Spark-ml-serving

Contextless ML implementation of Spark ML.

Proposal

To serve small ML pipelines there is no need to create SparkContext and use cluster-related features. In this project we made our implementations for ML Transformers. Some of them call context-independent Spark methods.

Structure

Instead of using DataFrames, we implemented simple LocalData class to get rid of SparkContext. All Transformers are rewritten to accept LocalData.

How to use

1. Import this project as dependency:

lazy val sparkMlServingDependency = RootProject(uri("git://github.com/Hydrospheredata/spark-ml-serving.git"))

project.in(file("."))
  // your project settings
  .dependsOn(sparkMlServingDependency)

2. Use it

import io.hydrosphere.mist.api.ml.{LocalPipelineModel, PipelineLoader}
import LocalPipelineModel._

// ....
val model = PipelineLoader.load("PATH_TO_MODEL") // Load
val columns = List(LocalDataColumn("text", Seq("Hello!")))
val localData = LocalData(columns)
val result = model.transform(localData) // Transformed result

More examples of different ML models are in tests.