Currently, Polars supports defining custom functions on the DataFrame namespace on the python side with pl.api.register_dataframe_namespace/pl.api.register_lazyframe_namespace. Would it be possible for pyo3-polars to support creating a custom DataFrame namespace with this same functionality, with the defined transformations living in rust?
For a somewhat abstract example, I have a function written in rust that takes a pl.DataFrame as an argument, loads a xgboost Booster, calculates a column, then returns the DataFrame. I want to expose this exact same functionality in Python by registering some custom namespace.
Currently, Polars supports defining custom functions on the DataFrame namespace on the python side with
pl.api.register_dataframe_namespace
/pl.api.register_lazyframe_namespace
. Would it be possible for pyo3-polars to support creating a custom DataFrame namespace with this same functionality, with the defined transformations living in rust?For a somewhat abstract example, I have a function written in rust that takes a pl.DataFrame as an argument, loads a xgboost Booster, calculates a column, then returns the DataFrame. I want to expose this exact same functionality in Python by registering some custom namespace.