"""Base class to manage the interaction with a running kernel""" # Copyright (c) Jupyter Development Team. # Distributed under the terms of the Modified BSD License. import asyncio import inspect import sys import time import typing as t from functools import partial from getpass import getpass from queue import Empty import zmq.asyncio from jupyter_core.utils import ensure_async from traitlets import Any, Bool, Instance, Type from .channels import major_protocol_version from .channelsabc import ChannelABC, HBChannelABC from .clientabc import KernelClientABC from .connect import ConnectionFileMixin from .session import Session # some utilities to validate message structure, these might get moved elsewhere # if they prove to have more generic utility def validate_string_dict(dct: t.Dict[str, str]) -> None: """Validate that the input is a dict with string keys and values. Raises ValueError if not.""" for k, v in dct.items(): if not isinstance(k, str): raise ValueError("key %r in dict must be a string" % k) if not isinstance(v, str): raise ValueError("value %r in dict must be a string" % v) def reqrep(wrapped: t.Callable, meth: t.Callable, channel: str = "shell") -> t.Callable: wrapped = wrapped(meth, channel) if not meth.__doc__: # python -OO removes docstrings, # so don't bother building the wrapped docstring return wrapped basedoc, _ = meth.__doc__.split("Returns\n", 1) parts = [basedoc.strip()] if "Parameters" not in basedoc: parts.append( """ Parameters ---------- """ ) parts.append( """ reply: bool (default: False) Whether to wait for and return reply timeout: float or None (default: None) Timeout to use when waiting for a reply Returns ------- msg_id: str The msg_id of the request sent, if reply=False (default) reply: dict The reply message for this request, if reply=True """ ) wrapped.__doc__ = "\n".join(parts) return wrapped class KernelClient(ConnectionFileMixin): """Communicates with a single kernel on any host via zmq channels. There are five channels associated with each kernel: * shell: for request/reply calls to the kernel. * iopub: for the kernel to publish results to frontends. * hb: for monitoring the kernel's heartbeat. * stdin: for frontends to reply to raw_input calls in the kernel. * control: for kernel management calls to the kernel. The messages that can be sent on these channels are exposed as methods of the client (KernelClient.execute, complete, history, etc.). These methods only send the message, they don't wait for a reply. To get results, use e.g. :meth:`get_shell_msg` to fetch messages from the shell channel. """ # The PyZMQ Context to use for communication with the kernel. context = Instance(zmq.Context) _created_context = Bool(False) def _context_default(self) -> zmq.Context: self._created_context = True return zmq.Context() # The classes to use for the various channels shell_channel_class = Type(ChannelABC) iopub_channel_class = Type(ChannelABC) stdin_channel_class = Type(ChannelABC) hb_channel_class = Type(HBChannelABC) control_channel_class = Type(ChannelABC) # Protected traits _shell_channel = Any() _iopub_channel = Any() _stdin_channel = Any() _hb_channel = Any() _control_channel = Any() # flag for whether execute requests should be allowed to call raw_input: allow_stdin: bool = True def __del__(self) -> None: """Handle garbage collection. Destroy context if applicable.""" if ( self._created_context and self.context is not None # type:ignore[redundant-expr] and not self.context.closed ): if self.channels_running: if self.log: self.log.warning("Could not destroy zmq context for %s", self) else: if self.log: self.log.debug("Destroying zmq context for %s", self) self.context.destroy() try: super_del = super().__del__ # type:ignore[misc] except AttributeError: pass else: super_del() # -------------------------------------------------------------------------- # Channel proxy methods # -------------------------------------------------------------------------- async def _async_get_shell_msg(self, *args: t.Any, **kwargs: t.Any) -> t.Dict[str, t.Any]: """Get a message from the shell channel""" return await ensure_async(self.shell_channel.get_msg(*args, **kwargs)) async def _async_get_iopub_msg(self, *args: t.Any, **kwargs: t.Any) -> t.Dict[str, t.Any]: """Get a message from the iopub channel""" return await ensure_async(self.iopub_channel.get_msg(*args, **kwargs)) async def _async_get_stdin_msg(self, *args: t.Any, **kwargs: t.Any) -> t.Dict[str, t.Any]: """Get a message from the stdin channel""" return await ensure_async(self.stdin_channel.get_msg(*args, **kwargs)) async def _async_get_control_msg(self, *args: t.Any, **kwargs: t.Any) -> t.Dict[str, t.Any]: """Get a message from the control channel""" return await ensure_async(self.control_channel.get_msg(*args, **kwargs)) async def _async_wait_for_ready(self, timeout: