Messaging in IPython
Versioning
The IPython message specification is versioned independently of IPython. The current version of the specification is 5.0.
Introduction
This document explains the basic communications design and messaging specification for how the various IPython objects interact over a network transport. The current implementation uses the ZeroMQ_ library for messaging within and between hosts.
Note
This document should be considered the authoritative description of the IPython messaging protocol, and all developers are strongly encouraged to keep it updated as the implementation evolves, so that we have a single common reference for all protocol details.
The basic design is explained in the following diagram:
![IPython kernel/frontend messaging architecture.](/upstream/ipython/raw/4017ee41f0b5a77b23a60f5c2986b9309de91c77/docs/source/development/figs/frontend-kernel.png)
A single kernel can be simultaneously connected to one or more frontends. The kernel has three sockets that serve the following functions:
Shell: this single ROUTER socket allows multiple incoming connections from frontends, and this is the socket where requests for code execution, object information, prompts, etc. are made to the kernel by any frontend. The communication on this socket is a sequence of request/reply actions from each frontend and the kernel.
IOPub: this socket is the 'broadcast channel' where the kernel publishes all side effects (stdout, stderr, etc.) as well as the requests coming from any client over the shell socket and its own requests on the stdin socket. There are a number of actions in Python which generate side effects: :func:`print` writes to sys.stdout, errors generate tracebacks, etc. Additionally, in a multi-client scenario, we want all frontends to be able to know what each other has sent to the kernel (this can be useful in collaborative scenarios, for example). This socket allows both side effects and the information about communications taking place with one client over the shell channel to be made available to all clients in a uniform manner.
stdin: this ROUTER socket is connected to all frontends, and it allows the kernel to request input from the active frontend when :func:`raw_input` is called. The frontend that executed the code has a DEALER socket that acts as a 'virtual keyboard' for the kernel while this communication is happening (illustrated in the figure by the black outline around the central keyboard). In practice, frontends may display such kernel requests using a special input widget or otherwise indicating that the user is to type input for the kernel instead of normal commands in the frontend.
All messages are tagged with enough information (details below) for clients to know which messages come from their own interaction with the kernel and which ones are from other clients, so they can display each type appropriately.
Control: This channel is identical to Shell, but operates on a separate socket, to allow important messages to avoid queueing behind execution requests (e.g. shutdown or abort).
The actual format of the messages allowed on each of these channels is specified below. Messages are dicts of dicts with string keys and values that are reasonably representable in JSON. Our current implementation uses JSON explicitly as its message format, but this shouldn't be considered a permanent feature. As we've discovered that JSON has non-trivial performance issues due to excessive copying, we may in the future move to a pure pickle-based raw message format. However, it should be possible to easily convert from the raw objects to JSON, since we may have non-python clients (e.g. a web frontend). As long as it's easy to make a JSON version of the objects that is a faithful representation of all the data, we can communicate with such clients.
Note
Not all of these have yet been fully fleshed out, but the key ones are, see kernel and frontend files for actual implementation details.
General Message Format
A message is defined by the following four-dictionary structure:
{ # The message header contains a pair of unique identifiers for the # originating session and the actual message id, in addition to the # username for the process that generated the message. This is useful in # collaborative settings where multiple users may be interacting with the # same kernel simultaneously, so that frontends can label the various # messages in a meaningful way. 'header' : { 'msg_id' : uuid, 'username' : str, 'session' : uuid, # All recognized message type strings are listed below. 'msg_type' : str, # the message protocol version 'version' : '5.0', }, # In a chain of messages, the header from the parent is copied so that # clients can track where messages come from. 'parent_header' : dict, # Any metadata associated with the message. 'metadata' : dict, # The actual content of the message must be a dict, whose structure # depends on the message type. 'content' : dict, }
The Wire Protocol
This message format exists at a high level, but does not describe the actual implementation at the wire level in zeromq. The canonical implementation of the message spec is our :class:`~IPython.kernel.zmq.session.Session` class.
