Research And Implementation Of Custom Test Report Generation Method Based On MDA

Automated testing is a very important technology in software testing,which can greatly improve the efficiency of testing work.Due to the complexity of software testing,each test may use multiple testing tools and get multiple test reports.The formats of these test reports are various,and the test report information is complex,and there is also a lot of repetitive and useless information.Consolidating information from multiple test reports can be time-consuming and labor-intensive for testers.This thesis mainly studies how to receive the test result data of heterogeneous test tools to generate a customized comprehensive test report.The problems to be solved include:1)Identify important data in heterogeneous and complex test reports,filter duplicate and irrelevant information automatically;2)Assemble the filtered information as needed to generate custom test reports.In order to solve the above problems,this thesis has done the following work:1)In order to unify the storage format of test result data,this thesis analyzes the test reports produced by the mainstream automated testing tools on the market,and summarizes the common basic elements in the reports into text elements,graph elements and chart elements,and designs the corresponding data storage format according to the characteristics of each element;2)This thesis proposes a method to generate custom test reports:firstly,use a formal language to describe the test report to get the generalized model of the report,and then design the model-to-code conversion methods,and finally input the model into the conversion method to get the engineering code of the test report When the test report is changed,only the model needs to be modified,which greatly reduces the coding effort when reporting updates.For the custom test report model,this thesis designs a Report Platform Independent Model(RPIM)that can be used to generate custom test reports according to the Model Driven Architecture(MDA)and executable UML.RPIM uses Action Specification Language to define the processing method of the test result data,and realizes automatic identification and filtering of heterogeneous and complex test result data.RPIM also abstracts elements such as text elements,graph elements and chart elements in the test report into classes in executable Unified Modeling Language(UML),and abstracts attributes such as the location,size,and content of each element into attributes of the class.By rationally creating class instances,it is achieved to generate custom reports through on-demand assembly of test report elements.For the model conversion method,this thesis implements a translation algorithm from the subset of Action Specification Language to JavaScript.The algorithm first extracts the valid information in the Action Specification Language code blocks with independent semantics,and then splices the extracted valid information according to the format to obtain the complete JavaScript code.Model-to-code conversion can be achieved by translating and executing the Action Specification Language code in RPIM;3)In order to reduce the difficulty of building RPIM,this thesis develops a Model Building System(MBS)for interactively building RPIM.On this system,users can design different RPIMs according to different requirements for report style and content.For the realization of MBS,firstly,this thesis analyzes the function of MBS with use case diagram and use case specification.Secondly,this thesis designs MBS,and describes the overall structure,interface design,database design and detailed design of each sub-module of the system.Thirdly,this thesis implements MBS based on vue.js,Element-UI and ECharts frameworks.Finally,this thesis designs and executes test cases for each function of MBS.