Implement TSV output mode. (#162)

* Implement TSV output mode.

* More doc.

Author: balhoff

build.sbt

@@ -67,7 +67,8 @@ lazy val cli = project
     publish / skip := true,
     libraryDependencies ++= Seq(
       "com.outr" %% "scribe-slf4j" % "3.10.4",
-      "com.github.alexarchambault" %% "case-app" % "2.0.6"
+      "com.github.alexarchambault" %% "case-app" % "2.0.6",
+      "io.circe" %% "circe-yaml" % "0.14.1",
     ),
     gitCommitString := git.gitHeadCommit.value.getOrElse("Not Set"),
     buildInfoKeys := Seq[BuildInfoKey](name, version, scalaVersion, sbtVersion, gitCommitString),

cli/src/main/scala/org/renci/relationgraph/Config.scala

@@ -3,8 +3,7 @@ package org.renci.relationgraph
 import caseapp._
 import caseapp.core.Error.MalformedValue
 import caseapp.core.argparser.{ArgParser, SimpleArgParser}
-import org.renci.relationgraph.Config.{BoolValue, FalseValue, TrueValue}
-import org.renci.relationgraph.RelationGraph.Config.{OWLMode, OutputMode, RDFMode}
+import org.renci.relationgraph.Config.{BoolValue, FalseValue, OutputMode, RDFMode, TrueValue}
 
 @AppName("relation-graph")
 @ProgName("relation-graph")
@@ -15,8 +14,8 @@ final case class Config(
                          @HelpMessage("File to stream output triples to.")
                          @ValueDescription("filename")
                          outputFile: String,
-                         @HelpMessage("Configure style of triples to be output. RDF mode is the default; each existential relation is collapsed to a single direct triple.")
-                         @ValueDescription("RDF|OWL")
+                         @HelpMessage("Configure style of triples to be output. RDF mode is the default; each existential relation is collapsed to a single direct triple. TSV mode outputs the same triples as RDF mode, but as TSV, compacting IRIs using an optional prefixes file.")
+                         @ValueDescription("RDF|OWL|TSV")
                          mode: OutputMode = RDFMode,
                          @HelpMessage("Property to restrict output relations to. Provide option multiple times for multiple properties. If no properties are provided (via CLI or file), then all properties found in the ontology will be used.")
                          @ValueDescription("IRI")
@@ -42,13 +41,23 @@ final case class Config(
                          @HelpMessage("Disable inference of unsatisfiable classes by the whelk reasoner (default false)")
                          @ValueDescription("bool")
                          disableOwlNothing: BoolValue = FalseValue,
+                         @HelpMessage("Prefix mappings to use for TSV output (YAML dictionary")
+                         @ValueDescription("filename")
+                         prefixes: Option[String],
+                         @HelpMessage("Compact OBO-style IRIs regardless of inclusion in prefixes file")
+                         @ValueDescription("bool")
+                         oboPrefixes: BoolValue = TrueValue,
                          @HelpMessage("Set log level to INFO")
                          @ValueDescription("bool")
                          verbose: Boolean = false) {
 
   def toRelationGraphConfig: RelationGraph.Config =
     RelationGraph.Config(
-      mode = this.mode,
+      mode = this.mode match {
+        case Config.RDFMode => RelationGraph.Config.RDFMode
+        case Config.OWLMode => RelationGraph.Config.OWLMode
+        case Config.TSVMode => RelationGraph.Config.RDFMode
+      },
       outputSubclasses = this.outputSubclasses.bool,
       reflexiveSubclasses = this.reflexiveSubclasses.bool,
       equivalenceAsSubclass = this.equivalenceAsSubclass.bool,
@@ -61,10 +70,19 @@ final case class Config(
 
 object Config {
 
+  sealed trait OutputMode
+
+  case object RDFMode extends OutputMode
+
+  case object OWLMode extends OutputMode
+
+  case object TSVMode extends OutputMode
+
   implicit val rdfModeParser: ArgParser[OutputMode] = SimpleArgParser.from[OutputMode]("output mode") { arg =>
     arg.toLowerCase match {
       case "rdf" => Right(RDFMode)
       case "owl" => Right(OWLMode)
+      case "tsv" => Right(TSVMode)
       case _     => Left(MalformedValue("output mode", arg))
     }
   }

cli/src/main/scala/org/renci/relationgraph/Main.scala

@@ -10,9 +10,10 @@ import org.semanticweb.owlapi.model._
 import scribe.Level
 import scribe.filter.{packageName, select}
 import zio._
-import Config._
+import io.circe.yaml.parser
+import org.renci.relationgraph.Config.{OWLMode, RDFMode, TSVMode}
 
