zhengruifeng commented on issue #25046: [SPARK-28243][PYSPARK][ML] Remove setFeatureSubsetStrategy and setSubsamplingRate from Python TreeEnsembleParams URL: https://github.com/apache/spark/pull/25046#issuecomment-520119051 @huaxingao @srowen @mgaido91 I agree that we should remove those setter from the py side. However, we should not directly touch `param/shared.py`, instead we have to modify `_shared_params_code_gen.py` and then run `python _shared_params_code_gen.py > shared.py`. This is caused by that the `_shared_params_code_gen.py` will automatic generate both the setter and the getter, while in the scala side, only getter is generated. And in the scala side, `DecisionTreeParams` is not placed in `shareParam.scala`. There are too many design conflicts between the class hierarchy of scala and py, it's too confusing that can not be maintained easily. Maybe it is time to re-org the py side to keep it in line the scala side. I found this when I'm adding `Implement Tree-Based Feature Transformation` https://github.com/apache/spark/pull/25383 in the py side.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
