ipc

Classes and utilities for communication between Executors and subprocesses.

Communicators manage message passing and parsing between subprocesses. They maintain a limited public interface of "read" and "write" operations. Behind this interface the methods of communication vary from serialization across pipes to Unix sockets, etc. All communicators pass a single object called a "Message" which contains an arbitrary "contents" field as well as an optional "signal" field.

Classes:

Name	Description
Party	Enum describing whether Communicator is on Task-side or Executor-side.
Message	A dataclass used for passing information from Task to Executor.
Communicator	Abstract base class for Communicator types.
PipeCommunicator	Manages communication between Task and Executor via pipes (stderr and stdout).
SocketCommunicator	Manages communication using sockets, either raw or using zmq. Supports both TCP and Unix sockets.

Communicator

Bases: ABC

Source code in lute/execution/ipc.py

class Communicator(ABC):
    def __init__(self, party: Party = Party.TASK, use_pickle: bool = True) -> None:
        """Abstract Base Class for IPC Communicator objects.

        Args:
            party (Party): Which object (side/process) the Communicator is
                managing IPC for. I.e., is this the "Task" or "Executor" side.
            use_pickle (bool): Whether to serialize data using pickle prior to
                sending it.
        """
        self._party = party
        self._use_pickle = use_pickle
        self.desc = "Communicator abstract base class."

    @abstractmethod
    def read(self, proc: subprocess.Popen) -> Message:
        """Method for reading data through the communication mechanism."""
        ...

    @abstractmethod
    def write(self, msg: Message) -> None:
        """Method for sending data through the communication mechanism."""
        ...

    def __str__(self) -> str:
        name: str = str(type(self)).split("'")[1].split(".")[-1]
        return f"{name}: {self.desc}"

    def __repr__(self) -> str:
        return self.__str__()

    def __enter__(self) -> Self:
        return self

    def __exit__(self) -> None: ...

    @property
    def has_messages(self) -> bool:
        """Whether the Communicator has remaining messages.

        The precise method for determining whether there are remaining messages
        will depend on the specific Communicator sub-class.
        """
        return False

    def stage_communicator(self):
        """Alternative method for staging outside of context manager."""
        self.__enter__()

    def clear_communicator(self):
        """Alternative exit method outside of context manager."""
        self.__exit__()

    def delayed_setup(self):
        """Any setup that should be done later than init."""
        ...

has_messages property

Whether the Communicator has remaining messages.

The precise method for determining whether there are remaining messages will depend on the specific Communicator sub-class.

__init__(party=Party.TASK, use_pickle=True)

Abstract Base Class for IPC Communicator objects.

Parameters:

Name	Type	Description	Default
party	Party	Which object (side/process) the Communicator is managing IPC for. I.e., is this the "Task" or "Executor" side.	TASK
use_pickle	bool	Whether to serialize data using pickle prior to sending it.	True

Source code in lute/execution/ipc.py

def __init__(self, party: Party = Party.TASK, use_pickle: bool = True) -> None:
    """Abstract Base Class for IPC Communicator objects.

    Args:
        party (Party): Which object (side/process) the Communicator is
            managing IPC for. I.e., is this the "Task" or "Executor" side.
        use_pickle (bool): Whether to serialize data using pickle prior to
            sending it.
    """
    self._party = party
    self._use_pickle = use_pickle
    self.desc = "Communicator abstract base class."

clear_communicator()

Alternative exit method outside of context manager.

Source code in lute/execution/ipc.py

def clear_communicator(self):
    """Alternative exit method outside of context manager."""
    self.__exit__()

delayed_setup()

Any setup that should be done later than init.

Source code in lute/execution/ipc.py

def delayed_setup(self):
    """Any setup that should be done later than init."""
    ...

read(proc) abstractmethod

Method for reading data through the communication mechanism.

Source code in lute/execution/ipc.py

@abstractmethod
def read(self, proc: subprocess.Popen) -> Message:
    """Method for reading data through the communication mechanism."""
    ...

stage_communicator()

Alternative method for staging outside of context manager.

Source code in lute/execution/ipc.py

def stage_communicator(self):
    """Alternative method for staging outside of context manager."""
    self.__enter__()

write(msg) abstractmethod

Method for sending data through the communication mechanism.

Source code in lute/execution/ipc.py

@abstractmethod
def write(self, msg: Message) -> None:
    """Method for sending data through the communication mechanism."""
    ...

Party

Bases: Enum

Identifier for which party (side/end) is using a communicator.

For some types of communication streams there may be different interfaces depending on which side of the communicator you are on. This enum is used by the communicator to determine which interface to use.

Source code in lute/execution/ipc.py

class Party(Enum):
    """Identifier for which party (side/end) is using a communicator.

    For some types of communication streams there may be different interfaces
    depending on which side of the communicator you are on. This enum is used
    by the communicator to determine which interface to use.
    """

    TASK = 0
    """
    The Task (client) side.
    """
    EXECUTOR = 1
    """
    The Executor (server) side.
    """

EXECUTOR = 1 class-attribute instance-attribute

The Executor (server) side.

TASK = 0 class-attribute instance-attribute

The Task (client) side.

PipeCommunicator

Bases: Communicator

Provides communication through pipes over stderr/stdout.

The implementation of this communicator has reading and writing ocurring on stderr and stdout. In general the Task will be writing while the Executor will be reading. stderr is used for sending signals.

Source code in lute/execution/ipc.py

class PipeCommunicator(Communicator):
    """Provides communication through pipes over stderr/stdout.

    The implementation of this communicator has reading and writing ocurring
    on stderr and stdout. In general the `Task` will be writing while the
    `Executor` will be reading. `stderr` is used for sending signals.
    """

    def __init__(self, party: Party = Party.TASK, use_pickle: bool = True) -> None:
        """IPC through pipes.

        Arbitrary objects may be transmitted using pickle to serialize the data.
        If pickle is not used

        Args:
            party (Party): Which object (side/process) the Communicator is
                managing IPC for. I.e., is this the "Task" or "Executor" side.
            use_pickle (bool): Whether to serialize data using Pickle prior to
                sending it. If False, data is assumed to be text whi
        """
        super().__init__(party=party, use_pickle=use_pickle)
        self.desc = "Communicates through stderr and stdout using pickle."

    def read(self, proc: subprocess.Popen) -> Message:
        """Read from stdout and stderr.

        Args:
            proc (subprocess.Popen): The process to read from.

        Returns:
            msg (Message): The message read, containing contents and signal.
        """
        signal: Optional[str]
        contents: Optional[str]
        raw_signal: Optional[bytes] = (
            proc.stderr.read() if proc.stderr is not None else None
        )
        raw_contents: Optional[bytes] = (
            proc.stdout.read() if proc.stdout is not None else None
        )
        if raw_signal is not None:
            signal = raw_signal.decode()
        else:
            signal = raw_signal
        if raw_contents:
            if self._use_pickle:
                try:
                    contents = pickle.loads(raw_contents)
                except (pickle.UnpicklingError, ValueError, EOFError):
                    logger.debug("PipeCommunicator (Executor) - Set _use_pickle=False")
                    self._use_pickle = False
                    contents = self._safe_unpickle_decode(raw_contents)
            else:
                try:
                    contents = raw_contents.decode()
                except UnicodeDecodeError:
                    logger.debug("PipeCommunicator (Executor) - Set _use_pickle=True")
                    self._use_pickle = True
                    contents = self._safe_unpickle_decode(raw_contents)
        else:
            contents = None

        if signal and signal not in LUTE_SIGNALS:
            # Some tasks write on stderr
            # If the signal channel has "non-signal" info, add it to
            # contents
            if not contents:
                contents = f"({signal})"
            else:
                contents = f"{contents} ({signal})"
            signal = None

        return Message(contents=contents, signal=signal)

    def _safe_unpickle_decode(self, maybe_mixed: bytes) -> Optional[str]:
        """This method is used to unpickle and/or decode a bytes object.

        It attempts to handle cases where contents can be mixed, i.e., part of
        the message must be decoded and the other part unpickled. It handles
        only two-way splits. If there are more complex arrangements such as:
        <pickled>:<unpickled>:<pickled> etc, it will give up.

        The simpler two way splits are unlikely to occur in normal usage. They
        may arise when debugging if, e.g., `print` statements are mixed with the
        usage of the `_report_to_executor` method.

