Changed HBaseSource and JDBCSource to allow testing with JobTest. Samples of tests included.

2 files changed, 111 insertions, 0 deletions

diff --git a/src/test/scala/parallelai/spyglass/hbase/example/SimpleHBaseSourceExampleTest.scala b/src/test/scala/parallelai/spyglass/hbase/example/SimpleHBaseSourceExampleTest.scala
new file mode 100644
index 0000000..e5b2cca
--- /dev/null
+++ b/src/test/scala/parallelai/spyglass/hbase/example/SimpleHBaseSourceExampleTest.scala
@@ -0,0 +1,57 @@
+package parallelai.spyglass.hbase.example
+
+import org.junit.runner.RunWith
+import com.twitter.scalding.{JobTest, TupleConversions}
+import org.scalatest.FunSpec
+import org.scalatest.junit.JUnitRunner
+import org.slf4j.LoggerFactory
+import org.apache.hadoop.hbase.io.ImmutableBytesWritable
+import cascading.tuple.{Tuple, Fields}
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+import org.apache.hadoop.hbase.util.Bytes
+import scala._
+import parallelai.spyglass.hbase.HBaseSource
+import com.twitter.scalding.Tsv
+
+/**
+ * Example of how to define tests for HBaseSource
+ */
+@RunWith(classOf[JUnitRunner])
+class SimpleHBaseSourceExampleTest extends FunSpec with TupleConversions {
+
+  val output = "/tmp/testOutput"
+
+  val log = LoggerFactory.getLogger(this.getClass.getName)
+
+  val sampleData = List(
+    List("1", "kk1", "pp1"),
+    List("2", "kk2", "pp2"),
+    List("3", "kk3", "pp3")
+  )
+
+  JobTest("parallelai.spyglass.hbase.example.SimpleHBaseSourceExample")
+    .arg("local", "")
+    .arg("app.conf.path", "app.conf")
+    .arg("output", output)
+    .arg("debug", "true")
+    .source[Tuple](
+    new HBaseSource(
+      "table_name",
+      "quorum_name:2181",
+      new Fields("key"),
+      List("column_family"),
+      List(new Fields("column_name1", "column_name2")),
+      sourceMode = SourceMode.GET_LIST, keyList = List("1", "2", "3")),
+    sampleData.map(l => new Tuple(l.map(s => {new ImmutableBytesWritable(Bytes.toBytes(s))}):_*)))
+    .sink[Tuple](Tsv(output format "get_list")) {
+      outputBuffer =>
+        log.debug("Output => " + outputBuffer)
+
+        it("should return the test data provided.") {
+          assert(outputBuffer.size === 3)
+        }
+    }
+    .run
+    .finish
+
+}
diff --git a/src/test/scala/parallelai/spyglass/jdbc/example/JdbcSourceExampleTest.scala b/src/test/scala/parallelai/spyglass/jdbc/example/JdbcSourceExampleTest.scala
new file mode 100644
index 0000000..98f275d
--- /dev/null
+++ b/src/test/scala/parallelai/spyglass/jdbc/example/JdbcSourceExampleTest.scala
@@ -0,0 +1,54 @@
+package parallelai.spyglass.jdbc.example
+
+import org.scalatest.junit.JUnitRunner
+import org.junit.runner.RunWith
+import org.scalatest.FunSpec
+import com.twitter.scalding.{Tsv, JobTest, TupleConversions}
+import org.slf4j.LoggerFactory
+import cascading.tuple.{Tuple, Fields}
+import parallelai.spyglass.jdbc.JDBCSource
+
+/**
+ * Simple example of JDBCSource testing
+ */
+@RunWith(classOf[JUnitRunner])
+class JdbcSourceExampleTest extends FunSpec with TupleConversions {
+
+  val output = "src/test/resources/outputs/testOutput"
+
+  val log = LoggerFactory.getLogger(this.getClass.getName)
+
+  val sampleData = List(
+    (1, "c11", "c12", 11),
+    (2, "c21", "c22", 22)
+  )
+
+  JobTest("parallelai.spyglass.jdbc.example.JdbcSourceExample")
+    .arg("local", "")
+    .arg("app.conf.path", "app.conf")
+    .arg("output", output)
+    .arg("debug", "true")
+    .source(
+    new JDBCSource(
+      "db_name",
+      "com.mysql.jdbc.Driver",
+      "jdbc:mysql://<hostname>:<port>/<db_name>?zeroDateTimeBehavior=convertToNull",
+      "user",
+      "password",
+      List("KEY_ID", "COL1", "COL2", "COL3"),
+      List("bigint(20)", "varchar(45)", "varchar(45)", "bigint(20)"),
+      List("key_id"),
+      new Fields("key_id", "col1", "col2", "col3")), sampleData)
+    .sink[Tuple](Tsv(output.format("get_list"))) {
+      outputBuffer =>
+        log.debug("Output => " + outputBuffer)
+
+        it("should return the mock data provided.") {
+          assert(outputBuffer.size === 2)
+        }
+    }
+    .run
+    .finish
+
+
+}