I am seeking help to implement my idea on developing an automation test framework that can fix locator related issues by itself. My abstract idea is as given below.
I will have a Selenium Test Automation framework that will run in three modes -
Analysis and Fix, and
Normal modes. When I run on Learning mode (assume that all test cases run fine on current build) the framework will run all test cases and it will capture screenshot of the web element it interact with (may be the html of the web page instead of screenshot).
Once a new build is released I will run the framework in Analysis and Fix mode. In this mode the framework will navigate through each pages in the application and check whether the elements I already captured are valid in the current build or not. If the element is no longer valid, it will try and automatic fix. Framework will check the screenshot it already captured (or html file) in Learning mode and it will learn how the element looks like and it will try to identify the element in current build. Once the element is identified it will automatically generate suitable locator for the element and update the element locator file.
I know the question is not specific to a problem but I am trying to apply AI techniques it Selenium Web Automation framework. Python has rich library for AI, I prefer Java though. How to implement this idea?.