quart-rpc 1.4.0

Turn Quart into a simple RPC server

📢 Quart-RPC

pip install quart-rpc

---

The Quart implementation of Flask-RPC

Quart-RPC currently only uses the weeRPC as its protocol, which is a micro JSON-based protocol that allows for easy communication between the client and server.

This extension is designed to stay slim and provides methods for generating requests and responses.

It does not enforce or validate the data passed in, or the data being sent back; this is left to you to implement in whatever way you feel comfortable (or not at all, if there's no need for it)

Quart-RPC does validate the version of weeRPC on an incoming request. This is to ensure that the request is structured in a way that the version of RPC you are using expects.

Other than that, you are free to use whatever data validation you feel comfortable with. Pydantic and Marshmallow are good choices.

The typical request/response cycle is as follows:

Request

{
  "wrpc": 1.0,
  "function": "add_numbers",
  "data": [
    1,
    2,
    3
  ]
}

Response

{
  "wrpc": 1.0,
  "ok": true,
  "message": "Function 'add_numbers' executed successfully",
  "data": 6
}

Usage

This repo contains a working example of Quart-RPC.

It also includes an example of using the JS library that helps in making requests via fetch to Quart-RPC.

Simplest example

from quart import Quart

from quart_rpc.latest import RPC, RPCResponse


def add_numbers(data):
    if isinstance(data, list):
        return RPCResponse.success(
            sum(data),
            "Function 'add_numbers' executed successfully"
        )


app = Quart(__name__)
rpc = RPC(app, url_prefix="/rpc")  # or RPC(blueprint, ...)
rpc.functions(
    add_numbers=add_numbers
)

or

...
RPC(
    app,   # or RPC(blueprint, ...)
    url_prefix="/rpc", 
    functions={
        "add_numbers": add_numbers
    }
)
...

RPC(...)

Will register a POST route with the app or blueprint that you pass in.

rpc.functions(...)

Will register the functions that you pass in to be called remotely. The argument names used will be the name of the function you will call remotely, for example:

rpc.functions(
    add_numbers=add_numbers,
    subtract=subtract_numbers
)

Calling subtract remotely will call the subtract_numbers function.

A request to the /rpc endpoint with the following JSON payload:

import requests
from quart_rpc import RPCRequest

response = requests.post(
    "http://localhost:5000/rpc",
    json=RPCRequest.build(
        function="add_numbers",
        data=[1, 2, 3]
    )
)

or, if you're using the JS library:

fetch("/rpc", {
    method: "POST",
    headers: {
        "Content-Type": "application/json"
    },
    body: wrpc(
        function_ = "add_numbers",
        data = [1, 2, 3]
    )
})

Will return:

{
  "wrpc": 1.0,
  "ok": true,
  "message": "Function 'add_numbers' executed successfully",
  "data": 6
}

Security

You can lock down RPC routes by using sessions and or host checking.

Session Auth

from quart_rpc.latest import RPCAuthSessionKey

...
RPC(
    app,   # or RPC(blueprint, ...)
    url_prefix="/rpc", 
    session_auth=RPCAuthSessionKey("logged_in", [True]),
    functions={
        "add_numbers": add_numbers
    }
)
...

or a list of RPCAuthSessionKey:

...
RPC(
    app,   # or RPC(blueprint, ...)
    url_prefix="/rpc", 
    session_auth=[
        RPCAuthSessionKey("logged_in", [True]),
        RPCAuthSessionKey("user_type", ["admin"])
    ],
    functions={
        "add_numbers": add_numbers
    }
)
...

Host Auth

In the following example, only requests from 127.0.0.1:5000 will be accepted.

...
RPC(
    app,   # or RPC(blueprint, ...)
    url_prefix="/rpc", 
    host_auth=["127.0.0.1:5000"],
    functions={
        "add_numbers": add_numbers
    }
)
...