Embedding Hadrian

The Hadrian Basic Use page provides enough information to embed the Hadrian library in an application. This tutorial acts as a check-list with an explicit example.

Before you begin...

This article was tested with Hadrian 0.8.1; newer versions should work with no modification. Scala >= 2.10 and sbt is required. Download sbt here and test it by typing sbt on the command line.

Also download the Iris dataset and a simple tree classification for it.

Example application

In this tutorial, we will write a Scala application that runs a PFA engine on data from a CSV file, writing the result as a CSV file.

Create directory minihadrian and add a file named build.sbt with the following contents:

libraryDependencies += "com.opendatagroup" % "hadrian" % "0.8.1"
resolvers += "opendatagroup" at "http://repository.opendatagroup.com/maven"

Next, add a file named 'minihadrian.scala' with the following contents:

import com.opendatagroup.hadrian.jvmcompiler.defaultPFAVersion

object MiniHadrian {
  def main(args: Array[String]) {
    println(s"PFA version $defaultPFAVersion")
  }
}

Run sbt in the minihadrian directory (close and restart it if it had been running before adding build.sbt). You should see

% sbt
[info] Set current project to minihadrian (in build file:/tmp/minihadrian/)
> run
[info] Running MiniHadrian 
PFA version 0.8.1
[success] Total time: 1 s, completed Nov 17, 2015 2:59:52 PM
> 

(possibly with different version numbers).

Adding a PFA engine reader