Kajal: Extracting Grammar of a Source Code Using Large Language Models

TL;DR

Kajal leverages large language models with prompt engineering to automatically infer domain-specific language grammars from code snippets, significantly reducing manual effort and improving accuracy.

Contribution

This paper introduces Kajal, a novel method that uses LLMs and few-shot learning to automatically extract DSL grammars from code, enhancing automation in software engineering tasks.

Findings

Kajal achieves 60% accuracy with few-shot learning.

Accuracy drops to 45% without few-shot learning.

Few-shot learning significantly improves grammar extraction performance.

Abstract

Understanding and extracting the grammar of a domain-specific language (DSL) is crucial for various software engineering tasks; however, manually creating these grammars is time-intensive and error-prone. This paper presents Kajal, a novel approach that automatically infers grammar from DSL code snippets by leveraging Large Language Models (LLMs) through prompt engineering and few-shot learning. Kajal dynamically constructs input prompts, using contextual information to guide the LLM in generating the corresponding grammars, which are iteratively refined through a feedback-driven approach. Our experiments show that Kajal achieves 60% accuracy with few-shot learning and 45% without it, demonstrating the significant impact of few-shot learning on the tool's effectiveness. This approach offers a promising solution for automating DSL grammar extraction, and future work will explore using smaller, open-source LLMs and testing on larger datasets to further validate Kajal's performance.