# Maintainer: Tao-Yi Lee <tylee@ieee.org>
pkgname=apache_spark
_pkgname=spark
pkgver=1.6.2
_fullpkgname="${_pkgname}-${pkgver}-bin-hadoop2.6"
pkgrel=1
pkgdesc="Apache Spark™ is a fast and general engine for large-scale data processing."
arch=('i686' 'x86_64')
url="http://spark.apache.org"
license=('apache')
groups=()
depends=('java-runtime-common')
provides=()
conflicts=()
replaces=()
backup=()
options=(!debug !strip)
install=
source=("http://mirror.nexcess.net/apache/$_pkgname/$_pkgname-$pkgver/$_fullpkgname.tgz"
"https://gist.githubusercontent.com/thrasibule/87a092d244d560aa5d081ea6a0a0db4c/raw/ef32a1fcad2306fbccca33a36c1fb6ccb3dee405/kernel.json")
noextract=("$_fullpkgname.tgz")
md5sums=('604936f2bd8af999d0d624b370f5c4b1'
         'e3137862727df13ccf1194248a4f50aa')
PKGEXT='.pkg.tar.gz'

package() {
  _etc_profiled=${pkgdir}/etc/profile.d
  mkdir -p "${pkgdir}/opt"
  mkdir -p $_etc_profiled

  # install spark
  msg "Spark will be installed under /opt/$_fullpkgname"
  tar -xf $_fullpkgname.tgz -C ${pkgdir}/opt
  pushd "${pkgdir}/opt/$_fullpkgname/python/lib/" > /dev/null
  rm py4j-0.9-src.zip PY4J_LICENSE.txt
  popd > /dev/null
  sed -i '/py4j/d' "${pkgdir}/opt/$_fullpkgname/bin/pyspark"
  echo "export PATH=\$PATH:/opt/$_fullpkgname/bin" >> \
    "${_etc_profiled}/${pkgname}.sh"
  echo "export SPARK_HOME=/opt/$_fullpkgname" >> \
    "${_etc_profiled}/${pkgname}.sh"
  sed -i "s|\$SPARK_HOME|${SPARK_HOME}|;s|\$pkgver|${pkgver}|" kernel.json
  mkdir -p "${pkgdir}/usr/share/jupyter/kernels/pyspark"
  cp kernel.json "${pkgdir}/usr/share/jupyter/kernels/pyspark"
}

# vim:set ts=2 sw=2 et: