I am looking for better name
Idea behind this tool, is to maintain snapshots of what your model can answer.
How it works: You define a set of questions. This tool run this questions against your model and save the answers. Then you can run this questions again and compare the answers with the saved ones.
pip install langsnapy
Check examples folder.
poetry install -E runtime-deps --with dev
poetry run pytest .
poetry run ruff .