Make sure the maxATAC python package can be installed on python 3.7/3.8/3.9 on both Ubuntu and macOS

[michael-kotliar]

maxATAC python package should be easily installed with all required dependencies on Ubuntu 18.04/20.04 and macOS with pre-installed python 3.7/3.8/3.9.

The task is finished, when:

• we can run maxatac --help command for any Python/OS combination with exit code 0
• we have a test for it and proper documentation

This will require:

• [x] Define the list of working dependencies for each python version. The list of working and tested dependencies should be saved in the same GitHub repo. When installing with pip, they will be provided through --constraint argument.
• [x] Find a solution for installing not python dependencies if they are also required (bedtools, samtools, etc).
• [ ] Make automatic tests for running the installation on every Python/OS combination. Here I can help because I've already implemented a similar thing here https://github.com/Barski-lab/cwl-airflow/blob/master/tests/run_conformance_tests_docker.sh

[FaizRizvi]

Hi Michael! I just wanted to ask why do we need to check for these three versions of python? We are currently operational on python 3.9 (functional for both gpu and non-gpu). Changing the version of python will result in conflicts with some packages - ie tensorflow-gpu, cuda, python compatibility (and possibly many other packages that I am un aware of right now.

[FaizRizvi]

So I chose python 3.9 as it is compatible with tensorflow-2.5, tensorflow_gpu-2.5 and these are compatible with CUDA 11.2. These are available on OSC and BMI clusters.

https://www.tensorflow.org/install/source#gpu

Once we get python 3.9 functional (i.e. install directions etc.), we can then test out python 3.7 and python 3.8.

[emiraldi]

I agree with Faiz. I can't think of a reason to include Python versions other than 3.9 -- although I might be missing something! Thanks!

On Tue, Oct 12, 2021 at 12:29 PM FaizRizvi @.***> wrote:

So I chose python 3.9 as it is compatible with tensorflow-2.5, tensorflow_gpu-2.5 and these are compatible with CUDA 11.2. These are available on OSC and BMI clusters. https://www.tensorflow.org/install/source#gpu

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/MiraldiLab/maxATAC/issues/63#issuecomment-941174006, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACUJXVKXK5J6QDGNM6KDF5LUGRO4ZANCNFSM5FY5HZOQ . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Emily Miraldi, Ph.D. Assistant Professor Divisions of Immunobiology and Biomedical Informatics Cincinnati Children's Hospital

[michael-kotliar]

Sure, we can start with any python version you are currently confident working with. Then we can just test it with the other not deprecated versions as well (3.7 and 3.8) if needed.

[tacazares]

@michael-kotliar Do you have any experience setting up CI and upload to pypi?