aboutsummaryrefslogtreecommitdiffstats
path: root/scalding/src
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2018-08-19 19:19:37 -0700
committerBryan Newbold <bnewbold@archive.org>2018-08-21 20:31:45 -0700
commite99700cb521d82a750497d58bcb04d8cf1abcd80 (patch)
tree24c006065c283fc8f09945f25bdf9c94f68b7176 /scalding/src
parent38a0a73e7b1aa6d0002f4f1807171fffac19542f (diff)
downloadsandcrawler-e99700cb521d82a750497d58bcb04d8cf1abcd80.tar.gz
sandcrawler-e99700cb521d82a750497d58bcb04d8cf1abcd80.zip
bibjson scorable class (no tests)
Diffstat (limited to 'scalding/src')
-rw-r--r--scalding/src/main/scala/sandcrawler/BibjsonScorable.scala50
1 files changed, 50 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..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)
+ }
+ }
+ }
+ }
+}