Skip to content

LogicHandler

faststream.rabbit.handler.LogicHandler #

LogicHandler(queue: RabbitQueue, log_context_builder: Callable[[StreamMessage[Any]], Dict[str, str]], graceful_timeout: Optional[float] = None, exchange: Optional[RabbitExchange] = None, consume_args: Optional[AnyDict] = None, description: Optional[str] = None, title: Optional[str] = None, include_in_schema: bool = True, virtual_host: str = '/')

Bases: AsyncHandler[IncomingMessage], BaseRMQInformation

A class to handle logic for RabbitMQ message consumption.

METHOD DESCRIPTION
__init__

Initializes the LogicHandler object

add_call

Adds a call to be handled by the LogicHandler

start

Starts consuming messages from the queue

close

Closes the consumer and cancels message consumption

Initialize a RabbitMQ consumer.

PARAMETER DESCRIPTION
queue

RabbitQueue object representing the queue to consume from

TYPE: RabbitQueue

log_context_builder

Callable that returns a dictionary with log context information

TYPE: Callable[[StreamMessage[Any]], Dict[str, str]]

graceful_timeout

Optional float representing the graceful timeout

TYPE: Optional[float] DEFAULT: None

exchange

RabbitExchange object representing the exchange to bind the queue to (optional)

TYPE: Optional[RabbitExchange] DEFAULT: None

consume_args

Additional arguments for consuming from the queue (optional)

TYPE: Optional[AnyDict] DEFAULT: None

description

Description of the consumer (optional)

TYPE: Optional[str] DEFAULT: None

title

Title of the consumer (optional)

TYPE: Optional[str] DEFAULT: None

include_in_schema

Whether to include the consumer in the API specification (optional)

TYPE: bool DEFAULT: True

virtual_host

Virtual host to connect to (optional)

TYPE: str DEFAULT: '/'

Source code in faststream/rabbit/handler.py
def __init__(
    self,
    queue: RabbitQueue,
    log_context_builder: Callable[[StreamMessage[Any]], Dict[str, str]],
    graceful_timeout: Optional[float] = None,
    # RMQ information
    exchange: Optional[RabbitExchange] = None,
    consume_args: Optional[AnyDict] = None,
    # AsyncAPI information
    description: Optional[str] = None,
    title: Optional[str] = None,
    include_in_schema: bool = True,
    virtual_host: str = "/",
) -> None:
    """Initialize a RabbitMQ consumer.

    Args:
        queue: RabbitQueue object representing the queue to consume from
        log_context_builder: Callable that returns a dictionary with log context information
        graceful_timeout: Optional float representing the graceful timeout
        exchange: RabbitExchange object representing the exchange to bind the queue to (optional)
        consume_args: Additional arguments for consuming from the queue (optional)
        description: Description of the consumer (optional)
        title: Title of the consumer (optional)
        include_in_schema: Whether to include the consumer in the API specification (optional)
        virtual_host: Virtual host to connect to (optional)

    """
    super().__init__(
        log_context_builder=log_context_builder,
        description=description,
        title=title,
        include_in_schema=include_in_schema,
        graceful_timeout=graceful_timeout,
    )

    self.queue = queue
    self.exchange = exchange
    self.virtual_host = virtual_host
    self.consume_args = consume_args or {}

    self._consumer_tag = None
    self._queue_obj = None

call_name property #

call_name: str

Returns the name of the handler call.

calls instance-attribute #

calls: List[Tuple[HandlerCallWrapper[MsgType, Any, SendableMessage], Callable[[StreamMessage[MsgType]], Awaitable[bool]], AsyncParser[MsgType, Any], AsyncDecoder[StreamMessage[MsgType]], Sequence[Callable[[Any], BaseMiddleware]], CallModel[Any, SendableMessage]]]

consume_args instance-attribute #

consume_args: AnyDict = consume_args or {}

description property #

description: Optional[str]

Returns the description of the handler.

exchange instance-attribute #

global_middlewares instance-attribute #

global_middlewares: Sequence[Callable[[Any], BaseMiddleware]] = []

graceful_timeout instance-attribute #

graceful_timeout = graceful_timeout

include_in_schema instance-attribute #

include_in_schema = include_in_schema

lock instance-attribute #

lock = MultiLock()

log_context_builder instance-attribute #

log_context_builder = log_context_builder

queue instance-attribute #

queue: RabbitQueue = queue

running instance-attribute #

running = False

virtual_host instance-attribute #

virtual_host = virtual_host

add_call #

add_call(*, handler: HandlerCallWrapper[IncomingMessage, P_HandlerParams, T_HandlerReturn], dependant: CallModel[P_HandlerParams, T_HandlerReturn], parser: Optional[CustomParser[IncomingMessage, RabbitMessage]], decoder: Optional[CustomDecoder[RabbitMessage]], filter: Filter[RabbitMessage], middlewares: Optional[Sequence[Callable[[IncomingMessage], BaseMiddleware]]]) -> None

Add a call to the handler.

