From e99700cb521d82a750497d58bcb04d8cf1abcd80 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Sun, 19 Aug 2018 19:19:37 -0700
Subject: bibjson scorable class (no tests)

---
 .../main/scala/sandcrawler/BibjsonScorable.scala   | 50 ++++++++++++++++++++++
 1 file changed, 50 insertions(+)
 create mode 100644 scalding/src/main/scala/sandcrawler/BibjsonScorable.scala

(limited to 'scalding')

diff --git a/scalding/src/main/scala/sandcrawler/BibjsonScorable.scala b/scalding/src/main/scala/sandcrawler/BibjsonScorable.scala
new file mode 100644
index 0000000..7221a66
--- /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 = new ScorableFeatures(title=title, doi=doi, sha1=sha1)
+            new MapFeatures(sf.toSlug, sf.toString)
+          }
+        } else {
+          new MapFeatures(Scorable.NoSlug, json)
+        }
+      }
+    }
+  }
+}
-- 
cgit v1.2.3