diff options
4 files changed, 226 insertions, 5 deletions
diff --git a/scalding/src/main/scala/sandcrawler/CrossrefScorable.scala b/scalding/src/main/scala/sandcrawler/CrossrefScorable.scala index 249c9ab..9842122 100644 --- a/scalding/src/main/scala/sandcrawler/CrossrefScorable.scala +++ b/scalding/src/main/scala/sandcrawler/CrossrefScorable.scala @@ -8,8 +8,9 @@ import com.twitter.scalding.typed.TDsl._ import parallelai.spyglass.hbase.HBaseConstants.SourceMode import parallelai.spyglass.hbase.HBasePipeConversions import parallelai.spyglass.hbase.HBaseSource +import TDsl._ -class CrossrefScorable extends Scorable { +class CrossrefScorable extends Scorable with HBasePipeConversions { // TODO: Generalize args so there can be multiple Grobid pipes in one job. def getSource(args : Args) : Source = { TextLine(args("crossref-input")) diff --git a/scalding/src/main/scala/sandcrawler/GrobidScorable.scala b/scalding/src/main/scala/sandcrawler/GrobidScorable.scala index 5c6b140..51e40f9 100644 --- a/scalding/src/main/scala/sandcrawler/GrobidScorable.scala +++ b/scalding/src/main/scala/sandcrawler/GrobidScorable.scala @@ -8,6 +8,7 @@ import com.twitter.scalding.typed.TDsl._ import parallelai.spyglass.hbase.HBaseConstants.SourceMode import parallelai.spyglass.hbase.HBasePipeConversions import parallelai.spyglass.hbase.HBaseSource +import TDsl._ class GrobidScorable extends Scorable with HBasePipeConversions { def getSource(args : Args) : Source = { @@ -15,10 +16,10 @@ class GrobidScorable extends Scorable with HBasePipeConversions { GrobidScorable.getHBaseSource(args("hbase-table"), args("zookeeper-hosts")) } - def getFeaturesPipe(pipe : Pipe) : TypedPipe[MapFeatures] = { + def getFeaturesPipe(pipe : cascading.pipe.Pipe) : TypedPipe[MapFeatures] = { pipe .fromBytesWritable(new Fields("key", "tei_json")) - .toTypedPipe[(String, String)](new Fields("key", "tei_json")) + .toTypedPipe[(String, String)](new Fields('key, 'tei_json)) .map { entry => val (key : String, json : String) = (entry._1, entry._2) GrobidScorable.grobidToSlug(json) match { diff --git a/scalding/src/main/scala/sandcrawler/HBaseCrossrefScoreJob.scala b/scalding/src/main/scala/sandcrawler/HBaseCrossrefScoreJob.scala new file mode 100644 index 0000000..725474d --- /dev/null +++ b/scalding/src/main/scala/sandcrawler/HBaseCrossrefScoreJob.scala @@ -0,0 +1,218 @@ +package sandcrawler + +import java.text.Normalizer +import java.util.Arrays +import java.util.Properties +import java.util.regex.Pattern + +import scala.math +import scala.util.parsing.json.JSON + +import cascading.tuple.Fields +import com.twitter.scalding._ +import com.twitter.scalding.typed.CoGrouped +import com.twitter.scalding.typed.Grouped +import com.twitter.scalding.typed.TDsl._ +import org.apache.hadoop.hbase.io.ImmutableBytesWritable +import org.apache.hadoop.hbase.util.Bytes +import parallelai.spyglass.base.JobBase +import parallelai.spyglass.hbase.HBaseConstants.SourceMode +import parallelai.spyglass.hbase.HBasePipeConversions +import parallelai.spyglass.hbase.HBaseSource + +class HBaseCrossrefScoreJob(args: Args) extends JobBase(args) with HBasePipeConversions { + val NoTitle = "NO TITLE" // Used for slug if title is empty or unparseable + + // key is SHA1 + val grobidSource = HBaseCrossrefScore.getHBaseSource( + args("hbase-table"), + args("zookeeper-hosts")) + + val temp : cascading.pipe.Pipe = grobidSource + .read + .fromBytesWritable(new Fields("key", "tei_json")) + val grobidPipe : TypedPipe[(String, String, String)] = temp + // .debug // Should be 4 tuples for mocked data + .toTypedPipe[(String, String)]('key, 'tei_json) + .map { entry => + val (key, json) = (entry._1, entry._2) + // TODO: Consider passing forward only a subset of JSON. + HBaseCrossrefScore.grobidToSlug(json) match { + case Some(slug) => (slug, key, json) + case None => (NoTitle, key, json) + } + } + .filter { entry => + val (slug, _, _) = entry + slug != NoTitle + } +// .debug // SHould be 3 tuples for mocked data + + val grobidGroup = grobidPipe + .groupBy { case (slug, key, json) => slug } + + val crossrefSource = TextLine(args("crossref-input")) + val temp2 : cascading.pipe.Pipe = crossrefSource.read + val crossrefPipe : TypedPipe[(String, String)] = temp2 + // .debug // Should be 4 tuples for mocked data + .toTypedPipe[String]('line) + .map{ json : String => + HBaseCrossrefScore.crossrefToSlug(json) match { + case Some(slug) => (slug, json) + case None => (NoTitle, json) + } + } + .filter { entry => + val (slug, json) = entry + slug != NoTitle + } + + val crossrefGroup = crossrefPipe + .groupBy { case (slug, json) => slug } + + val theJoin : CoGrouped[String, ((String, String, String), (String, String))] = + grobidGroup.join(crossrefGroup) + + theJoin.map{ entry => + val (slug : String, + ((slug0: String, sha1 : String, grobidJson : String), + (slug1 : String, crossrefJson : String))) = entry + HBaseCrossrefScore.computeOutput(sha1, grobidJson, crossrefJson)} + // Output: score, sha1, doi, grobid title, crossref title + .write(TypedTsv[(Int, String, String, String, String)](args("output"))) + +} + +object HBaseCrossrefScore { + def getHBaseSource(hbaseTable: String, zookeeperHosts: String) : HBaseSource = HBaseBuilder.build( + hbaseTable, // HBase Table Name + zookeeperHosts, // HBase Zookeeper server (to get runtime config info; can be array?) + List("grobid0:tei_json"), + SourceMode.SCAN_ALL) + + def jsonToMap(json : String) : Option[Map[String, Any]] = { + // https://stackoverflow.com/a/32717262/631051 + val jsonObject = JSON.parseFull(json) + if (jsonObject == None) { + None + } else { + Some(jsonObject.get.asInstanceOf[Map[String, Any]]) + } + } + + def grobidToSlug(json : String) : Option[String] = { + jsonToMap(json) match { + case None => None + case Some(map) => { + if (map contains "title") { + titleToSlug(map("title").asInstanceOf[String]) + } else { + None + } + } + } + } + + def crossrefToSlug(json : String) : Option[String] = { + jsonToMap(json) match { + case None => None + case Some(map) => { + if (map contains "title") { + // TODO: Don't ignore titles after the first. + titleToSlug(map("title").asInstanceOf[List[String]](0)) + } else { + None + } + } + } + } + + def titleToSlug(title : String) : Option[String] = { + val slug = removeAccents(title).split(":")(0).toLowerCase() + if (slug.isEmpty) { + None + } else { + Some(slug) + } + } + + val MaxScore = 1000 + + def computeOutput(sha1 : String, grobidJson : String, crossrefJson : String) : + // (score, sha1, doi, grobidTitle, crossrefTitle) + (Int, String, String, String, String) = { + jsonToMap(grobidJson) match { + case None => (0, "", "", "", "") // This can't happen, because grobidJson already validated in earlier stage + case Some(grobid) => { + val grobidTitle = grobid("title").asInstanceOf[String].toLowerCase() + + jsonToMap(crossrefJson) match { + case None => (0, "", "", "", "") // This can't happen, because crossrefJson already validated in earlier stage + case Some(crossref) => { + val crossrefTitle = crossref("title").asInstanceOf[List[String]](0).toLowerCase() + + (similarity(removeAccents(grobidTitle), removeAccents(crossrefTitle)), + sha1, + crossref("DOI").asInstanceOf[String], + "'" + grobidTitle + "'", + "'" + crossrefTitle + "'") + } + } + } + } + } + + // Adapted from https://git-wip-us.apache.org/repos/asf?p=commons-lang.git;a=blob;f=src/main/java/org/apache/commons/lang3/StringUtils.java;h=1d7b9b99335865a88c509339f700ce71ce2c71f2;hb=HEAD#l934 + def removeAccents(s : String) : String = { + val replacements = Map( + '\u0141' -> 'L', + '\u0142' -> 'l', // Letter ell + '\u00d8' -> 'O', + '\u00f8' -> 'o' + ) + val sb = new StringBuilder(Normalizer.normalize(s, Normalizer.Form.NFD)) + for (i <- 0 to sb.length - 1) { + for (key <- replacements.keys) { + if (sb(i) == key) { + sb.deleteCharAt(i); + sb.insert(i, replacements(key)) + } + } + } + val pattern = Pattern.compile("\\p{InCombiningDiacriticalMarks}+") + pattern.matcher(sb).replaceAll("") + } + + // Adapted from: https://stackoverflow.com/a/16018452/631051 + def similarity(s1 : String, s2 : String) : Int = { + val longer : String = if (s1.length > s2.length) s1 else s2 + val shorter : String = if (s1.length > s2.length) s2 else s1 + if (longer.length == 0) { + // Both strings are empty. + MaxScore + } else { + (longer.length - stringDistance(longer, shorter)) * MaxScore / longer.length + } + } + + // Source: // https://oldfashionedsoftware.com/2009/11/19/string-distance-and-refactoring-in-scala/ + def stringDistance(s1: String, s2: String): Int = { + val memo = scala.collection.mutable.Map[(List[Char],List[Char]),Int]() + def min(a:Int, b:Int, c:Int) = Math.min( Math.min( a, b ), c) + def sd(s1: List[Char], s2: List[Char]): Int = { + if (!memo.contains((s1, s2))) { + memo((s1,s2)) = (s1, s2) match { + case (_, Nil) => s1.length + case (Nil, _) => s2.length + case (c1::t1, c2::t2) => + min( sd(t1,s2) + 1, sd(s1,t2) + 1, + sd(t1,t2) + (if (c1==c2) 0 else 1) ) + } + } + memo((s1,s2)) + } + + sd( s1.toList, s2.toList ) + } +} + diff --git a/scalding/src/main/scala/sandcrawler/Scorable.scala b/scalding/src/main/scala/sandcrawler/Scorable.scala index 92b61bc..bd03d57 100644 --- a/scalding/src/main/scala/sandcrawler/Scorable.scala +++ b/scalding/src/main/scala/sandcrawler/Scorable.scala @@ -6,13 +6,14 @@ import scala.util.parsing.json.JSON import cascading.flow.FlowDef import com.twitter.scalding._ import com.twitter.scalding.typed.TDsl._ +import TDsl._ case class MapFeatures(slug : String, json : String) case class ReduceFeatures(json : String) case class ReduceOutput(val slug : String, score : Int, json1 : String, json2 : String) abstract class Scorable { - def getInputPipe(pipe : Pipe) : TypedPipe[(String, ReduceFeatures)] = + def getInputPipe(pipe : cascading.pipe.Pipe) : TypedPipe[(String, ReduceFeatures)] = { getFeaturesPipe(pipe) .filter { entry => Scorable.isValidSlug(entry.slug) } @@ -25,7 +26,7 @@ abstract class Scorable { // abstract methods def getSource(args : Args) : Source - def getFeaturesPipe(pipe : Pipe) : TypedPipe[MapFeatures] + def getFeaturesPipe(pipe : cascading.pipe.Pipe) : TypedPipe[MapFeatures] } object Scorable { |