Upgrade numba version to support python 3.12

[danclaytondev]

It would be great to run Anonymeter with Python 3.12.

It was previously constrained by numba python support, so this is a fairly conservative update from numba 0.58 to 0.59 which moves anonymeter support to <=3.9, <3.13.
Python 3.8 is already end of life so I don't think new versions of Anonymeter need to be supporting anything lower than 3.9.

I have left other dependency versions the same.

Tests passing on 3.12:

============================= test session starts ==============================
platform darwin -- Python 3.12.8, pytest-7.4.4, pluggy-1.6.0
rootdir: /Users/daniel.clayton/code/anonymeter
configfile: pyproject.toml
testpaths: tests
collected 152 items

tests/test_confidence.py ....................                            [ 13%]
tests/test_inference_evaluator.py ................................       [ 34%]
tests/test_linkability_evaluator.py .................................... [ 57%]
.........                                                                [ 63%]
tests/test_mixed_types_kneigbors.py ...........                          [ 71%]
tests/test_singling_out_evaluator.py ...............................     [ 91%]
tests/test_transformations.py .......                                    [ 96%]
tests/test_type_detection.py ......                                      [100%]

============================= 152 passed in 3.50s ==============================

[CLAassistant]

All committers have signed the CLA.

[MatteoGiomi]

Thanks @danclaytondev for taking care of this! Feel free to merge this PR!

[MatteoGiomi]

Hi @danclaytondev, do you need help with this?

[danclaytondev]

Hi @MatteoGiomi sorry I left this for a while, I've updated the type hints to remove deprecated usage (now that we aren't supporting 3.8, ruff will not let us use them).

[MatteoGiomi]

@danclaytondev it seems that there is some incompatibility between pandas and numpy. Probably installing a lower version of the latter would solve the problem.

[MatteoGiomi]

Thanks a lot @danclaytondev for this PR, I fixed the CI by capping the numpy version and it works ok now. I'll merge it!

[danclaytondev]

Hi @MatteoGiomi thanks very much for picking this up, and apologies that I left it for so long unresolved. Cheers!

[MatteoGiomi]

Hi @MatteoGiomi thanks very much for picking this up, and apologies that I left it for so long unresolved. Cheers!

Thanks to your work, I had very little left to do. I wanted to merge this because another PR was waiting and didn't want to leave the rebase to this one. They are always fun.