[ 
https://issues.apache.org/jira/browse/SPARK-61633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Ruifeng Zheng resolved SPARK-61633.
-----------------------------------
    Resolution: Fixed

Issue resolved by pull request 637
[https://github.com/apache/spark-connect-rust/pull/637]

> Support the remaining native clustering algorithms: LDA and 
> PowerIterationClustering
> ------------------------------------------------------------------------------------
>
>                 Key: SPARK-61633
>                 URL: https://issues.apache.org/jira/browse/SPARK-61633
>             Project: Spark
>          Issue Type: Improvement
>          Components: Connect
>    Affects Versions: connect-rust-4.3.0
>            Reporter: Ruifeng Zheng
>            Assignee: Ruifeng Zheng
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: connect-rust-4.3.0
>
>
> Complete native clustering algorithm execution by adding the remaining 
> PySpark 4.3.0 algorithms: LDA and PowerIterationClustering.
> LDA should support both online and EM optimizers, all 14 constructor 
> parameters and setters, cached fitted models and DataFrame transforms. 
> PowerIterationClustering should expose its six parameters and assignClusters, 
> using Spark Connect's powerIterationClusteringAssignClusters helper instead 
> of an estimator fit API. Implement both the Rust API and native Python 
> bindings, preserving the Spark-DataFrame-only input contract.
> Validate parameter serialization and helper requests against official PySpark 
> fixtures, and run Rust and native Python live tests for inference, graph 
> assignments, parameter validation and model lifetime. This completes 
> algorithm execution coverage; richer model inspection and persistence remain 
> separate work.



--
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