Add support to Python 3.10 and 3.11. Before that – update documentation

[GegznaV]

Describe the bug

auto-sklearn does not install in Python 3.10 and 3.11-based environments.

Before this issue is fixed (as it will take some time), please update the documentation, stating which exact Python versions are really supported (now it is stated that "Python (>=3.7)" which implies that there are should be no issues with 3.10 and 3.11).

To Reproduce

1. Create a new Python 3.10 or Python 3.11-based conda environment.

conda create -n py310 python=3.10

2. Activate the environment:

conda activate py310

3. Try to install auto-sklearn there. I use mamba here as it gives the easiest-to-understand error message in a reasonable amount of time (I tired pip and conda too with no success):

Expected behavior

Install without errors with Python 3.100 and 3.11.

Actual behavior, stacktrace or logfile

/ ... /

        mamba (1.4.2) supported by @QuantStack

        GitHub:  https://github.com/mamba-org/mamba
        Twitter: https://twitter.com/QuantStack

█████████████████████████████████████████████████████████████


Looking for: ['auto-sklearn']

pkgs/r/linux-64                                               No change
pkgs/main/noarch                                              No change
pkgs/r/noarch                                                 No change
pkgs/main/linux-64                                            No change
conda-forge/noarch                                            No change
conda-forge/linux-64                                34.3MB @   3.3MB/s 10.9s

Pinned packages:
  - python 3.10.*
  
  Could not solve for environment specs
The following package could not be installed
└─ auto-sklearn   is installable and it requires
   └─ scikit-learn >=0.24.0,<0.25.0  with the potential options
      ├─ scikit-learn [0.24.0|0.24.1|0.24.2] would require
      │  └─ python >=3.6,<3.7.0a0 , which can be installed;
      ├─ scikit-learn [0.24.0|0.24.1|0.24.2] would require
      │  └─ python >=3.7,<3.8.0a0 , which can be installed;
      ├─ scikit-learn [0.24.0|0.24.1|0.24.2] would require
      │  └─ python >=3.8,<3.9.0a0 , which can be installed;
      ├─ scikit-learn [0.24.0|0.24.1|0.24.2] would require
      │  └─ python >=3.9,<3.10.0a0 , which can be installed;
      └─ scikit-learn [0.24.1|0.24.2] conflicts with any installable versions previously reported.

Environment and installation:

Please give details about your installation:

• OS: Ubuntu Ubuntu 20.04 via WSL on Windows 10
• Is your installation in a virtual environment or conda environment? conda environment
• Python version: either 3.10 or 3.11
• Auto-sklearn version: the one that is tried to be installed by these tools (the current version, I guess 0.15.0)

Possibly related issues:

• #1690
• #1670

[GegznaV]

Using a Python 3.9-based environment, the installation of auto-sklearn was without errors.

[connorkeevill]

Also having the same issue - have a project running 3.10 and can't downgrade to 3.9 as the project uses 3.10 language features. A shame as this looks like a very powerful project!

[wencan]

Using a Python 3.8-based environment, the installation of auto-sklearn was without errors.

[namelessperson0]

in python3.12, getting the below error in pip ModuleNotFoundError: No module named 'numpy.distutils'

[namelessperson0]

In python3.10, getting below error in pip Cython.Compiler.Errors.CompileError: sklearn/ensemble/_hist_gradient_boosting/splitting.pyx