to use the gevent-enabled Poller, so you should be able to use the ZMQStream interface in gevent apps as well, though using two eventloops simultaneously (tornado + gevent) is not recommended.

Warning

There is a known issue in gevent ≤ 1.0 or libevent, which can cause zeromq socket events to be missed. PyZMQ works around this by adding a timeout so it will not wait forever for gevent to notice events. The only known solution for this is to use gevent ≥ 1.0, which is currently at 1.0b3, and does not exhibit this behavior.

See also

zmq.green examples on GitHub.

zmq.green is simply gevent_zeromq, merged into the pyzmq project.

Asynchronous Logging via PyZMQ¶

See also

The ØMQ guide coverage of PUB/SUB messaging
Python logging module documentation

Python provides extensible logging facilities through its logging module. This module allows for easily extensible logging functionality through the use of Handler objects. The most obvious case for hooking up pyzmq to logging would be to broadcast log messages over a PUB socket, so we have provided a PUBHandler class for doing just that.

PUB/SUB and Topics¶

The ØMQ PUB/SUB pattern consists of a PUB socket broadcasting messages, and a collection of SUB sockets that receive those messages. Each PUB message is a multipart-message, where the first part is interpreted as a topic. SUB sockets can subscribe to topics by setting their SUBSCRIBE sockopt, e.g.:

sub = ctx.socket(zmq.SUB)
sub.setsockopt(zmq.SUBSCRIBE, 'topic1')
sub.setsockopt(zmq.SUBSCRIBE, 'topic2')

When subscribed, the SUB socket will only receive messages where the first part starts with one of the topics set via SUBSCRIBE. The default behavior is to exclude all messages, and subscribing to the empty string ‘’ will receive all messages.

PUBHandler¶

The PUBHandler object is created for allowing the python logging to be emitted on a PUB socket. The main difference between a PUBHandler and a regular logging Handler is the inclusion of topics. For the most basic logging, you can simply create a PUBHandler with an interface or a configured PUB socket, and just let it go:

pub = context.socket(zmq.PUB)
pub.bind('tcp://*:12345')
handler = PUBHandler(pub)
logger = logging.getLogger()
logger.addHandler(handler)

At this point, all messages logged with the default logger will be broadcast on the pub socket.

the PUBHandler does work with topics, and the handler has an attribute root_topic:

handler.root_topic = 'myprogram'

Python loggers also have loglevels. The base topic of messages emitted by the PUBHandler will be of the form: <handler.root_topic>.<loglevel>, e.g. ‘myprogram.INFO’ or ‘whatever.ERROR’. This way, subscribers can easily subscribe to subsets of the logging messages. Log messages are always two-part, where the first part is the topic tree, and the second part is the actual log message.

>>> logger.info('hello there')
>>> print sub.recv_multipart()
['myprogram.INFO', 'hello there']

Subtopics¶

You can also add to the topic tree below the loglevel on an individual message basis. Assuming your logger is connected to a PUBHandler, you can add as many additional topics on the front of the message, which will be added always after the loglevel. A special delimiter defined at zmq.log.handlers.TOPIC_DELIM is scanned by the PUBHandler, so if you pass your own subtopics prior to that symbol, they will be stripped from the message and added to the topic tree:

>>> log_msg  = "hello there"
>>> subtopic = "sub.topic"
>>> msg = zmq.log.handlers.TOPIC_DELIM.join([subtopic, log_msg])
>>> logger.warn(msg)
>>> print sub.recv_multipart()
['myprogram.WARN.sub.topic', 'hello there']

Tunneling PyZMQ Connections with SSH¶

New in version 2.1.9.

You may want to connect ØMQ sockets across machines, or untrusted networks. One common way to do this is to tunnel the connection via SSH. IPython introduced some tools for tunneling ØMQ connections over ssh in simple cases. These functions have been brought into pyzmq as zmq.ssh under IPython’s BSD license.

PyZMQ will use the shell ssh command via pexpect by default, but it also supports using paramiko for tunnels, so it should work on Windows.

