With both Python code and TS, LLMs are in my experience very good at generating test code from e.g. markdown files of test cases.