search

wolberlab / dynophores

Dynamic pharmacophore modeling of molecular interactions
https://dynophores.readthedocs.io
MIT License
31 stars 7 forks source link

Docker container for dynophores? #41

Open cap-jmk opened 3 years ago

cap-jmk commented 3 years ago

The setup.py does not install all necessary libraries. Maybe add Docker distribution?

dominiquesydow commented 3 years ago

Hi @MQSchleich,

I have been working on an updated conda package today; could you try that? https://dynophores.readthedocs.io/en/latest/installing.html#install-from-the-conda-package

Adding a docker distribution is also a great idea; I am not familiar with the setup, yet, but I'll put that on my list.

Just a heads-up: This repo is still quite heavily under constructions.

cap-jmk commented 3 years ago

Thanks, @dominiquesydow. I just wanted to have it in sight. I got everything working up to now. Still, do you know how to fix the nglviewer in Jupyter Lab? By default, it says : Error displaying widget: model not found

I tried setting it up with: 

python -m ipykernel install --sys-prefix
jupyter nbextension enable --py --sys-prefix widgetsnbextension
jupyter nbextension enable --py --sys-prefix nglview

That's why I think it might be beneficial to have a docker, such that everything works out of the box. Maybe I missed installing a certain widget?

dominiquesydow commented 3 years ago

Thanks for reporting this; I had hoped we got rid of that problem by using Jupyter Lab 3, sorry.

Can you check your version with conda list jupyterlab? If you have installed the dynophores conda package, you should have >=3 but let's double-check.

In case you do have Jupyter Lab 2, you might need this:

jupyter labextension install @jupyter-widgets/jupyterlab-manager nglview-js-widgets

Re docker - I will look into it but I won't manage to get it done quickly, sorry for that!

cap-jmk commented 3 years ago

You are welcome. I can take care of Docker as well, if you'd like, or we tinker together. Regarding the visualisation: 1.) Good Job! ;) 2.) The mistake was definitely on my side (as I was too lazy to install mamba and did not take care about installing Jupyter Lab 3), My output for conda list jupyterlab reads: 

# Name                    Version                   Build  Channel
jupyterlab                3.1.9              pyhd8ed1ab_0    conda-forge
jupyterlab_pygments       0.1.2                      py_0
jupyterlab_server         2.7.2              pyhd8ed1ab_0    conda-forge
jupyterlab_widgets        1.0.0              pyhd3eb1b0_1

After running the aforementioned commands, I could display your beautiful visuals.

dominiquesydow commented 3 years ago

I can take care of Docker as well, if you'd like, or we tinker together.

Great, yes, feel free to create a PR about this when you have time. I have really no clue right now how to set this up, so I would learn from what you propose and add to it where I can.

cap-jmk commented 3 years ago

Okay, great! Once I included the packaging from the deep-learning models, I will start setting up the docker. :)