Research On Refactoring Technology Of Open Source Software Oriented To Gaze Analysis Application

With the emergence and rapid development of the open source community,more and more software chooses to use open source code,and software development based on open source code has gradually become a new software development phenomenon.However,open source software rarely performs refactoring activities in version iterations,which leads to more and more complex software structures and lower software quality,which brings difficulties to software development based on open source codes.Not only that,the functions that software developers want to refactor often come from multiple open source software with different architectures.How to quickly refactor the required functions is an urgent problem to be solved in software refactoring.So open source software refactoring is an important research direction in software refactoring technology.Gaze estimation and analysis originated from psychology and visual research,and the research results is paid close attention to day by day,and have been widely used in sports,education,online advertising and other fields.This article starts from the search-based software refactoring method,and studies the open source software refactoring technology for the field of gaze analysis.Open source software has many functions and it is difficult to implement refactoring operations.We propose a feasible refactoring technology that separates the software's functions at the method level,extracts the functions in the first step,and then reorganizes and optimizes it.Genetic algorithm(GA)is easy to implement,has the advantages of good robustness,objective function guidance and strong optimization ability,etc.The evolution direction of GA can be customized by changing the definition of objective method and constraint rules of genetic algorithm.Therefore,this paper uses two different goal-oriented genetic algorithms to realize the refactoring algorithm of decomposition algorithm and combination algorithm.The experiment combined with the gaze estimation code in the open source community verifies the feasibility and effectiveness of the algorithm,and analyzes its performance.We use refactoring technology to design and implement the gaze analysis system,and propose functionas improvement methods,use performance testing and functionas testing to verify the robustness of the system,and contribute to the experience of refactoring software based on open source code.The system is oriented to the two application fields of gaze analysis for athletes and gaze analysis for future MOOC online teaching.It involves the current mainstream gaze estimation technology,enriches the field of software engineering development cases,and further expands the application of gaze analysis.