Skip to content

Latest commit

 

History

History
77 lines (50 loc) · 2.69 KB

install.md

File metadata and controls

77 lines (50 loc) · 2.69 KB
Raw

Installation Guide

The most part of OpenASCE is written in Python, and the installation process of this part is quite simple. While the causal tree algorithms rely on C++ libraries, currently we only provide pre-built binary wheels for Linux(x86_64). We also provide a Docker image which includes all the necessary libraries for users to build and install OpenASCE from the source.

Pip Installation

Linux

For Linux (x86_64) with Python 3.11, the pre-built wheel is uploaded to PyPI (Python Package Index), users could simply install it as follows.

pip install openasce

We will support more platforms later.

Other Platforms

For other platforms, users can install from the source code as follows.

git clone https://github.com/Open-All-Scale-Causal-Engine/OpenASCE.git
cd OpenASCE
pip install .

After the installation, you can use all the algorithms except causal tree algorithms.

Docker Installation

We provide a docker image that contains all the dependencies required to build the C++ code and run OpenASCE. Users can practice the system using the image without environmental issues 1.

Before getting the image, the user should log in to the Aliyun Registry first. For users without an Aliyun account, please refer to this link for registration.

docker login --username=${your_user_name} registry.cn-hangzhou.aliyuncs.com
docker pull registry.cn-hangzhou.aliyuncs.com/openasce/openasce:gcc9.4-py3.11

You may need to use sudo if you are using a non-root user.

The users can also build their image for OpenASCE according to the Dockerfile.

docker build --network host -f Dockerfile -t openasce:gcc9.4-py3.11 .

1. Start Docker

After getting the image, the user can run the container as follows.

# pull from the registry
docker run --net=host --rm -it -m 16g --name openasce_env registry.cn-hangzhou.aliyuncs.com/openasce/openasce:gcc9.4-py3.11 "/usr/bin/bash"
# build from the local
docker run --net=host --rm -it -m 16g --name openasce_env openasce:gcc9.4-py3.11 "/usr/bin/bash"

2. Clone the Code

git clone https://github.com/Open-All-Scale-Causal-Engine/OpenASCE.git
cd OpenASCE
git submodule update --recursive --init

3. Compile the Source Code and Install the Package

You can use the build.sh script to build the C++ code and install OpenASCE in the current environment.

bash scripts/build.sh dev

The package can be found in the dist directory and is typically named openasce-0.1.0-cp311-none-linux_x86_64.whl for reinstallation.

Footnotes

  1. Currently it doesn't work well on Mac with a Silicon chip. ↩