t.Optional[float] = None) -> None: """Waits for a response when a client is blocked - Sets future time for timeout - Blocks on shell channel until a message is received - Exit if the kernel has died - If client times out before receiving a message from the kernel, send RuntimeError - Flush the IOPub channel """ if timeout is None: timeout = float("inf") abs_timeout = time.time() + timeout from .manager import KernelManager if not isinstance(self.parent, KernelManager): # This Client was not created by a KernelManager, # so wait for kernel to become responsive to heartbeats # before checking for kernel_info reply while not await self._async_is_alive(): if time.time() > abs_timeout: raise RuntimeError( "Kernel didn't respond to heartbeats in %d seconds and timed out" % timeout ) await asyncio.sleep(0.2) # Wait for kernel info reply on shell channel while True: self.kernel_info() try: msg = await ensure_async(self.shell_channel.get_msg(timeout=1)) except Empty: pass else: if msg["msg_type"] == "kernel_info_reply": # Checking that IOPub is connected. If it is not connected, start over. try: await ensure_async(self.iopub_channel.get_msg(timeout=0.2)) except Empty: pass else: self._handle_kernel_info_reply(msg) break if not await self._async_is_alive(): msg = "Kernel died before replying to kernel_info" raise RuntimeError(msg) # Check if current time is ready check time plus timeout if time.time() > abs_timeout: raise RuntimeError("Kernel didn't respond in %d seconds" % timeout) # Flush IOPub channel while True: try: msg = await ensure_async(self.iopub_channel.get_msg(timeout=0.2)) except Empty: break async def _async_recv_reply( self, msg_id: str, timeout: t.Optional[float] = None, channel: str = "shell" ) -> t.Dict[str, t.Any]: """Receive and return the reply for a given request""" if timeout is not None: deadline = time.monotonic() + timeout while True: if timeout is not None: timeout = max(0, deadline - time.monotonic()) try: if channel == "control": reply = await self._async_get_control_msg(timeout=timeout) else: reply = await self._async_get_shell_msg(timeout=timeout) except Empty as e: msg = "Timeout waiting for reply" raise TimeoutError(msg) from e if reply["parent_header"].get("msg_id") != msg_id: # not my reply, someone may have forgotten to retrieve theirs continue return reply async def _stdin_hook_default(self, msg: t.Dict[str, t.Any]) -> None: """Handle an input request""" content = msg["content"] prompt = getpass if content.get("password", False) else input try: raw_data = prompt(content["prompt"]) # type:ignore[operator] except EOFError: # turn EOFError into EOF character raw_data = "\x04" except KeyboardInterrupt: sys.stdout.write("\n") return # only send stdin reply if there *was not* another request # or execution finished while we were reading. if not (await self.stdin_channel.msg_ready() or await self.shell_channel.msg_ready()): self.input(raw_data) def _output_hook_default(self, msg: t.Dict[str, t.Any]) -> None: """Default hook for redisplaying plain-text output""" msg_type = msg["header"]["msg_type"] content = msg["content"] if msg_type == "stream": stream = getattr(sys, content["name"]) stream.write(content["text"]) elif msg_type in ("display_data", "execute_result"): sys.stdout.write(content["data"].get("text/plain", "")) elif msg_type == "error": sys.stderr.write("\n".join(content["traceback"])) def _output_hook_kernel( self, session: Session, socket: zmq.sugar.socket.Socket, parent_header: t.Any, msg: t.Dict[str, t.Any], ) -> None: """Output hook when running inside an IPython kernel adds rich output support. """ msg_type = msg["header"]["msg_type"] if msg_type in ("display_data", "execute_result", "error"): session.send(socket, msg_type, msg["content"], parent=parent_header) else: self._output_hook_default(msg) # -------------------------------------------------------------------------- # Channel management methods # -------------------------------------------------------------------------- def start_channels( self, shell: bool = True, iopub: bool = True, stdin: bool = True, hb: bool = True, control: bool = True, ) -> None: """Starts the channels for this kernel. This will create the channels if they do not exist and then start them (their activity runs in a thread). If port numbers of 0 are being used (random ports) then you must first call :meth:`start_kernel`. If the channels have been stopped and you call this, :class:`RuntimeError` will be raised. """ if iopub: self.iopub_channel.start() if shell: self.shell_channel.start() if stdin: self.stdin_channel.start() self.allow_stdin = True else: self.allow_stdin = False if hb: self.hb_channel.start() if control: self.control_channel.start() def stop_channels(self) -> None: """Stops all the running channels for this kernel. This stops their event loops and joins their threads. """ if