arXiv subjects

Xunjin Zheng

Publications and source records attributed to Xunjin Zheng.

6 recordsLinked to original sources

arXiv2026

Principles and Practices of Large-Scale Code Analysis at Ant Group: A Data- and Logic-Oriented Approach

Large-scale software development requires dynamic and multifaceted static code analysis that extends beyond the capabilities of traditional tools. Existing tools like CodeQL lack cross-language analysis capabilities and can be time-consuming and resource-intensive. We present CodeFuse-Query, a data system tailored for large-scale code analysis. First, CodeFuse-Query adopts a Logic-Oriented Computation Design, employing Datalog with a two-tiered schema, COREF, to convert source code into data facts, and Godel to express complex analysis tasks in logical terms. Furthermore, CodeFuse-Query adopts a Domain-Optimized System Design. This approach optimizes resource utilization, prioritizes data reusability, applies incremental code extraction, and introduces task-type characteristics specifically for code changes, underscoring its domain-optimized design. We present empirical results demonstrating CodeFuse-Query's robustness, scalability, and efficiency in large-scale real-world scenarios at Ant Group, where it serves as a core static analysis infrastructure. Deployed in production environments, CodeFuse-Query processes up to 10 billion lines of code daily across more than 300,000 distinct analysis tasks. CodeFuse-Query has been open-sourced.

cs.SE↗

arXiv2025

CodeFuse-CR-Bench: A Comprehensiveness-aware Benchmark for End-to-End Code Review Evaluation in Python Projects

Automated code review (CR) is a key application for Large Language Models (LLMs), but progress is hampered by a "reality gap": existing benchmarks evaluate models on isolated sub-tasks using simplified, context-poor data. This fails to reflect the holistic context-rich nature of real-world CR. To bridge this gap, we introduce CodeFuse-CR-Bench, the first comprehensiveness-aware benchmark for repository-level CR evaluation. CodeFuse-CR-Bench comprises 601 high-quality instances from 70 Python projects covering nine Pull-Request (PR) problem domains, where each instance provides rich, multi-faceted context including the associated issue, PR details, and repository state, enabling end-to-end evaluation. Beyond superficial metrics, we also propose a novel evaluation framework that combines rule-based checks for location and syntax with model-based judgments of review quality. We present the first large-scale assessment of state-of-the-art LLMs on this comprehensive CR task. Our results establish crucial baselines and reveal that (1) no single LLM dominates all aspects of CR; (2) Gemini 2.5 Pro achieves the highest comprehensive performance; and (3) different LLMs exhibit varying robustness to redundant context. These findings highlight the necessity of holistic, multi-dimensional evaluation and provide actionable insights for advancing truly intelligent yet practical CR assistants.

cs.SE↗

arXiv2025

Every Sample Matters: Leveraging Mixture-of-Experts and High-Quality Data for Efficient and Accurate Code LLM

Recent advancements in code large language models (LLMs) have demonstrated remarkable capabilities in code generation and understanding. It is still challenging to build a code LLM with comprehensive performance yet ultimate efficiency. Many attempts have been released in the open source community to break the trade-off between performance and efficiency, such as the Qwen Coder series and the DeepSeek Coder series. This paper introduces yet another attempt in this area, namely Ling-Coder-Lite. We leverage the efficient Mixture-of-Experts (MoE) architecture along with a set of high-quality data curation methods (especially those based on program analytics) to build an efficient yet powerful code LLM. Ling-Coder-Lite exhibits on-par performance on 12 representative coding benchmarks compared to state-of-the-art models of similar size, such as Qwen2.5-Coder-7B and DeepSeek-Coder-V2-Lite, while offering competitive latency and throughput. In practice, we achieve a 50\% reduction in deployment resources compared to the similar-sized dense model without performance loss. To facilitate further research and development in this area, we open-source our models as well as a substantial portion of high-quality data for the annealing and post-training stages. The models and data can be accessed at~\url{https://huggingface.co/inclusionAI/Ling-Coder-lite}.

