Added biobb-md

Hi,

For this package to work it is required the 'gromacs' spack package, which is in the spack upstream already. Would it be possible to install it as well?

Thank you :)

merged

mentioned in commit ad440c06

Hi @dbeltran,

I merged your MR, but the build for the packages failed. You can check it out here: https://gitlab.ebrains.eu/technical-coordination/project-internal/devops/platform/ebrains-spack-builds/-/jobs/27434

The full build log with the error is here spack-build-out.txt

It seems to be because you have specified (only) python3.7 as a dependency of your packages (both biobb_common and biobb_md - see setuptools documentation) and as a result they can't be installed in the Spack environment, where only python 3.8.11 is available.

Please check it out and let me know when you find a solution.

Thank you!

Hi Eleni,

How is this possible? biobb_common has the python 3.7 dependency and it was working already before my merge request, right?

Would it be possible to install python 3.7 back in the spack environment?

Thanks!

Hi @dbeltran and @elmath,

Python 3.7 was never installed in the JupyterLab kernel (and consequently in the spack environment) we are preparing for the Collaboratory Lab.
The Jupyterlab kernel that is providing the EBRAINS tools for the Lab was built since the beginning for python 3.8.11.
This is a hard requirement and all components need to comply with it.

I am pretty certain that regarding the biobb_common package it had been built successfully in the past due to some error in the build pipeline's configuration which allowed it to install in the environment the python 3.7 version when it should not have done so.

biobb_common has not been promoted to any release as a result. It is not installed in production and not in the experimental releases due to this fact.

Apologies for not informing you about this sooner, i thought that i had done so but most probably i never pressed the "send" button.

Hi @akarmas,

does this mean that any library implemented in Python versions lower that 3.8.11 is completely incompatible with the Collaboratory lab? That would basically mean that we will not be able to integrate any of our tools/workflows in the platform. Have you thought about an alternative for these cases? As you know, we are typically using Conda packages and environments to solve this kind of issues. Any other option?

Hi @adamhospital,

No, it does NOT mean that any library implemented in Python versions lower that 3.8.11 is completely incompatible with the Collaboratory lab.

It only means that any library implemented in Python versions lower that 3.8.11 is incompatible with the JupyterLab kernels prepared with Spack for the Lab with the present delivery scheme (i.e., EBRAINS_release_v0.1, EBRAINS_experimental_release).

Regarding your case and the use of Conda packages, a suggestion could be (as mentioned in the past), that you open a card in the TC Gitlab board to briefly describe the needs, and request the installation and inclusion of Conda in the base docker image of the Collab Lab (mentioning that it was available in Collab v.1 in case this is the case). The discussion can be picked up from there with the Collab team to see what is possible/reasonable to proceed.