Note
This section should only be relevant to non-Python consumers of the protocol. Python consumers should simply import and use IPython's own implementation of the wire protocol in the :class:`IPython.kernel.zmq.session.Session` object.
Every message is serialized to a sequence of at least six blobs of bytes:
[ b'u-u-i-d', # zmq identity(ies) b'<IDS|MSG>', # delimiter b'baddad42', # HMAC signature b'{header}', # serialized header dict b'{parent_header}', # serialized parent header dict b'{metadata}', # serialized metadata dict b'{content}, # serialized content dict b'blob', # extra raw data buffer(s) ... ]
The front of the message is the ZeroMQ routing prefix, which can be zero or more socket identities. This is every piece of the message prior to the delimiter key <IDS|MSG>. In the case of IOPub, there should be just one prefix component, which is the topic for IOPub subscribers, e.g. execute_result, display_data.
Note
In most cases, the IOPub topics are irrelevant and completely ignored, because frontends just subscribe to all topics. The convention used in the IPython kernel is to use the msg_type as the topic, and possibly extra information about the message, e.g. execute_result or stream.stdout
After the delimiter is the HMAC signature of the message, used for authentication. If authentication is disabled, this should be an empty string. By default, the hashing function used for computing these signatures is sha256.
Note
To disable authentication and signature checking, set the key field of a connection file to an empty string.
The signature is the HMAC hex digest of the concatenation of:
- A shared key (typically the key field of a connection file)
- The serialized header dict
- The serialized parent header dict
- The serialized metadata dict
- The serialized content dict
In Python, this is implemented via:
# once: digester = HMAC(key, digestmod=hashlib.sha256) # for each message d = digester.copy() for serialized_dict in (header, parent, metadata, content): d.update(serialized_dict) signature = d.hexdigest()
After the signature is the actual message, always in four frames of bytes. The four dictionaries that compose a message are serialized separately, in the order of header, parent header, metadata, and content. These can be serialized by any function that turns a dict into bytes. The default and most common serialization is JSON, but msgpack and pickle are common alternatives.
After the serialized dicts are zero to many raw data buffers, which can be used by message types that support binary data (mainly apply and data_pub).
Python functional API
As messages are dicts, they map naturally to a func(**kw) call form. We should develop, at a few key points, functional forms of all the requests that take arguments in this manner and automatically construct the necessary dict for sending.
In addition, the Python implementation of the message specification extends messages upon deserialization to the following form for convenience:
{ 'header' : dict, # The msg's unique identifier and type are always stored in the header, # but the Python implementation copies them to the top level. 'msg_id' : uuid, 'msg_type' : str, 'parent_header' : dict, 'content' : dict, 'metadata' : dict, }
All messages sent to or received by any IPython process should have this extended structure.
Messages on the shell ROUTER/DEALER sockets
Execute
This message type is used by frontends to ask the kernel to execute code on behalf of the user, in a namespace reserved to the user's variables (and thus separate from the kernel's own internal code and variables).
Message type: execute_request:
content = { # Source code to be executed by the kernel, one or more lines. 'code' : str, # A boolean flag which, if True, signals the kernel to execute # this code as quietly as possible. # silent=True forces store_history to be False, # and will *not*: # - broadcast output on the IOPUB channel # - have an execute_result # The default is False. 'silent' : bool, # A boolean flag which, if True, signals the kernel to populate history # The default is True if silent is False. If silent is True, store_history # is forced to be False. 'store_history' : bool, # A dict mapping names to expressions to be evaluated in the # user's dict. The rich display-data representation of each will be evaluated after execution. # See the display_data content for the structure of the representation data. 'user_expressions' : dict, # Some frontends do not support stdin requests. # If raw_input is called from code executed from such a frontend, # a StdinNotImplementedError will be raised. 'allow_stdin' : True, }
The code field contains a single string (possibly multiline) to be executed.
