CoChat
Sign Up
Cited by 18Software Engineering ResearchTopic ModelingInformation Retrieval and Search Behavior

cAST: Enhancing Code Retrieval-Augmented Generation with Structural Chunking via Abstract Syntax Tree

Open access
Zhang, Yilin, Zhao, Xinran, Wang, Zora Zhiruo, Yang, Chenyang, Wei, Jiayi, Wu, Tongshuang·2025·Findings of the Association for Computational Linguistics: EMNLP 2025
DOI 10.18653/v1/2025.findings-emnlp.430
Open this paper in CoChat — free accountOpen PDF

Abstract

Retrieval-Augmented Generation (RAG) has become essential for large-scale code generation, grounding predictions in external code corpora to improve factuality.However, a critical yet underexplored aspect of RAG pipelines is chunking-the process of dividing documents into retrievable units.Existing linebased chunking heuristics often break semantic structures, splitting functions or merging unrelated code, which can degrade generation quality.We propose chunking via Abstract Syntax Trees (CAST), a structure-aware method that recursively breaks large AST nodes into smaller chunks and merges sibling nodes while respecting size limits.This approach generates selfcontained, semantically coherent units across programming languages and tasks, improving performance on diverse code generation tasks, e.g., boosting Recall@5 by 4.3 points on Re-poEval retrieval and Pass@1 by 2.67 points on SWE-bench generation.Our work highlights the importance of structure-aware chunking for scaling retrieval-enhanced code intelligence.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.18653/v1/2025.findings-emnlp.430

Cites

0 citation errors · 4 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • LongCodeZip consistently outperforms advanced RAG methods (A3-CodGen, cAST, RepoGenix, RLCoder) on Long Code Completion across SeedCoder and Claude-3.7-Sonnet.

    “As shown in Table VI, LongCodeZip consistently outperforms these advanced RAG methods across the most competitive open-source and closed-source models, SeedCoder and Claude-3.7-Sonnet.”

    From LongCodeZip: Compress Long Context for Code Language Models (2025) · cites this paper as [40]

    The Metrics section identifies cAST’s generation models, including Claude-3.7-Sonnet, and Table 7 reports cAST results on SWE-bench. Our copy does not report a LongCodeZip comparison with cAST or the other named methods across SeedCoder and Claude-3.7-Sonnet (searched for LongCodeZip and SeedCoder). (LLM, conf=0.76)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

The publisher hasn’t deposited this paper’s reference list.

Research works printed but not linked: 11

  1. Daya Guo, Shuai Lu, Nan Duan, Yanlin Wang, Ming Zhou, and Jian Yin. 2022. UniXcoder: Unified cross- modal pre-training for code representation. In Pro- ceedings of the 60th Annual Meeting of the Associ- ation for Computational Linguistics (ACL), pages 7212-7225.
  2. PMLR.
  3. Carlos E Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, and Karthik R Narasimhan. 2024. SWE-bench: Can language mod- els resolve real-world github issues? In The Twelfth International Conference on Learning Representa- tions.
  4. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Hein- rich Küttler, Mike Lewis, Wen-tau Yih, Tim Rock- täschel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-augmented generation for knowledge- intensive NLP tasks. In Advances in Neural Infor- mation Processing Systems (NeurIPS), pages 9459- 9474.
  5. Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Köpf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Te- jani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, and 2 others. 2019. Pytorch: An imperative style, high-performance deep learning library. In Advances 8112 in Neural Information Processing Systems 32: An- nual Conference on Neural Information Processing Systems 2019, NeurIPS 2019, December 8-14, 2019, Vancouver, BC, Canada, pages 8024-8035.
  6. Parth Sarthi, Salman Abdullah, Aditi Tuli, Shubh Khanna, Anna Goldie, and Christopher D. Manning. 2024. Raptor: Recursive abstractive processing for tree-organized retrieval. In International Conference on Learning Representations (ICLR).
  7. Yewei Song, Cedric Lothritz, Xunzhu Tang, Tegawendé Bissyandé, and Jacques Klein. 2024. Revisiting code similarity evaluation with abstract syntax tree edit distance. In Proceedings of the 62nd Annual Meet- ing of the Association for Computational Linguistics (Volume 2: Short Papers), pages 38-46, Bangkok, Thailand. Association for Computational Linguistics. · DOI
  8. Jiayi Wei, Greg Durrett, and Isil Dillig. 2023. TypeT5: Seq2seq type inference using static analysis. In In- ternational Conference on Learning Representations (ICLR).
  9. Shitao Xiao, Zheng Liu, Peitian Zhang, and Niklas Muennighoff. 2023. C-pack: Packaged resources to advance general chinese embedding. arXiv.
  10. Fengji Zhang, Bei Chen, Yue Zhang, Jacky Keung, Jin Liu, Daoguang Zan, Yi Mao, Jian-Guang Lou, and Weizhu Chen. 2023b. RepoCoder: Repository-level code completion through iterative retrieval and gen- eration. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 2471-2484, Singapore. Association for Com- putational Linguistics. · DOI · Paper page
  11. Daniel Zügner, Tobias Kirschstein, Michele Catasta, Jure Leskovec, and Stephan Günnemann. 2021. Language-agnostic representation learning of source code from structure and context. In International Conference on Learning Representations.

Cited by

Sorted by citations

Related

Authors
Zhang, Y., Zhao, X., Wang, Z. Z., Yang, C., Wei, J., & Wu, T.
Affiliations
Carnegie Mellon University (US)
ORCID
Venue
Findings of the Association for Computational Linguistics: EMNLP 2025
Publisher
Association for Computational Linguistics
Published
2025
Pages
8106-8116
DOI
10.18653/v1/2025.findings-emnlp.430
OpenAlex
W4416034875
Cited by
11
References
0
Citations by year
2026: 9 · 2025: 2
FWCI
12.58
Type
conference paper
Language
en
Open access
Licence
cc-by
Locations
  • Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) (publishedVersion, cc-by)
  • arXiv (Cornell University) (submittedVersion, cc-by-sa)
  • arXiv (Cornell University)
Funders
Office of Naval Research
Topics
  • Physical Sciences › Computer Science › Information Systems › Software Engineering Research
  • Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
  • Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
Keywords
retrieval-augmented generation, abstract syntax tree, code generation, code search, large language models, semantic chunking, SWE-bench

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review