Boosting Code Search with Structural Code Annotation

Code search is a process that takes given query as input and retrieves relevant code snippets from base. The relationship between commonly built on annotation, which extracted comments or other documents. current studies approximately treat annotation common natural language, regardless of its hidden structural information. To address the information loss, this work proposes model to extract features five perspectives, further conduct engine, i.e., CodeHunter. CodeHunter evaluated dataset 7 million descriptions. experimental results show obtains more effective than Lucene DeepCS. And we also prove effectiveness comes rich models, can well with different sizes