Learning to Represent Patches

TL;DR

Patcherizer introduces a comprehensive patch representation method combining code context, semantic intent, and structural modifications, leveraging graph neural networks and transformers to improve patch understanding and related tasks.

Contribution

It presents a novel holistic patch representation approach that captures semantic and structural intentions, outperforming existing methods in multiple software engineering tasks.

Findings

Outperforms state-of-the-art in patch description generation

Achieves significant improvements in BLEU, ROUGE-L, and METEOR scores

Effective across patch description, accuracy prediction, and intention identification

Abstract

Patch representation is crucial in automating various software engineering tasks, like determining patch accuracy or summarizing code changes. While recent research has employed deep learning for patch representation, focusing on token sequences or Abstract Syntax Trees (ASTs), they often miss the change's semantic intent and the context of modified lines. To bridge this gap, we introduce a novel method, Patcherizer. It delves into the intentions of context and structure, merging the surrounding code context with two innovative representations. These capture the intention in code changes and the intention in AST structural modifications pre and post-patch. This holistic representation aptly captures a patch's underlying intentions. Patcherizer employs graph convolutional neural networks for structural intention graph representation and transformers for intention sequence representation. We evaluated Patcherizer's embeddings' versatility in three areas: (1) Patch description generation, (2) Patch accuracy prediction, and (3) Patch intention identification. Our experiments demonstrate the representation's efficacy across all tasks, outperforming state-of-the-art methods. For example, in patch description generation, Patcherizer excels, showing an average boost of 19.39% in BLEU, 8.71% in ROUGE-L, and 34.03% in METEOR scores.