        Note that this method works because ONLY text data is assumed to be
        sent via the pipes. The method needs to be revised to handle non-text
        data if the `Task` is modified to also send that via PipeCommunicator.
        The use of pickle is supported to provide for this option if it is
        necessary. It may be deprecated in the future.

        Be careful when making changes. This method has seemingly redundant
        checks because unpickling will not throw an error if a full object can
        be retrieved. That is, the library will ignore extraneous bytes. This
        method attempts to retrieve that information if the pickled data comes
        first in the stream.

        Args:
            maybe_mixed (bytes): A bytes object which could require unpickling,
                decoding, or both.

        Returns:
            contents (Optional[str]): The unpickled/decoded contents if possible.
                Otherwise, None.
        """
        contents: Optional[str]
        try:
            contents = pickle.loads(maybe_mixed)
            repickled: bytes = pickle.dumps(contents)
            if len(repickled) < len(maybe_mixed):
                # Successful unpickling, but pickle stops even if there are more bytes
                try:
                    additional_data: str = maybe_mixed[len(repickled) :].decode()
                    contents = f"{contents}{additional_data}"
                except UnicodeDecodeError:
                    # Can't decode the bytes left by pickle, so they are lost
                    missing_bytes: int = len(maybe_mixed) - len(repickled)
                    logger.debug(
                        f"PipeCommunicator has truncated message. Unable to retrieve {missing_bytes} bytes."
                    )
        except (pickle.UnpicklingError, ValueError, EOFError):
            # Pickle may also throw a ValueError, e.g. this bytes: b"Found! \n"
            # Pickle may also throw an EOFError, eg. this bytes: b"F0\n"
            try:
                contents = maybe_mixed.decode()
            except UnicodeDecodeError as err2:
                try:
                    contents = maybe_mixed[: err2.start].decode()
                    contents = f"{contents}{pickle.loads(maybe_mixed[err2.start:])}"
                except Exception as err3:
                    logger.debug(
                        f"PipeCommunicator unable to decode/parse data! {err3}"
                    )
                    contents = None
        return contents

    def write(self, msg: Message) -> None:
        """Write to stdout and stderr.

         The signal component is sent to `stderr` while the contents of the
         Message are sent to `stdout`.

        Args:
            msg (Message): The Message to send.
        """
        if self._use_pickle:
            signal: bytes
            if msg.signal:
                signal = msg.signal.encode()
            else:
                signal = b""

            contents: bytes = pickle.dumps(msg.contents)

            sys.stderr.buffer.write(signal)
            sys.stdout.buffer.write(contents)

            sys.stderr.buffer.flush()
            sys.stdout.buffer.flush()
        else:
            raw_signal: str
            if msg.signal:
                raw_signal = msg.signal
            else:
                raw_signal = ""

            raw_contents: str
            if isinstance(msg.contents, str):
                raw_contents = msg.contents
            elif msg.contents is None:
                raw_contents = ""
            else:
                raise ValueError(
                    f"Cannot send msg contents of type: {type(msg.contents)} when not using pickle!"
                )
            sys.stderr.write(raw_signal)
            sys.stdout.write(raw_contents)

__init__(party=Party.TASK, use_pickle=True)

IPC through pipes.

Arbitrary objects may be transmitted using pickle to serialize the data. If pickle is not used

Parameters:

Name	Type	Description	Default
party	Party	Which object (side/process) the Communicator is managing IPC for. I.e., is this the "Task" or "Executor" side.	TASK
use_pickle	bool	Whether to serialize data using Pickle prior to sending it. If False, data is assumed to be text whi	True

Source code in lute/execution/ipc.py

def __init__(self, party: Party = Party.TASK, use_pickle: bool = True) -> None:
    """IPC through pipes.

    Arbitrary objects may be transmitted using pickle to serialize the data.
    If pickle is not used

    Args:
        party (Party): Which object (side/process) the Communicator is
            managing IPC for. I.e., is this the "Task" or "Executor" side.
        use_pickle (bool): Whether to serialize data using Pickle prior to
            sending it. If False, data is assumed to be text whi
    """
    super().__init__(party=party, use_pickle=use_pickle)
    self.desc = "Communicates through stderr and stdout using pickle."

_safe_unpickle_decode(maybe_mixed)

This method is used to unpickle and/or decode a bytes object.

It attempts to handle cases where contents can be mixed, i.e., part of the message must be decoded and the other part unpickled. It handles only two-way splits. If there are more complex arrangements such as: :: etc, it will give up.

The simpler two way splits are unlikely to occur in normal usage. They may arise when debugging if, e.g., print statements are mixed with the usage of the _report_to_executor method.

Note that this method works because ONLY text data is assumed to be sent via the pipes. The method needs to be revised to handle non-text data if the Task is modified to also send that via PipeCommunicator. The use of pickle is supported to provide for this option if it is necessary. It may be deprecated in the future.

Be careful when making changes. This method has seemingly redundant checks because unpickling will not throw an error if a full object can be retrieved. That is, the library will ignore extraneous bytes. This method attempts to retrieve that information if the pickled data comes first in the stream.

Parameters:

Name	Type	Description	Default
maybe_mixed	bytes	A bytes object which could require unpickling, decoding, or both.	required

Returns:

Name	Type	Description
contents	Optional[str]	The unpickled/decoded contents if possible. Otherwise, None.

Source code in lute/execution/ipc.py

def _safe_unpickle_decode(self, maybe_mixed: bytes) -> Optional[str]:
    """This method is used to unpickle and/or decode a bytes object.

    It attempts to handle cases where contents can be mixed, i.e., part of
    the message must be decoded and the other part unpickled. It handles
    only two-way splits. If there are more complex arrangements such as:
    <pickled>:<unpickled>:<pickled> etc, it will give up.

    The simpler two way splits are unlikely to occur in normal usage. They
    may arise when debugging if, e.g., `print` statements are mixed with the
    usage of the `_report_to_executor` method.

    Note that this method works because ONLY text data is assumed to be
    sent via the pipes. The method needs to be revised to handle non-text
    data if the `Task` is modified to also send that via PipeCommunicator.
    The use of pickle is supported to provide for this option if it is
    necessary. It may be deprecated in the future.

    Be careful when making changes. This method has seemingly redundant
    checks because unpickling will not throw an error if a full object can
    be retrieved. That is, the library will ignore extraneous bytes. This
    method attempts to retrieve that information if the pickled data comes
    first in the stream.

    Args:
        maybe_mixed (bytes): A bytes object which could require unpickling,
            decoding, or both.

    Returns:
        contents (Optional[str]): The unpickled/decoded contents if possible.
            Otherwise, None.
    """
    contents: Optional[str]
    try:
        contents = pickle.loads(maybe_mixed)
        repickled: bytes = pickle.dumps(contents)
        if len(repickled) < len(maybe_mixed):
            # Successful unpickling, but pickle stops even if there are more bytes
            try:
                additional_data: str = maybe_mixed[len(repickled) :].decode()
                contents = f"{contents}{additional_data}"
            except UnicodeDecodeError:
                # Can't decode the bytes left by pickle, so they are lost
                missing_bytes: int = len(maybe_mixed) - len(repickled)
                logger.debug(
                    f"PipeCommunicator has truncated message. Unable to retrieve {missing_bytes} bytes."
                )
    except (pickle.UnpicklingError, ValueError, EOFError):
        # Pickle may also throw a ValueError, e.g. this bytes: b"Found! \n"
        # Pickle may also throw an EOFError, eg. this bytes: b"F0\n"
        try:
            contents = maybe_mixed.decode()
        except UnicodeDecodeError as err2:
            try:
                contents = maybe_mixed[: err2.start].decode()
                contents = f"{contents}{pickle.loads(maybe_mixed[err2.start:])}"
            except Exception as err3:
                logger.debug(
                    f"PipeCommunicator unable to decode/parse data! {err3}"
                )
                contents = None
    return contents

read(proc)

Read from stdout and stderr.

Parameters:

Name	Type	Description	Default
proc	Popen	The process to read from.	required

Returns:

Name	Type	Description
msg	Message	The message read, containing contents and signal.

Source code in lute/execution/ipc.py

def read(self, proc: subprocess.Popen) -> Message:
    """Read from stdout and stderr.

    Args:
        proc (subprocess.Popen): The process to read from.

