[SPARK-1267]Adding a pip installer setup file for PySpark.

python/pyspark/__init__.py

@@ -53,8 +53,27 @@
 
 import sys
 import os
-sys.path.insert(0, os.path.join(os.environ["SPARK_HOME"], "python/lib/py4j-0.8.1-src.zip"))
+import xml.etree.ElementTree as ET
+
+if (os.environ.get("SPARK_HOME", "not found") == "not found"):
+	raise ImportError("Environment variable SPARK_HOME is undefined.")
+
+spark_home = os.environ['SPARK_HOME']
+pom_xml_file_path = spark_home + '/pom.xml'
 
+try:
+	tree = ET.parse(pom_xml_file_path)
+	root = tree.getroot()
+	version_tag = root[4].text
+	snapshot_version = version_tag[:5]
+except:
+	raise ImportError("Could not read the spark version, because pom.xml file is not found in SPARK_HOME(%s) directory." %(spark_home))
+
+from pyspark.pyspark_version import __version__
+if (snapshot_version != __version__):
+	raise ImportError("Incompatible version of Spark(%s) and PySpark(%s)." %(snapshot_version, __version__))
+
+sys.path.insert(0, os.path.join(os.environ["SPARK_HOME"], "python/lib/py4j-0.8.1-src.zip"))
 
 from pyspark.conf import SparkConf
 from pyspark.context import SparkContext

python/pyspark/pyspark_version.py

@@ -0,0 +1 @@
+__version__='1.0.0'

python/setup.py

@@ -0,0 +1,19 @@
+#!/usr/bin/env python
+
+from setuptools import setup
+
+exec(compile(open("pyspark/pyspark_version.py").read(), 
+	"pyspark/pyspark_version.py", 'exec'))
+VERSION = __version__
+
+setup(name = 'pyspark',
+	version = VERSION,
+	description = 'Apache Spark Python API',
+	author = 'Prabin Banka',
+	author_email = '[email protected]',
+	url = 'https://github.com/apache/spark/tree/master/python',
+	packages = ['pyspark', 'pyspark.mllib'],
+	data_files = [('pyspark', ['pyspark/pyspark_version.py'])],
+	install_requires = ['numpy>=1.7', 'py4j==0.8.1'],
+	license = 'http://www.apache.org/licenses/LICENSE-2.0',
+	)