Add a stage (an Estimator or a Transformer) to an existing spark.ml Pipeline
Add a stage (an Estimator or a Transformer) to an existing spark.ml Pipeline
a spark.ml PipelineStage (an Estimator or Transformer)
the input Pipeline, with the added stage
Apply a spark.ml PipelineModel to a DataFrame to make a prediction.
Apply a spark.ml PipelineModel to a DataFrame to make a prediction. This op is intended to be used after merging two Pipes, one providng the spark.ml PipelineModel, and the other providing the DataFrame. For example, Pipe(mlModelPipe, dataPipe).to(ops.core.ml.applyModel)
a Tuple2 including a spark.ml PipelineModel and a DataFrame
the resultant DataFrame
Stub object necessary due to https://issues.scala-lang.org/browse/SI-8124
Stub object necessary due to https://issues.scala-lang.org/browse/SI-8124
Documentation for ops.core.ml.pipeline
can be found at software.uncharted.sparkpipe.ops.core.ml.pipeline
Fit a spark.ml pipeline to a DataFrame to produce a PipelineModel.
Fit a spark.ml pipeline to a DataFrame to produce a PipelineModel. This op is intended to be used after merging two Pipes, one providng the spark.ml Pipeline, and the other providing the DataFrame. For example, Pipe(mlPipelinePipe, dataPipe).to(ops.core.ml.fit)
a Tuple2 including a spark.ml Pipeline and a DataFrame
the fitted spark.ml PipelineModel
Load a spark.ml Pipeline from a file
Load a spark.ml Pipeline from a file
the SparkContext
the path to the persisted Pipeline
a spark.ml Pipeline constructed from the given file
java.lang.UnsupportedOperationException
on spark version < 1.6.0
Persist a spark.ml Pipeline to a file
Persist a spark.ml Pipeline to a file
the SparkContext
the path for the persisted Pipeline file
the input spark.ml Pipeline, unchanged
java.lang.UnsupportedOperationException
on spark version < 1.6.x
Lightweight helpers for using spark.ml Pipelines with sparkpipe. The goal is not to replace the spark.ml pipeline, but rather to smooth its integration with other logic and libraries supporting the sparkpipe ops format.