From a5a3df30bad2250b4f50e6234dc810efaeaceb2f Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Tue, 22 May 2018 12:07:01 -0700
Subject: more tests (failing)

---
 .../example/SimpleHBaseSourceExampleTest.scala     |  2 +-
 .../test/scala/sandcrawler/HBaseRowCountTest.scala | 55 ++++++++++++++++++++++
 2 files changed, 56 insertions(+), 1 deletion(-)
 create mode 100644 scald-mvp/src/test/scala/sandcrawler/HBaseRowCountTest.scala

diff --git a/scald-mvp/src/test/scala/example/SimpleHBaseSourceExampleTest.scala b/scald-mvp/src/test/scala/example/SimpleHBaseSourceExampleTest.scala
index d416af8..41c297d 100644
--- a/scald-mvp/src/test/scala/example/SimpleHBaseSourceExampleTest.scala
+++ b/scald-mvp/src/test/scala/example/SimpleHBaseSourceExampleTest.scala
@@ -29,7 +29,7 @@ class SimpleHBaseSourceExampleTest extends FunSpec with TupleConversions {
     List("3", "kk3", "pp3")
   )
 
-  JobTest("parallelai.spyglass.hbase.example.SimpleHBaseSourceExample")
+  JobTest("example.SimpleHBaseSourceExample")
     .arg("test", "")
     .arg("app.conf.path", "app.conf")
     .arg("output", output)
diff --git a/scald-mvp/src/test/scala/sandcrawler/HBaseRowCountTest.scala b/scald-mvp/src/test/scala/sandcrawler/HBaseRowCountTest.scala
new file mode 100644
index 0000000..c46b0fc
--- /dev/null
+++ b/scald-mvp/src/test/scala/sandcrawler/HBaseRowCountTest.scala
@@ -0,0 +1,55 @@
+package 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 org.apache.hadoop.hbase.util.Bytes
+import scala._
+import com.twitter.scalding.Tsv
+import parallelai.spyglass.hbase.HBaseSource
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+
+/**
+ * Example of how to define tests for HBaseSource
+ */
+@RunWith(classOf[JUnitRunner])
+class HBaseRowCountTest extends FunSpec with TupleConversions {
+
+  val output = "/tmp/testOutput"
+
+  val log = LoggerFactory.getLogger(this.getClass.getName)
+
+  val sampleData = List(
+    List("sha1:K2DKSSVTXWPRMFDTWSTCQW3RVWRIOV3Q"),
+    List("sha1:C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ3WU")
+  )
+
+  JobTest("sandcrawler.HBaseRowCountJob")
+    .arg("test", "")
+    .arg("app.conf.path", "app.conf")
+    .arg("output", output)
+    .arg("debug", "true")
+    .source[Tuple](
+    new HBaseSource(
+      "table_name",
+      "quorum_name:2181",
+      new Fields("key"),
+      sourceMode = SourceMode.GET_LIST, keyList = List("sha1:K2DKSSVTXWPRMFDTWSTCQW3RVWRIOV3Q", "sha1:C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ3WU", "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.") {
+          println("outputBuffer.size => " + outputBuffer.size)
+          assert(outputBuffer.size === 3)
+        }
+    }
+    .run
+    .finish
+
+}
-- 
cgit v1.2.3