Note

pexpect has no Python3 support at this time, so Python 3 users should get Thomas Kluyver’s pexpect-u fork.

An SSH tunnel has five basic components:

server : the SSH server through which the tunnel will be created
remote ip : the IP of the remote machine as seen from the server (remote ip may be, but is not not generally the same machine as server).
remote port : the port on the remote machine that you want to connect to.
local ip : the interface on your local machine you want to use (default: 127.0.0.1)
local port : the local port you want to forward to the remote port (default: high random)

So once you have established the tunnel, connections to localip:localport will actually be connections to remoteip:remoteport.

In most cases, you have a zeromq url for a remote machine, but you need to tunnel the connection through an ssh server. This is

So if you would use this command from the same LAN as the remote machine:

sock.connect("tcp://10.0.1.2:5555")

to make the same connection from another machine that is outside the network, but you have ssh access to a machine server on the same LAN, you would simply do:

from zmq import ssh
ssh.tunnel_connection(sock, "tcp://10.0.1.2:5555", "server")

Note that "server" can actually be a fully specified "user@server:port" ssh url. Since this really just launches a shell command, all your ssh configuration of usernames, aliases, keys, etc. will be respected. If necessary, tunnel_connection() does take arguments for specific passwords, private keys (the ssh -i option), and non-default choice of whether to use paramiko.

If you are on the same network as the machine, but it is only listening on localhost, you can still connect by making the machine itself the server, and using loopback as the remote ip:

from zmq import ssh
ssh.tunnel_connection(sock, "tcp://127.0.0.1:5555", "10.0.1.2")

The tunnel_connection() function is a simple utility that forwards a random localhost port to the real destination, and connects a socket to the new local url, rather than the remote one that wouldn’t actually work.

See also

A short discussion of ssh tunnels: http://www.revsys.com/writings/quicktips/ssh-tunnel.html

Notes from developing PyZMQ¶

PyZMQ, Python2.5, and Python3¶

PyZMQ is a fairly light, low-level library, so supporting as many versions as is reasonable is our goal. Currently, we support at least Python 2.5-3.1. Making the changes to the codebase required a few tricks, which are documented here for future reference, either by us or by other developers looking to support several versions of Python.

Note

It is far simpler to support 2.6-3.x than to include 2.5. Many of the significant syntax changes have been backported to 2.6, so just writing new-style code would work in many cases. I will try to note these points as they come up.

pyversion_compat.h¶

Many functions we use, primarily involved in converting between C-buffers and Python objects, are not available on all supported versions of Python. In order to resolve missing symbols, we added a header utils/pyversion_compat.h that defines missing symbols with macros. Some of these macros alias new names to old functions (e.g. PyBytes_AsString), so that we can call new-style functions on older versions, and some simply define the function as an empty exception raiser. The important thing is that the symbols are defined to prevent compiler warnings and linking errors. Everywhere we use C-API functions that may not be available in a supported version, at the top of the file is the code:

cdef extern from "pyversion_compat.h":
    pass

This ensures that the symbols are defined in the Cython generated C-code. Higher level switching logic exists in the code itself, to prevent actually calling unavailable functions, but the symbols must still be defined.

Bytes and Strings¶

Note

If you are using Python >= 2.6, to prepare your PyZMQ code for Python3 you should use the b'message' syntax to ensure all your string literal messages will still be bytes after you make the upgrade.

The most cumbersome part of PyZMQ compatibility from a user’s perspective is the fact that, since ØMQ uses C-strings, and would like to do so without copying, we must use the Py3k bytes object, which is backported to 2.6. In order to do this in a Python-version independent way, we added a small utility that unambiguously defines the string types: bytes, unicode, basestring. This is important, because str means different things on 2.x and 3.x, and bytes is undefined on 2.5, and both unicode and basestring are undefined on 3.x. All typechecking in PyZMQ is done against these types:

Explicit Type	2.x	3.x
`bytes`	`str`	`bytes`
`unicode`	`unicode`	`str`
`basestring`	`basestring`	`(str, bytes)`

Note

2.5 specific

Where we really noticed the issue of bytes vs strings coming up for users was in updating the tests to run on every version. Since the b'bytes literal' syntax was not backported to 2.5, we must call "message".encode() for every string in the test suite.

