From 28c0518379d226ac25597c2840c5c81bd8551487 Mon Sep 17 00:00:00 2001 From: Ellen Spertus Date: Thu, 9 Aug 2018 20:26:31 -0700 Subject: WIP --- scalding/src/main/scala/sandcrawler/CrossrefScorable.scala | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'scalding/src/main/scala/sandcrawler/CrossrefScorable.scala') diff --git a/scalding/src/main/scala/sandcrawler/CrossrefScorable.scala b/scalding/src/main/scala/sandcrawler/CrossrefScorable.scala index b221718..249c9ab 100644 --- a/scalding/src/main/scala/sandcrawler/CrossrefScorable.scala +++ b/scalding/src/main/scala/sandcrawler/CrossrefScorable.scala @@ -10,10 +10,13 @@ import parallelai.spyglass.hbase.HBasePipeConversions import parallelai.spyglass.hbase.HBaseSource class CrossrefScorable extends Scorable { - def getFeaturesPipe(args : Args) : TypedPipe[MapFeatures] = { - // TODO: Generalize args so there can be multiple Grobid pipes in one job. + // TODO: Generalize args so there can be multiple Grobid pipes in one job. + def getSource(args : Args) : Source = { TextLine(args("crossref-input")) - .read + } + + def getFeaturesPipe(pipe : Pipe) : TypedPipe[MapFeatures] = { + pipe .toTypedPipe[String](new Fields("line")) .map{ json : String => CrossrefScorable.crossrefToSlug(json) match { -- cgit v1.2.3