RPG: A Repository Planning Graph for Unified and Scalable Codebase Generation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908830795825152 |
|---|---|
| author | Luo, Jane Zhang, Xin Liu, Steven Wu, Jie Liu, Jianfeng Huang, Yiming Huang, Yangyu Yin, Chengyu Xin, Ying Zhan, Yuefeng Sun, Hao Chen, Qi Li, Scarlett Yang, Mao |
| author_facet | Luo, Jane Zhang, Xin Liu, Steven Wu, Jie Liu, Jianfeng Huang, Yiming Huang, Yangyu Yin, Chengyu Xin, Ying Zhan, Yuefeng Sun, Hao Chen, Qi Li, Scarlett Yang, Mao |
| contents | Large language models excel at generating individual functions or single files of code, yet generating complete repositories from scratch remains a fundamental challenge. This capability is key to building coherent software systems from high-level specifications and realizing the full potential of automated code generation. The process requires planning at two levels: deciding what features and modules to build (proposal stage) and defining their implementation details (implementation stage). Current approaches rely on natural language planning, which often produces unclear specifications, misaligned components, and brittle designs due to its inherent ambiguity and lack of structure. To address these limitations, we introduce the Repository Planning Graph (RPG), a structured representation that encodes capabilities, file structures, data flows, and functions in a unified graph. By replacing free-form natural language with an explicit blueprint, RPG enables consistent long-horizon planning for repository generation. Building on RPG, we develop ZeroRepo, a graph-driven framework that operates in three stages: proposal-level planning, implementation-level construction, and graph-guided code generation with test validation. To evaluate, we construct RepoCraft, a benchmark of six real-world projects with 1,052 tasks. On RepoCraft, ZeroRepo produces nearly 36K Code Lines and 445K Code Tokens, on average 3.9$\times$ larger than the strongest baseline (Claude Code), and 68$\times$ larger than other baselines. It achieves 81.5% coverage and 69.7% test accuracy, improving over Claude Code by 27.3 and 35.8 points. Further analysis shows that RPG models complex dependencies, enables more sophisticated planning through near-linear scaling, and improves agent understanding of repositories, thus accelerating localization. Our data and code are available at https://github.com/microsoft/RPG-ZeroRepo. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_16198 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RPG: A Repository Planning Graph for Unified and Scalable Codebase Generation Luo, Jane Zhang, Xin Liu, Steven Wu, Jie Liu, Jianfeng Huang, Yiming Huang, Yangyu Yin, Chengyu Xin, Ying Zhan, Yuefeng Sun, Hao Chen, Qi Li, Scarlett Yang, Mao Computation and Language Artificial Intelligence Software Engineering Large language models excel at generating individual functions or single files of code, yet generating complete repositories from scratch remains a fundamental challenge. This capability is key to building coherent software systems from high-level specifications and realizing the full potential of automated code generation. The process requires planning at two levels: deciding what features and modules to build (proposal stage) and defining their implementation details (implementation stage). Current approaches rely on natural language planning, which often produces unclear specifications, misaligned components, and brittle designs due to its inherent ambiguity and lack of structure. To address these limitations, we introduce the Repository Planning Graph (RPG), a structured representation that encodes capabilities, file structures, data flows, and functions in a unified graph. By replacing free-form natural language with an explicit blueprint, RPG enables consistent long-horizon planning for repository generation. Building on RPG, we develop ZeroRepo, a graph-driven framework that operates in three stages: proposal-level planning, implementation-level construction, and graph-guided code generation with test validation. To evaluate, we construct RepoCraft, a benchmark of six real-world projects with 1,052 tasks. On RepoCraft, ZeroRepo produces nearly 36K Code Lines and 445K Code Tokens, on average 3.9$\times$ larger than the strongest baseline (Claude Code), and 68$\times$ larger than other baselines. It achieves 81.5% coverage and 69.7% test accuracy, improving over Claude Code by 27.3 and 35.8 points. Further analysis shows that RPG models complex dependencies, enables more sophisticated planning through near-linear scaling, and improves agent understanding of repositories, thus accelerating localization. Our data and code are available at https://github.com/microsoft/RPG-ZeroRepo. |
| title | RPG: A Repository Planning Graph for Unified and Scalable Codebase Generation |
| topic | Computation and Language Artificial Intelligence Software Engineering |
| url | https://arxiv.org/abs/2509.16198 |