Do Not Treat Code as Natural Language: Implications for Repository-Level Code Generation and Beyond

TL;DR

This paper introduces Hydra, a structure-aware, dependency-focused retrieval framework for repository-level code generation that significantly improves performance over existing methods by preserving code structure and dependencies.

Contribution

Hydra presents a novel hierarchy-based indexing and dependency-aware retrieval approach tailored for repository-level code generation, addressing limitations of traditional NLP-inspired methods.

Findings

Hydra achieves over 5% improvement in Pass@1 on DevEval and RepoExec benchmarks.

Hydra enables smaller models to match or outperform larger models with existing retrieval methods.

The approach significantly enhances code generation quality by preserving structural dependencies.

Abstract

Large language models for code (CodeLLMs) have demonstrated remarkable success in standalone code completion and generation, sometimes even surpassing human performance, yet their effectiveness diminishes in repository-level settings where cross-file dependencies and structural context are essential. Existing Retrieval-Augmented Generation (RAG) approaches often borrow strategies from NLP, relying on chunking-based indexing and similarity-based retrieval. Chunking results in the loss of coherence between code units and overlooks structural relationships, while similarity-driven methods frequently miss functionally relevant dependencies such as helper functions, classes, or global variables. To address these limitations, we present Hydra, a repository-level code generation framework that treats code as structured code rather than natural language. Our approach introduces (i) a structure-aware indexing strategy that represents repositories as hierarchical trees of functions, classes, and variables, preserving code structure and dependencies, (ii) a lightweight dependency-aware retriever (DAR) that explicitly identifies and retrieves the true dependencies required by a target function, and (iii) a hybrid retrieval mechanism that combines DAR with similarity-based retrieval to provide both essential building blocks and practical usage examples. Extensive experiments on the challenging DevEval and RepoExec benchmarks, both requiring function implementation from real-world repositories with complex large repository context, show that Hydra achieves state-of-the-art performance across open- and closed-source CodeLLMs. Notably, our method establishes a new state of the art in repository-level code generation, surpassing strongest baseline by over 5% in Pass@1 and even enabling smaller models to match or exceed the performance of much larger ones that rely on existing retrievers.