The Design And Implementation Of The Analysis Service In Web Application Automation Testing System

With the gradual deepening of the Internet degree,web applications are widely used in all walks of life and daily life due to their simplicity and convenience,and have become an integral part of the current information age.Faced with the current situa-tion of the ever-changing needs and agile development iterations,the web applications quality has also become a core area that developers,testers and users are extremely con-cerned about.However,the existing web application quality detection methods mainly rely on manual testing or script testing,which not only consumes a lot of manpower and time,but also has low testing efficiency and poor repeatability.On the other hand,as a regression test,although these testing methods can also find some system defects with obvious characterization,they still require users to participate in defect positioning and test report filling.At the same time,some non-characteristic defects with certain risks may not be find.Therefore,it is very necessary to study the automated testing and testing analysis of web applications.This thesis designs and implements the analysis service part of the Web applica-tion automation testing system,mainly for the defect scanning and defect analysis of web application testing,and provides intelligent solutions.The system relies on multi-source "black box" data of the system under testing,such as related logs,screenshots and so on,which are recorded by the execution service and scheduling component re-sponsible for automated traversal testing.Using methods such as clustering algorithms and feature detection,a web application related defect system is constructed,realizing and integrating some system defect detectors corresponding to relevant data.Finally,the system discovers defects through detection,and constructs test report of the appli-cation to be tested.At the same time,it supports the expansion and compatibility need of other analysis data and possible defect categories.On the other hand,the system has also established a preliminary web application evaluation system to achieve multi-dimensional evaluation of web application quality.The entire system exposes services for downstream use based on the SpringBoot framework,uses MongoDB,Redis,and OSS as data centers,uses RabbitMQ as a message queue to implement asynchronous calls,and uses Celery to schedule and distribute subtasks.At present,the system has been deployed and provides differentiated support for enterprises and universities.Experiments based on 50 real online websites also show that the system can effectively discover various web application defects such as broken links,resource loading failures,server-side errors,JavaScript errors and so on.At the same time,both the accuracy rate of system defect classification and the correct rate of export test script execution exceeds 95%based on manual sample and review,and system is not limited to the selection of diverse web application front-end development frameworks.It can provide automated testing and evaluation services for most industry websites.Together with a variety of testing methods such as manual testing,script testing,and crowdsourcing testing,it escorts web application quality.