A black-box test generation framework for test libraries. ATAG employs machine learning algorithms to construct test cases based on testing goals specified by the user. This version, created for a master's thesis, is currently in the prototype phase. Instructions will be updated at a later time.
Algorithms ready for test generation:
- Proximal Policy Optimization (PPO)
- Online Decision Transformer (ODT)
- Local Large Language Models (LLMs)
Control interfaces implemented for the following test automation libraries:
- Browser Library
More info: www.rikulehtonen.fi/projects/atag
Detailed instructions are updated later.
-
Install pytorch https://pytorch.org/get-started/locally/
-
Install requirements
pip install -r requirements.txtBenchmarking tools are moved to https://github.com/rikulehtonen/testgen-benchmarks
This is a repository for a master thesis project. Pull requests are not possible for the time being.