aboutsummaryrefslogtreecommitdiffstats
path: root/scalding/src/main
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2018-09-13 21:11:57 -0700
committerBryan Newbold <bnewbold@archive.org>2018-09-13 21:11:57 -0700
commit5521b1b520a550373369da8b9cbd36148e071115 (patch)
tree058c20660cb6b1c2ac0a8c23070681092e6fbc7a /scalding/src/main
parent6681ee7d699fc481b3dc0e1e8f905395a0b42a3b (diff)
downloadsandcrawler-5521b1b520a550373369da8b9cbd36148e071115.tar.gz
sandcrawler-5521b1b520a550373369da8b9cbd36148e071115.zip
new simple file metadata dump script
Diffstat (limited to 'scalding/src/main')
-rw-r--r--scalding/src/main/scala/sandcrawler/DumpFileMetaJob.scala36
1 files changed, 36 insertions, 0 deletions
diff --git a/scalding/src/main/scala/sandcrawler/DumpFileMetaJob.scala b/scalding/src/main/scala/sandcrawler/DumpFileMetaJob.scala
new file mode 100644
index 0000000..b3734f0
--- /dev/null
+++ b/scalding/src/main/scala/sandcrawler/DumpFileMetaJob.scala
@@ -0,0 +1,36 @@
+package sandcrawler
+
+import java.util.Properties
+
+import cascading.property.AppProps
+import cascading.tuple.Fields
+import com.twitter.scalding._
+import com.twitter.scalding.typed.TDsl._
+import org.apache.hadoop.hbase.io.ImmutableBytesWritable
+import org.apache.hadoop.hbase.util.Bytes
+import parallelai.spyglass.base.JobBase
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+import parallelai.spyglass.hbase.HBasePipeConversions
+import parallelai.spyglass.hbase.HBaseSource
+
+// Dumps all the info needed to insert a file entity in Fatcat. Useful for
+// joining.
+class DumpFileMetaJob(args: Args) extends JobBase(args) with HBasePipeConversions {
+
+ val metaPipe : TypedPipe[(String, String, String, Long)] = HBaseBuilder.build(args("hbase-table"),
+ args("zookeeper-hosts"),
+ List("file:cdx", "file:mime", "file:size"),
+ SourceMode.SCAN_ALL)
+ .read
+ .toTypedPipe[(ImmutableBytesWritable,ImmutableBytesWritable,ImmutableBytesWritable,ImmutableBytesWritable)](new Fields("key", "cdx", "mime", "size"))
+ .filter { case (_, cdx, mime, size) => cdx != null && mime != null && size != null }
+ .map { case (key, cdx, mime, size) =>
+ (Bytes.toString(key.copyBytes()),
+ Bytes.toString(cdx.copyBytes()),
+ Bytes.toString(mime.copyBytes()),
+ Bytes.toLong(size.copyBytes()))
+ };
+
+ metaPipe.write(TypedTsv[(String,String,String,Long)](args("output")))
+
+}