The user_expressions field deserves a detailed explanation. In the past, IPython had the notion of a prompt string that allowed arbitrary code to be evaluated, and this was put to good use by many in creating prompts that displayed system status, path information, and even more esoteric uses like remote instrument status acquired over the network. But now that IPython has a clean separation between the kernel and the clients, the kernel has no prompt knowledge; prompts are a frontend feature, and it should be even possible for different frontends to display different prompts while interacting with the same kernel. user_expressions can be used to retrieve this information.
Any error in evaluating any expression in user_expressions will result in only that key containing a standard error message, of the form:
{ 'status' : 'error', 'ename' : 'NameError', 'evalue' : 'foo', 'traceback' : ... }
Note
In order to obtain the current execution counter for the purposes of displaying input prompts, frontends may make an execution request with an empty code string and silent=True.
Upon completion of the execution request, the kernel always sends a reply, with a status code indicating what happened and additional data depending on the outcome. See :ref:`below <execution_results>` for the possible return codes and associated data.
Execution counter (prompt number)
The kernel should have a single, monotonically increasing counter of all execution requests that are made with store_history=True. This counter is used to populate the In[n] and Out[n] prompts. The value of this counter will be returned as the execution_count field of all execute_reply and execute_input messages.
Execution results
Message type: execute_reply:
content = { # One of: 'ok' OR 'error' OR 'abort' 'status' : str, # The global kernel counter that increases by one with each request that # stores history. This will typically be used by clients to display # prompt numbers to the user. If the request did not store history, this will # be the current value of the counter in the kernel. 'execution_count' : int, }
When status is 'ok', the following extra fields are present:
{ # 'payload' will be a list of payload dicts. # Each execution payload is a dict with string keys that may have been # produced by the code being executed. It is retrieved by the kernel at # the end of the execution and sent back to the front end, which can take # action on it as needed. # The only requirement of each payload dict is that it have a 'source' key, # which is a string classifying the payload (e.g. 'pager'). 'payload' : list(dict), # Results for the user_expressions. 'user_expressions' : dict, }
Execution payloads
The notion of an 'execution payload' is different from a return value of a given set of code, which normally is just displayed on the execute_result stream through the PUB socket. The idea of a payload is to allow special types of code, typically magics, to populate a data container in the IPython kernel that will be shipped back to the caller via this channel. The kernel has an API for this in the PayloadManager:
ip.payload_manager.write_payload(payload_dict)
which appends a dictionary to the list of payloads.
The payload API is not yet stabilized, and should probably not be supported by non-Python kernels at this time. In such cases, the payload list should always be empty.
When status is 'error', the following extra fields are present:
{ 'ename' : str, # Exception name, as a string 'evalue' : str, # Exception value, as a string # The traceback will contain a list of frames, represented each as a # string. For now we'll stick to the existing design of ultraTB, which # controls exception level of detail statefully. But eventually we'll # want to grow into a model where more information is collected and # packed into the traceback object, with clients deciding how little or # how much of it to unpack. But for now, let's start with a simple list # of strings, since that requires only minimal changes to ultratb as # written. 'traceback' : list, }
When status is 'abort', there are for now no additional data fields. This happens when the kernel was interrupted by a signal.
Introspection
Code can be inspected to show useful information to the user. It is up to the Kernel to decide what information should be displayed, and its formatting.
Message type: inspect_request:
content = { # The code context in which introspection is requested # this may be up to an entire multiline cell. 'code' : str, # The cursor position within 'code' (in unicode characters) where inspection is requested 'cursor_pos' : int, # The level of detail desired. In IPython, the default (0) is equivalent to typing # 'x?' at the prompt, 1 is equivalent to 'x??'. # The difference is up to kernels, but in IPython level 1 includes the source code # if available. 'detail_level' : 0 or 1, }
The reply is a mime-bundle, like a display_data message, which should be a formatted representation of information about the context. In the notebook, this is used to show tooltips over function calls, etc.
