Snowpark Migration Accelerator: Project Overview¶
The Snowpark Migration Accelerator (SMA) is a simple tool that is easy to run. By using this accelerator, you can learn more about the code in an existing codebase and convert references from the Spark API to the Snowpark API.
So how does it work? That’s what this section is all about. Topics include:
Before we get started, there are two key terms that we need to understand: project and readiness. When you use the tool, your execution of the tool is organized by projects. When analyzing the output, the key indicator is a “readiness” score. Let’s look closer at both of these.
What is a SnowConvert Project?¶
To run this accelerator, you need to create a project. A project is what connects your runs of the tool and the settings that you configure. The tool does this by creating a .snowct file in the source code directory that stores all of your project information locally. This includes your chosen source platform, conversion settings, and status.
What is the Readiness Score?¶
The readiness score is an indicator of how compatible (or “ready”) the references to the Spark API are to the Snowpark API in a scanned codebase. It is not a measure of how much of that codebase can be run in Snowflake. There are other factors to consider when evaluating whether your use case is ready for Snowpark, but the Readiness Score is a good first indicator that there is compatibility between the aspects of your code using Spark and Snowpark.
For more definitions, check out the glossary.
On to the project setup…