nest-rpc-client 1.3.0b0

Async Python client for RPC transport interoperability with NestJS microservices.

nest-rpc-client

Async Python client for RPC transport interoperability with NestJS microservices. Supports RabbitMQ, Redis, NATS, TCP (and planned support for Kafka, MQTT).

Features

• Fully async transports
• Pluggable transport architecture
• Compatible with NestJS microservices patterns (send, emit)
• Extras-based installation

Installation

Install with a specific transport:

# TCP uses asyncio
pip install "nest-rpc-client[rabbitmq]"
pip install "nest-rpc-client[redis]"
pip install "nest-rpc-client[nats]"
pip install "nest-rpc-client[all]"

Usage

from nest_rpc_client.client import Client
from nest_rpc_client.transports.rabbitmq import RabbitMQTransport
from nest_rpc_client.config.rabbitmq import RabbitMQConfig

transport = RabbitMQTransport(RabbitMQConfig(url="amqp://guest:guest@localhost/", queue="rpc_queue"))

async with Client(transport) as client:
    result = await client.send("get_user", {"id": 123})
    await client.emit("user_created", {"id": 123})

You can also use only transport:

from nest_rpc_client.transports.rabbitmq import RabbitMQTransport
from nest_rpc_client.config.rabbitmq import RabbitMQConfig

transport = RabbitMQTransport(RabbitMQConfig(url="amqp://guest:guest@localhost/", queue="rpc_queue"))

await transport.connect()

result = await transport.send("get_user", {"id": 123})
await transport.emit("user_created", {"id": 123})

await transport.close()

Pattern usage with dictionaries

If your pattern is a dictionary (object in js, for example: {cmd: "sum"}), you need to serialize it to a string before using it. Example:

pattern = json.dumps({"cmd": "sum"}) # '{"cmd": "sum"}'
await client.send(pattern, data)

Custom Transport

nest-rpc-client allows you to use your own custom transports. To do this, inherit from the Transport base class and implement its abstract methods:

from nest_rpc_client.transport import Transport

class MyCustomTransport(Transport):
    async def connect(self) -> None:
        # Initialize any connections or resources
        ...

    async def close(self) -> None:
        # Clean up resources or close connections
        ...

    async def send(self, pattern: str, data: dict) -> dict:
        # Implement the RPC request (expects a response)
        ...

    async def emit(self, pattern: str, data: dict) -> None:
        # Implement the fire-and-forget event
        ...

You can then use your custom transport exactly like the built-in ones:

transport = MyCustomTransport()
async with Client(transport) as client:
    result = await client.send("my_pattern", {"foo": "bar"})

Exception handling

nest-rpc-client raises RpcException whenever the RPC server returns an error. Example:

from nest_rpc_client.client import Client
from nest_rpc_client.transports.rabbitmq import RabbitMQTransport
from nest_rpc_client.config.rabbitmq import RabbitMQConfig
from nest_rpc_client.exceptions import RpcException

transport = RabbitMQTransport(RabbitMQConfig(url="amqp://guest:guest@localhost/", queue="rpc_queue"))

async with Client(transport) as client:
    try:
        result = await client.send("get_user", {"id": 123})
    except RpcException as e:
        print(f"RPC Error: {e.err}")

When the RPC response includes an err field (following the NestJS microservice pattern), the client automatically raises RpcException with the contents of err as its attribute:

{
    "id": "correlation-id",
    "err": {
        "message": "User not found",
        "code": 404
    }
}

This allows consistent error handling across all transports.

Tests:

This project includes both unit tests and full integration tests.

• Unit tests are included in this repository and can be run with: pytest
• Full integration tests are available in a separate repository: nest-rpc-client-tests

TODO:

• Implement kafka transport
• Implement mqtt transport