Message type: inspect_reply:
content = { # 'ok' if the request succeeded or 'error', with error information as in all other replies. 'status' : 'ok', # data can be empty if nothing is found 'data' : dict, 'metadata' : dict, }
Completion
Message type: complete_request:
content = { # The code context in which completion is requested # this may be up to an entire multiline cell, such as # 'foo = a.isal' 'code' : str, # The cursor position within 'code' (in unicode characters) where completion is requested 'cursor_pos' : int, }
Message type: complete_reply:
content = { # The list of all matches to the completion request, such as # ['a.isalnum', 'a.isalpha'] for the above example. 'matches' : list, # The range of text that should be replaced by the above matches when a completion is accepted. # typically cursor_end is the same as cursor_pos in the request. 'cursor_start' : int, 'cursor_end' : int, # Information that frontend plugins might use for extra display information about completions. 'metadata' : dict, # status should be 'ok' unless an exception was raised during the request, # in which case it should be 'error', along with the usual error message content # in other messages. 'status' : 'ok' }
History
For clients to explicitly request history from a kernel. The kernel has all the actual execution history stored in a single location, so clients can request it from the kernel when needed.
Message type: history_request:
content = { # If True, also return output history in the resulting dict. 'output' : bool, # If True, return the raw input history, else the transformed input. 'raw' : bool, # So far, this can be 'range', 'tail' or 'search'. 'hist_access_type' : str, # If hist_access_type is 'range', get a range of input cells. session can # be a positive session number, or a negative number to count back from # the current session. 'session' : int, # start and stop are line numbers within that session. 'start' : int, 'stop' : int, # If hist_access_type is 'tail' or 'search', get the last n cells. 'n' : int, # If hist_access_type is 'search', get cells matching the specified glob # pattern (with * and ? as wildcards). 'pattern' : str, # If hist_access_type is 'search' and unique is true, do not # include duplicated history. Default is false. 'unique' : bool, }
Message type: history_reply:
content = { # A list of 3 tuples, either: # (session, line_number, input) or # (session, line_number, (input, output)), # depending on whether output was False or True, respectively. 'history' : list, }
Connect
When a client connects to the request/reply socket of the kernel, it can issue a connect request to get basic information about the kernel, such as the ports the other ZeroMQ sockets are listening on. This allows clients to only have to know about a single port (the shell channel) to connect to a kernel.
Message type: connect_request:
content = { }
Message type: connect_reply:
content = { 'shell_port' : int, # The port the shell ROUTER socket is listening on. 'iopub_port' : int, # The port the PUB socket is listening on. 'stdin_port' : int, # The port the stdin ROUTER socket is listening on. 'hb_port' : int, # The port the heartbeat socket is listening on. }
Kernel info
If a client needs to know information about the kernel, it can make a request of the kernel's information. This message can be used to fetch core information of the kernel, including language (e.g., Python), language version number and IPython version number, and the IPython message spec version number.
Message type: kernel_info_request:
content = { }
Message type: kernel_info_reply:
content = { # Version of messaging protocol. # The first integer indicates major version. It is incremented when # there is any backward incompatible change. # The second integer indicates minor version. It is incremented when # there is any backward compatible change. 'protocol_version': 'X.Y.Z', # The kernel implementation name # (e.g. 'ipython' for the IPython kernel) 'implementation': str, # Implementation version number. # The version number of the kernel's implementation # (e.g. IPython.__version__ for the IPython kernel) 'implementation_version': 'X.Y.Z', # Programming language in which kernel is implemented. # Kernel included in IPython returns 'python'. 'language': str, # Language version number. # It is Python version number (e.g., '2.7.3') for the kernel # included in IPython. 'language_version': 'X.Y.Z', # A banner of information about the kernel, # which may be desplayed in console environments. 'banner' : str, }
Kernel shutdown
The clients can request the kernel to shut itself down; this is used in multiple cases:
- when the user chooses to close the client application via a menu or window control.
- when the user types 'exit' or 'quit' (or their uppercase magic equivalents).
- when the user chooses a GUI method (like the 'Ctrl-C' shortcut in the IPythonQt client) to force a kernel restart to get a clean kernel without losing client-side state like history or inlined figures.
The client sends a shutdown request to the kernel, and once it receives the reply message (which is otherwise empty), it can assume that the kernel has completed shutdown safely.