-import java.io.{File, FileOutputStream}
+import java.io.{File, FileOutputStream, FileReader}
 import scala.io.Source
 
 object Main extends ZCaseApp[Config] {
@@ -27,7 +28,7 @@ object Main extends ZCaseApp[Config] {
         .replace()
     }
     val program = ZIO.scoped {
-      createStreamRDF(config.outputFile).flatMap { rdfWriter =>
+      createStream(config).flatMap { rdfWriter =>
         for {
           fileProperties <- config.propertiesFile.map(readPropertiesFile).getOrElse(ZIO.succeed(Set.empty[AtomicConcept]))
           specifiedProperties = fileProperties ++ config.property.map(prop => AtomicConcept(prop)).to(Set)
@@ -47,6 +48,15 @@ object Main extends ZCaseApp[Config] {
       }.exitCode
   }
 
+  def createStream(config: Config): ZIO[Scope, Throwable, StreamRDF] = config.mode match {
+    case RDFMode => createStreamRDF(config.outputFile)
+    case OWLMode => createStreamRDF(config.outputFile)
+    case TSVMode =>
+      ZIO.foreach(config.prefixes)(readPrefixesFile).flatMap { maybePrefixes =>
+        createStreamTSV(config.outputFile, maybePrefixes.getOrElse(Map.empty), config.oboPrefixes.bool)
+      }
+  }
+
   def createStreamRDF(path: String): ZIO[Scope, Throwable, StreamRDF] = {
     ZIO.acquireRelease(ZIO.attempt(new FileOutputStream(new File(path))))(stream => ZIO.succeed(stream.close())).flatMap { outputStream =>
       ZIO.acquireRelease(ZIO.attempt {
@@ -57,6 +67,16 @@ object Main extends ZCaseApp[Config] {
     }
   }
 
+  def createStreamTSV(path: String, prefixes: Map[String, String], oboPrefixes: Boolean): ZIO[Scope, Throwable, StreamRDF] = {
+    ZIO.attempt(new File(path)).flatMap { file =>
+      ZIO.acquireRelease(ZIO.attempt {
+        val stream = new TSVStreamRDF(file, prefixes, oboPrefixes)
+        stream.start()
+        stream
+      })(stream => ZIO.succeed(stream.finish()))
+    }
+  }
+
   def loadOntology(path: String): Task[OWLOntology] = for {
     manager <- ZIO.attempt(OWLManager.createOWLOntologyManager())
     ontology <- ZIO.attemptBlocking(manager.loadOntologyFromOntologyDocument(new File(path)))
@@ -67,4 +87,13 @@ object Main extends ZCaseApp[Config] {
       ZIO.attemptBlocking(source.getLines().map(_.trim).filter(_.nonEmpty).map(line => AtomicConcept(line)).to(Set))
     }
 
+  def readPrefixesFile(filename: String): ZIO[Any, Throwable, Map[String, String]] =
+    ZIO.attemptBlocking(new FileReader(new File(filename))).acquireReleaseWithAuto { reader =>
+      ZIO.fromEither {
+        parser.parse(reader).flatMap { json =>
+          json.as[Map[String, String]]
+        }
+      }
+    }
+
 }

cli/src/main/scala/org/renci/relationgraph/TSVStreamRDF.scala

@@ -0,0 +1,56 @@
+package org.renci.relationgraph
+
+import org.apache.jena.graph.Triple
+import org.apache.jena.riot.system.StreamRDF
+import org.apache.jena.shared.PrefixMapping
+import org.apache.jena.shared.impl.PrefixMappingImpl
+import org.apache.jena.sparql.core.Quad
+
+import java.io.{File, PrintWriter}
+import scala.jdk.CollectionConverters._
+
+class TSVStreamRDF(file: File, prefixes: Map[String, String], oboPrefixes: Boolean) extends StreamRDF {
+
+  private val prefixMapping: PrefixMapping = {
+    val pm = new PrefixMappingImpl() {
+      override def shortForm(uri: String): String = {
+        val shortForm = super.shortForm(uri)
+        if (oboPrefixes && (shortForm == uri) && (uri.startsWith("http://purl.obolibrary.org/obo/"))) {
+          val tail = uri.replace("http://purl.obolibrary.org/obo/", "")
+          tail.split("_", 2).mkString(":")
+        } else shortForm
+      }
+    }
+    pm.setNsPrefixes(prefixes.asJava).withDefaultMappings(PrefixMapping.Standard)
+  }
+
+  private var writer: PrintWriter = _
+
+  override def start(): Unit = {
+    writer = new PrintWriter(file, "utf-8")
+  }
+
+  override def triple(triple: Triple): Unit = {
+    val s = triple.getSubject.toString(prefixMapping, true)
+    val p = triple.getPredicate.toString(prefixMapping, true)
+    val o = triple.getObject.toString(prefixMapping, true)
+    writer.println(s"$s\t$p\t$o")
+  }
+
+  override def quad(quad: Quad): Unit = {
+    val s = quad.getSubject.toString(prefixMapping, true)
+    val p = quad.getPredicate.toString(prefixMapping, true)
+    val o = quad.getObject.toString(prefixMapping, true)
+    val g = quad.getGraph.toString(prefixMapping, true)
+    writer.println(s"$s\t$p\t$o\t$g")
+  }
+
+  override def base(base: String): Unit = ()
+
+  override def prefix(prefix: String, iri: String): Unit = ()
+
+  override def finish(): Unit = {
+    writer.close()
+  }
+
+}

