TornadoFX-Suite (Stage 2)
Automated UI test generation for TornadoFX. A study in Kotlin & Metaprogramming and how metaprogramming might be able to eventually further machine learning.
Note: For the time being, the scope of this project is kept small. A user should be able to input their TornadoFX project and have tests generated for inputs like textfields, checkboxes, buttons, and so on
Testing has always been a challenge for engineers - many of us are opiniated on how testing should affect our development process and what ought to be tested in applications. While we will do our best to update as we go both in our issues and on here, TornadoFX-Suite is a work in progress:
- Stage 1: Locating UI Nodes Scanning and parsing uploaded code to locate UI Nodes and create a basic class breakdown.
- Stage 2: Generating UI Tests finding ways for Kotlin to create code-that-creates-code
- Stage 2: Abstract Compositional Contracting Post-parsing analysis to map associate functions to nodes and create functional composition for nodes. Creating a DSL to create an easy structure for writing tests.
- Stage 3: Creating FINITE State Machines
- Stage 4: TBD Data collection/analysis using machine-learning to learn what makes more useful testing rules
Interested in the project? It's open source and anyone is welcome to contribute! Feel free to head over to the Issues page to see what's going on.
Data Science Courses
You'll notice some classifications for the way the issues are labeled. Ideally, every issue should be formatted a certain way. I just made this up. If you guys have something you like better let me know. You'll often see the issues necessary for this project to move forward in the format of
category is just an arbitary set of concerns [
INVESTIGATION] I made to address different kinds of issues as this project progresses.
And when it comes to naming branches to address these issues:
bug_[bugNumber]/[name of bug]
enhancement_[enhancement]/[name of enhancement]
and so on.
Features are issues that are required to complete a stage.
I like to go with Yegor Bugayenko's definition, as quoted in his blog:
Lack of documentation. If, say, a Javadoc block for a class does not clearly explain to you how to use the class, or the entire module is not documented well (yegor256/takes#790), it’s a bug.
Suboptimal implementation. If a piece of code doesn’t look good to you, and you think it can be refactored to look better, it’s a bug.
Design inconsistency. If the design doesn’t look logical to you (yegor256/cactoos#436) and you know how it can be improved, it’s a bug.
Naming is weird. If class, variable or package names don’t look consistent and obvious to you, and you know how they can be fixed ([yegor256/cactoos#274] (https://github.com/yegor256/cactoos/issues/274)), it’s a bug.
Is there something down the road that might be relevant to later stages? Go ahead in put in a issue for discussion/invesigation.
Improve on a feature/functionality.