Upon their own shutdown, client applications will typically execute a last minute sanity check and forcefully terminate any kernel that is still alive, to avoid leaving stray processes in the user's machine.
Message type: shutdown_request:
content = { 'restart' : bool # whether the shutdown is final, or precedes a restart }
Message type: shutdown_reply:
content = { 'restart' : bool # whether the shutdown is final, or precedes a restart }
Note
When the clients detect a dead kernel thanks to inactivity on the heartbeat socket, they simply send a forceful process termination signal, since a dead process is unlikely to respond in any useful way to messages.
Messages on the PUB/SUB socket
Streams (stdout, stderr, etc)
Message type: stream:
content = { # The name of the stream is one of 'stdout', 'stderr' 'name' : str, # The data is an arbitrary string to be written to that stream 'data' : str, }
Display Data
This type of message is used to bring back data that should be displayed (text, html, svg, etc.) in the frontends. This data is published to all frontends. Each message can have multiple representations of the data; it is up to the frontend to decide which to use and how. A single message should contain all possible representations of the same information. Each representation should be a JSON'able data structure, and should be a valid MIME type.
Some questions remain about this design:
- Do we use this message type for execute_result/displayhook? Probably not, because the displayhook also has to handle the Out prompt display. On the other hand we could put that information into the metadata section.
Message type: display_data:
content = { # Who create the data 'source' : str, # The data dict contains key/value pairs, where the keys are MIME # types and the values are the raw data of the representation in that # format. 'data' : dict, # Any metadata that describes the data 'metadata' : dict }
The metadata contains any metadata that describes the output. Global keys are assumed to apply to the output as a whole. The metadata dict can also contain mime-type keys, which will be sub-dictionaries, which are interpreted as applying only to output of that type. Third parties should put any data they write into a single dict with a reasonably unique name to avoid conflicts.
The only metadata keys currently defined in IPython are the width and height of images:
metadata = { 'image/png' : { 'width': 640, 'height': 480 } }
Raw Data Publication
display_data lets you publish representations of data, such as images and html. This data_pub message lets you publish actual raw data, sent via message buffers.
data_pub messages are constructed via the :func:`IPython.lib.datapub.publish_data` function:
from IPython.kernel.zmq.datapub import publish_data ns = dict(x=my_array) publish_data(ns)
Message type: data_pub:
content = { # the keys of the data dict, after it has been unserialized 'keys' : ['a', 'b'] } # the namespace dict will be serialized in the message buffers, # which will have a length of at least one buffers = [b'pdict', ...]
The interpretation of a sequence of data_pub messages for a given parent request should be to update a single namespace with subsequent results.
Note
No frontends directly handle data_pub messages at this time. It is currently only used by the client/engines in :mod:`IPython.parallel`, where engines may publish data to the Client, of which the Client can then publish representations via display_data to various frontends.
Code inputs
To let all frontends know what code is being executed at any given time, these messages contain a re-broadcast of the code portion of an :ref:`execute_request <execute>`, along with the :ref:`execution_count <execution_counter>`.
Message type: execute_input:
content = { 'code' : str, # Source code to be executed, one or more lines # The counter for this execution is also provided so that clients can # display it, since IPython automatically creates variables called _iN # (for input prompt In[N]). 'execution_count' : int }
Execution results
Results of an execution are published as an execute_result. These are identical to display_data messages, with the addition of an execution_count key.
Results can have multiple simultaneous formats depending on its configuration. A plain text representation should always be provided in the text/plain mime-type. Frontends are free to display any or all of these according to its capabilities. Frontends should ignore mime-types they do not understand. The data itself is any JSON object and depends on the format. It is often, but not always a string.
Message type: execute_result:
content = { # The counter for this execution is also provided so that clients can # display it, since IPython automatically creates variables called _N # (for prompt N). 'execution_count' : int, # data and metadata are identical to a display_data message. # the object being displayed is that passed to the display hook, # i.e. the *result* of the execution. 'data' : dict, 'metadata' : dict, }
Execution errors
When an error occurs during code execution
Message type: error:
content = { # Similar content to the execute_reply messages for the 'error' case, # except the 'status' field is omitted. }
Kernel status
This message type is used by frontends to monitor the status of the kernel.
