SkCoder: A Sketch-based Approach for Automatic Code Generation

TL;DR

SkCoder introduces a sketch-based method for automatic code generation that retrieves, extracts, and edits code sketches to produce more accurate programs, outperforming existing models on multiple datasets.

Contribution

The paper presents SkCoder, a novel approach that leverages code sketches for improved accuracy in code generation, addressing limitations of previous copy-based methods.

Findings

SkCoder outperforms state-of-the-art models by over 30% on three datasets.

The approach enhances multiple code generation models, increasing Pass@1 by up to 120%.

Manual evaluation confirms SkCoder's superior code quality.

Abstract

Recently, deep learning techniques have shown great success in automatic code generation. Inspired by the code reuse, some researchers propose copy-based approaches that can copy the content from similar code snippets to obtain better performance. Practically, human developers recognize the content in the similar code that is relevant to their needs, which can be viewed as a code sketch. The sketch is further edited to the desired code. However, existing copy-based approaches ignore the code sketches and tend to repeat the similar code without necessary modifications, which leads to generating wrong results. In this paper, we propose a sketch-based code generation approach named SkCoder to mimic developers' code reuse behavior. Given a natural language requirement, SkCoder retrieves a similar code snippet, extracts relevant parts as a code sketch, and edits the sketch into the desired code. Our motivations are that the extracted sketch provides a well-formed pattern for telling models "how to write". The post-editing further adds requirement-specific details to the sketch and outputs the complete code. We conduct experiments on two public datasets and a new dataset collected by this work. We compare our approach to 20 baselines using 5 widely used metrics. Experimental results show that (1) SkCoder can generate more correct programs, and outperforms the state-of-the-art - CodeT5-base by 30.30%, 35.39%, and 29.62% on three datasets. (2) Our approach is effective to multiple code generation models and improves them by up to 120.1% in Pass@1. (3) We investigate three plausible code sketches and discuss the importance of sketches. (4) We manually evaluate the generated code and prove the superiority of our SkCoder in three aspects.