Unity3D-Python-Communication
⚡️ A very fast, simple, and general inter-process communication example between Unity3D C# and Python, using ZeroMQ.
PS. It looks slow in the GIF above because I put a delay of one second between each message so that you can see it working.
Core Pillars
- very fast — ZeroMQ is a networking library that allows you to send huge amount of data from server to client in a short period of time. I'm talking about casually sending/receiving 10,000 requests per second.
- simple — You don't have to explicitly open and bind a socket or know anything about low-level networking.
- general — You can use this to send/receive any kind of data request. You can send image, video, text, JSON, file, or whatever you want. In this example, we are sending text.
- inter-process — The communication is done inside the same machine. Which means very low-latency.
Introduction
- Have you ever tried to communicate C# code in Unity3D with Python before but could not find a satisfying solution?
- Have you ever tried implementing communication protocol using file read/write and found out that it's a stupid approach?
- Have you ever tried communicating using Web HTTP request and found out that it's stupidly slow and high latency?
- Have you ever tried communicating using socket/TCP/UDP stuff, but it feels like you are reinventing the wheel and you are becoming a network engineer?
- Have you ever tried to communicate by emulating a serial port, and found out that it's not how cool guys do work?
- Have you ever tried to send Unity input to python and do some scientific work (maybe even machine learning task) and return the output to Unity?
- Have you ever tried to build a
.dll
from python or even rewrite everything inC#
because you don't know how to communicate between python and C# processes? - Have you ever tried to embed
IronPython
orPython.NET
inside Unity but it doesn't allow you to install your amazing external python libraries? (And its minimal power is pretty ridiculous compared to your external python) - Have you ever tried to export a
TensorFlow Protobuf Graph
(Deep learning model) and useTensorFlowSharp
orOpenCVForUnity
to import the graph inside Unity because you want to use the model to predict stuff in Unity, but it doesn't allow you to use/utilize your new NVIDIA GeForce RTX 2080Ti, and it's also hard to code? - Tried
MLAgents
, anyone?
Getting Started
- Clone this repository using
git clone https://github.com/offchan42/Unity3D-Python-Communication.git
command. - Open UnityProject (its
dll
files are targeting .NET 4.x version) and runAssets/NetMQExample/Scenes/SampleScene
. - Run python file
PythonFiles/server.py
using commandpython server.py
on a command prompt. - You should start seeing messages being logged inside Unity and the command prompt.
Specifically, Unity will send request with a message Hello
10 times, and Python will simply reply World
10 times. There is a one second sleep between each reply on the server (to simulate long processing time of the request).
Please read the comments inside PythonFiles/server.py
and UnityProject/Assets/NetMQExample/Scripts/
and you will understand everything more deeply.
The most important thing is that you should follow the 4 getting started steps first. Don't skip it! ❣️
After you've understood most of the stuff but it's not advanced enough, you should consult the official ØMQ - The Guide.