Prompt Based Intelligent Service Method for Public Code Review

Public Code Review (PCR) can be implemented through a Software Question Answering (SQA) community, which facilitates high knowledge dissemination. Quality assurance in PCR considers two links from which developers start: review necessity prediction and tag recommendation subtasks. To this end, we propose a intelligent service called UniPCR to complete developer-based request quality assurance (i.e., predicting request necessity and recommending tags subtask) under a Masked Language Model (MLM). Specifically, we reformulate both subtasks via text prompt tuning, which converts two subtasks into MLM by constructing a task constructive prompt templates using hard prompt. Through this intelligent service the request quality of public code review can modeling in an unified method. Experimental results demonstrate that our UniPCR framework adapts to the two subtasks and outperforms comparable accuracy-based results with state-of-the-art methods for the request quality assurance.