See also

Unicode discussion for more information on strings/bytes.

`PyBytes_*`¶

The standard C-API function for turning a C-string into a Python string was a set of functions with the prefix PyString_*. However, with the Unicode changes made in Python3, this was broken into PyBytes_* for bytes objects and PyUnicode_* for unicode objects. We changed all our PyString_* code to PyBytes_*, which was backported to 2.6.

Note

2.5 Specific:

Since Python 2.5 doesn’t support the PyBytes_* functions, we had to alias them to the PyString_* methods in utils/pyversion_compat.h.

#define PyBytes_FromStringAndSize PyString_FromStringAndSize
#define PyBytes_FromString PyString_FromString
#define PyBytes_AsString PyString_AsString
#define PyBytes_Size PyString_Size

Buffers¶

The layer that is most complicated for developers, but shouldn’t trouble users, is the Python C-Buffer APIs. These are the methods for converting between Python objects and C buffers. The reason it is complicated is that it keeps changing.

There are two buffer interfaces for converting an object to a C-buffer, known as new-style and old-style. Old-style buffers were introduced long ago, but the new-style is only backported to 2.6. The old-style buffer interface is not available in 3.x. There is also an old- and new-style interface for creating Python objects that view C-memory. The old-style object is called a buffer, and the new-style object is memoryview. Unlike the new-style buffer interface for objects, memoryview has only been backported to 2.7. This means that the available buffer-related functions are not the same in any two versions of Python 2.5, 2.6, 2.7, or 3.1.

We have a utils/buffers.pxd file that defines our asbuffer() and frombuffer() functions. utils/buffers.pxd was adapted from mpi4py‘s asbuffer.pxi. The frombuffer() functionality was added. These functions internally switch based on Python version to call the appropriate C-API functions.

See also

Python Buffer API

`str`¶

As discussed, str is not a platform independent type. The two places where we are required to return native str objects are error.strerror(), and Message.__str__(). In both of these cases, the natural return is actually a bytes object. In the methods, the native str type is checked, and if the native str is actually unicode, then we decode the bytes into unicode:

# ...
b = natural_result()
if str is unicode:
    return b.decode()
else:
    return b

Exceptions¶

Note

This section is only relevant for supporting Python 2.5 and 3.x, not for 2.6-3.x.

The syntax for handling exceptions has changed in Python 3. The old syntax:

try:
    s.send(msg)
except zmq.ZMQError, e:
    handle(e)

is no longer valid in Python 3. Instead, the new syntax for this is:

try:
    s.send(msg)
except zmq.ZMQError as e:
    handle(e)

This new syntax is backported to Python 2.6, but is invalid on 2.5. For 2.6-3.x compatible code, we could just use the new syntax. However, the only method we found to catch an exception for handling on both 2.5 and 3.1 is to get the exception object inside the exception block:

try:
    s.send(msg)
except zmq.ZMQError:
    e = sys.exc_info()[1]
    handle(e)

This is certainly not as elegant as either the old or new syntax, but it’s the only way we have found to work everywhere.

See also

PEP-3110

PyZMQ and Unicode¶

PyZMQ is built with an eye towards an easy transition to Python 3, and part of that is dealing with unicode strings. This is an overview of some of what we found, and what it means for PyZMQ.

First, Unicode in Python 2 and 3¶

In Python < 3, a str object is really a C string with some sugar - a specific series of bytes with some fun methods like endswith() and split(). In 2.0, the unicode object was added, which handles different methods of encoding. In Python 3, however, the meaning of str changes. A str in Python 3 is a full unicode object, with encoding and everything. If you want a C string with some sugar, there is a new object called bytes, that behaves much like the 2.x str. The idea is that for a user, a string is a series of characters, not a series of bytes. For simple ascii, the two are interchangeable, but if you consider accents and non-Latin characters, then the character meaning of byte sequences can be ambiguous, since it depends on the encoding scheme. They decided to avoid the ambiguity by forcing users who want the actual bytes to specify the encoding every time they want to convert a string to bytes. That way, users are aware of the difference between a series of bytes and a collection of characters, and don’t confuse the two, as happens in Python 2.x.

