SparkSession Creation fix for compatibility with pySpark (bigdatageno…

…mics#21)

bin/pygnocchi

@@ -30,14 +30,14 @@ else
   ASSEMBLY_DIR="$SCRIPT_DIR/gnocchi-assembly/target"
 fi
 
-num_jars="$(ls -1 "$ASSEMBLY_DIR" | grep "^gnocchi-assembly_[0-9A-Za-z\.-]*\.jar$" | grep -v javadoc | grep -v sources | wc -l)"
+num_jars="$(ls -1 "$ASSEMBLY_DIR" | grep "^gnocchi[0-9A-Za-z\.\_\-]*\.jar$" | grep -v javadoc | grep -v sources | wc -l)"
 if [ "$num_jars" -eq "0" ]; then
-  echo "Failed to find Gnocchi assembly in $ASSEMBLY_DIR." 1>&2
+  echo "Failed to find Gnocchi cli assembly in $ASSEMBLY_DIR." 1>&2
   echo "You need to build Gnocchi before running this program." 1>&2
   exit 1
 fi
 
-ASSEMBLY_JARS="$(ls -1 "$ASSEMBLY_DIR" | grep "^gnocchi-assembly_[0-9A-Za-z\.-]*\.jar$" | grep -v javadoc | grep -v sources || true)"
+ASSEMBLY_JARS="$(ls -1 "$ASSEMBLY_DIR" | grep "^gnocchi[0-9A-Za-z\.\_\-]*\.jar$" | grep -v javadoc | grep -v sources || true)"
 if [ "$num_jars" -gt "1" ]; then
   echo "Found multiple Gnocchi cli assembly jars in $ASSEMBLY_DIR:" 1>&2
   echo "$ASSEMBLY_JARS" 1>&2

gnocchi-core/src/main/scala/net/fnothaft/gnocchi/sql/GnocchiSession.scala

@@ -32,9 +32,22 @@ import scala.collection.JavaConversions._
 object GnocchiSession {
 
   // Add GnocchiContext methods
-  implicit def sparkContextToGnocchiContext(sc: SparkContext): GnocchiSession =
+  implicit def sparkContextToGnocchiSession(sc: SparkContext): GnocchiSession =
     new GnocchiSession(sc)
 
+  /**
+   * Creates a GnocchiSession from SparkSession. Sets the active session to the
+   * input session.
+   *
+   *
+   * @param ss SparkSession
+   * @return GnocchiSession
+   */
+  def GnocchiSessionFromSession(ss: SparkSession): GnocchiSession = {
+    SparkSession.setActiveSession(ss)
+    new GnocchiSession(ss.sparkContext)
+  }
+
 }
 
 class GnocchiSession(@transient val sc: SparkContext) extends Serializable with Logging {
@@ -293,4 +306,4 @@ class GnocchiSession(@transient val sc: SparkContext) extends Serializable with
 
     phenoCovarDF.withColumn("phenoName", lit(phenoName)).as[Phenotype].collect().map(x => (x.sampleId, x)).toMap
   }
-}
+}

gnocchi-python/bdgenomics/gnocchi/gnocchiSession.py

@@ -20,11 +20,11 @@ class GnocchiSession(object):
     """
     """
 
-    def __init__(self, sc):
-        self._sc = sc
-        self._jvm = sc._jvm
-        session = sc._jvm.net.fnothaft.gnocchi.sql.GnocchiSession(sc._jsc.sc())
-        self.__jgs = sc._jvm.net.fnothaft.gnocchi.api.java.JavaGnocchiSession(session)
+    def __init__(self, ss):
+        self._sc = ss.sparkContext
+        self._jvm = self._sc._jvm
+        session = self._jvm.net.fnothaft.gnocchi.sql.GnocchiSession.GnocchiSessionFromSession(ss._jsparkSession)
+        self.__jgs = self._jvm.net.fnothaft.gnocchi.api.java.JavaGnocchiSession(session)
 
     def filterSamples(self, genotypesDataset, mind, ploidy):
         dataset = self.__jgs.filterSamples(genotypesDataset.get(), mind, ploidy)