Wiki2GH: A Recommendation Service to Link Software Engineering Knowledge to Practical Development

High-quality software applications are built on the knowledge and best practices of software engineering. However, there is a huge gap between the theoretical knowledge of software engineering and its practical implementation for software developers, especially beginners of software development. This paper proposes an innovative recommendation service named Wiki2GH to link Wikipedia software engineering knowledge to GitHub frameworks and libraries, aiming to help software developers systematically learn knowledge and apply it into practice. To implement the Wiki2GH recommendation service, this paper first uses the public data of Wikipedia to build a hierarchy software development knowledge catalog as the entrance of the service. Next, we propose a deep learning method to identify frameworks and libraries and construct a knowledge graph of open source software repositories based on dependency graph. Then, we propose a combined application recommendation service of frameworks and libraries in the development practice. The methods proposed in this paper are all evaluated, including the repository type classification method, the recommendation method, and the knowledge graph construction method. Experimental results show the effectiveness of the recommendation service proposed in this paper.