Reproduce “EM-Assist: Safe Automated ExtractMethod Refactoring with LLMs” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Dorin Pomian, Abhiram Bellur, Malinda Dilhara, Zarina Kurbatova, Egor Bogomolov, Andrey Sokolov, Timofey Bryksin, Danny Dig
EM-Assist helps developers extract methods from long code by suggesting candidate fragments with a language model, then checking, refining, and ranking them through IntelliJ analysis. Developers review and choose a suggestion, which IntelliJ applies. On 1,752 historical refactorings, it matched an oracle in Recall@5 53.4% of the time, versus 39.4% for JExtract, within this corpus and task.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.