English

DocPrompting: Generating Code by Retrieving the Docs

Computation and Language 2023-02-21 v3 Artificial Intelligence Software Engineering

Abstract

Publicly available source-code libraries are continuously growing and changing. This makes it impossible for models of code to keep current with all available APIs by simply training these models on existing code repositories. Thus, existing models inherently cannot generalize to using unseen functions and libraries, because these would never appear in the training data. In contrast, when human programmers use functions and libraries for the first time, they frequently refer to textual resources such as code manuals and documentation, to explore and understand the available functionality. Inspired by this observation, we introduce DocPrompting: a natural-language-to-code generation approach that explicitly leverages documentation by (1) retrieving the relevant documentation pieces given an NL intent, and (2) generating code based on the NL intent and the retrieved documentation. DocPrompting is general: it can be applied to any programming language and is agnostic to the underlying neural model. We demonstrate that DocPrompting consistently improves NL-to-code models: DocPrompting improves strong base models such as CodeT5 by 2.85% in pass@1 (52% relative gain) and 4.39% in pass@10 (30% relative gain) in execution-based evaluation on the popular Python CoNaLa benchmark; on a new Bash dataset tldr, DocPrompting improves CodeT5 and GPT-Neo1.3B by up to absolute 6.9% exact match.

Keywords

Cite

@article{arxiv.2207.05987,
  title  = {DocPrompting: Generating Code by Retrieving the Docs},
  author = {Shuyan Zhou and Uri Alon and Frank F. Xu and Zhiruo Wang and Zhengbao Jiang and Graham Neubig},
  journal= {arXiv preprint arXiv:2207.05987},
  year   = {2023}
}

Comments

ICLR 2023 (notable-top-25%); code and data are available at https://github.com/shuyanzhou/docprompting