complete_example.scala

import java.sql.Timestamp
import java.time.{LocalDateTime, ZoneId, ZonedDateTime}
import com.cloudera.sparkts._
import org.apache.spark.sql.SparkSession
import org.apache.spark.sql.functions.udf
import org.apache.spark.sql.functions._

/**
  * Created by josep2 on 12/5/16.
  */


case class TimeSeries(id: String, week_beginning: String, sales: Double)
object TimeSeriesBlog extends App {

val sparkSession = SparkSession.builder
  .master("local[4]")
  .appName("Panel Engine")
  .config("spark.driver.memory", "1g")
  .config("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
  .getOrCreate()

import sparkSession.implicits._



var TimeSeriesData = sparkSession.sparkContext.textFile("./data.txt")
  .map(_.split(","))
  .map(row => TimeSeries(row(0), row(1),  row(2).toDouble))
  .toDF()

def makeDate = udf((s: String) => Timestamp.from(ZonedDateTime.of(LocalDateTime.parse(s + "T00:00:00"), ZoneId.systemDefault()).toInstant))

TimeSeriesData = TimeSeriesData.withColumn("week_beginning", makeDate(TimeSeriesData("week_beginning")))


val zone = ZoneId.systemDefault()
val dtIndex = DateTimeIndex.uniformFromInterval(
    ZonedDateTime.of(LocalDateTime.parse("2016-01-04T00:00:00"), zone),
    ZonedDateTime.of(LocalDateTime.parse("2016-04-04T00:00:00"), zone),
    new DayFrequency(7))


val dataRdd = TimeSeriesRDD.timeSeriesRDDFromObservations(dtIndex, TimeSeriesData,
    "week_beginning", "id", "sales")
  
  sparkSession.sparkContext.stop()
}