From df341a68459829380f1f01015768acee5642f15b Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Wed, 15 Aug 2018 20:20:43 -0700 Subject: grobid scoring: status_code as signed int, not string --- scalding/src/main/scala/sandcrawler/GrobidScorable.scala | 9 +++++++-- scalding/src/test/scala/sandcrawler/ScoreJobTest.scala | 5 +++-- 2 files changed, 10 insertions(+), 4 deletions(-) (limited to 'scalding') diff --git a/scalding/src/main/scala/sandcrawler/GrobidScorable.scala b/scalding/src/main/scala/sandcrawler/GrobidScorable.scala index c319fe6..f484fad 100644 --- a/scalding/src/main/scala/sandcrawler/GrobidScorable.scala +++ b/scalding/src/main/scala/sandcrawler/GrobidScorable.scala @@ -6,6 +6,8 @@ import cascading.flow.FlowDef 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.hbase.HBaseConstants.SourceMode import parallelai.spyglass.hbase.HBasePipeConversions import parallelai.spyglass.hbase.HBaseSource @@ -21,8 +23,11 @@ class GrobidScorable extends Scorable with HBasePipeConversions { def getFeaturesPipe(args : Args)(implicit mode : Mode, flowDef : FlowDef) : TypedPipe[MapFeatures] = { getSource(args) .read - .fromBytesWritable(new Fields("key", "tei_json", "status_code")) - .toTypedPipe[(String, String, Int)](new Fields("key", "tei_json", "status_code")) + // Can't just "fromBytesWritable" because we have multiple types? + .toTypedPipe[(ImmutableBytesWritable,ImmutableBytesWritable,ImmutableBytesWritable)](new Fields("key", "tei_json", "status_code")) + .map { case (key, tei_json, status_code) => + (Bytes.toString(key.copyBytes()), Bytes.toString(tei_json.copyBytes()), Bytes.toLong(status_code.copyBytes())) + } // TODO: Should I combine next two stages for efficiency? .collect { case (key, json, StatusOK) => (key, json) } .map { entry : (String, String) => GrobidScorable.jsonToMapFeatures(entry._1, entry._2) } diff --git a/scalding/src/test/scala/sandcrawler/ScoreJobTest.scala b/scalding/src/test/scala/sandcrawler/ScoreJobTest.scala index 34081a5..f68ee1d 100644 --- a/scalding/src/test/scala/sandcrawler/ScoreJobTest.scala +++ b/scalding/src/test/scala/sandcrawler/ScoreJobTest.scala @@ -150,8 +150,9 @@ class ScoreJobTest extends FlatSpec with Matchers { JsonString.replace("<>", "Title 2") ) - val Ok = "200" - val Bad = "400" + // bnewbold: status codes aren't strings, they are uint64 + val Ok : Long = 200 + val Bad : Long = 400 val StatusCodes = List(Ok, Ok, Ok, Bad, Ok, Bad) val SampleData : List[List[Array[Byte]]] = (Sha1Strings, JsonStrings, StatusCodes) -- cgit v1.2.3