SweRank: Software Issue Localization with Code Ranking

2025-05-07

AI Summary

Precise localization of software bugs requires mapping verbose, ambiguous natural language bug reports to specific code locations; however, existing approaches suffer from high latency, prohibitive cost (relying on proprietary large language models), or poor retrieval adaptability (conventional code ranking models are not optimized for bug localization). Method: We propose a lightweight, efficient retrieve-then-rerank framework featuring: (i) the first two-stage paradigm explicitly optimized for bug localization; (ii) SweLoc—the first large-scale, real-world GitHub issue-patch dataset; and (iii) an open-model stack comprising the Contriever dense retriever and a fine-tuned cross-encoder reranker. Contribution/Results: Our method achieves state-of-the-art performance on SWE-Bench-Lite and LocBench, significantly outperforming LLM-based agents (e.g., Claude-3.5) and mainstream code retrieval models—while relying solely on open-source models and incurring substantially lower computational overhead.

Technology Category

• Search and Optimization: Local Search
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

Software issue localization, the task of identifying the precise code locations (files, classes, or functions) relevant to a natural language issue description (e.g., bug report, feature request), is a critical yet time-consuming aspect of software development. While recent LLM-based agentic approaches demonstrate promise, they often incur significant latency and cost due to complex multi-step reasoning and relying on closed-source LLMs. Alternatively, traditional code ranking models, typically optimized for query-to-code or code-to-code retrieval, struggle with the verbose and failure-descriptive nature of issue localization queries. To bridge this gap, we introduce SweRank, an efficient and effective retrieve-and-rerank framework for software issue localization. To facilitate training, we construct SweLoc, a large-scale dataset curated from public GitHub repositories, featuring real-world issue descriptions paired with corresponding code modifications. Empirical results on SWE-Bench-Lite and LocBench show that SweRank achieves state-of-the-art performance, outperforming both prior ranking models and costly agent-based systems using closed-source LLMs like Claude-3.5. Further, we demonstrate SweLoc's utility in enhancing various existing retriever and reranker models for issue localization, establishing the dataset as a valuable resource for the community.

Problem

• Identifying code locations from issue descriptions efficiently
• Reducing latency and cost of LLM-based localization methods
• Improving accuracy of traditional code ranking models for verbose queries

Innovation

• SweRank uses retrieve-and-rerank framework
• Constructs SweLoc dataset from GitHub
• Outperforms prior models and agent-based systems