org.apache.spark.ml.bundle.BundleHelper.scala Maven / Gradle / Ivy
The newest version!
package org.apache.spark.ml.bundle
/**
* Created by hollinwilkins on 5/25/17.
*/
object BundleHelper {
def sampleDataframeMessage(klazz: Class[_]): String = {
s"""
|*****************************************************************************************************
|Must provide a sample data frame for the ${klazz.getCanonicalName} transformer.
|
|See more information here:
|https://combust.github.io/mleap-docs//troubleshooting.html#must-provide-a-sample-dataset-for-the-x-transformer
|
|// Use your Spark ML Pipeline to transform the Spark DataFrame
|val transformedDataset = sparkTransformer.transform(sparkDataset)
|
|// Create a custom SparkBundleContext and provide the transformed DataFrame
|implicit val sbc = SparkBundleContext().withDataset(transformedDataset)
|
|// Serialize the pipeline as you would normally
|Using(BundleFile(file))) { bf =>
| sparkTransformer.writeBundle.save(bf).get
|}.get
|*****************************************************************************************************
""".stripMargin
}
}