cs.LG↗

arXiv2024

REPOFUSE: Repository-Level Code Completion with Fused Dual Context

The success of language models in code assistance has spurred the proposal of repository-level code completion as a means to enhance prediction accuracy, utilizing the context from the entire codebase. However, this amplified context can inadvertently increase inference latency, potentially undermining the developer experience and deterring tool adoption - a challenge we termed the Context-Latency Conundrum. This paper introduces REPOFUSE, a pioneering solution designed to enhance repository-level code completion without the latency trade-off. REPOFUSE uniquely fuses two types of context: the analogy context, rooted in code analogies, and the rationale context, which encompasses in-depth semantic relationships. We propose a novel rank truncated generation (RTG) technique that efficiently condenses these contexts into prompts with restricted size. This enables REPOFUSE to deliver precise code completions while maintaining inference efficiency. Through testing with the CrossCodeEval suite, REPOFUSE has demonstrated a significant leap over existing models, achieving a 40.90% to 59.75% increase in exact match (EM) accuracy for code completions and a 26.8% enhancement in inference speed. Beyond experimental validation, REPOFUSE has been integrated into the workflow of a large enterprise, where it actively supports various coding tasks.

cs.SE↗

arXiv2024

CodeFuse-13B: A Pretrained Multi-lingual Code Large Language Model

Code Large Language Models (Code LLMs) have gained significant attention in the industry due to their wide applications in the full lifecycle of software engineering. However, the effectiveness of existing models in understanding non-English inputs for multi-lingual code-related tasks is still far from well studied. This paper introduces CodeFuse-13B, an open-sourced pre-trained code LLM. It is specifically designed for code-related tasks with both English and Chinese prompts and supports over 40 programming languages. CodeFuse achieves its effectiveness by utilizing a high quality pre-training dataset that is carefully filtered by program analyzers and optimized during the training process. Extensive experiments are conducted using real-world usage scenarios, the industry-standard benchmark HumanEval-x, and the specially designed CodeFuseEval for Chinese prompts. To assess the effectiveness of CodeFuse, we actively collected valuable human feedback from the AntGroup's software development process where CodeFuse has been successfully deployed. The results demonstrate that CodeFuse-13B achieves a HumanEval pass@1 score of 37.10%, positioning it as one of the top multi-lingual code LLMs with similar parameter sizes. In practical scenarios, such as code generation, code translation, code comments, and testcase generation, CodeFuse performs better than other models when confronted with Chinese prompts.

cs.SE↗

arXiv2023

Static Code Analysis in the AI Era: An In-depth Exploration of the Concept, Function, and Potential of Intelligent Code Analysis Agents

The escalating complexity of software systems and accelerating development cycles pose a significant challenge in managing code errors and implementing business logic. Traditional techniques, while cornerstone for software quality assurance, exhibit limitations in handling intricate business logic and extensive codebases. To address these challenges, we introduce the Intelligent Code Analysis Agent (ICAA), a novel concept combining AI models, engineering process designs, and traditional non-AI components. The ICAA employs the capabilities of large language models (LLMs) such as GPT-3 or GPT-4 to automatically detect and diagnose code errors and business logic inconsistencies. In our exploration of this concept, we observed a substantial improvement in bug detection accuracy, reducing the false-positive rate to 66\% from the baseline's 85\%, and a promising recall rate of 60.8\%. However, the token consumption cost associated with LLMs, particularly the average cost for analyzing each line of code, remains a significant consideration for widespread adoption. Despite this challenge, our findings suggest that the ICAA holds considerable potential to revolutionize software quality assurance, significantly enhancing the efficiency and accuracy of bug detection in the software development process. We hope this pioneering work will inspire further research and innovation in this field, focusing on refining the ICAA concept and exploring ways to mitigate the associated costs.

cs.SE↗