lexigram-queue

Message bus and queue with Named DI multi-backend support for the Lexigram Framework.

Overview

lexigram-queue provides async message queue and bus functionality with Redis, RabbitMQ, Kafka, SQS, and in-memory backends. It includes consumer discovery, dead-letter queue routing after max retries, transactional outbox for atomic DB+message publishing, and composable middleware pipelines — all wired through the DI container.

Full documentation: docs.lexigram.dev

Install

uv add lexigram lexigram-queue

# With Redis support
uv add "lexigram-queue[redis]"

# With RabbitMQ support
uv add "lexigram-queue[rabbitmq]"

# With Kafka support
uv add "lexigram-queue[kafka]"

# With AWS SQS support
uv add "lexigram-queue[sqs]"

Quick Start

from lexigram import Application
from lexigram.di.module import Module, module
from lexigram.queue import MessageConsumer, QueueModule
from lexigram.queue.config import KafkaDriverConfig, NamedQueueConfig, QueueConfig
from lexigram.contracts.queue.protocols import QueueProtocol


class OrderConsumer(MessageConsumer):
    topic = "orders"

    async def handle(self, message: dict) -> None:
        print(f"Processing order: {message}")


@module(
    imports=[
        QueueModule.configure(
            QueueConfig(
                backends=[
                    NamedQueueConfig(
                        name="primary",
                        primary=True,
                        driver="kafka",
                        kafka=KafkaDriverConfig(
                            bootstrap_servers="localhost:9092",
                        ),
                    )
                ]
            )
        ),
        QueueModule.scope(OrderConsumer),
    ]
)
class AppModule(Module):
    pass


async def main() -> None:
    async with Application.boot(modules=[AppModule]) as app:
        queue = await app.container.resolve(QueueProtocol)

        await queue.publish("orders", {"order_id": "12345", "total": 99.99})


if __name__ == "__main__":
    import asyncio
    asyncio.run(main())

Configuration

Zero-config usage: Call QueueModule.configure() with no arguments to use all defaults (memory backend).

Option 1 — YAML file

# application.yaml
queue:
  backends:
    - name: default
      primary: true
      driver: kafka
      max_retries: 3
      kafka:
        bootstrap_servers: "localhost:9092"
        group_id: "lexigram-consumers"

Option 2 — Profiles + Environment Variables (recommended)

export LEX_QUEUE__ENABLED=true
export LEX_QUEUE__BACKENDS__0__DRIVER=kafka

Option 3 — Python

from lexigram.queue import QueueModule
from lexigram.queue.config import QueueConfig, NamedQueueConfig, KafkaDriverConfig

QueueModule.configure(
    QueueConfig(
        backends=[
            NamedQueueConfig(
                name="default",
                primary=True,
                driver="kafka",
                kafka=KafkaDriverConfig(
                    bootstrap_servers="localhost:9092",
                ),
            ),
        ]
    )
)

Config reference

Module Factory Methods

Key Features

• Multi-backend messaging — Redis Pub/Sub, RabbitMQ, Kafka, AWS SQS, and in-memory
• Consumer discovery — MessageConsumer subclasses auto-discovered and registered
• Dead-letter queue (DLQ) — failed messages routed to DLQ after max_retries
• Transactional outbox — atomic DB transaction + message publish via TransactionalOutbox
• Message middleware pipelines — composable logging, validation, transformation middleware
• Named DI multi-backend — Annotated[QueueProtocol, Named("events")] for multiple backends
• Retry logic — exponential backoff before DLQ routing
• Consumer groups — Kafka consumer groups and RabbitMQ consumer tags for load balancing

Backend Trade-offs

Testing

from lexigram import Application
from lexigram.queue import QueueModule
from lexigram.contracts.queue.protocols import QueueProtocol

async def test_message_consumer():
    async with Application.boot(
        modules=[QueueModule.stub()]
    ) as app:
        queue = await app.container.resolve(QueueProtocol)
        await queue.publish("test-topic", {"key": "value"})
        # Test with in-memory backend

Key Source Files