self.shell_channel.is_alive(): self.shell_channel.stop() if self.iopub_channel.is_alive(): self.iopub_channel.stop() if self.stdin_channel.is_alive(): self.stdin_channel.stop() if self.hb_channel.is_alive(): self.hb_channel.stop() if self.control_channel.is_alive(): self.control_channel.stop() @property def channels_running(self) -> bool: """Are any of the channels created and running?""" return ( (self._shell_channel and self.shell_channel.is_alive()) or (self._iopub_channel and self.iopub_channel.is_alive()) or (self._stdin_channel and self.stdin_channel.is_alive()) or (self._hb_channel and self.hb_channel.is_alive()) or (self._control_channel and self.control_channel.is_alive()) ) ioloop = None # Overridden in subclasses that use pyzmq event loop @property def shell_channel(self) -> t.Any: """Get the shell channel object for this kernel.""" if self._shell_channel is None: url = self._make_url("shell") self.log.debug("connecting shell channel to %s", url) socket = self.connect_shell(identity=self.session.bsession) self._shell_channel = self.shell_channel_class( # type:ignore[call-arg,abstract] socket, self.session, self.ioloop ) return self._shell_channel @property def iopub_channel(self) -> t.Any: """Get the iopub channel object for this kernel.""" if self._iopub_channel is None: url = self._make_url("iopub") self.log.debug("connecting iopub channel to %s", url) socket = self.connect_iopub() self._iopub_channel = self.iopub_channel_class( # type:ignore[call-arg,abstract] socket, self.session, self.ioloop ) return self._iopub_channel @property def stdin_channel(self) -> t.Any: """Get the stdin channel object for this kernel.""" if self._stdin_channel is None: url = self._make_url("stdin") self.log.debug("connecting stdin channel to %s", url) socket = self.connect_stdin(identity=self.session.bsession) self._stdin_channel = self.stdin_channel_class( # type:ignore[call-arg,abstract] socket, self.session, self.ioloop ) return self._stdin_channel @property def hb_channel(self) -> t.Any: """Get the hb channel object for this kernel.""" if self._hb_channel is None: url = self._make_url("hb") self.log.debug("connecting heartbeat channel to %s", url) self._hb_channel = self.hb_channel_class( # type:ignore[call-arg,abstract] self.context, self.session, url ) return self._hb_channel @property def control_channel(self) -> t.Any: """Get the control channel object for this kernel.""" if self._control_channel is None: url = self._make_url("control") self.log.debug("connecting control channel to %s", url) socket = self.connect_control(identity=self.session.bsession) self._control_channel = self.control_channel_class( # type:ignore[call-arg,abstract] socket, self.session, self.ioloop ) return self._control_channel async def _async_is_alive(self) -> bool: """Is the kernel process still running?""" from .manager import KernelManager if isinstance(self.parent, KernelManager): # This KernelClient was created by a KernelManager, # we can ask the parent KernelManager: return await self.parent._async_is_alive() if self._hb_channel is not None: # We don't have access to the KernelManager, # so we use the heartbeat. return self._hb_channel.is_beating() # no heartbeat and not local, we can't tell if it's running, # so naively return True return True async def _async_execute_interactive( self, code: str, silent: bool = False, store_history: bool = True, user_expressions: t.Optional[t.Dict[str, t.Any]] = None, allow_stdin: t.Optional[bool] = None, stop_on_error: bool = True, timeout: t.Optional[float] = None, output_hook: t.Optional[t.Callable] = None, stdin_hook: t.Optional[t.Callable] = None, ) -> t.Dict[str, t.Any]: """Execute code in the kernel interactively Output will be redisplayed, and stdin prompts will be relayed as well. If an IPython kernel is detected, rich output will be displayed. You can pass a custom output_hook callable that will be called with every IOPub message that is produced instead of the default redisplay. .. versionadded:: 5.0 Parameters ---------- code : str A string of code in the kernel's language. silent : bool, optional (default False) If set, the kernel will execute the code as quietly possible, and will force store_history to be False. store_history : bool, optional (default True) If set, the kernel will store command history. This is forced to be False if silent is True. user_expressions : dict, optional A dict mapping names to expressions to be evaluated in the user's dict. The expression values are returned as strings formatted using :func:`repr`. allow_stdin : bool, optional (default self.allow_stdin) Flag for whether the kernel can send stdin requests to frontends. Some frontends (e.g. the Notebook) do not support stdin requests. If raw_input is called from code executed from such a frontend, a StdinNotImplementedError will be raised. stop_on_error: