Hyukjin Kwon created SPARK-61762:
------------------------------------

             Summary: Re-export PySpark's readers and writers of the Python 
format in the native ML modules
                 Key: SPARK-61762
                 URL: https://issues.apache.org/jira/browse/SPARK-61762
             Project: Spark
          Issue Type: Improvement
          Components: Connect
    Affects Versions: connect-rust-4.3.0
            Reporter: Hyukjin Kwon


PySpark's modules pyspark.ml.pipeline, pyspark.ml.tuning and 
pyspark.ml.classification have readers and writers of the Python format that 
their __all__ does not name (PipelineReader, PipelineWriter, 
PipelineModelReader, PipelineModelWriter, PipelineSharedReadWrite; the 
CrossValidator and TrainValidationSplit readers and writers of the estimators 
and the models; the OneVsRest ones), which the native modules of 
pyspark.ml.connect lack.

Over the param maps of the native classes (SPARK-61752), PySpark's own classes 
save and load the native objects, so this re-exports them in 
pyspark.ml.connect.pipeline, tuning and classification.




--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to