fix pytests failing

[ChristopherMancuso]

Right now I have some tests commented out. The ones that used the model weights or the predictions are failing as I think there are some differences in the model training, especially for low probability genes across species. Might need a different pytest input and results.

[ChristopherMancuso]

These include

• model probs
• model similarities
• genes in the edgelist

[ChristopherMancuso]

another thing is that @falquaddoomi mentioned is that the way were are running the tests might be redundant, i.e. we have tests set up fro multiple python versions and OSs in GitHub actions, but then possibly also doing that in our tox configuration. Cutting down on too many tests will help with the tests taking so long to run.

[ChristopherMancuso]

Also, try to figure out a way to just get all the pytest data into cache, I think each test is downloading from Zenodo. If there is a pytest for downloading, probably best to remove that.

[ChristopherMancuso]

seems like removing the scaling in code and PreTrainedWeights makes the tests act like they should

[ChristopherMancuso]

I feel like #364 fixed the main reason the tests failed