bool, optional (default True) Flag whether to abort the execution queue, if an exception is encountered. timeout: float or None (default: None) Timeout to use when waiting for a reply output_hook: callable(msg) Function to be called with output messages. If not specified, output will be redisplayed. stdin_hook: callable(msg) Function or awaitable to be called with stdin_request messages. If not specified, input/getpass will be called. Returns ------- reply: dict The reply message for this request """ if not self.iopub_channel.is_alive(): emsg = "IOPub channel must be running to receive output" raise RuntimeError(emsg) if allow_stdin is None: allow_stdin = self.allow_stdin if allow_stdin and not self.stdin_channel.is_alive(): emsg = "stdin channel must be running to allow input" raise RuntimeError(emsg) msg_id = await ensure_async( self.execute( code, silent=silent, store_history=store_history, user_expressions=user_expressions, allow_stdin=allow_stdin, stop_on_error=stop_on_error, ) ) if stdin_hook is None: stdin_hook = self._stdin_hook_default # detect IPython kernel if output_hook is None and "IPython" in sys.modules: from IPython import get_ipython ip = get_ipython() # type:ignore[no-untyped-call] in_kernel = getattr(ip, "kernel", False) if in_kernel: output_hook = partial( self._output_hook_kernel, ip.display_pub.session, ip.display_pub.pub_socket, ip.display_pub.parent_header, ) if output_hook is None: # default: redisplay plain-text outputs output_hook = self._output_hook_default # set deadline based on timeout if timeout is not None: deadline = time.monotonic() + timeout else: timeout_ms = None poller = zmq.asyncio.Poller() iopub_socket = self.iopub_channel.socket poller.register(iopub_socket, zmq.POLLIN) if allow_stdin: stdin_socket = self.stdin_channel.socket poller.register(stdin_socket, zmq.POLLIN) else: stdin_socket = None # wait for output and redisplay it while True: if timeout is not None: timeout = max(0, deadline - time.monotonic()) timeout_ms = int(1000 * timeout) events = dict(await poller.poll(timeout_ms)) if not events: emsg = "Timeout waiting for output" raise TimeoutError(emsg) if stdin_socket in events: req = await ensure_async(self.stdin_channel.get_msg(timeout=0)) res = stdin_hook(req) if inspect.isawaitable(res): await res continue if iopub_socket not in events: continue msg = await ensure_async(self.iopub_channel.get_msg(timeout=0)) if msg["parent_header"].get("msg_id") != msg_id: # not from my request continue output_hook(msg) # stop on idle if ( msg["header"]["msg_type"] == "status" and msg["content"]["execution_state"] == "idle" ): break # output is done, get the reply if timeout is not None: timeout = max(0, deadline - time.monotonic()) return await self._async_recv_reply(msg_id, timeout=timeout) # Methods to send specific messages on channels def execute( self, code: str, silent: bool = False, store_history: bool = True, user_expressions: t.Optional[t.Dict[str, t.Any]] = None, allow_stdin: t.Optional[bool] = None, stop_on_error: bool = True, ) -> str: """Execute code in the kernel. Parameters ---------- code : str A string of code in the kernel's language. silent : bool, optional (default False) If set, the kernel will execute the code as quietly possible, and will force store_history to be False. store_history : bool, optional (default True) If set, the kernel will store command history. This is forced to be False if silent is True. user_expressions : dict, optional A dict mapping names to expressions to be evaluated in the user's dict. The expression values are returned as strings formatted using :func:`repr`. allow_stdin : bool, optional (default self.allow_stdin) Flag for whether the kernel can send stdin requests to frontends. Some frontends (e.g. the Notebook) do not support stdin requests. If raw_input is called from code executed from such a frontend, a StdinNotImplementedError will be raised. stop_on_error: bool, optional (default True) Flag whether to abort the execution queue, if an exception is encountered. Returns ------- The msg_id of the message sent. """ if user_expressions is None: user_expressions = {} if allow_stdin is None: allow_stdin = self.allow_stdin # Don't waste network traffic if inputs are invalid if not isinstance(code, str): raise ValueError("code %r must be a string" % code) validate_string_dict(user_expressions) # Create class for content/msg creation. Related to, but possibly # not in Session. content = { "code": code, "silent": silent, "store_history": store_history, "user_expressions": user_expressions, "allow_stdin": allow_stdin, "stop_on_error": stop_on_error, } msg = self.session.msg("execute_request", content) self.shell_channel.send(msg) return msg["header"]["msg_id"] def complete(self, code: str, cursor_pos: t.Optional[int] = None) -> str: """Tab complete text in the kernel's