The problems (on both sides) come from the fact that regardless of the language design, users are mostly going to use str objects to represent collections of characters, and the behavior of that object is dramatically different in certain aspects between the 2.x bytes approach and the 3.x unicode approach. The unicode approach has the advantage of removing byte ambiguity - it’s a list of characters, not bytes. However, if you really do want the bytes, it’s very inefficient to get them. The bytes approach has the advantage of efficiency. A bytes object really is just a char* pointer with some methods to be used on it, so when interacting with, so interacting with C code, etc is highly efficient and straightforward. However, understanding a bytes object as a string with extended characters introduces ambiguity and possibly confusion.

To avoid ambiguity, hereafter we will refer to encoded C arrays as ‘bytes’ and abstract unicode objects as ‘strings’.

Unicode Buffers¶

Since unicode objects have a wide range of representations, they are not stored as the bytes according to their encoding, but rather in a format called UCS (an older fixed-width Unicode format). On some platforms (OS X, Windows), the storage is UCS-2, which is 2 bytes per character. On most *ix systems, it is UCS-4, or 4 bytes per character. The contents of the buffer of a unicode object are not encoding dependent (always UCS-2 or UCS-4), but they are platform dependent. As a result of this, and the further insistence on not interpreting unicode objects as bytes without specifying encoding, str objects in Python 3 don’t even provide the buffer interface. You simply cannot get the raw bytes of a unicode object without specifying the encoding for the bytes. In Python 2.x, you can get to the raw buffer, but the platform dependence and the fact that the encoding of the buffer is not the encoding of the object makes it very confusing, so this is probably a good move.

The efficiency problem here comes from the fact that simple ascii strings are 4x as big in memory as they need to be (on most Linux, 2x on other platforms). Also, to translate to/from C code that works with char*, you always have to copy data and encode/decode the bytes. This really is horribly inefficient from a memory standpoint. Essentially, Where memory efficiency matters to you, you should never ever use strings; use bytes. The problem is that users will almost always use str, and in 2.x they are efficient, but in 3.x they are not. We want to make sure that we don’t help the user make this mistake, so we ensure that zmq methods don’t try to hide what strings really are.

What This Means for PyZMQ¶

PyZMQ is a wrapper for a C library, so it really should use bytes, since a string is not a simple wrapper for char * like it used to be, but an abstract sequence of characters. The representations of bytes in Python are either the bytes object itself, or any object that provides the buffer interface (aka memoryview). In Python 2.x, unicode objects do provide the buffer interface, but as they do not in Python 3, where pyzmq requires bytes, we specifically reject unicode objects.

The relevant methods here are socket.send/recv, socket.get/setsockopt, socket.bind/connect. The important consideration for send/recv and set/getsockopt is that when you put in something, you really should get the same object back with its partner method. We can easily coerce unicode objects to bytes with send/setsockopt, but the problem is that the pair method of recv/getsockopt will always be bytes, and there should be symmetry. We certainly shouldn’t try to always decode on the retrieval side, because if users just want bytes, then we are potentially using up enormous amounts of excess memory unnecessarily, due to copying and larger memory footprint of unicode strings.

Still, we recognize the fact that users will quite frequently have unicode strings that they want to send, so we have added socket.<method>_string() wrappers. These methods simply wrap their bytes counterpart by encoding to/decoding from bytes around them, and they all take an encoding keyword argument that defaults to utf-8. Since encoding and decoding are necessary to translate between unicode and bytes, it is impossible to perform non-copying actions with these wrappers.

socket.bind/connect methods are different from these, in that they are strictly setters and there is not corresponding getter method. As a result, we feel that we can safely coerce unicode objects to bytes (always to utf-8) in these methods.

Note

For cross-language symmetry (including Python 3), the _unicode methods are now _string. Many languages have a notion of native strings, and the use of _unicode was wedded too closely to the name of such objects in Python 2. For the time being, anywhere you see _string, _unicode also works, and is the only option in pyzmq ≤ 2.1.11.

