{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Spark Magic\n",
    "\n",
    "BeakerX has a Spark magic that provides deeper integration with Spark.  It provides a GUI dialog for connecting to a cluster, a progress meter that shows how your job is working and links to the regular Spark UI, and it forwards kernel interrupt messages onto the cluster so you can stop a job without leaving the notebook, and it automatically displays Datasets using an interactive widget.  Finally, it automatically closes the Spark session when the notebook is closed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "%%classpath add mvn\n",
    "org.apache.spark spark-sql_2.11 2.2.1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The spark cell magic can be run all by itself in a cell.  It produces a GUI dialog you fill out to connect to your cluster. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%spark"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Optionally, the contents of the cell can produce a Spark session to fill out default values for the GUI.  Only one spark magic can be active at a time."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%spark\n",
    "SparkSession.builder()\n",
    "      .appName(\"BeakerX Demo\")\n",
    "      .master(\"local[4]\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can also provide a `--connect` (or `-c`) option to automatically connect with the cluster."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%spark --connect\n",
    "SparkSession.builder().master(\"local[100]\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "val NUM_SAMPLES = 10000000\n",
    "\n",
    "val count2 = spark.sparkContext.parallelize(1 to NUM_SAMPLES).map{i =>\n",
    "  val x = Math.random()\n",
    "  val y = Math.random()\n",
    "  if (x*x + y*y < 1) 1 else 0\n",
    "}.reduce(_ + _)\n",
    "\n",
    "println(\"Pi is roughly \" + 4.0 * count2 / NUM_SAMPLES)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default the first 1000 rows are materialized to preview a dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "val tornadoesPath = java.nio.file.Paths.get(\"../resources/data/tornadoes_2014.csv\").toAbsolutePath()\n",
    "\n",
    "val ds = spark.read.format(\"csv\").option(\"header\", \"true\").load(\"file://\" + tornadoesPath)\n",
    "ds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Or you can use the display method to specify any number of rows."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ds.display(1)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Scala",
   "language": "scala",
   "name": "scala"
  },
  "language_info": {
   "codemirror_mode": "text/x-scala",
   "file_extension": ".scala",
   "mimetype": "",
   "name": "Scala",
   "nbconverter_exporter": "",
   "version": "2.11.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}