aboutsummaryrefslogtreecommitdiffstats
path: root/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala
diff options
context:
space:
mode:
Diffstat (limited to 'scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala')
-rw-r--r--scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala61
1 files changed, 61 insertions, 0 deletions
diff --git a/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala b/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala
new file mode 100644
index 0000000..8b5c3d6
--- /dev/null
+++ b/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala
@@ -0,0 +1,61 @@
+package sandcrawler
+
+import cascading.tuple.{Tuple, Fields}
+import com.twitter.scalding.{JobTest, Tsv, TupleConversions}
+import org.apache.hadoop.hbase.io.ImmutableBytesWritable
+import org.apache.hadoop.hbase.util.Bytes
+import org.junit.runner.RunWith
+import org.scalatest.FunSpec
+import org.scalatest.junit.JUnitRunner
+import org.slf4j.LoggerFactory
+import parallelai.spyglass.hbase.HBaseSource
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+import scala._
+
+@RunWith(classOf[JUnitRunner])
+class HBaseStatusCountTest extends FunSpec with TupleConversions {
+
+ val output = "/tmp/testOutput"
+
+ val log = LoggerFactory.getLogger(this.getClass.getName)
+
+ val statusType1 = "200"
+ val statusType2 = "404"
+
+ val sampleData = List(
+ List("sha1:K2DKSSVTXWPRMFDTWSTCQW3RVWRIOV3Q", statusType1),
+ List("sha1:C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ3WU", statusType1),
+ List("sha1:SDKUVHC3YNNEGH5WAG5ZAAXWAEBNX4WT", statusType2),
+ List("sha1:35985C3YNNEGH5WAG5ZAAXWAEBNXJW56", statusType2),
+ List("sha1:885C3YNNEGH5WAG5ZAAXWA8BNXJWT6CZ", statusType2),
+ List("sha1:00904C3YNNEGH5WAG5ZA9XWAEBNXJWT6", statusType2),
+ List("sha1:249C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ", statusType1),
+ List("sha1:095893C3YNNEGH5WAG5ZAAXWAEBNXJWT", statusType2)
+ )
+
+ val statusType1Count = sampleData.count(lst => lst(1) == statusType1)
+ val statusType2Count = sampleData.count(lst => lst(1) == statusType2)
+
+ JobTest("sandcrawler.HBaseStatusCountJob")
+ .arg("test", "")
+ .arg("app.conf.path", "app.conf")
+ .arg("output", output)
+ .arg("debug", "true")
+ .source[Tuple](HBaseStatusCountJob.getHBaseSource,
+ sampleData.map(l => new Tuple(l.map(s => {new ImmutableBytesWritable(Bytes.toBytes(s))}):_*)))
+ .sink[Tuple](Tsv(output)) {
+ outputBuffer =>
+ it("should return a 2-element list.") {
+ assert(outputBuffer.size === 2)
+ }
+
+ // Convert List[Tuple] to Map[String, Integer].
+ val counts = outputBuffer.map(t => (t.getString(0), t.getInteger(1))).toMap
+ it("should have the appropriate number of each status type") {
+ assert(counts(statusType1) == statusType1Count)
+ assert(counts(statusType2) == statusType2Count)
+ }
+ }
+ .run
+ .finish
+}