Dev and Test

This document describes how to build and test GraphScope Learning Engine from source code.

Dev Environment

To build the learning engine from source code, you need to prepare a development environment with many dependencies and build toolchains. You have two options to prepare the development environment, install all tools and dependencies on your local machine, or build it in our provided docker image.

Install deps on local

To install all dependencies on your local, use the GraphScope command-line utility gs with the subcommand install-deps like this

python3 gsctl.py install-deps dev

# for more usage, try
# python3 gsctl.py install-deps -h

Dev on docker container

We provided a docker image graphscope-dev with all tools and dependencies included.

# Use a mirror in HK aliyun to speed up the download if in need.
# export REGISTRY=registry.cn-hongkong.aliyuncs.com/

docker run --rm -it --shm-size=4096m REGISTRY/graphscope/graphscope-dev:latest

Since we are going to build GraphScope within the container, here we assign shm-size of 4G, which refers to the amount of shared memory allocated to a docker container. More options about docker command can be found here.

Build Learning Engine

You can build all targets for GraphScope Learning Engine with a single command.

python3 gsctl.py make learning

You could install it to a location by:

python3 gsctl.py make learning-install --install-prefix /opt/graphscope

How to Test

Test with the new artifacts.