What type of study is this?

This is a Quantitative Study study.

October 3, 2025Open Access

CodeRAG: Finding Relevant and Necessary Knowledge for Retrieval-Augmented Repository-Level Code Completion

Puntos clave

CodeRAG significantly improves repository-level code completion by refining code retrieval and query processes.
Benchmark tests on ReccEval and CCEval show that CodeRAG outperforms existing state-of-the-art methods in 95% of cases.
The framework incorporates multi-path code retrieval and log probability guided query construction for efficient results.
This approach highlights the importance of aligning code retrieval methods with code LLMs to enhance performance.

Resumen

Repository-level code completion automatically predicts the unfinished code based on the broader information from the repository. Recent strides in Code Large Language Models (code LLMs) have spurred the development of repository-level code completion methods, yielding promising results. Nevertheless, they suffer from issues such as inappropriate query construction, single-path code retrieval, and misalignment between code retriever and code LLM. To address these problems, we introduce CodeRAG, a framework tailored to identify relevant and necessary knowledge for retrieval-augmented repository-level code completion. Its core components include log probability guided query construction, multi-path code retrieval, and preference-aligned BestFit reranking. Extensive experiments on benchmarks ReccEval and CCEval demonstrate that CodeRAG significantly and consistently outperforms state-of-the-art methods. The implementation of CodeRAG is available at https://github.com/KDEGroup/CodeRAG.

Leer artículo completoexternamente

Me gusta

Guardar

Ver artículo completo