WARP: Wrapper around Replicated Processes

WARP is a light-weight manager for background processes.

Functionality

• Easily spawn any process as a managed subprocess
• Automatic recover from failed process
• Runs completely async
• Cleanly exits in a safe order (subprocess, WARP manager, then main process)

Overview

So you need to run a subprocess in the background of your Python application. How do you do it?

• If you simply start the process in an async way, how can you ensure that the process can is restarted if it exits?
• If you start the process in a synchronous way, how do you ensure that the rest of your program continues to execute?
• If you use supervisord, how do you ensure that all your users have supervisord installed ahead of time?
• If you use celery or some kind of heavy weight solution, you now need standalone infra and need to manage a lot of complexity

WARP makes it very simple to manage a background process in a self-healing way. The subprocess runs in a background thread that continually monitors the health of the subprocess. WARP is also able to intelligently support interrupt signals.

Installation

pip install warp

Usage

Import WARP

from warp import Warp

Create a WARP object

wp = Warp(command="dummy_server.py")

Start the WARP process

wp.start()

The process will automatically be killed (cleanly) when the main process exits

Alternatively you may also forcefully stop this background process

wp.stop()

Testing

Run the following command from the repo root

pytest

Python Compatibility

WARP only supports the following Python versions

• 3.7
• 3.8
• 3.9

Contributing

WARP is a community project. Feel free to submit a PR and/or an issue if you are looking for more functionality or found a bug!