Code Reviewer Recommendation Based on a Hypergraph with Multiplex Relationships

TL;DR

The paper tackles the problem of assigning code reviewers efficiently amid large volumes of pull requests by introducing MIRRec, a hypergraph-based method that captures high-order, multiplex interactions between developers and PRs across multiple roles. MIRRec constructs a multiplex-relationship hypergraph with six PR–developer relations and PR similarity, then applies hypergraph learning and a ranking framework to produce reviewer scores for new PRs. Key contributions include the formalization of a multiplex hypergraph with time-aware, role-specific edges, a ranking-based objective that leverages high-order connectivity, and extensive evaluation on ten OSS projects showing superior ACC and MRR compared to state-of-the-art baselines, along with ablation analyses that highlight the importance of PR-Reviewers and PR-Committers relations. The work demonstrates that incorporating diverse interaction types and high-order relationships can significantly improve reviewer recommendations, enabling faster, more accurate code reviews in large-scale open-source ecosystems.

Abstract

Code review is an essential component of software development, playing a vital role in ensuring a comprehensive check of code changes. However, the continuous influx of pull requests and the limited pool of available reviewer candidates pose a significant challenge to the review process, making the task of assigning suitable reviewers to each review request increasingly difficult. To tackle this issue, we present MIRRec, a novel code reviewer recommendation method that leverages a hypergraph with multiplex relationships. MIRRec encodes high-order correlations that go beyond traditional pairwise connections using degree-free hyperedges among pull requests and developers. This way, it can capture high-order implicit connectivity and identify potential reviewers. To validate the effectiveness of MIRRec, we conducted experiments using a dataset comprising 48,374 pull requests from ten popular open-source software projects hosted on GitHub. The experiment results demonstrate that MIRRec, especially without PR-Review Commenters relationship, outperforms existing stateof-the-art code reviewer recommendation methods in terms of ACC and MRR, highlighting its significance in improving the code review process.

Figures (4)

• Figure 1: Multiplex interactions of developers in PR-based review process
• Figure 2: Overview of approach
• Figure 3: ACC and MRR performance of MIRRec under different $\mu$
• Figure 4: Training time and ACC performance of MIRRec under different K