Research On Web Application Automated Test Based On Selenium

Web application automation testing often uses a data-driven or keyword-driven automated testing framework and uses DOM-based methods to locate and identify Web controls.However,there is a problem that the coverage of the use case is not high in the data driver.The application of the keyword driver for the data input has the problem of high test script development time,and the DOM-based control identification method cannot identify the non-Web control.In view of the above problems,this paper combines data-driven and keyword-driven models,and uses image matching to identify non-Web controls.A hybrid drive automation testing framework that can identify non-Web controls is designed.This article analyzes the digital pool virtual experiment Web system and makes a comparative study of the mainstream Web automation testing framework and tools.For this system has the characteristics of large data input,diversified and complex page controls and diversified functions,this paper designs and implements a hybrid drive automation testing framework based on image matching to identify non-Web controls.The hybrid drive test framework designed in this paper contains five modules: function library,configuration file module,execution module,Page-Object module and drive module.The driver module is a framework core module that integrates data drivers,keyword drivers,and image matching drivers.Data drivers and keyword drivers enable this framework to use data and keyword hybrid drive automation testing.The image matching module uses the SIFT algorithm to extract the feature vectors of the control screenshot and the tested page to perform image matching,and to improve the non-Web element control of Selenium Web Driver to identify the tested page.Finally,in the case of the same test cases,this paper compares the mainstream automated testing framework and the hybrid driven automated testing framework designed in this paper on control coverage,functional coverage,comprehensive satisfaction and testing costs.Experimental results show that the hybrid driving test framework designed by this paper combined with image matching algorithm to identify non-Web element controls has higher control coverage,functional coverage and comprehensive satisfaction,and lower test time cost,and is more suitable for Web applications.High practical value in automated testing.