    Returns:
        msg (Message): The message read, containing contents and signal.
    """
    signal: Optional[str]
    contents: Optional[str]
    raw_signal: Optional[bytes] = (
        proc.stderr.read() if proc.stderr is not None else None
    )
    raw_contents: Optional[bytes] = (
        proc.stdout.read() if proc.stdout is not None else None
    )
    if raw_signal is not None:
        signal = raw_signal.decode()
    else:
        signal = raw_signal
    if raw_contents:
        if self._use_pickle:
            try:
                contents = pickle.loads(raw_contents)
            except (pickle.UnpicklingError, ValueError, EOFError):
                logger.debug("PipeCommunicator (Executor) - Set _use_pickle=False")
                self._use_pickle = False
                contents = self._safe_unpickle_decode(raw_contents)
        else:
            try:
                contents = raw_contents.decode()
            except UnicodeDecodeError:
                logger.debug("PipeCommunicator (Executor) - Set _use_pickle=True")
                self._use_pickle = True
                contents = self._safe_unpickle_decode(raw_contents)
    else:
        contents = None

    if signal and signal not in LUTE_SIGNALS:
        # Some tasks write on stderr
        # If the signal channel has "non-signal" info, add it to
        # contents
        if not contents:
            contents = f"({signal})"
        else:
            contents = f"{contents} ({signal})"
        signal = None

    return Message(contents=contents, signal=signal)

write(msg)

Write to stdout and stderr.

The signal component is sent to stderr while the contents of the Message are sent to stdout.

Parameters:

Name	Type	Description	Default
msg	Message	The Message to send.	required

Source code in lute/execution/ipc.py

def write(self, msg: Message) -> None:
    """Write to stdout and stderr.

     The signal component is sent to `stderr` while the contents of the
     Message are sent to `stdout`.

    Args:
        msg (Message): The Message to send.
    """
    if self._use_pickle:
        signal: bytes
        if msg.signal:
            signal = msg.signal.encode()
        else:
            signal = b""

        contents: bytes = pickle.dumps(msg.contents)

        sys.stderr.buffer.write(signal)
        sys.stdout.buffer.write(contents)

        sys.stderr.buffer.flush()
        sys.stdout.buffer.flush()
    else:
        raw_signal: str
        if msg.signal:
            raw_signal = msg.signal
        else:
            raw_signal = ""

        raw_contents: str
        if isinstance(msg.contents, str):
            raw_contents = msg.contents
        elif msg.contents is None:
            raw_contents = ""
        else:
            raise ValueError(
                f"Cannot send msg contents of type: {type(msg.contents)} when not using pickle!"
            )
        sys.stderr.write(raw_signal)
        sys.stdout.write(raw_contents)

SocketCommunicator

Bases: Communicator

Provides communication over Unix or TCP sockets.

Communication is provided either using sockets with the Python socket library or using ZMQ. The choice of implementation is controlled by the global bool USE_ZMQ.

Whether to use TCP or Unix sockets is controlled by the environment

LUTE_USE_TCP=1

If defined, TCP sockets will be used, otherwise Unix sockets will be used.

Regardless of socket type, the environment variable LUTE_EXECUTOR_HOST=<hostname> will be defined by the Executor-side Communicator.

For TCP sockets: The Executor-side Communicator should be run first and will bind to all interfaces on the port determined by the environment variable: LUTE_PORT=### If no port is defined, a port scan will be performed and the Executor-side Communicator will bind the first one available from a random selection. It will then define the environment variable so the Task-side can pick it up.

For Unix sockets: The path to the Unix socket is defined by the environment variable: LUTE_SOCKET=/path/to/socket This class assumes proper permissions and that this above environment variable has been defined. The Task is configured as what would commonly be referred to as the client, while the Executor is configured as the server.

If the Task process is run on a different machine than the Executor, the Task-side Communicator will open a ssh-tunnel to forward traffic from a local Unix socket to the Executor Unix socket. Opening of the tunnel relies on the environment variable: LUTE_EXECUTOR_HOST=<hostname> to determine the Executor's host. This variable should be defined by the Executor and passed to the Task process automatically, but it can also be defined manually if launching the Task process separately. The Task will use the local socket <LUTE_SOCKET>.task{##}. Multiple local sockets may be created. Currently, it is assumed that the user is identical on both the Task machine and Executor machine.

Source code in lute/execution/ipc.py

class SocketCommunicator(Communicator):
    """Provides communication over Unix or TCP sockets.

    Communication is provided either using sockets with the Python socket library
    or using ZMQ. The choice of implementation is controlled by the global bool
    `USE_ZMQ`.

    Whether to use TCP or Unix sockets is controlled by the environment:
                           `LUTE_USE_TCP=1`
    If defined, TCP sockets will be used, otherwise Unix sockets will be used.

    Regardless of socket type, the environment variable
                      `LUTE_EXECUTOR_HOST=<hostname>`
    will be defined by the Executor-side Communicator.


    For TCP sockets:
    The Executor-side Communicator should be run first and will bind to all
    interfaces on the port determined by the environment variable:
                            `LUTE_PORT=###`
    If no port is defined, a port scan will be performed and the Executor-side
    Communicator will bind the first one available from a random selection. It
    will then define the environment variable so the Task-side can pick it up.

    For Unix sockets:
    The path to the Unix socket is defined by the environment variable:
                      `LUTE_SOCKET=/path/to/socket`
    This class assumes proper permissions and that this above environment
    variable has been defined. The `Task` is configured as what would commonly
    be referred to as the `client`, while the `Executor` is configured as the
    server.

    If the Task process is run on a different machine than the Executor, the
    Task-side Communicator will open a ssh-tunnel to forward traffic from a local
    Unix socket to the Executor Unix socket. Opening of the tunnel relies on the
    environment variable:
                      `LUTE_EXECUTOR_HOST=<hostname>`
    to determine the Executor's host. This variable should be defined by the
    Executor and passed to the Task process automatically, but it can also be
    defined manually if launching the Task process separately. The Task will use
    the local socket `<LUTE_SOCKET>.task{##}`. Multiple local sockets may be
    created. Currently, it is assumed that the user is identical on both the Task
    machine and Executor machine.
    """

    ACCEPT_TIMEOUT: float = 0.01
    """
    Maximum time to wait to accept connections. Used by Executor-side.
    """
    MSG_HEAD: bytes = b"MSG"
    """
    Start signal of a message. The end of a message is indicated by MSG_HEAD[::-1].
    """
    MSG_SEP: bytes = b";;;"
    """
    Separator for parts of a message. Messages have a start, length, message and end.
    """

    def __init__(self, party: Party = Party.TASK, use_pickle: bool = True) -> None:
        """IPC over a TCP or Unix socket.

        Unlike with the PipeCommunicator, pickle is always used to send data
        through the socket.

        Args:
            party (Party): Which object (side/process) the Communicator is
                managing IPC for. I.e., is this the "Task" or "Executor" side.

            use_pickle (bool): Whether to use pickle. Always True currently,
                passing False does not change behaviour.
        """
        super().__init__(party=party, use_pickle=use_pickle)

    def delayed_setup(self) -> None:
        """Delays the creation of socket objects.

        The Executor initializes the Communicator when it is created. Since
        all Executors are created and available at once we want to delay
        acquisition of socket resources until a single Executor is ready
        to use them.
        """
        self._data_socket: Union[socket.socket, zmq.sugar.socket.Socket]
        self.desc: str
        use_tcp: Optional[str] = os.getenv("LUTE_USE_TCP")
        sock_type: str
        if use_tcp is not None:
            sock_type = "TCP"
        else:
            sock_type = "Unix"
        if USE_ZMQ:
            self.desc = f"Communicates using ZMQ through {sock_type} sockets."
            self._context: zmq.Context = zmq.Context()
            self._data_socket = self._create_socket_zmq()
        else:
            self.desc = f"Communicates through {sock_type} sockets."
            self._data_socket = self._create_socket_raw()
            self._data_socket.settimeout(SocketCommunicator.ACCEPT_TIMEOUT)

        if self._party == Party.EXECUTOR:
            # Executor created first so we can define the hostname env variable
            os.environ["LUTE_EXECUTOR_HOST"] = socket.gethostname()
            # Setup reader thread
            self._reader_thread: threading.Thread = threading.Thread(
                target=self._read_socket
            )
            self._msg_queue: queue.Queue = queue.Queue()
            self._partial_msg: Optional[bytes] = None
            self._stop_thread: bool = False
            self._reader_thread.start()
        else:
            # Only used by Party.TASK
            self._use_ssh_tunnel: bool = False
            self._ssh_proc: Optional[subprocess.Popen] = None
            self._local_socket_path: Optional[str] = None

    # Read
    ############################################################################

    def read(self, proc: subprocess.Popen) -> Message:
        """Return a message from the queue if available.

