GLIF Kernel

This repository contains the code for a re-implementation of the GLIF kernel for Jupyter.

Online Demo

You can try out GLIF online at https://mybinder.org/v2/gh/jfschaefer/GlifBinder/version2.

What is GLIF?

TODO

Running the Kernel Locally with Docker

If you have docker installed, you should be able to simply run

docker run -p 8888:8888 jfschaefer/glif:2.0.1

Note that running the command again will create a new container, i.e. your changes won't persist. However, you can "link" (mount) a folder of your file system into the container, which will allow you to keep files stored there:

docker run -v "/path/to/my/folder:/home/worker/glifkernel/notebooks/persistent" -p 8888:8888 jfschaefer/glif:2.0.1

/path/to/my/folder should be a folder on your local file system. It will be synchronized with the persistent folder.

Installing the Kernel locally

Installing the kernel directly on your machine has several advantages and is recommended for serious users. Some of the reasons are:

You can develop components in specialized IDEs (e.g. for MMT) rather than in the notebooks, which is particularly useful for larger projects
You can work with your own build of GF/MMT/ELPI
You can develop demos with custom code using the GLIF python package

Prerequisites

A recent Python version (at least 3.9, you can check with python3 --version)
git (currently required for the installation with pip from Github) Optional:
GF, MMT and ELPI, which GLIF is based on. Note that you only need to install the frameworks you actually want to use (i.e. you can e.g. not install ELPI if you don't plan to use it). To help GLIF find MMT, you should set the MMT_JAR environment variable to the installation destination (export MMT_JAR=/path/to/mmt.jar on unix).
Graphviz, specifically dot, if you want to visualize ASTs.

Actual Installation

python3 -m pip install git+https://github.com/jfschaefer/glifcore.git#egg=glif
python3 -m pip install git+https://github.com/jfschaefer/GLIFkernel.git#egg=glif_kernel
python3 -m glif_kernel.install

Running the Kernel

Simply start Jupyter with

python3 -m jupyter notebook

and open an existing notebook or make a new one with the kernel GLIF.