Unable to install easymode on Python 3.11 and higher after commit 989e77b

[tieneupin]

Thanks to the addition of Ais-cryoET as a package dependency in commit 989e77b, it is no longer possible to install easymode on Python 3.11 and higher. This is due to Ais-cryoET being locked to tensorflow==2.8.0, which is not available for Python 3.11 and greater (see Issue bionanopatterning/Ais#11).

This will become a greater issue going forwards, as Python 3.10 will reach end-of-life status this October, so this package and its dependencies need to be able to run on package versions available in Python 3.11 and above.

[mgflast]

Hey,

Sorry for the very late reply, I really thought I replied to this issue a while ago.
You're right about the breaking dependency. I know it's inconvenient and I want to fix it eventually, but because the compatibily of Python / CUDA / cuDNN / tensorflow is always a bit horrible to untangle, I'm sticking to 3.9 during development. I want to do a proper upgrade in both easymode and Ais once I'm ready for a first 'official' release.

For the time being, would it be an option for you to make a separate env with Python 3.9?

[tieneupin]

Thanks for the reply, @mgflast . Our team pinned the repo at the commit right before Ais-cryoET was introduced, so we are good for now. It's good to know that you're aware of the dependency issue. Looking forward to once it's addressed!