PARAMETER DESCRIPTION
handler

The handler for the call.

TYPE: HandlerCallWrapper[IncomingMessage, P_HandlerParams, T_HandlerReturn]

dependant

The dependant for the call.

TYPE: CallModel[P_HandlerParams, T_HandlerReturn]

parser

Optional custom parser for the call.

TYPE: Optional[CustomParser[IncomingMessage, RabbitMessage]]

decoder

Optional custom decoder for the call.

TYPE: Optional[CustomDecoder[RabbitMessage]]

filter

The filter for the call.

TYPE: Filter[RabbitMessage]

middlewares

Optional sequence of middlewares for the call.

TYPE: Optional[Sequence[Callable[[IncomingMessage], BaseMiddleware]]]

RETURNS DESCRIPTION
None

None

Source code in faststream/rabbit/handler.py
def add_call(
    self,
    *,
    handler: HandlerCallWrapper[
        aio_pika.IncomingMessage, P_HandlerParams, T_HandlerReturn
    ],
    dependant: CallModel[P_HandlerParams, T_HandlerReturn],
    parser: Optional[CustomParser[aio_pika.IncomingMessage, RabbitMessage]],
    decoder: Optional[CustomDecoder[RabbitMessage]],
    filter: Filter[RabbitMessage],
    middlewares: Optional[
        Sequence[Callable[[aio_pika.IncomingMessage], BaseMiddleware]]
    ],
) -> None:
    """Add a call to the handler.

    Args:
        handler: The handler for the call.
        dependant: The dependant for the call.
        parser: Optional custom parser for the call.
        decoder: Optional custom decoder for the call.
        filter: The filter for the call.
        middlewares: Optional sequence of middlewares for the call.

    Returns:
        None

    """
    super().add_call(
        handler=handler,
        parser=resolve_custom_func(parser, AioPikaParser.parse_message),
        decoder=resolve_custom_func(decoder, AioPikaParser.decode_message),
        filter=filter,  # type: ignore[arg-type]
        dependant=dependant,
        middlewares=middlewares,
    )

close async #

close() -> None
Source code in faststream/rabbit/handler.py
async def close(self) -> None:
    await super().close()

    if self._queue_obj is not None:
        if self._consumer_tag is not None:  # pragma: no branch
            if not self._queue_obj.channel.is_closed:
                await self._queue_obj.cancel(self._consumer_tag)
            self._consumer_tag = None
        self._queue_obj = None

consume async #

consume(msg: MsgType) -> SendableMessage

Consume a message asynchronously.

PARAMETER DESCRIPTION
msg

The message to be consumed.

TYPE: MsgType

RETURNS DESCRIPTION
SendableMessage

The sendable message.

RAISES DESCRIPTION
StopConsume

If the consumption needs to be stopped.

RAISES DESCRIPTION
Exception

If an error occurs during consumption.

Source code in faststream/broker/handler.py
@override
async def consume(self, msg: MsgType) -> SendableMessage:  # type: ignore[override]
    """Consume a message asynchronously.

    Args:
        msg: The message to be consumed.

    Returns:
        The sendable message.

    Raises:
        StopConsume: If the consumption needs to be stopped.

    Raises:
        Exception: If an error occurs during consumption.

