Jupyter notebook setup

You'll first need to run the recommended, simple installation of Imandra.

After running the simple installer, you should see imandra-jupyter-kernel on your path. You'll now need to install the kernel as part of your local Jupyter setup in order to invoke this.

To install the Imandra and Imandra-reason kernels, run the following commands:

(Note: if you have a more intricate Jupyter environment setup or if you use conda, you'll need to adjust the commands accordingly to install things into the correct location).

# install imandra and imandra kernels
jupyter kernelspec install /usr/local/var/imandra/_opam/share/jupyter/kernelspec/imandra
jupyter kernelspec install /usr/local/var/imandra/_opam/share/jupyter/kernelspec/imandra-reason

#install nbimandra notebook extensions as a python package (assumes this is in the same package target as jupyter itself)
pip install /usr/local/var/imandra/_opam/share/jupyter/nbextensions/nbimandra

#install nbimandra as an extension from the installed python package
jupyter nbextension install --py nbimandra
jupyter nbextension enable --py nbimandra

Docker image

The notebook is also available as a docker image: imandra/imandra-client-notebook.

If you haven't authenticated yet and don't have an ~/.imandra folder on your docker host, first authenticate using the regular imandra/imandra-client image:

docker pull imandra/imandra-client:latest && \
  docker run -it --rm -v ${HOME}/.imandra:/root/.imandra -p 8000:8000 \

If everything succeeds you should be dropped into an imandra-repl session. Exit this with C-d.

Then to launch the notebook, mount your ~/.imandra folder as a volume in the container. You may find it useful to mount the current directory in the container too so you can save notebooks you've edited:

docker pull imandra/imandra-client-notebook:latest && \
  docker run -it --rm -v ${HOME}/.imandra:/home/jovyan/.imandra -p 8888:8888 \
  -v `pwd`:/home/jovyan/workdir \