This system is designed to showcase the capabilities of a Retrieval-Augmented Generation (RAG) system with a small local LLM (Llama.cpp) in generating structured and detailed test cases (and code, in different frameworks) using design documents and HTML files.
Parses software design documents to extract relevant features and functionalities.
Analyzes the HTML structure of the mock website to identify key elements such as buttons, forms, and links.
Uses a Retrieval-Augmented Generation model to create detailed and structured test scenarios.
Supports generating example QA pairs and step-by-step guidance for test cases.
Code can be generated as well (Playwright or other, with given documentation or sample cases added to RAG)
Example results and progress can be seen on the wiki page: https://github.com/busraoguzoglu/Software-Test-RAG/wiki/Example-QA