fugue-project / fugue

A unified interface for distributed computing. Fugue executes SQL, Python, Pandas, and Polars code on Spark, Dask and Ray without any rewrites.
https://fugue-tutorials.readthedocs.io/
Apache License 2.0
1.92k stars 94 forks source link

[BUG] Make Fugue compatible with Ray 2.5.0 #485

Closed goodwanghan closed 1 year ago

goodwanghan commented 1 year ago

Ray 2.5.0 had break changes again! So the dataset_format function becomes deprecated. But in lower version, we rely on it to determine whether the dataframe is empty and also to check whether it is arrow or pandas dataset. In addition, they start to introduce their own Schema class, and changed the return values of a few functions.

It breaks most of the functionalities of Fugue on Ray, so we have to make Fugue compatible from lower version Ray (~2.1.0) to the latest version.