initial commit for release CD

[bernt-matthias]

TODO:

• [ ] switch to on: release (https://docs.github.com/en/actions/reference/events-that-trigger-workflows#release)
• [ ] fix tensorflow base image to a version? i.e. 2.3.0-gpu instead of latest
• [ ] deploy the container somewhere
• [ ] add test task for container

[halirutan]

@bernt-matthias Your initial intention to install the dfpl directly from the source was a good one! You just missed one important point: You need to map the deepFPlearn directory into the container so that it has read access. To make it work, you'd just need fix the following thing in the singularity def file: Here you remove the package again and put the file system mapping (source -> container destination) in. So since we're calling singularity from the singularity_container directory, it is

.. /deepFPlearn

The parent dir is mapped to /deepFPlearn in the container. Then we can do exactly what you initially had:

/bin/bash -c ". $CONDA_DIR/etc/profile.d/conda.sh; conda activate rdkit2019; python -m pip install /deepFPlearn"

[halirutan]

Alright, I pulled your branch, fixed the singularity built and created a new feat_singularity_CD branch with all additions on this repo.

@mai00fti Can we make Matt a developer of this repository so that we can work together on this?

[halirutan]

@bernt-matthias OK, check here:

FATAL:   While performing build: 
while creating SIF: while creating container: writing data object for SIF file:
copying data object file to SIF file:
write /__w/deepFPlearn/deepFPlearn/singularity_container/deepFPlearn.sif: no space left on device

What now?

The container builds fine but when finally storing the sif file it runs out of memory. I don't understand why because there seems to be plenty of space

[mai00fti]

@mai00fti Can we make Matt a developer of this repository so that we can work together on this?

done

[halirutan]

@bernt-matthias Let us close this PR and work together on https://github.com/yigbt/deepFPlearn/tree/feat_singularity_CD

[bernt-matthias]

I just took you commits into my branch. I would prefer to keep working on this PR and you push to the branch in my form.

Because leaving the PR also gives us a means of communication.

[halirutan]

Yes, I understand. We open a new WIP PR. The big advantage is that we don't need to pull and push between repos. Give me a second.

[halirutan]

Here you go: https://github.com/yigbt/deepFPlearn/pull/2