search

EpistasisLab / tpot

A Python Automated Machine Learning tool that optimizes machine learning pipelines using genetic programming.
http://epistasislab.github.io/tpot/
GNU Lesser General Public License v3.0
9.57k stars 1.55k forks source link

Update requirements #1335

Closed gatl closed 6 months ago

gatl commented 6 months ago

Due to the use of the decorator sklearn.utils.metaestimators.available_if() by TPOT, declare as the minimum version required of scikit-learn the version that introduced this decorator, that was 1.0.

What does this PR do?

Update the requirements for TPOT.

Where should the reviewer start?

This is just a value change.

How should this PR be tested?

Verify that TPOT no longer works with scikit-learn version 0.22.0 or any other version prior to 1.0, but it works with version 1.0 and above.

Any background context you want to provide?

I think none is needed.

What are the relevant issues?

There are none, other than TPOT failing to work when all its stated requirements are satisfied.

Screenshots (if appropriate)

None is provided.

Questions:

I think there is one reference that would benefit of attention in the installation section, despite referring to dask.

No, it does not.

perib commented 6 months ago

The setup.py file should also be updated to match. If you can make that change, I can merge this. Thanks for the contribution!

gatl commented 6 months ago

Oh, sorry. Right. But now I see you did it already. Thanks.