        Socket(s) are continuously monitored, and read from when new data is
        available.

        Args:
            proc (subprocess.Popen): The process to read from. Provided for
                compatibility with other Communicator subtypes. Is ignored.

        Returns:
             msg (Message): The message read, containing contents and signal.
        """
        msg: Message
        try:
            msg = self._msg_queue.get(timeout=SocketCommunicator.ACCEPT_TIMEOUT)
        except queue.Empty:
            msg = Message()

        return msg

    def _read_socket(self) -> None:
        """Read data from a socket.

        Socket(s) are continuously monitored, and read from when new data is
        available.

        Calls an underlying method for either raw sockets or ZMQ.
        """

        while True:
            if self._stop_thread:
                logger.debug("Stopping socket reader thread.")
                break
            if USE_ZMQ:
                self._read_socket_zmq()
            else:
                self._read_socket_raw()

    def _read_socket_raw(self) -> None:
        """Read data from a socket.

        Raw socket implementation for the reader thread.
        """
        assert isinstance(self._data_socket, socket.socket)
        connection: socket.socket
        _: Union[str, Tuple[str, int]]
        try:
            connection, _ = self._data_socket.accept()
            full_data: bytes = b""
            while True:
                data: bytes = connection.recv(8192)
                if data:
                    full_data += data
                else:
                    break
            connection.close()
            self._unpack_messages(full_data)
        except socket.timeout:
            pass

    def _read_socket_zmq(self) -> None:
        """Read data from a socket.

        ZMQ implementation for the reader thread.
        """
        try:
            full_data: bytes = self._data_socket.recv(0)
            self._unpack_messages(full_data)
        except zmq.ZMQError:
            pass

    def _unpack_messages(self, data: bytes) -> None:
        """Unpacks a byte stream into individual messages.

        Messages are encoded in the following format:
                 <HEAD><SEP><len(msg)><SEP><msg><SEP><HEAD[::-1]>
        The items between <> are replaced as follows:
            - <HEAD>: A start marker
            - <SEP>: A separator for components of the message
            - <len(msg)>: The length of the message payload in bytes.
            - <msg>: The message payload in bytes
            - <HEAD[::-1]>: The start marker in reverse to indicate the end.

        Partial messages (a series of bytes which cannot be converted to a full
        message) are stored for later. An attempt is made to reconstruct the
        message with the next call to this method.

        Args:
            data (bytes): A raw byte stream containing anywhere from a partial
                message to multiple full messages.
        """
        msg: Message
        working_data: bytes
        if self._partial_msg:
            # Concatenate the previous partial message to the beginning
            working_data = self._partial_msg + data
            self._partial_msg = None
        else:
            working_data = data
        while working_data:
            try:
                # Message encoding: <HEAD><SEP><len><SEP><msg><SEP><HEAD[::-1]>
                end = working_data.find(
                    SocketCommunicator.MSG_SEP + SocketCommunicator.MSG_HEAD[::-1]
                )
                msg_parts: List[bytes] = working_data[:end].split(
                    SocketCommunicator.MSG_SEP
                )
                if len(msg_parts) != 3:
                    self._partial_msg = working_data
                    break

                cmd: bytes
                nbytes: bytes
                raw_msg: bytes
                cmd, nbytes, raw_msg = msg_parts
                if len(raw_msg) != int(nbytes):
                    self._partial_msg = working_data
                    break
                msg = pickle.loads(raw_msg)
                self._msg_queue.put(msg)
            except pickle.UnpicklingError:
                self._partial_msg = working_data
                break
            if end < len(working_data):
                # Add len(SEP+HEAD) since end marks the start of <SEP><HEAD[::-1]
                offset: int = len(
                    SocketCommunicator.MSG_SEP + SocketCommunicator.MSG_HEAD
                )
                working_data = working_data[end + offset :]
            else:
                working_data = b""

    # Write
    ############################################################################

    def _write_socket(self, msg: Message) -> None:
        """Sends data over a socket from the 'client' (Task) side.

        Messages are encoded in the following format:
                 <HEAD><SEP><len(msg)><SEP><msg><SEP><HEAD[::-1]>
        The items between <> are replaced as follows:
            - <HEAD>: A start marker
            - <SEP>: A separator for components of the message
            - <len(msg)>: The length of the message payload in bytes.
            - <msg>: The message payload in bytes
            - <HEAD[::-1]>: The start marker in reverse to indicate the end.

        This structure is used for decoding the message on the other end.
        """
        data: bytes = pickle.dumps(msg)
        cmd: bytes = SocketCommunicator.MSG_HEAD
        size: bytes = b"%d" % len(data)
        end: bytes = SocketCommunicator.MSG_HEAD[::-1]
        sep: bytes = SocketCommunicator.MSG_SEP
        packed_msg: bytes = cmd + sep + size + sep + data + sep + end
        if USE_ZMQ:
            self._data_socket.send(packed_msg)
        else:
            assert isinstance(self._data_socket, socket.socket)
            self._data_socket.sendall(packed_msg)

    def write(self, msg: Message) -> None:
        """Send a single Message.

        The entire Message (signal and contents) is serialized and sent through
        a connection over Unix socket.

        Args:
            msg (Message): The Message to send.
        """
        self._write_socket(msg)

    # Generic create
    ############################################################################

    def _create_socket_raw(self) -> socket.socket:
        """Create either a Unix or TCP socket.

        If the environment variable:
                              `LUTE_USE_TCP=1`
        is defined, a TCP socket is returned, otherwise a Unix socket.

        Refer to the individual initialization methods for additional environment
        variables controlling the behaviour of these two communication types.

        Returns:
            data_socket (socket.socket): TCP or Unix socket.
        """
        import struct

        use_tcp: Optional[str] = os.getenv("LUTE_USE_TCP")
        sock: socket.socket
        if use_tcp is not None:
            if self._party == Party.EXECUTOR:
                logger.info("Will use raw TCP sockets.")
            sock = self._init_tcp_socket_raw()
        else:
            if self._party == Party.EXECUTOR:
                logger.info("Will use raw Unix sockets.")
            sock = self._init_unix_socket_raw()
        sock.setsockopt(
            socket.SOL_SOCKET, socket.SO_LINGER, struct.pack("ii", 1, 10000)
        )
        return sock

    def _create_socket_zmq(self) -> zmq.sugar.socket.Socket:
        """Create either a Unix or TCP socket.

        If the environment variable:
                              `LUTE_USE_TCP=1`
        is defined, a TCP socket is returned, otherwise a Unix socket.

        Refer to the individual initialization methods for additional environment
        variables controlling the behaviour of these two communication types.

        Returns:
            data_socket (socket.socket): Unix socket object.
        """
        socket_type: int  # zmq.SocketType - either zmq.PULL or zmq.PUSH
        if self._party == Party.EXECUTOR:
            socket_type = zmq.PULL
        else:
            socket_type = zmq.PUSH

        data_socket: zmq.sugar.socket.Socket = self._context.socket(socket_type)
        data_socket.set_hwm(160000)
        # Need to multiply by 1000 since ZMQ uses ms
        data_socket.setsockopt(
            zmq.RCVTIMEO, int(SocketCommunicator.ACCEPT_TIMEOUT * 1000)
        )
        # Try TCP first
        use_tcp: Optional[str] = os.getenv("LUTE_USE_TCP")
        if use_tcp is not None:
            if self._party == Party.EXECUTOR:
                logger.info("Will use TCP (ZMQ).")
            self._init_tcp_socket_zmq(data_socket)
        else:
            if self._party == Party.EXECUTOR:
                logger.info("Will use Unix sockets (ZMQ).")
            self._init_unix_socket_zmq(data_socket)

        return data_socket

    # TCP Init
    ############################################################################

    def _find_random_port(
        self, min_port: int = 41923, max_port: int = 64324, max_tries: int = 100
    ) -> Optional[int]:
        """Find a random open port to bind to if using TCP."""
        from random import choices

        sock: socket.socket
        ports: List[int] = choices(range(min_port, max_port), k=max_tries)
        for port in ports:
            sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
            try:
                sock.bind(("", port))
                sock.close()
                del sock
                return port
            except Exception:
                continue
        return None

    def _init_tcp_socket_raw(self) -> socket.socket:
        """Initialize a TCP socket.

