Konveyor’s core power lies in its holistic method to migration and modernization. On the coronary heart of Konveyor’s performance is its highly effective analytics engine. This engine analyzes static supply code, figuring out anti-patterns and points which will hinder the applying’s operation on the goal platform. Leveraging neighborhood requirements such because the Language Server Protocol, Konveyor’s evaluation engine makes use of guidelines designed to help varied migration eventualities. Customers also can create customized guidelines to satisfy particular migration wants, growing the pliability and flexibility of the conveyor.
Konveyor AI – Enter “KAI”.
Konveyor AI, the newest growth beneath the Konveyor umbrella, leverages generative AI to additional streamline utility modernization. The first aim of Konveyor AI (or Kai) is to automate supply code modifications, thereby enhancing the economics of migration and modernization efforts.
Our method is to make use of static code evaluation to search out areas within the supply code that should be modified. ‘kai’ will iterate via the evaluation info and work with the LLMs to create code modifications to deal with incidents recognized from the evaluation. This method doesn’t require fine-tuning of LLMs, we improve the data of LLMs via prompts, such because the method with RAG from contained in the conveyor and leveraging evaluation rules to present LLM higher outcomes. Assist to do. For instance, Analyzer-LSP guidelines reminiscent of this ( Quarkus Rulesets from Java EE ) are used to assist information LLMs to replace a Java EE utility in Quarkus.
John Matthews, one of many venture’s lead engineers, has an amazing weblog put up that explains “moss” in depth.
BYO huge language mannequin
One of many distinguishing options of Kai is its model-agnostic method. Not like different options, Kai doesn’t bundle a particular massive language mannequin (LLM). As a substitute, it extends Konveyor to work together with completely different LLMs, offering the pliability to make use of the best-suited mannequin for every particular migration context. This method ensures that organizations can optimize their migration technique with out locking right into a single know-how.
Sensible Use Circumstances: From Java EE to Quarkus
The brief demonstration exhibits how Konveyor AI facilitates migration from legacy frameworks like Java EE to fashionable options like Quarkus. The method begins with a static code evaluation utilizing the Konveyor CLI, which identifies migration points inside the code base. After the evaluation is full, Konveyor AI steps in to develop patches for recognized points, leveraging LLMs to counsel the proper code modifications.
For instance, older Java EE annotations have been seamlessly changed with their fashionable Jakarta EE counterparts, and JMS-based message-driven beans have been transformed to JakartaEE annotations, and EJBs Comparable legacy applied sciences have developed into REST endpoints. These modifications may be validated by a developer inside an built-in growth atmosphere (IDE), demonstrating how Konveyor AI integrates into present developer workflows. Konveyor AI does this by utilizing the KAI extension.
Take a look at this episode of OpenShift Commons for an in depth demo and presentation the place Ramón and I’m going via the main points.
Why ought to I exploit Konveyor AI's or how? The facility of “Moss” lies in its capability to automate repetitive migration duties, leveraging data accrued from earlier migrations. By specializing in code modifications reasonably than architectural modifications, Kai gives a strong device for effectively modernizing functions. It empowers builders to make knowledgeable choices, enabling seamless migration to fashionable frameworks and cloud-native environments.
Go to the directions right here to attempt the newest construct.
Be a part of the Conveyor neighborhood via their mailing lists, Slack channels, and common meetups.
