FPGrowth Performance Test Main

import org.apache.log4j.{Level, Logger}
import org.apache.spark.ml.fpm.FPGrowth
import org.apache.spark.sql.SparkSession

object KosarakTest {

  def main(args: Array[String]): Unit = {

    Logger.getLogger("org").setLevel(Level.WARN)

    val spark = SparkSession.builder().master("local[4]").appName("sss").getOrCreate()
    val sc = spark.sparkContext
    import spark.implicits._

    //http://fimi.ua.ac.be/data/kosarak.dat
    val df = sc.textFile("/home/yuhao/workspace/data/fpm/kosarak.dat", 4)
      .map(_.trim().split(" ").map(_.toInt).distinct).map(a => (a)).toDF()
    df.show()
    val model = new FPGrowth()
      .setFeaturesCol("value")
      .setMinSupport(0.1)
      .fit(df)

    model.freqItemsets.show()
    model.associationRules.show()

    val st = System.nanoTime()
    model.transform(df).show()
    println((System.nanoTime() - st) / 1e9)

    spark.stop()
  }
}