        Executor-side code should always be run first. It checks to see if
        the environment variable
                                `LUTE_PORT=###`
        is defined, if so binds it, otherwise find a free port from a selection
        of random ports. If a port search is performed, the `LUTE_PORT` variable
        will be defined so it can be picked up by the the Task-side Communicator.

        In the event that no port can be bound on the Executor-side, or the port
        and hostname information is unavailable to the Task-side, the program
        will exit.

        Returns:
            data_socket (socket.socket): TCP socket object.
        """
        data_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        port: Optional[Union[str, int]] = os.getenv("LUTE_PORT")
        if self._party == Party.EXECUTOR:
            if port is None:
                # If port is None find one
                # Executor code executes first
                port = self._find_random_port()
                if port is None:
                    # Failed to find a port to bind
                    logger.info(
                        "Executor failed to bind a port. "
                        "Try providing a LUTE_PORT directly! Exiting!"
                    )
                    sys.exit(-1)
                # Provide port env var for Task-side
                os.environ["LUTE_PORT"] = str(port)
            data_socket.bind(("", int(port)))
            data_socket.listen()
        else:
            hostname: str = socket.gethostname()
            executor_hostname: Optional[str] = os.getenv("LUTE_EXECUTOR_HOST")
            if executor_hostname is None or port is None:
                logger.info(
                    "Task-side does not have host/port information!"
                    " Check environment variables! Exiting!"
                )
                sys.exit(-1)
            if hostname == executor_hostname:
                data_socket.connect(("localhost", int(port)))
            else:
                data_socket.connect((executor_hostname, int(port)))
        return data_socket

    def _init_tcp_socket_zmq(self, data_socket: zmq.sugar.socket.Socket) -> None:
        """Initialize a TCP socket using ZMQ.

        Equivalent as the method above but requires passing in a ZMQ socket
        object instead of returning one.

        Args:
            data_socket (zmq.socket.Socket): Socket object.
        """
        port: Optional[Union[str, int]] = os.getenv("LUTE_PORT")
        if self._party == Party.EXECUTOR:
            if port is None:
                new_port: int = data_socket.bind_to_random_port("tcp://*")
                if new_port is None:
                    # Failed to find a port to bind
                    logger.info(
                        "Executor failed to bind a port. "
                        "Try providing a LUTE_PORT directly! Exiting!"
                    )
                    sys.exit(-1)
                port = new_port
                os.environ["LUTE_PORT"] = str(port)
            else:
                data_socket.bind(f"tcp://*:{port}")
            logger.debug(f"Executor bound port {port}")
        else:
            executor_hostname: Optional[str] = os.getenv("LUTE_EXECUTOR_HOST")
            if executor_hostname is None or port is None:
                logger.info(
                    "Task-side does not have host/port information!"
                    " Check environment variables! Exiting!"
                )
                sys.exit(-1)
            data_socket.connect(f"tcp://{executor_hostname}:{port}")

    # Unix Init
    ############################################################################

    def _get_socket_path(self) -> str:
        """Return the socket path, defining one if it is not available.

        Returns:
            socket_path (str): Path to the Unix socket.
        """
        socket_path: str
        try:
            socket_path = os.environ["LUTE_SOCKET"]
        except KeyError:
            import uuid
            import tempfile

            # Define a path, and add to environment
            # Executor-side always created first, Task will use the same one
            socket_path = f"{tempfile.gettempdir()}/lute_{uuid.uuid4().hex}.sock"
            os.environ["LUTE_SOCKET"] = socket_path
            logger.debug(f"SocketCommunicator defines socket_path: {socket_path}")
        if USE_ZMQ:
            return f"ipc://{socket_path}"
        else:
            return socket_path

    def _init_unix_socket_raw(self) -> socket.socket:
        """Returns a Unix socket object.

        Executor-side code should always be run first. It checks to see if
        the environment variable
                                `LUTE_SOCKET=XYZ`
        is defined, if so binds it, otherwise it will create a new path and
        define the environment variable for the Task-side to find.

        On the Task (client-side), this method will also open a SSH tunnel to
        forward a local Unix socket to an Executor Unix socket if the Task and
        Executor processes are on different machines.

        Returns:
            data_socket (socket.socket): Unix socket object.
        """
        socket_path: str = self._get_socket_path()
        data_socket = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM)
        if self._party == Party.EXECUTOR:
            if os.path.exists(socket_path):
                os.unlink(socket_path)
            data_socket.bind(socket_path)
            data_socket.listen()
        elif self._party == Party.TASK:
            hostname: str = socket.gethostname()
            executor_hostname: Optional[str] = os.getenv("LUTE_EXECUTOR_HOST")
            if executor_hostname is None:
                logger.info("Hostname for Executor process not found! Exiting!")
                data_socket.close()
                sys.exit(-1)
            if hostname == executor_hostname:
                data_socket.connect(socket_path)
            else:
                self._local_socket_path = self._setup_unix_ssh_tunnel(
                    socket_path, hostname, executor_hostname
                )
                while 1:
                    # Keep trying reconnect until ssh tunnel works.
                    try:
                        data_socket.connect(self._local_socket_path)
                        break
                    except FileNotFoundError:
                        continue

        return data_socket

    def _init_unix_socket_zmq(self, data_socket: zmq.sugar.socket.Socket) -> None:
        """Initialize a Unix socket object, using ZMQ.

        Equivalent as the method above but requires passing in a ZMQ socket
        object instead of returning one.

        Args:
            data_socket (socket.socket): ZMQ object.
        """
        socket_path = self._get_socket_path()
        if self._party == Party.EXECUTOR:
            if os.path.exists(socket_path):
                os.unlink(socket_path)
            data_socket.bind(socket_path)
        elif self._party == Party.TASK:
            hostname: str = socket.gethostname()
            executor_hostname: Optional[str] = os.getenv("LUTE_EXECUTOR_HOST")
            if executor_hostname is None:
                logger.info("Hostname for Executor process not found! Exiting!")
                self._data_socket.close()
                sys.exit(-1)
            if hostname == executor_hostname:
                data_socket.connect(socket_path)
            else:
                # Need to remove ipc:// from socket_path for forwarding
                self._local_socket_path = self._setup_unix_ssh_tunnel(
                    socket_path[6:], hostname, executor_hostname
                )
                # Need to add it back
                path: str = f"ipc://{self._local_socket_path}"
                data_socket.connect(path)

    def _setup_unix_ssh_tunnel(
        self, socket_path: str, hostname: str, executor_hostname: str
    ) -> str:
        """Prepares an SSH tunnel for forwarding between Unix sockets on two hosts.

        An SSH tunnel is opened with `ssh -L <local>:<remote> sleep 2`.
        This method of communication is slightly slower and incurs additional
        overhead - it should only be used as a backup. If communication across
        multiple hosts is required consider using TCP.  The Task will use
        the local socket `<LUTE_SOCKET>.task{##}`. Multiple local sockets may be
        created. It is assumed that the user is identical on both the
        Task machine and Executor machine.

        Returns:
            local_socket_path (str): The local Unix socket to connect to.
        """
        if "uuid" not in globals():
            import uuid
        local_socket_path = f"{socket_path}.task{uuid.uuid4().hex[:4]}"
        self._use_ssh_tunnel = True
        ssh_cmd: List[str] = [
            "ssh",
            "-o",
            "LogLevel=quiet",
            "-L",
            f"{local_socket_path}:{socket_path}",
            executor_hostname,
            "sleep",
            "2",
        ]
        logger.debug(f"Opening tunnel from {hostname} to {executor_hostname}")
        self._ssh_proc = subprocess.Popen(ssh_cmd)
        time.sleep(0.4)  # Need to wait... -> Use single Task comm at beginning?
        return local_socket_path

    # Clean up and properties
    ############################################################################

    def _clean_up(self) -> None:
        """Clean up connections."""
        if self._party == Party.EXECUTOR:
            self._stop_thread = True
            self._reader_thread.join()
            logger.debug("Closed reading thread.")

        self._data_socket.close()
        if USE_ZMQ:
            self._context.term()
        else:
            ...

        if os.getenv("LUTE_USE_TCP"):
            return
        else:
            if self._party == Party.EXECUTOR:
                os.unlink(cast(str, os.getenv("LUTE_SOCKET")))  # Should be defined
                return
            elif self._use_ssh_tunnel:
                if self._ssh_proc is not None:
                    self._ssh_proc.terminate()

    @property
    def has_messages(self) -> bool:
        if self._party == Party.TASK:
            # Shouldn't be called on Task-side
            return False

        if self._msg_queue.qsize() > 0:
            return True
        return False

    def __exit__(self):
        self._clean_up()

ACCEPT_TIMEOUT = 0.01 class-attribute instance-attribute

Maximum time to wait to accept connections. Used by Executor-side.

