Code Retrieval Technology And Its Application In Online Teaching System

Online education has experienced rapid development in recent years.Currently,the scale of international students in Chinese universities is increasing,resulting in a significant demand for courses tailored to international students.Moreover,the ongoing pandemic has prevented a large number of students from entering the country,forcing them to engage in online learning activities through e-learning platforms.Although there are numerous well-established online teaching platforms,such as Chaoxing and Zhishu Tree,they are primarily designed for general student populations and lack platforms specifically tailored to the preferences and language needs of international students.For students in information engineering disciplines,studying excellent code snippets from others is an effective way to enhance their programming skills.Typically,students in programming and development courses are at the beginner level and lack knowledge of related and advanced concepts.Consequently,students often have incomplete understanding of their programming problems and struggle to express them adequately.Traditional search engines fail to provide satisfactory results in such cases.Therefore,the field of vertical domain code retrieval has emerged as a specialized search engine for code retrieval.By using this technology,students can input natural language expressions,and the underlying algorithm processes their queries to provide relevant code method snippets.Based on the above research background,this thesis aims to investigate how to design and implement an online teaching system based on microservice architecture,and also apply code retrieval technology to the online teaching system,the main work of this paper includes:(1)A code retrieval model based on multi-feature modelling of code(abbreviated as: MFCRM in this paper)is proposed to exploit the deep semantic features of code fragments and query statements into vectors in a unified vector space by differentially processing and semantically enhancing code fragments and natural language utterances,combining the deep learning network model Trasnsformer and the self-attention mechanism to achieve semantic association between code and natural language queries by calculating similarity to measure the degree of similarity between them.(2)Analyse the business process of the online teaching system for international students,carry out requirements analysis accordingly,adopt the system architecture of microservices,split the different services of the online teaching system into multiple small services,each of which can be deployed,extended and maintained independently,and carry out the overall design of the system and the detailed design of the basic functional modules of the system,as well as the design of the code retrieval sub-module of the application module.(3)Using technologies such as Spring Cloud and Vue,the basic functional modules of the online teaching system,such as the user management submodule,course management submodule,exam management submodule,etc.,were developed.Additionally,the code retrieval technology was applied to the online teaching system.Furthermore,a reasonable testing plan was designed to validate whether the online teaching system meets the design requirements in various aspects.