a lot of symbol merge logic improvements

buildAndRunOnProject.sh

@@ -0,0 +1,5 @@
+sbt +publishLocal
+cd sbt-plugin-integration-test/target/scala-2.11/classes/integration-test-projects/pipeline
+rc=$?; if [[ $rc != 0 ]]; then exit $rc; fi
+echo "Running for target project..."
+sbt canve > out.txt && subl out.txt

compiler-plugin/src/main/scala/org/canve/compilerPlugin/Classes.scala → compiler-plugin/src/main/scala/org/canve/compilerPlugin/DataClasses.scala

@@ -1,29 +1,41 @@
 package org.canve.compilerPlugin
 import tools.nsc.Global
 
-/*
- * types for whether a symbol is defined in 
- * the current project, or is an external one
- */
-
-abstract class DefiningProject
-object ProjectDefined    extends DefiningProject
-object ExternallyDefined extends DefiningProject
-
 case class ExtractedSymbol 
   (symbolCompilerId: SymbolCompilerId,
    name: String,
    kind: String,
-   notSynthetic: Boolean,
    qualifiedId: QualifiedID,
+
+   signatureString: Option[String] = None,  
+
+   notSynthetic: Boolean,
    definingProject: DefiningProject) extends SymbolSerialization {
 
-     var ownersTraversed = false
+     val qualifiedIdAndSignature: FurtherQualifiedID = 
+       qualifiedId.pickle + (signatureString match {
+         case Some(s) => " * " + s
+         case None => "" 
+       }) 
 
      def definitionCode(implicit extractedModel: ExtractedModel): Option[Code] = 
        extractedModel.codes.get.get(symbolCompilerId)
+
+     /* more inclusive serialization for this class - for logging */
+     override def toString = 
+       List(symbolCompilerId, 
+            name, 
+            kind, 
+            qualifiedId, 
+            signatureString, 
+            notSynthetic, 
+            definingProject,
+            qualifiedIdAndSignature).map(_.toString).mkString(",")
 
-     def toJoinedString(implicit extractedModel: ExtractedModel) = this.toString + definitionCode.toString  
+     /* symbol and its code info joined into a string - for logging */
+     def toJoinedString(implicit extractedModel: ExtractedModel) = toString + ",code: " + definitionCode.toString
+
+     var ownersTraversed = false
 }
 
 /*
@@ -46,8 +58,18 @@ case class Span(start: Int, end: Int) extends Position
 case class Point(init: Int) extends Position { def apply = init }
 
 /*
- * Simply a join of a symbol and its extracted code, wherever helpful
+ * types for whether a symbol is defined in 
+ * the current project, or is an external one
  */
+
+abstract class DefiningProject
+object ProjectDefined    extends DefiningProject
+object ExternallyDefined extends DefiningProject
+
+/*
+ * a lame join type of a symbol and its extracted code, wherever helpful
+ */
+
 case class SymbolCodeJoin(
   extractedModel: ExtractedModel, 
   symbol: ExtractedSymbol) {

compiler-plugin/src/main/scala/org/canve/compilerPlugin/ExtractedModel.scala

@@ -38,11 +38,22 @@ class ExtractedModel(global: Global) extends ContainsExtractedGraph {
           case _    => ProjectDefined
         }
 
-
         /*
          * add the symbol to the extracted model
          */
-        val extractedSymbol = ExtractedSymbol(s.id, s.nameString, s.kindString, !(s.isSynthetic), qualifiedId, definingProject)
+        val extractedSymbol = 
+          ExtractedSymbol(
+            symbolCompilerId = s.id, 
+            name = s.nameString, 
+            kind = s.kindString, 
+            qualifiedId = qualifiedId, 
+            notSynthetic = !(s.isSynthetic), 
+            definingProject = definingProject,
+            signatureString = s.signatureString match {
+              case("<_>") => None
+              case signatureString@_ => Some(signatureString)
+            })
+
         graph ++ graph.Vertex(extractedSymbol.symbolCompilerId, extractedSymbol) 
 
         /*

compiler-plugin/src/main/scala/org/canve/compilerPlugin/Output.scala

@@ -6,21 +6,21 @@ object Output {
 
   def quote(int: Int) = "\"" + int.toString() + "\""
 
-  def write(model: ExtractedModel) = {
+  def write(implicit extractedModel: ExtractedModel) = {
     Log(s"writing extracted type relations and call graph for project ${PluginArgs.projectName}...")
 
     writeOutputFile(PluginArgs.projectName, "symbols", 
-                    "definition,notSynthetic,id,name,kind,qualifiedId\n" +
-                    (model.graph.vertexIterator.map(_.data.toCsvRow)).mkString("\n"))
+                    "definition,notSynthetic,id,name,kind,qualifiedId,signature\n" +
+                    (extractedModel.graph.vertexIterator.map(_.data.toCsvRow)).mkString("\n"))
 
     println("before output")
     writeOutputFile(PluginArgs.projectName, "relations", 
         "id1,relation,id2\n" +
-        model.graph.edgeIterator.map ( e => 
+        extractedModel.graph.edgeIterator.map ( e => 
           List(e.v1, e.data, e.v2).mkString(","))
           .mkString("\n"))
 