MSG_HEAD = b'MSG' class-attribute instance-attribute

Start signal of a message. The end of a message is indicated by MSG_HEAD[::-1].

MSG_SEP = b';;;' class-attribute instance-attribute

Separator for parts of a message. Messages have a start, length, message and end.

__init__(party=Party.TASK, use_pickle=True)

IPC over a TCP or Unix socket.

Unlike with the PipeCommunicator, pickle is always used to send data through the socket.

Parameters:

Name	Type	Description	Default
party	Party	Which object (side/process) the Communicator is managing IPC for. I.e., is this the "Task" or "Executor" side.	TASK
use_pickle	bool	Whether to use pickle. Always True currently, passing False does not change behaviour.	True

Source code in lute/execution/ipc.py

def __init__(self, party: Party = Party.TASK, use_pickle: bool = True) -> None:
    """IPC over a TCP or Unix socket.

    Unlike with the PipeCommunicator, pickle is always used to send data
    through the socket.

    Args:
        party (Party): Which object (side/process) the Communicator is
            managing IPC for. I.e., is this the "Task" or "Executor" side.

        use_pickle (bool): Whether to use pickle. Always True currently,
            passing False does not change behaviour.
    """
    super().__init__(party=party, use_pickle=use_pickle)

_clean_up()

Clean up connections.

Source code in lute/execution/ipc.py

def _clean_up(self) -> None:
    """Clean up connections."""
    if self._party == Party.EXECUTOR:
        self._stop_thread = True
        self._reader_thread.join()
        logger.debug("Closed reading thread.")

    self._data_socket.close()
    if USE_ZMQ:
        self._context.term()
    else:
        ...

    if os.getenv("LUTE_USE_TCP"):
        return
    else:
        if self._party == Party.EXECUTOR:
            os.unlink(cast(str, os.getenv("LUTE_SOCKET")))  # Should be defined
            return
        elif self._use_ssh_tunnel:
            if self._ssh_proc is not None:
                self._ssh_proc.terminate()

_create_socket_raw()

Create either a Unix or TCP socket.

If the environment variable

LUTE_USE_TCP=1

is defined, a TCP socket is returned, otherwise a Unix socket.

Refer to the individual initialization methods for additional environment variables controlling the behaviour of these two communication types.

Returns:

Name	Type	Description
data_socket	socket	TCP or Unix socket.

Source code in lute/execution/ipc.py

def _create_socket_raw(self) -> socket.socket:
    """Create either a Unix or TCP socket.

    If the environment variable:
                          `LUTE_USE_TCP=1`
    is defined, a TCP socket is returned, otherwise a Unix socket.

    Refer to the individual initialization methods for additional environment
    variables controlling the behaviour of these two communication types.

    Returns:
        data_socket (socket.socket): TCP or Unix socket.
    """
    import struct

    use_tcp: Optional[str] = os.getenv("LUTE_USE_TCP")
    sock: socket.socket
    if use_tcp is not None:
        if self._party == Party.EXECUTOR:
            logger.info("Will use raw TCP sockets.")
        sock = self._init_tcp_socket_raw()
    else:
        if self._party == Party.EXECUTOR:
            logger.info("Will use raw Unix sockets.")
        sock = self._init_unix_socket_raw()
    sock.setsockopt(
        socket.SOL_SOCKET, socket.SO_LINGER, struct.pack("ii", 1, 10000)
    )
    return sock

_create_socket_zmq()

Create either a Unix or TCP socket.

If the environment variable

LUTE_USE_TCP=1

is defined, a TCP socket is returned, otherwise a Unix socket.

Refer to the individual initialization methods for additional environment variables controlling the behaviour of these two communication types.

Returns:

Name	Type	Description
data_socket	socket	Unix socket object.

Source code in lute/execution/ipc.py

def _create_socket_zmq(self) -> zmq.sugar.socket.Socket:
    """Create either a Unix or TCP socket.

    If the environment variable:
                          `LUTE_USE_TCP=1`
    is defined, a TCP socket is returned, otherwise a Unix socket.

    Refer to the individual initialization methods for additional environment
    variables controlling the behaviour of these two communication types.

    Returns:
        data_socket (socket.socket): Unix socket object.
    """
    socket_type: int  # zmq.SocketType - either zmq.PULL or zmq.PUSH
    if self._party == Party.EXECUTOR:
        socket_type = zmq.PULL
    else:
        socket_type = zmq.PUSH

    data_socket: zmq.sugar.socket.Socket = self._context.socket(socket_type)
    data_socket.set_hwm(160000)
    # Need to multiply by 1000 since ZMQ uses ms
    data_socket.setsockopt(
        zmq.RCVTIMEO, int(SocketCommunicator.ACCEPT_TIMEOUT * 1000)
    )
    # Try TCP first
    use_tcp: Optional[str] = os.getenv("LUTE_USE_TCP")
    if use_tcp is not None:
        if self._party == Party.EXECUTOR:
            logger.info("Will use TCP (ZMQ).")
        self._init_tcp_socket_zmq(data_socket)
    else:
        if self._party == Party.EXECUTOR:
            logger.info("Will use Unix sockets (ZMQ).")
        self._init_unix_socket_zmq(data_socket)

    return data_socket

_find_random_port(min_port=41923, max_port=64324, max_tries=100)

Find a random open port to bind to if using TCP.

Source code in lute/execution/ipc.py

def _find_random_port(
    self, min_port: int = 41923, max_port: int = 64324, max_tries: int = 100
) -> Optional[int]:
    """Find a random open port to bind to if using TCP."""
    from random import choices

    sock: socket.socket
    ports: List[int] = choices(range(min_port, max_port), k=max_tries)
    for port in ports:
        sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        try:
            sock.bind(("", port))
            sock.close()
            del sock
            return port
        except Exception:
            continue
    return None

_get_socket_path()

Return the socket path, defining one if it is not available.

Returns:

Name	Type	Description
socket_path	str	Path to the Unix socket.

Source code in lute/execution/ipc.py

def _get_socket_path(self) -> str:
    """Return the socket path, defining one if it is not available.

    Returns:
        socket_path (str): Path to the Unix socket.
    """
    socket_path: str
    try:
        socket_path = os.environ["LUTE_SOCKET"]
    except KeyError:
        import uuid
        import tempfile

        # Define a path, and add to environment
        # Executor-side always created first, Task will use the same one
        socket_path = f"{tempfile.gettempdir()}/lute_{uuid.uuid4().hex}.sock"
        os.environ["LUTE_SOCKET"] = socket_path
        logger.debug(f"SocketCommunicator defines socket_path: {socket_path}")
    if USE_ZMQ:
        return f"ipc://{socket_path}"
    else:
        return socket_path

_init_tcp_socket_raw()

Initialize a TCP socket.

Executor-side code should always be run first. It checks to see if the environment variable LUTE_PORT=### is defined, if so binds it, otherwise find a free port from a selection of random ports. If a port search is performed, the LUTE_PORT variable will be defined so it can be picked up by the the Task-side Communicator.

In the event that no port can be bound on the Executor-side, or the port and hostname information is unavailable to the Task-side, the program will exit.

Returns:

Name	Type	Description
data_socket	socket	TCP socket object.

Source code in lute/execution/ipc.py

def _init_tcp_socket_raw(self) -> socket.socket:
    """Initialize a TCP socket.

    Executor-side code should always be run first. It checks to see if
    the environment variable
                            `LUTE_PORT=###`
    is defined, if so binds it, otherwise find a free port from a selection
    of random ports. If a port search is performed, the `LUTE_PORT` variable
    will be defined so it can be picked up by the the Task-side Communicator.

    In the event that no port can be bound on the Executor-side, or the port
    and hostname information is unavailable to the Task-side, the program
    will exit.