namespace. Parameters ---------- code : str The context in which completion is requested. Can be anything between a variable name and an entire cell. cursor_pos : int, optional The position of the cursor in the block of code where the completion was requested. Default: ``len(code)`` Returns ------- The msg_id of the message sent. """ if cursor_pos is None: cursor_pos = len(code) content = {"code": code, "cursor_pos": cursor_pos} msg = self.session.msg("complete_request", content) self.shell_channel.send(msg) return msg["header"]["msg_id"] def inspect(self, code: str, cursor_pos: t.Optional[int] = None, detail_level: int = 0) -> str: """Get metadata information about an object in the kernel's namespace. It is up to the kernel to determine the appropriate object to inspect. Parameters ---------- code : str The context in which info is requested. Can be anything between a variable name and an entire cell. cursor_pos : int, optional The position of the cursor in the block of code where the info was requested. Default: ``len(code)`` detail_level : int, optional The level of detail for the introspection (0-2) Returns ------- The msg_id of the message sent. """ if cursor_pos is None: cursor_pos = len(code) content = { "code": code, "cursor_pos": cursor_pos, "detail_level": detail_level, } msg = self.session.msg("inspect_request", content) self.shell_channel.send(msg) return msg["header"]["msg_id"] def history( self, raw: bool = True, output: bool = False, hist_access_type: str = "range", **kwargs: t.Any, ) -> str: """Get entries from the kernel's history list. Parameters ---------- raw : bool If True, return the raw input. output : bool If True, then return the output as well. hist_access_type : str 'range' (fill in session, start and stop params), 'tail' (fill in n) or 'search' (fill in pattern param). session : int For a range request, the session from which to get lines. Session numbers are positive integers; negative ones count back from the current session. start : int The first line number of a history range. stop : int The final (excluded) line number of a history range. n : int The number of lines of history to get for a tail request. pattern : str The glob-syntax pattern for a search request. Returns ------- The ID of the message sent. """ if hist_access_type == "range": kwargs.setdefault("session", 0) kwargs.setdefault("start", 0) content = dict(raw=raw, output=output, hist_access_type=hist_access_type, **kwargs) msg = self.session.msg("history_request", content) self.shell_channel.send(msg) return msg["header"]["msg_id"] def kernel_info(self) -> str: """Request kernel info Returns ------- The msg_id of the message sent """ msg = self.session.msg("kernel_info_request") self.shell_channel.send(msg) return msg["header"]["msg_id"] def comm_info(self, target_name: t.Optional[str] = None) -> str: """Request comm info Returns ------- The msg_id of the message sent """ content = {} if target_name is None else {"target_name": target_name} msg = self.session.msg("comm_info_request", content) self.shell_channel.send(msg) return msg["header"]["msg_id"] def _handle_kernel_info_reply(self, msg: t.Dict[str, t.Any]) -> None: """handle kernel info reply sets protocol adaptation version. This might be run from a separate thread. """ adapt_version = int(msg["content"]["protocol_version"].split(".")[0]) if adapt_version != major_protocol_version: self.session.adapt_version = adapt_version def is_complete(self, code: str) -> str: """Ask the kernel whether some code is complete and ready to execute. Returns ------- The ID of the message sent. """ msg = self.session.msg("is_complete_request", {"code": code}) self.shell_channel.send(msg) return msg["header"]["msg_id"] def input(self, string: str) -> None: """Send a string of raw input to the kernel. This should only be called in response to the kernel sending an ``input_request`` message on the stdin channel. Returns ------- The ID of the message sent. """ content = {"value": string} msg = self.session.msg("input_reply", content) self.stdin_channel.send(msg) def shutdown(self, restart: bool = False) -> str: """Request an immediate kernel shutdown on the control channel. Upon receipt of the (empty) reply, client code can safely assume that the kernel has shut down and it's safe to forcefully terminate it if it's still alive. The kernel will send the reply via a function registered with Python's atexit module, ensuring it's truly done as the kernel is done with all normal operation. Returns ------- The msg_id of the message sent """ # Send quit message to kernel. Once we implement kernel-side setattr, # this should probably be done that way, but for now this will do. msg = self.session.msg("shutdown_request", {"restart": restart}) self.control_channel.send(msg) return msg["header"]["msg_id"] KernelClientABC.register(KernelClient)