IPyC - Python InterProcess Communication

An elegant and modern Python IPC implementation using multiprocessing and asyncio. IPyC comes in two flavors, synchronous and asynchronous, both using the same backend allowing you to pick and chose to your needs.

Send builtins, custom objects, and more!

Key Features

• Uses the modern async and await AsyncIO Python API.
• Includes a synchronous version for backward compatibility.
• Flexible, easy to install, setup, and use.
• Can transfer custom objects and classes at runtime!

Installing

Python 3.5.3 or higher is required if you use the asynchronous version

To install the library you can just run the following command:

# In general
pip3 install IPyC

# Linux/macOS
python3 -m pip install -U IPyC

# Windows
py -3 -m pip install -U IPyC

Hello World Example (using the asynchronous version)

Client

import asyncio
from ipyc import AsyncIPyCClient

async def hello_world():
    client = AsyncIPyCClient()  # Create a client
    link = await client.connect()  # Connect to the host
    await link.send("Hello World!")  # Send a string
    await client.close()  # Close the connection

loop = asyncio.get_event_loop()
loop.run_until_complete(hello_world())

Host

import datetime
from ipyc import AsyncIPyCHost, AsyncIPyCLink

host = AsyncIPyCHost()

@host.on_connect
async def on_client_connect(connection: AsyncIPyCLink):
    while connection.is_active():
        message = await connection.receive()
        if message:
            print(f"[{datetime.datetime.now()}] - Client says: {message}")
    print(f"[{datetime.datetime.now()}] - Connection was closed!")

host.run()

Custom Objects Example (using the synchronous version)

Custom Object

class CustomObject:
    def __init__(self, arg1: int, arg2: float, arg3: str, arg4: set):
        self.a1 = arg1
        self.a2 = arg2
        self.a3 = arg3
        self.a4 = arg4

    def __str__(self):
        return f"<CustomObject:a1={self.a1},a2={self.a2},a3={self.a3},a4={self.a4}>"

    # Define a serializer that returns a string/str representation of the object
    def serialize(self):
        return f"{self.a1}|{self.a2}|{self.a3}|{self.a4}"

    # Define a deserializer that undoes what the serializer did and returns the object
    @staticmethod
    def deserialize(serialization):
        a1, a2, a3, a4 = serialization.split('|')
        return CustomObject(int(a1), float(a2), str(a3), eval(a4))

Client

from ipyc import IPyCClient, IPyCSerialization

custom_object = CustomObject(42, 3.1415926535897932, "Lorem ipsum dolor sit amet", {'s', 'e', 't'})
IPyCSerialization.add_custom_serialization(CustomObject, CustomObject.serialize)

client = IPyCClient()
link = client.connect()
link.send(custom_object)
client.close()

Host

import datetime
from ipyc import IPyCHost, IPyCLink, IPyCSerialization

host = IPyCHost()
IPyCSerialization.add_custom_deserialization(CustomObject, CustomObject.deserialize)

while not host.is_closed():
    connection = host.wait_for_client()

    while connection.is_active():
        message = connection.receive()
        if message:
            print(f"[{datetime.datetime.now()}] - Client sent us a {type(message)} and it was {message}")
            print(message.a1, message.a2, message.a3, message.a4)
    print(f"[{datetime.datetime.now()}] - Connection was closed!")

You can find more examples in the examples directory.

Useful Links

• Documentation
• AsyncIO Documentation
• Multiprocessing Communication Documentation