    Returns:
        data_socket (socket.socket): TCP socket object.
    """
    data_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    port: Optional[Union[str, int]] = os.getenv("LUTE_PORT")
    if self._party == Party.EXECUTOR:
        if port is None:
            # If port is None find one
            # Executor code executes first
            port = self._find_random_port()
            if port is None:
                # Failed to find a port to bind
                logger.info(
                    "Executor failed to bind a port. "
                    "Try providing a LUTE_PORT directly! Exiting!"
                )
                sys.exit(-1)
            # Provide port env var for Task-side
            os.environ["LUTE_PORT"] = str(port)
        data_socket.bind(("", int(port)))
        data_socket.listen()
    else:
        hostname: str = socket.gethostname()
        executor_hostname: Optional[str] = os.getenv("LUTE_EXECUTOR_HOST")
        if executor_hostname is None or port is None:
            logger.info(
                "Task-side does not have host/port information!"
                " Check environment variables! Exiting!"
            )
            sys.exit(-1)
        if hostname == executor_hostname:
            data_socket.connect(("localhost", int(port)))
        else:
            data_socket.connect((executor_hostname, int(port)))
    return data_socket

_init_tcp_socket_zmq(data_socket)

Initialize a TCP socket using ZMQ.

Equivalent as the method above but requires passing in a ZMQ socket object instead of returning one.

Parameters:

Name	Type	Description	Default
data_socket	Socket	Socket object.	required

Source code in lute/execution/ipc.py

def _init_tcp_socket_zmq(self, data_socket: zmq.sugar.socket.Socket) -> None:
    """Initialize a TCP socket using ZMQ.

    Equivalent as the method above but requires passing in a ZMQ socket
    object instead of returning one.

    Args:
        data_socket (zmq.socket.Socket): Socket object.
    """
    port: Optional[Union[str, int]] = os.getenv("LUTE_PORT")
    if self._party == Party.EXECUTOR:
        if port is None:
            new_port: int = data_socket.bind_to_random_port("tcp://*")
            if new_port is None:
                # Failed to find a port to bind
                logger.info(
                    "Executor failed to bind a port. "
                    "Try providing a LUTE_PORT directly! Exiting!"
                )
                sys.exit(-1)
            port = new_port
            os.environ["LUTE_PORT"] = str(port)
        else:
            data_socket.bind(f"tcp://*:{port}")
        logger.debug(f"Executor bound port {port}")
    else:
        executor_hostname: Optional[str] = os.getenv("LUTE_EXECUTOR_HOST")
        if executor_hostname is None or port is None:
            logger.info(
                "Task-side does not have host/port information!"
                " Check environment variables! Exiting!"
            )
            sys.exit(-1)
        data_socket.connect(f"tcp://{executor_hostname}:{port}")

_init_unix_socket_raw()

Returns a Unix socket object.

Executor-side code should always be run first. It checks to see if the environment variable LUTE_SOCKET=XYZ is defined, if so binds it, otherwise it will create a new path and define the environment variable for the Task-side to find.

On the Task (client-side), this method will also open a SSH tunnel to forward a local Unix socket to an Executor Unix socket if the Task and Executor processes are on different machines.

Returns:

Name	Type	Description
data_socket	socket	Unix socket object.

Source code in lute/execution/ipc.py

def _init_unix_socket_raw(self) -> socket.socket:
    """Returns a Unix socket object.

    Executor-side code should always be run first. It checks to see if
    the environment variable
                            `LUTE_SOCKET=XYZ`
    is defined, if so binds it, otherwise it will create a new path and
    define the environment variable for the Task-side to find.

    On the Task (client-side), this method will also open a SSH tunnel to
    forward a local Unix socket to an Executor Unix socket if the Task and
    Executor processes are on different machines.

    Returns:
        data_socket (socket.socket): Unix socket object.
    """
    socket_path: str = self._get_socket_path()
    data_socket = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM)
    if self._party == Party.EXECUTOR:
        if os.path.exists(socket_path):
            os.unlink(socket_path)
        data_socket.bind(socket_path)
        data_socket.listen()
    elif self._party == Party.TASK:
        hostname: str = socket.gethostname()
        executor_hostname: Optional[str] = os.getenv("LUTE_EXECUTOR_HOST")
        if executor_hostname is None:
            logger.info("Hostname for Executor process not found! Exiting!")
            data_socket.close()
            sys.exit(-1)
        if hostname == executor_hostname:
            data_socket.connect(socket_path)
        else:
            self._local_socket_path = self._setup_unix_ssh_tunnel(
                socket_path, hostname, executor_hostname
            )
            while 1:
                # Keep trying reconnect until ssh tunnel works.
                try:
                    data_socket.connect(self._local_socket_path)
                    break
                except FileNotFoundError:
                    continue

    return data_socket

_init_unix_socket_zmq(data_socket)

Initialize a Unix socket object, using ZMQ.

Equivalent as the method above but requires passing in a ZMQ socket object instead of returning one.

Parameters:

Name	Type	Description	Default
data_socket	socket	ZMQ object.	required

Source code in lute/execution/ipc.py

def _init_unix_socket_zmq(self, data_socket: zmq.sugar.socket.Socket) -> None:
    """Initialize a Unix socket object, using ZMQ.

    Equivalent as the method above but requires passing in a ZMQ socket
    object instead of returning one.

    Args:
        data_socket (socket.socket): ZMQ object.
    """
    socket_path = self._get_socket_path()
    if self._party == Party.EXECUTOR:
        if os.path.exists(socket_path):
            os.unlink(socket_path)
        data_socket.bind(socket_path)
    elif self._party == Party.TASK:
        hostname: str = socket.gethostname()
        executor_hostname: Optional[str] = os.getenv("LUTE_EXECUTOR_HOST")
        if executor_hostname is None:
            logger.info("Hostname for Executor process not found! Exiting!")
            self._data_socket.close()
            sys.exit(-1)
        if hostname == executor_hostname:
            data_socket.connect(socket_path)
        else:
            # Need to remove ipc:// from socket_path for forwarding
            self._local_socket_path = self._setup_unix_ssh_tunnel(
                socket_path[6:], hostname, executor_hostname
            )
            # Need to add it back
            path: str = f"ipc://{self._local_socket_path}"
            data_socket.connect(path)

_read_socket()

Read data from a socket.

Socket(s) are continuously monitored, and read from when new data is available.

Calls an underlying method for either raw sockets or ZMQ.

Source code in lute/execution/ipc.py

def _read_socket(self) -> None:
    """Read data from a socket.

    Socket(s) are continuously monitored, and read from when new data is
    available.

    Calls an underlying method for either raw sockets or ZMQ.
    """

    while True:
        if self._stop_thread:
            logger.debug("Stopping socket reader thread.")
            break
        if USE_ZMQ:
            self._read_socket_zmq()
        else:
            self._read_socket_raw()

_read_socket_raw()

Read data from a socket.

Raw socket implementation for the reader thread.

Source code in lute/execution/ipc.py

def _read_socket_raw(self) -> None:
    """Read data from a socket.

    Raw socket implementation for the reader thread.
    """
    assert isinstance(self._data_socket, socket.socket)
    connection: socket.socket
    _: Union[str, Tuple[str, int]]
    try:
        connection, _ = self._data_socket.accept()
        full_data: bytes = b""
        while True:
            data: bytes = connection.recv(8192)
            if data:
                full_data += data
            else:
                break
        connection.close()
        self._unpack_messages(full_data)
    except socket.timeout:
        pass

_read_socket_zmq()

Read data from a socket.

ZMQ implementation for the reader thread.

Source code in lute/execution/ipc.py

def _read_socket_zmq(self) -> None:
    """Read data from a socket.

    ZMQ implementation for the reader thread.
    """
    try:
        full_data: bytes = self._data_socket.recv(0)
        self._unpack_messages(full_data)
    except zmq.ZMQError:
        pass

_setup_unix_ssh_tunnel(socket_path, hostname, executor_hostname)

Prepares an SSH tunnel for forwarding between Unix sockets on two hosts.

An SSH tunnel is opened with ssh -L <local>:<remote> sleep 2. This method of communication is slightly slower and incurs additional overhead - it should only be used as a backup. If communication across multiple hosts is required consider using TCP. The Task will use the local socket <LUTE_SOCKET>.task{##}. Multiple local sockets may be created. It is assumed that the user is identical on both the Task machine and Executor machine.

Returns:

Name	Type	Description
local_socket_path	str	The local Unix socket to connect to.

Source code in lute/execution/ipc.py

def _setup_unix_ssh_tunnel(
    self, socket_path: str, hostname: str, executor_hostname: str
) -> str:
    """Prepares an SSH tunnel for forwarding between Unix sockets on two hosts.