core/src/main/scala/org/renci/relationgraph/RelationGraph.scala

@@ -6,7 +6,7 @@ import org.apache.jena.sys.JenaSystem
 import org.apache.jena.vocabulary.{OWL2, RDF, RDFS}
 import org.geneontology.whelk.BuiltIn.{Bottom, Top}
 import org.geneontology.whelk._
-import org.renci.relationgraph.RelationGraph.Config.{OWLMode, OutputMode, RDFMode}
+import org.renci.relationgraph.RelationGraph.Config.{OWLMode, RDFMode, TriplesMode}
 import org.semanticweb.owlapi.apibinding.OWLFunctionalSyntaxFactory.{OWLNothing, OWLThing}
 import org.semanticweb.owlapi.model.parameters.Imports
 import org.semanticweb.owlapi.model._
@@ -32,7 +32,7 @@ object RelationGraph extends StrictLogging {
   private val OWLOntology = OWL2.Ontology.asNode
 
   final case class Config(
-                           mode: OutputMode = RDFMode,
+                           mode: TriplesMode = RDFMode,
                            outputSubclasses: Boolean = false,
                            reflexiveSubclasses: Boolean = true,
                            equivalenceAsSubclass: Boolean = true,
@@ -43,11 +43,11 @@ object RelationGraph extends StrictLogging {
 
   object Config {
 
-    sealed trait OutputMode
+    sealed trait TriplesMode
 
-    case object RDFMode extends OutputMode
+    case object RDFMode extends TriplesMode
 
-    case object OWLMode extends OutputMode
+    case object OWLMode extends TriplesMode
 
   }
 
@@ -101,7 +101,7 @@ object RelationGraph extends StrictLogging {
     } yield ()
   }
 
-  def processRestrictionAndExtendQueue(restriction: Restriction, properties: Hierarchy, classes: Hierarchy, whelk: IndexedReasonerState, mode: Config.OutputMode, descendProperties: Boolean, outputClasses: Boolean, outputIndividuals: Boolean, queue: Queue[Restriction], activeRestrictions: Ref[Int], seenRefs: Map[Role, Ref[Set[AtomicConcept]]]): UIO[TriplesGroup] = {
+  def processRestrictionAndExtendQueue(restriction: Restriction, properties: Hierarchy, classes: Hierarchy, whelk: IndexedReasonerState, mode: Config.TriplesMode, descendProperties: Boolean, outputClasses: Boolean, outputIndividuals: Boolean, queue: Queue[Restriction], activeRestrictions: Ref[Int], seenRefs: Map[Role, Ref[Set[AtomicConcept]]]): UIO[TriplesGroup] = {
     val triples = processRestriction(restriction, whelk, mode, outputClasses, outputIndividuals)
     val continue = triples.redundant.nonEmpty
     for {
@@ -134,7 +134,7 @@ object RelationGraph extends StrictLogging {
     } yield triples
   }
 
-  def processRestriction(restriction: Restriction, whelk: IndexedReasonerState, mode: Config.OutputMode, outputClasses: Boolean, outputIndividuals: Boolean): TriplesGroup = {
+  def processRestriction(restriction: Restriction, whelk: IndexedReasonerState, mode: Config.TriplesMode, outputClasses: Boolean, outputIndividuals: Boolean): TriplesGroup = {
     val subConcepts = queryExistentialSubclasses(restriction, whelk)
     val subclasses = if (outputClasses) (subConcepts - Bottom).collect { case AtomicConcept(id) => id } else Set.empty[String]
     val instances = if (outputIndividuals) subConcepts.collect { case Nominal(Individual(id)) => id } else Set.empty[String]