Software used in LBS

Disclaimer: Dieser Thread wurde aus dem alten Forum importiert. Daher werden eventuell nicht alle Formatierungen richtig angezeigt. Der ursprüngliche Thread beginnt im zweiten Post dieses Threads.

Software used in LBS
Dear all,

here is the promised post about the software we will be using in the LBS lecture.

We will use

More concretely, we will combine GF, MMT and ELPI into one framework that we call GLIF.

In the next lecture, I will simply share my screen. But eventually you will also get GLIF-based homework assignments.

There are several options:

1. Use GLIF Online
You can use GLIF online under this address: https://mybinder.org/v2/gh/jfschaefer/GlifBinder/master

Advantages:

  • nothing to install :slight_smile:

Disadvantages:

  • might take a minute or so to load
  • you always have to download/upload your notebooks

2. Use the Docker image
After installing docker, you just have to run

docker run -p 8888:8888 jfschaefer/glif:1.2

to get a local GLIF Jupyter server that you can access from your browser.

Advantages:

  • You only have to install docker
  • You can easily remove GLIF when you don’t need it anymore

Disadvantages:

  • We’ve never really used GLIF like this
  • I don’t know much about docker :smiley:

3. Install GLIF on your machine
For that you will have to install GF, MMT and ELPI locally. Note that (as far as I know) ELPI doesn’t support Windows. You may use the subsystem for linux though.
Furthermore, you need to install something to combine them (see below).

Advantages:

  • You can use GF, MMT and ELPI on their own (not in the context of GLIF). In particular for MMT it can be beneficial to use an IDE.
  • Might make debugging easier.

Disadvantages:

  • A lot to install
  • Need to make sure that everything works together

3.1 GLIF through Jupyter
This is the recommended way if you want to use GLIF locally.
In the Jupyter interface you can easily implement GLIF pipelines and test them.
For installation instructions see https://github.com/kwarc/glif
Make sure you have a recent Python version (at least 3.7).

Advantages:

  • Intuitive to use
  • You can open and modify the notebooks we create in class

Disadvantages:

  • You have to install Jupyter
  • Some people (myself included) generally prefer programming in an editor. You can still do that and use the notebooks only for testing your implementation. In my personal experience, implementing GLIF pipelines in Jupyter is more convenient, unless you are working on a larger project. The reason is that a simple GLIF pipeline would require many small files in different languages and it’s just easier to have everything in one place. Also note that Jupyter Lab has advanced editor support (vim/emacs/sublime text keybindings).

3.2 GLIF through Custom Scripts
Just use a custom Python script to combine GF, MMT and ELPI.
I can provide you with one.

Advantages:

  • You don’t have to install Jupyter

Disadvantages:

  • Has to be customized for every project
  • Rather inconvenient to use (at least for now)

Did anyone else have immense trouble installing the grammatical framework on ubuntu20? If so, i think found a way to do it.