    An SSH tunnel is opened with `ssh -L <local>:<remote> sleep 2`.
    This method of communication is slightly slower and incurs additional
    overhead - it should only be used as a backup. If communication across
    multiple hosts is required consider using TCP.  The Task will use
    the local socket `<LUTE_SOCKET>.task{##}`. Multiple local sockets may be
    created. It is assumed that the user is identical on both the
    Task machine and Executor machine.

    Returns:
        local_socket_path (str): The local Unix socket to connect to.
    """
    if "uuid" not in globals():
        import uuid
    local_socket_path = f"{socket_path}.task{uuid.uuid4().hex[:4]}"
    self._use_ssh_tunnel = True
    ssh_cmd: List[str] = [
        "ssh",
        "-o",
        "LogLevel=quiet",
        "-L",
        f"{local_socket_path}:{socket_path}",
        executor_hostname,
        "sleep",
        "2",
    ]
    logger.debug(f"Opening tunnel from {hostname} to {executor_hostname}")
    self._ssh_proc = subprocess.Popen(ssh_cmd)
    time.sleep(0.4)  # Need to wait... -> Use single Task comm at beginning?
    return local_socket_path

_unpack_messages(data)

Unpacks a byte stream into individual messages.

Messages are encoded in the following format

The items between <> are replaced as follows: - : A start marker - : A separator for components of the message - : The length of the message payload in bytes. - : The message payload in bytes - : The start marker in reverse to indicate the end.

Partial messages (a series of bytes which cannot be converted to a full message) are stored for later. An attempt is made to reconstruct the message with the next call to this method.

Parameters:

Name	Type	Description	Default
data	bytes	A raw byte stream containing anywhere from a partial message to multiple full messages.	required

Source code in lute/execution/ipc.py

def _unpack_messages(self, data: bytes) -> None:
    """Unpacks a byte stream into individual messages.

    Messages are encoded in the following format:
             <HEAD><SEP><len(msg)><SEP><msg><SEP><HEAD[::-1]>
    The items between <> are replaced as follows:
        - <HEAD>: A start marker
        - <SEP>: A separator for components of the message
        - <len(msg)>: The length of the message payload in bytes.
        - <msg>: The message payload in bytes
        - <HEAD[::-1]>: The start marker in reverse to indicate the end.

    Partial messages (a series of bytes which cannot be converted to a full
    message) are stored for later. An attempt is made to reconstruct the
    message with the next call to this method.

    Args:
        data (bytes): A raw byte stream containing anywhere from a partial
            message to multiple full messages.
    """
    msg: Message
    working_data: bytes
    if self._partial_msg:
        # Concatenate the previous partial message to the beginning
        working_data = self._partial_msg + data
        self._partial_msg = None
    else:
        working_data = data
    while working_data:
        try:
            # Message encoding: <HEAD><SEP><len><SEP><msg><SEP><HEAD[::-1]>
            end = working_data.find(
                SocketCommunicator.MSG_SEP + SocketCommunicator.MSG_HEAD[::-1]
            )
            msg_parts: List[bytes] = working_data[:end].split(
                SocketCommunicator.MSG_SEP
            )
            if len(msg_parts) != 3:
                self._partial_msg = working_data
                break

            cmd: bytes
            nbytes: bytes
            raw_msg: bytes
            cmd, nbytes, raw_msg = msg_parts
            if len(raw_msg) != int(nbytes):
                self._partial_msg = working_data
                break
            msg = pickle.loads(raw_msg)
            self._msg_queue.put(msg)
        except pickle.UnpicklingError:
            self._partial_msg = working_data
            break
        if end < len(working_data):
            # Add len(SEP+HEAD) since end marks the start of <SEP><HEAD[::-1]
            offset: int = len(
                SocketCommunicator.MSG_SEP + SocketCommunicator.MSG_HEAD
            )
            working_data = working_data[end + offset :]
        else:
            working_data = b""

_write_socket(msg)

Sends data over a socket from the 'client' (Task) side.

Messages are encoded in the following format

The items between <> are replaced as follows: - : A start marker - : A separator for components of the message - : The length of the message payload in bytes. - : The message payload in bytes - : The start marker in reverse to indicate the end.

This structure is used for decoding the message on the other end.

Source code in lute/execution/ipc.py

def _write_socket(self, msg: Message) -> None:
    """Sends data over a socket from the 'client' (Task) side.

    Messages are encoded in the following format:
             <HEAD><SEP><len(msg)><SEP><msg><SEP><HEAD[::-1]>
    The items between <> are replaced as follows:
        - <HEAD>: A start marker
        - <SEP>: A separator for components of the message
        - <len(msg)>: The length of the message payload in bytes.
        - <msg>: The message payload in bytes
        - <HEAD[::-1]>: The start marker in reverse to indicate the end.

    This structure is used for decoding the message on the other end.
    """
    data: bytes = pickle.dumps(msg)
    cmd: bytes = SocketCommunicator.MSG_HEAD
    size: bytes = b"%d" % len(data)
    end: bytes = SocketCommunicator.MSG_HEAD[::-1]
    sep: bytes = SocketCommunicator.MSG_SEP
    packed_msg: bytes = cmd + sep + size + sep + data + sep + end
    if USE_ZMQ:
        self._data_socket.send(packed_msg)
    else:
        assert isinstance(self._data_socket, socket.socket)
        self._data_socket.sendall(packed_msg)

delayed_setup()

Delays the creation of socket objects.

The Executor initializes the Communicator when it is created. Since all Executors are created and available at once we want to delay acquisition of socket resources until a single Executor is ready to use them.

Source code in lute/execution/ipc.py

def delayed_setup(self) -> None:
    """Delays the creation of socket objects.

    The Executor initializes the Communicator when it is created. Since
    all Executors are created and available at once we want to delay
    acquisition of socket resources until a single Executor is ready
    to use them.
    """
    self._data_socket: Union[socket.socket, zmq.sugar.socket.Socket]
    self.desc: str
    use_tcp: Optional[str] = os.getenv("LUTE_USE_TCP")
    sock_type: str
    if use_tcp is not None:
        sock_type = "TCP"
    else:
        sock_type = "Unix"
    if USE_ZMQ:
        self.desc = f"Communicates using ZMQ through {sock_type} sockets."
        self._context: zmq.Context = zmq.Context()
        self._data_socket = self._create_socket_zmq()
    else:
        self.desc = f"Communicates through {sock_type} sockets."
        self._data_socket = self._create_socket_raw()
        self._data_socket.settimeout(SocketCommunicator.ACCEPT_TIMEOUT)

    if self._party == Party.EXECUTOR:
        # Executor created first so we can define the hostname env variable
        os.environ["LUTE_EXECUTOR_HOST"] = socket.gethostname()
        # Setup reader thread
        self._reader_thread: threading.Thread = threading.Thread(
            target=self._read_socket
        )
        self._msg_queue: queue.Queue = queue.Queue()
        self._partial_msg: Optional[bytes] = None
        self._stop_thread: bool = False
        self._reader_thread.start()
    else:
        # Only used by Party.TASK
        self._use_ssh_tunnel: bool = False
        self._ssh_proc: Optional[subprocess.Popen] = None
        self._local_socket_path: Optional[str] = None

read(proc)

Return a message from the queue if available.

Socket(s) are continuously monitored, and read from when new data is available.

Parameters:

Name	Type	Description	Default
proc	Popen	The process to read from. Provided for compatibility with other Communicator subtypes. Is ignored.	required

Returns:

Name	Type	Description
msg	Message	The message read, containing contents and signal.

Source code in lute/execution/ipc.py

def read(self, proc: subprocess.Popen) -> Message:
    """Return a message from the queue if available.

    Socket(s) are continuously monitored, and read from when new data is
    available.

    Args:
        proc (subprocess.Popen): The process to read from. Provided for
            compatibility with other Communicator subtypes. Is ignored.

    Returns:
         msg (Message): The message read, containing contents and signal.
    """
    msg: Message
    try:
        msg = self._msg_queue.get(timeout=SocketCommunicator.ACCEPT_TIMEOUT)
    except queue.Empty:
        msg = Message()

    return msg

write(msg)

Send a single Message.

The entire Message (signal and contents) is serialized and sent through a connection over Unix socket.

Parameters:

Name	Type	Description	Default
msg	Message	The Message to send.	required

Source code in lute/execution/ipc.py

def write(self, msg: Message) -> None:
    """Send a single Message.

    The entire Message (signal and contents) is serialized and sent through
    a connection over Unix socket.

    Args:
        msg (Message): The Message to send.
    """
    self._write_socket(msg)