sktime/sktime

[MNT] Migration to newer skforecast interface

オープン

#7,451 opened on 2024/11/26

 (11 件のコメント) (0 件のリアクション) (0 人の担当者)Python (1,192 件のフォーク)batch import
good first issueinterfacing algorithmsmodule:forecasting

Repository metrics

Stars
 (7,162 個のスター)
PR merge metrics
 (平均マージ 26d 10h) (30d で 86 merged PRs)

説明

Originally posted by @JavierEscobarOrtiz in https://github.com/sktime/sktime/issues/6531#issuecomment-2499989752

Hello @fkiraly @Abhay-Lejith @yarnabrina

I’m Javier, one of the co-authors of skforecast. First of all, thank you so much for including our library as a wrapper in sktime. It’s amazing to see how our work helps other teams!

I wanted to let you know that in skforecast 0.14.0, we’ve implemented a major refactoring aimed at improving performance and usability. For instance, the class ForecasterAutoreg has been renamed to ForecasterRecursive:

# Before Skforecast 0.14
from skforecast.ForecasterAutoreg import ForecasterAutoreg

# From skforecast 0.14.0 onward 
from skforecast.recursive import ForecasterRecursive

Same Forecaster but new name and new features.

If you need help with the integration, just ping us! 😄 We have also created a migration guide section for anyone who wants to update to the de version.

Best,

Javi, @JoaquinAmatRodrigo

コントリビューターガイド