Release Notes: v0.2.0

The GraphScope team is pleased to announce the 0.2.0 release after two-months development. The 0.2.0 release is focused on better getting started experience for end-users and we have make a lot of improvements since the last minor release. We have improved our documentation a lot, and made the kubernetes integration work for more settings. We have also brought the support for various I/O to make GraphScope suitable for more production environments.

Along with the 0.2.0 release we also bring GraphScope Playground to users for getting started with GraphScope easier.

We highlights the following improvements included in this releases:

1. Better Kubernetes integration.

   • LoadBalancer service type has been supported in this release, which brings GraphScope to vendored could environments like Alibabacloud, AWS and Azure.
   • A volumes parameter has been added to graphscope.session() to allow users mount their volumes (host path, or cloud volumes, etc.) to worker pods for input graph data and retrieve results.

2. Enhancement for interactive query.

   • GraphScope v0.2.0 added a InteractiveQuery.traversal_source() method to allows an interative query seamlessly works with the gremlinpython SDK.

3. Better I/O and serialization support.

   • GraphScope v0.2.0 started to support various I/O, including Aliyun OSS, AWS S3, HDFS and other storage services that the filesystem-spec could support, for input datasets and result sink.
   • GraphScope v0.2.0 added a Graph.serialize() method to allow users serialize a graph to external storage and restore it back via Graph.deserailize().

4. Bugfix for learning engine.

   • A bug in closing learning instance of a graph has been fixed along this release.

For more detailed bugfixes and improvements that have been made in this version, please refer to the complete changelog.