The Methods¶

Overview of the relevant methods:

socket.bind(self, addr)¶: addr is bytes or unicode. If unicode, encoded to utf-8 bytes

socket.connect(self, addr)¶: addr is bytes or unicode. If unicode, encoded to utf-8 bytes

socket.send(self, object obj, flags=0, copy=True)¶

obj is bytes or provides buffer interface.

if obj is unicode, raise TypeError

socket.recv(self, flags=0, copy=True)¶

returns bytes if copy=True

returns zmq.Message if copy=False:

message.buffer is a buffer view of the bytes

str(message) provides the bytes

unicode(message) decodes message.buffer with utf-8

socket.send_string(self, unicode s, flags=0, encoding='utf-8')¶

takes a unicode string s, and sends the bytes after encoding without an extra copy, via:

socket.send(s.encode(encoding), flags, copy=False)

socket.recv_string(self, flags=0, encoding='utf-8')¶

always returns unicode string

there will be a UnicodeError if it cannot decode the buffer

performs non-copying recv, and decodes the buffer with encoding

socket.setsockopt(self, opt, optval)¶

only accepts bytes for optval (or int, depending on opt)

TypeError if unicode or anything else

socket.getsockopt(self, opt)¶: returns bytes (or int), never unicode

socket.setsockopt_string(self, opt, unicode optval, encoding='utf-8')¶

accepts unicode string for optval

encodes optval with encoding before passing the bytes to setsockopt

socket.getsockopt_string(self, opt, encoding='utf-8')¶

always returns unicode string, after decoding with encoding

note that zmq.IDENTITY is the only sockopt with a string value that can be queried with getsockopt

Indices and tables¶

Links¶

ØMQ Home
The ØMQ Guide
PyZMQ Installation notes on the ZeroMQ website
PyZMQ on GitHub
Issue Tracker

PyZMQ Documentation¶

Supported LibZMQ¶

Using PyZMQ¶

The PyZMQ API¶

zmq¶

Basic Classes¶

Polling¶

Exceptions¶

Functions¶

devices¶

Functions¶

Module: zmq.devices¶

Base Devices¶

Proxy Devices¶

MonitoredQueue Devices¶

green¶

Module: green¶

Usage¶

auth¶

Module: auth¶

Functions¶

auth.thread¶

Module: auth.thread¶

Classes¶

auth.ioloop¶

Module: auth.ioloop¶

utils.win32¶

Module: zmq.utils.win32¶

More Than Just Bindings¶

The Core as Bindings¶

Thread Safety¶

Socket Options as Attributes¶

Default Options on the Context¶

Core Extensions¶

Builtin Serialization¶

MessageTracker¶

Extensions¶

Serializing messages with PyZMQ¶

Builtin serialization¶

Using your own serialization¶

Devices in PyZMQ¶

BackgroundDevices¶

MonitoredQueue¶

Eventloops and PyZMQ¶

Tornado IOLoop¶

install()¶

send()¶

on_recv()¶

on_recv_stream()¶

flush()¶

PyZMQ and gevent¶

Asynchronous Logging via PyZMQ¶

PUB/SUB and Topics¶

PUBHandler¶

Subtopics¶

Tunneling PyZMQ Connections with SSH¶

Notes from developing PyZMQ¶

PyZMQ, Python2.5, and Python3¶

pyversion_compat.h¶

Bytes and Strings¶

PyBytes_*¶

Buffers¶

__str__¶

Exceptions¶

PyZMQ and Unicode¶

First, Unicode in Python 2 and 3¶

Unicode Buffers¶

What This Means for PyZMQ¶

The Methods¶

Indices and tables¶

Links¶

Module: `zmq.devices`¶

Module: `green`¶

Module: `auth`¶

Module: `auth.thread`¶

Module: `auth.ioloop`¶

Module: `zmq.utils.win32`¶

`install()`¶

`send()`¶

`on_recv()`¶

`on_recv_stream()`¶

`flush()`¶

`PyBytes_*`¶

`str`¶