Improving Code Localization with Repository Memory

2025-10-01

AI Summary

Existing code localization methods ignore developers’ long-term memory of codebases—such as module functionality and defect–fix location associations—forcing task-specific reasoning from scratch. Method: We propose a non-parametric, repository-level long-term memory mechanism grounded in commit history. By analyzing historical commits, issue–commit links, and active module summaries, it automatically discovers code evolution patterns and defect distribution regularities, constructing a retrievable external memory store. This memory is integrated into the LocAgent framework for the first time, enabling memory-augmented localization reasoning. Contribution/Results: Our approach achieves significant improvements over state-of-the-art methods on both SWE-bench-verified and SWE-bench-live benchmarks, empirically demonstrating that long-term memory is critical for enhancing performance in complex software engineering tasks.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Search and Optimization: Local Search
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Code localization is a fundamental challenge in repository-level software engineering tasks such as bug fixing. While existing methods equip language agents with comprehensive tools/interfaces to fetch information from the repository, they overlook the critical aspect of memory, where each instance is typically handled from scratch assuming no prior repository knowledge. In contrast, human developers naturally build long-term repository memory, such as the functionality of key modules and associations between various bug types and their likely fix locations. In this work, we augment language agents with such memory by leveraging a repository's commit history - a rich yet underutilized resource that chronicles the codebase's evolution. We introduce tools that allow the agent to retrieve from a non-parametric memory encompassing recent historical commits and linked issues, as well as functionality summaries of actively evolving parts of the codebase identified via commit patterns. We demonstrate that augmenting such a memory can significantly improve LocAgent, a state-of-the-art localization framework, on both SWE-bench-verified and the more recent SWE-bench-live benchmarks. Our research contributes towards developing agents that can accumulate and leverage past experience for long-horizon tasks, more closely emulating the expertise of human developers.

Problem

• Enhancing code localization using repository commit history memory
• Addressing the lack of prior knowledge in repository-level software tasks
• Improving bug fixing by leveraging historical commits and issue links

Innovation

• Leverages repository commit history for memory
• Retrieves from non-parametric memory of commits
• Uses commit patterns to identify active code parts