Bug fix: TypeError due to np.matrix use

[ManonSabot]

Hi Jacob,
I've been trying out your ET partitioning algorithm and I've run into issues because np.matrix has been deprecated in sklearn in python 3.9.
I thought I'd flag that converting the matrices to arrays fixes the issue and allows the partitioning algorithm to run.

[jnelson18]

Hey Manon,

Thanks for working on this. In the end, I updated a few things as well as the np.matrix depreciation in version 1.10, so hopefully everything is in working order now with python >3.7. Let me know if there are any issues.

-Jake

[ManonSabot]

Hi Jake,

Thanks for updating the repo! Yes, the code runs without a hitch now if I clone the TEA and the pyquantrf repos.

The conda install seems to be broken though (I've tried all sorts of things -- including reinstalling conda and nothing works). As far as I can tell, that might come from a compatibility issue / clashing python version requirements in your .yaml file in pyquantrf. No matter what I do, this error pops up:

UnsatisfiableError: The following specifications were found to be incompatible with each other:

Output in format: Requested package -> Available versions
The following specifications were found to be incompatible with your system:

• feature:/linux-64::__glibc==2.27=0
• feature:|@/linux-64::__glibc==2.27=0

Your installed version is: 2.27

As you can see, my installed version is the right one, so that error is nonsense. I don't know if it helps, but I've tracked down a similar issue in a conda thread, where it was fixed by changing some parts of the .yaml file: KevinMusgrave/pytorch-metric-learning#55

Personally, I'm happy with the git clones as they're working fine. Just letting you know.

Manon