Message type: status:
content = { # When the kernel starts to handle a message, it will enter the 'busy' # state and when it finishes, it will enter the 'idle' state. # The kernel will publish state 'starting' exactly once at process startup. execution_state : ('busy', 'idle', 'starting') }
Clear output
This message type is used to clear the output that is visible on the frontend.
Message type: clear_output:
content = { # Wait to clear the output until new output is available. Clears the # existing output immediately before the new output is displayed. # Useful for creating simple animations with minimal flickering. 'wait' : bool, }
Messages on the stdin ROUTER/DEALER sockets
This is a socket where the request/reply pattern goes in the opposite direction: from the kernel to a single frontend, and its purpose is to allow raw_input and similar operations that read from sys.stdin on the kernel to be fulfilled by the client. The request should be made to the frontend that made the execution request that prompted raw_input to be called. For now we will keep these messages as simple as possible, since they only mean to convey the raw_input(prompt) call.
Message type: input_request:
content = { # the text to show at the prompt 'prompt' : str, # Is the request for a password? # If so, the frontend shouldn't echo input. 'password' : bool }
Message type: input_reply:
content = { 'value' : str }
When password is True, the frontend should not echo the input as it is entered.
Note
The stdin socket of the client is required to have the same zmq IDENTITY as the client's shell socket. Because of this, the input_request must be sent with the same IDENTITY routing prefix as the execute_reply in order for the frontend to receive the message.
Note
We do not explicitly try to forward the raw sys.stdin object, because in practice the kernel should behave like an interactive program. When a program is opened on the console, the keyboard effectively takes over the stdin file descriptor, and it can't be used for raw reading anymore. Since the IPython kernel effectively behaves like a console program (albeit one whose "keyboard" is actually living in a separate process and transported over the zmq connection), raw stdin isn't expected to be available.
Heartbeat for kernels
Clients send ping messages on a REQ socket, which are echoed right back from the Kernel's REP socket. These are simple bytestrings, not full JSON messages described above.
Custom Messages
IPython 2.0 (msgspec v4.1) adds a messaging system for developers to add their own objects with Frontend and Kernel-side components, and allow them to communicate with each other. To do this, IPython adds a notion of a Comm, which exists on both sides, and can communicate in either direction.
These messages are fully symmetrical - both the Kernel and the Frontend can send each message, and no messages expect a reply. The Kernel listens for these messages on the Shell channel, and the Frontend listens for them on the IOPub channel.
Opening a Comm
Opening a Comm produces a comm_open message, to be sent to the other side:
{ 'comm_id' : 'u-u-i-d', 'target_name' : 'my_comm', 'data' : {} }
Every Comm has an ID and a target name. The code handling the message on the receiving side is responsible for maintaining a mapping of target_name keys to constructors. After a comm_open message has been sent, there should be a corresponding Comm instance on both sides. The data key is always a dict and can be any extra JSON information used in initialization of the comm.
If the target_name key is not found on the receiving side, then it should immediately reply with a comm_close message to avoid an inconsistent state.
Comm Messages
Comm messages are one-way communications to update comm state, used for synchronizing widget state, or simply requesting actions of a comm's counterpart.
Essentially, each comm pair defines their own message specification implemented inside the data dict.
There are no expected replies (of course, one side can send another comm_msg in reply).
Message type: comm_msg:
{ 'comm_id' : 'u-u-i-d', 'data' : {} }
Tearing Down Comms
Since comms live on both sides, when a comm is destroyed the other side must be notified. This is done with a comm_close message.
Message type: comm_close:
{ 'comm_id' : 'u-u-i-d', 'data' : {} }
Output Side Effects
Since comm messages can execute arbitrary user code, handlers should set the parent header and publish status busy / idle, just like an execute request.
To Do
Missing things include:
- Important: finish thinking through the payload concept and API.