Graphistry 2.33.17 introduces a powerful piece in our mission to bring powerful visual investigations technology: Dashboarding. With StreamLit, embedded in our open-source graph-app-kit project, you can quickly create interactive graph dashboards and share them. In addition, 2.33 includes the big RAPIDS 0.16 release and continued GPU infrastructure improvements. Read on for these below, and check out the changelog for additional release items.
As usual, you can try the release live on Graphistry Hub, download a private version as an enterprise user, and upon approval, one-click launch from the cloud marketplaces.
Graph dashboarding with StreamLit using graph-app-kit
Turning graph data into interactive dashboards is a powerful way to enable more of your team and organization to benefit from working with the relationships in their data. StreamLit is a new and incredibly popular Python-based data science dashboarding system that is poised to fill the dashboarding gap in our ecosystem. It should be natural for Graphistry users already comfortable with Python or data science notebooks. To ease the journey from idea prototype to deployment, we are excited to announce graph-app-kit, which is a batteries-included dashboarding toolkit for combining Graphistry, StreamLit, RAPIDS, and various graph technologies.
In the coming months, you’ll be hearing more about graph-app-kit. We are already deploying it across projects, and are excited by the momentum in all the communities. Meanwhile, head over to graph-app-kit repo to try for yourself, and stop by our Slack channel and we’d be happy to help.
RAPIDS 0.16
With the stable release of RAPIDS 0.16, we are excited to bring the latest GPU ecosystem advances to our community. It comes with a lot to like. For example, GPU-accelerated SHAP explanations of models are now possible, and multi-GPU computing keeps getting easier with advances in dask-cudf and blazingsql.
Look forward to some coming blogposts about working with these as well. Graphistry servers come with RAPIDS already integrated into the notebooks, so you can launch, `import cudf`, and go!
Infrastructure
We spent significant energy on infrastructure-level improvements and various fixes. Areas that improved include:
- Running concurrent Graphistry instances
- Optimized Docker configuration: Smaller + faster!
- Self tests: APIs, GPU environment, and more
Graphistry 2.33.17 is a great introduction to new users and contains useful upgrades for existing ones, so we hope you enjoy it as much as we do.