RLCoder: Reinforcement Learning for Repository-Level Code Completion

Repository-level code completion aims to generate code for unfinished code snippets within the context of a specified repository. Existing approaches mainly rely on retrievalaugmented generation strategies due to limitations in input sequence length. However, traditional lexical-based retrieval meth...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Proceedings / International Conference on Software Engineering s. 1140 - 1152
Hlavní autori: Wang, Yanlin, Wang, Yanli, Guo, Daya, Chen, Jiachi, Zhang, Ruikai, Ma, Yuchi, Zheng, Zibin
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 26.04.2025
Predmet:
Codes
Data models
Faces
Iterative methods
Perplexity
Programming
Reinforcement learning
Repository-Level Code Completion
Semantics
Software engineering
Stop Signal Mechanism
Training
Windows
ISSN:1558-1225
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Repository-level code completion aims to generate code for unfinished code snippets within the context of a specified repository. Existing approaches mainly rely on retrievalaugmented generation strategies due to limitations in input sequence length. However, traditional lexical-based retrieval methods like BM25 struggle to capture code semantics, while model-based retrieval methods face challenges due to the lack of labeled data for training. Therefore, we propose RLCoder, a novel reinforcement learning framework, which can enable the retriever to learn to retrieve useful content for code completion without the need for labeled data. Specifically, we iteratively evaluate the usefulness of retrieved content based on the perplexity of the target code when provided with the retrieved content as additional context, and provide feedback to update the retriever parameters. This iterative process enables the retriever to learn from its successes and failures, gradually improving its ability to retrieve relevant and high-quality content. Considering that not all situations require information beyond code files and not all retrieved context is helpful for generation, we also introduce a stop signal mechanism, allowing the retriever to decide when to retrieve and which candidates to retain autonomously. Extensive experimental results demonstrate that RLCoder consistently outperforms state-of-the-art methods on CrossCodeEval and RepoEval, achieving 12.2% EM improvement over previous methods. Moreover, experiments show that our framework can generalize across different programming languages and further improve previous methods like RepoCoder. We provide the code and data at https://github.com/DeepSoftwareAnalytics/RLCoder.
ISSN:1558-1225
DOI:10.1109/ICSE55347.2025.00014