aboutsummaryrefslogtreecommitdiffstats
path: root/scalding/src
diff options
context:
space:
mode:
Diffstat (limited to 'scalding/src')
-rw-r--r--scalding/src/main/scala/sandcrawler/BibjsonScorable.scala50
-rw-r--r--scalding/src/main/scala/sandcrawler/MatchBenchmarkJob.scala29
2 files changed, 79 insertions, 0 deletions
diff --git a/scalding/src/main/scala/sandcrawler/BibjsonScorable.scala b/scalding/src/main/scala/sandcrawler/BibjsonScorable.scala
new file mode 100644
index 0000000..cdd598f
--- /dev/null
+++ b/scalding/src/main/scala/sandcrawler/BibjsonScorable.scala
@@ -0,0 +1,50 @@
+package sandcrawler
+
+import scala.math
+import scala.util.parsing.json.JSON
+import scala.util.parsing.json.JSONObject
+
+import cascading.flow.FlowDef
+import cascading.tuple.Fields
+import com.twitter.scalding._
+import com.twitter.scalding.typed.TDsl._
+// XXX: import parallelai.spyglass.hbase.HBasePipeConversions
+
+// XXX: class BibjsonScorable extends Scorable with HBasePipeConversions {
+
+class BibjsonScorable extends Scorable {
+
+ def getSource(args : Args) : Source = {
+ TextLine(args("bibjson-input"))
+ }
+
+ def getFeaturesPipe(args : Args)(implicit mode : Mode, flowDef : FlowDef) : TypedPipe[MapFeatures] = {
+ getSource(args).read
+ .toTypedPipe[String](new Fields("line"))
+ .map { BibjsonScorable.bibjsonToMapFeatures(_) }
+ }
+}
+
+object BibjsonScorable {
+ def bibjsonToMapFeatures(json : String) : MapFeatures = {
+ Scorable.jsonToMap(json) match {
+ case None => MapFeatures(Scorable.NoSlug, json)
+ case Some(map) => {
+ if (map contains "title") {
+ val title = Scorable.getString(map, "title")
+ val doi = Scorable.getString(map, "doi")
+ val sha1 = Scorable.getString(map, "sha")
+ // TODO: year, authors (if available)
+ if (title == null || title.isEmpty) {
+ new MapFeatures(Scorable.NoSlug, json)
+ } else {
+ val sf : ScorableFeatures = ScorableFeatures.create(title=title, doi=doi, sha1=sha1)
+ new MapFeatures(sf.toSlug, sf.toString)
+ }
+ } else {
+ new MapFeatures(Scorable.NoSlug, json)
+ }
+ }
+ }
+ }
+}
diff --git a/scalding/src/main/scala/sandcrawler/MatchBenchmarkJob.scala b/scalding/src/main/scala/sandcrawler/MatchBenchmarkJob.scala
new file mode 100644
index 0000000..1578258
--- /dev/null
+++ b/scalding/src/main/scala/sandcrawler/MatchBenchmarkJob.scala
@@ -0,0 +1,29 @@
+package sandcrawler
+
+import cascading.pipe.Pipe
+import com.twitter.scalding.Args
+import com.twitter.scalding.TypedPipe
+import com.twitter.scalding.TypedTsv
+import parallelai.spyglass.base.JobBase
+
+class MatchBenchmarkJob(args: Args) extends JobBase(args) {
+ // TODO: Instantiate any subclass of Scorable specified in args.
+ val sc1 : Scorable = new BibjsonScorable()
+ val sc2 : Scorable = new BibjsonScorable()
+ val leftArgs = args + ("bibjson-input" -> List(args("left-bibjson")))
+ val rightArgs = args + ("bibjson-input" -> List(args("right-bibjson")))
+ val pipe1 : TypedPipe[(String, ReduceFeatures)] = sc1.getInputPipe(leftArgs)
+ val pipe2 : TypedPipe[(String, ReduceFeatures)] = sc2.getInputPipe(rightArgs)
+
+ pipe1.join(pipe2).map { entry =>
+ val (slug : String, (features1 : ReduceFeatures, features2 : ReduceFeatures)) = entry
+ new ReduceOutput(
+ slug,
+ Scorable.computeSimilarity(features1, features2),
+ features1.json,
+ features2.json)
+ }
+ //TypedTsv doesn't work over case classes.
+ .map { entry => (entry.slug, entry.score, entry.json1, entry.json2) }
+ .write(TypedTsv[(String, Int, String, String)](args("output")))
+}