    """
    result: Optional[WrappedReturn[SendableMessage]] = None
    result_msg: SendableMessage = None

    if not self.running:
        return result_msg

    log_context_tag: Optional["Token[Any]"] = None
    async with AsyncExitStack() as stack:
        stack.enter_context(self.lock)

        stack.enter_context(context.scope("handler_", self))

        gl_middlewares: List[BaseMiddleware] = [
            await stack.enter_async_context(m(msg)) for m in self.global_middlewares
        ]

        logged = False
        processed = False
        for handler, filter_, parser, decoder, middlewares, _ in self.calls:
            local_middlewares: List[BaseMiddleware] = [
                await stack.enter_async_context(m(msg)) for m in middlewares
            ]

            all_middlewares = gl_middlewares + local_middlewares

            # TODO: add parser & decoder caches
            message = await parser(msg)

            if not logged:  # pragma: no branch
                log_context_tag = context.set_local(
                    "log_context",
                    self.log_context_builder(message),
                )

            message.decoded_body = await decoder(message)
            message.processed = processed

            if await filter_(message):
                assert (  # nosec B101
                    not processed
                ), "You can't process a message with multiple consumers"

                try:
                    async with AsyncExitStack() as consume_stack:
                        for m_consume in all_middlewares:
                            message.decoded_body = (
                                await consume_stack.enter_async_context(
                                    m_consume.consume_scope(message.decoded_body)
                                )
                            )

                        result = await cast(
                            Awaitable[Optional[WrappedReturn[SendableMessage]]],
                            handler.call_wrapped(message),
                        )

                    if result is not None:
                        result_msg, pub_response = result

                        # TODO: suppress all publishing errors and raise them after all publishers will be tried
                        for publisher in (pub_response, *handler._publishers):
                            if publisher is not None:
                                async with AsyncExitStack() as pub_stack:
                                    result_to_send = result_msg

                                    for m_pub in all_middlewares:
                                        result_to_send = (
                                            await pub_stack.enter_async_context(
                                                m_pub.publish_scope(result_to_send)
                                            )
                                        )

                                    await publisher.publish(
                                        message=result_to_send,
                                        correlation_id=message.correlation_id,
                                    )

                except StopConsume:
                    await self.close()
                    handler.trigger()

                except HandlerException as e:  # pragma: no cover
                    handler.trigger()
                    raise e

                except Exception as e:
                    handler.trigger(error=e)
                    raise e

                else:
                    handler.trigger(result=result[0] if result else None)
                    message.processed = processed = True
                    if IS_OPTIMIZED:  # pragma: no cover
                        break

        assert not self.running or processed, "You have to consume message"  # nosec B101

    if log_context_tag is not None:
        context.reset_local("log_context", log_context_tag)

    return result_msg

get_payloads #

get_payloads() -> List[Tuple[AnyDict, str]]

Get the payloads of the handler.

Source code in faststream/broker/handler.py
def get_payloads(self) -> List[Tuple[AnyDict, str]]:
    """Get the payloads of the handler."""
    payloads: List[Tuple[AnyDict, str]] = []

    for h, _, _, _, _, dep in self.calls:
        body = parse_handler_params(
            dep,
            prefix=f"{self._title or self.call_name}:Message",
        )
        payloads.append(
            (
                body,
                to_camelcase(unwrap(h._original_call).__name__),
            ),
        )

    return payloads

name #

name() -> str

Returns the name of the API operation.

Source code in faststream/asyncapi/base.py
@abstractproperty
def name(self) -> str:
    """Returns the name of the API operation."""
    raise NotImplementedError()

schema #

schema() -> Dict[str, Channel]

Returns the schema of the API operation as a dictionary of channel names and channel objects.

Source code in faststream/asyncapi/base.py
def schema(self) -> Dict[str, Channel]:  # pragma: no cover
    """Returns the schema of the API operation as a dictionary of channel names and channel objects."""
    return {}

start async #

start(declarer: RabbitDeclarer) -> None

Starts the consumer for the RabbitMQ queue.

PARAMETER DESCRIPTION
declarer

RabbitDeclarer object used to declare the queue and exchange

TYPE: RabbitDeclarer

RETURNS DESCRIPTION
None

None

Source code in faststream/rabbit/handler.py
@override
async def start(self, declarer: RabbitDeclarer) -> None:  # type: ignore[override]
    """Starts the consumer for the RabbitMQ queue.

    Args:
        declarer: RabbitDeclarer object used to declare the queue and exchange

    Returns:
        None

    """
    self._queue_obj = queue = await declarer.declare_queue(self.queue)

    if self.exchange is not None:
        exchange = await declarer.declare_exchange(self.exchange)
        await queue.bind(
            exchange,
            routing_key=self.queue.routing,
            arguments=self.queue.bind_arguments,
        )

    self._consumer_tag = await queue.consume(
        # NOTE: aio-pika expects AbstractIncomingMessage, not IncomingMessage
        self.consume,  # type: ignore[arg-type]
        arguments=self.consume_args,
    )

    await super().start()