sktime / skpro

A unified framework for tabular probabilistic regression, time-to-event prediction, and probability distributions in python
https://skpro.readthedocs.io/en/latest
BSD 3-Clause "New" or "Revised" License
250 stars 46 forks source link

[MNT] `numpy 2` compatibility #394

Closed fkiraly closed 4 months ago

fkiraly commented 5 months ago

numpy 2 has been released. We should test framework and all estimators for compatibility, and bump the version with the next MINOR release.

fkiraly commented 5 months ago

From https://github.com/sktime/skpro/pull/393, framework is compatible without any issues.

fkiraly commented 5 months ago

From https://github.com/sktime/skpro/pull/395, everything including all estimators is compatible without any issues. Minor restriction is that, in case of soft dependencies, we know this only for packages that already support numpy 2, but I would not expect any major problems as interface code does not use numpy typically, and if it does, only very superficially.

fkiraly commented 5 months ago

Meaning, we can bump the numpy bound in 2.4.0.