stormvogel 0.12.0

Probabilistic Model Checking for almost everyone

Stormvogel 🐦: An interactive approach to probabilistic model checking in Python

The state-of-the-art model checking tools that are currently available are optimized to be efficient. The result of this is that they are quite hard to learn and use. Stormvogel flattens the learning cuve by providing easy and user-friendly APIs for creating probabilistic Markov models, and tools to visualize and debug them. It supports seemless conversion to the powerful Storm(py) model checker out of the box.

Features

• Easy APIs for constructing Markov models in dedicated data structures. Currently, DTMCs, MDPs, CTMCs, POMDPs and Markov Automata are supported. This also includes parametric and interval models.

• Seamless conversion between stormvogel and stormpy models with some runtime overhead. This allows, e.g., also using formats such as JANI and PRISM that are not supported by stormvogel directly. It is also possible to add support for a different model checker.

• Visualization of Markov models as an interactive graph and into SVGs via dot. This includes extensive layout options, and displaying model checking results and simulations in an interactive way.

• Support for gymnasium environments

• An extensive documentation with clear examples.

Check out the the stormvogel documentation for examples of how to use stormvogel.

Installation

Pip (release version, recommended for users)

1. Run pip install stormvogel.
2. To also install stormpy, run pip install stormpy.
3. Run jupyter lab
4. Now a browser window should open that runs jupyter lab with stormvogel installed.

Docker (release version)

1. Install docker. Run:
2. docker run -it -p 8080:8080 stormvogel/stormvogel
3. Now a browser window should open that runs jupyter lab with stormvogel and stormpy installed.

For contributors (latest version)

Note that you might have to tweak these steps a bit to get it to work on your particular system, but here is an outline. Contributors need Poetry to manage the development environment.

1. Clone the stormvogel repo (or your own fork) in a separate folder
2. In the stormvogel folder:

   poetry install
   poetry shell # Activate poetry virtual environment
   pip install stormpy
   pip install . # Install stormvogel
   

   If installing stormpy fails in poetry, you can also try to follow the official stormpy installation instructions, and run poetry shell on top of the virtualenv environment that they describe there.
3. Install pre-commit hook: pre-commit install

Testing

nox -s tests   # run test suite
nox -s lint    # ruff + pyright
nox -s docs    # sphinx-build (executes doc notebooks)

Or run pytest directly without nox.

Notice that part of the tests will be skipped if stormpy is not installed.

Authors

Stormvogel was mainly developed at Radboud University by Linus Heck, Pim Leerkes, and Ivo Melse under supervision from Sebastian Junges and Matthias Volk.

Thank you to our contributors: Luko van der Maas, Nicklas Osmers.

License

Stormvogel is licenced under the GPL-3.0 license.