EVOR: Evolving Retrieval for Code Generation

TL;DR

EVOR introduces a dynamic retrieval pipeline that evolves both queries and knowledge bases in real-time, significantly improving code generation accuracy over static methods by leveraging diverse, updated information sources.

Contribution

The paper presents EVOR, a novel approach that synchronously evolves queries and knowledge bases, enhancing retrieval-augmented code generation beyond static knowledge approaches.

Findings

EVOR achieves 2-4x higher accuracy than existing methods.

Synchronous evolution of queries and knowledge bases improves performance.

Diverse information sources contribute to better code generation results.

Abstract

Recently the retrieval-augmented generation (RAG) has been successfully applied in code generation. However, existing pipelines for retrieval-augmented code generation (RACG) employ static knowledge bases with a single source, limiting the adaptation capabilities of Large Language Models (LLMs) to domains they have insufficient knowledge of. In this work, we develop a novel pipeline, EVOR, that employs the synchronous evolution of both queries and diverse knowledge bases. On two realistic settings where the external knowledge is required to solve code generation tasks, we compile four new datasets associated with frequently updated libraries and long-tail programming languages, named EVOR-BENCH. Extensive experiments demonstrate that EVOR achieves two to four times of execution accuracy compared to other methods such as Reflexion (Shinn et al., 2024), DocPrompting (Zhou et al., 2023), etc. We demonstrate that EVOR is flexible and can be easily combined with them to achieve further improvement. Further analysis reveals that EVOR benefits from the synchronous evolution of queries and documents and the diverse information sources in the knowledge base. We hope that our studies will inspire more insights into the design of advanced RACG pipelines in future research. Our model, code, and data are available at https://arks-codegen.github.io.