-    model.codes.get.foreach { keyVal => 
+    extractedModel.codes.get.foreach { keyVal => 
       val extractedCode = keyVal._2
       if (extractedCode.code.isDefined)
         writeOutputFile(

compiler-plugin/src/main/scala/org/canve/compilerPlugin/QualiferID.scala

@@ -22,14 +22,21 @@ object QualifiedID {
 
     assert(kindNameList.head.kind == "package")
     assert(kindNameList.head.name == "<root>")
-
+
+    /*
+     * TODO: remove this commented out code
+     * 
     (s.kindString == "method") match {
       case true => 
         val paramListList: List[List[global.Symbol]] = s.paramss
         QualifiedID(kindNameList.drop(1) ++ s.paramss.flatten.map(p => KindAndName(global)(p)))
       case false => 
         QualifiedID(kindNameList.drop(1))
     }
+    */
+
+    QualifiedID(kindNameList.drop(1))
+
   }
 
   def unpickle(s: String) = 

compiler-plugin/src/main/scala/org/canve/compilerPlugin/Serialization.scala

@@ -1,12 +1,13 @@
 package org.canve.compilerPlugin
 
 /*
- * helper for writing node data to csv 
+ * Serialize extracted symbol into CSV data row
  */
 trait SymbolSerialization {
   self: ExtractedSymbol =>
 
-  def toCsvRow: String = {
+  def toCsvRow(implicit extractedModel: ExtractedModel): String = {
+    // Note: escaping for csv is for now handled here by hand (likely a tad faster)
     List(
       definingProject match {
         case ProjectDefined    => "project"
@@ -15,7 +16,8 @@ trait SymbolSerialization {
       symbolCompilerId, 
       name, 
       kind,
-      qualifiedId.pickle
+      qualifiedId.pickle,
+      "\"" + signatureString + "\"" // escaped as it contains, typically, commas
     ).mkString(",")
   }
 }
@@ -24,28 +26,26 @@ trait SymbolSerialization {
  * Symbol constructor from CANVE CSV data row
  */
 object SymbolFromCsvRow {
+  import Util._
   def apply(projectName: String, rowMap: Map[String, String]): ExtractedSymbol = { 
      ExtractedSymbol(symbolCompilerId = rowMap("id").toInt,
           name = rowMap("name"),
           kind = rowMap("kind"),
           notSynthetic = rowMap("notSynthetic").toBoolean,
           qualifiedId = QualifiedID.unpickle(rowMap("qualifiedId")),
+          signatureString = deSerializeOption(rowMap("signature")),
           definingProject = rowMap("definition") match {
             case "project" => ProjectDefined
             case "external" => ExternallyDefined
           })
   }
 }
 
-/*
-/*
- * Symbol relation from CANVE CSV data row
- */
-object SymbolRelationFromCsvRow {
-  def apply(rowMap: Map[String, String]) = 
-    ExtractedSymbolRelation(
-      rowMap("id1").toInt,
-      rowMap("id2").toInt,
-      rowMap("relation"))
-}
-*/
+object Util {
+  def deSerializeOption[T](s: String): Option[T] = {
+    s match {
+      case "None" => None
+      case s@_ => Some(s.drop("Some(".length).dropRight(1).asInstanceOf[T])
+    }
+  }
+}

compiler-plugin/src/main/scala/org/canve/compilerPlugin/TraversalExtraction.scala

@@ -33,8 +33,14 @@ object TraversalExtraction {
 
                 extractedModel.add(global)(select.symbol)
 
-                if (defParent.isDefined) extractedModel.add(defParent.get.id, "uses", select.symbol.id)
-
+                if (defParent.isDefined) {
+                  extractedModel.add(defParent.get.id, "uses", select.symbol.id)
+                  println
+                  println(s"Method call. Symbol owner chain: ${select.symbol.ownerChain.reverse}, \nParams: ${select.symbol.paramss}")
+                  println("signatureString: " + select.symbol.signatureString)
+                  println
+                }
+
                 // record the source code location where the symbol is being used by the user 
                 // this is a proof of concept, that only doesn't propagate its information
                 // to the UI in any way yet.
@@ -90,7 +96,11 @@ object TraversalExtraction {
 
             extractedModel.add(global)(symbol)
             extractedModel.addIfUnique(defParent.get.id, "declares member", symbol.id)
-
+            println
+            println(s"Method definition. Symbol owner chain: ${symbol.ownerChain.reverse}, \nParams: ${symbol.paramss}")
+            println("signatureString: " + symbol.signatureString)
+            println
+
             val traverser = new ExtractionTraversal(Some(tree.symbol))
             if (symbol.nameString == "get") {
               //val tracer = new TraceTree
@@ -115,8 +125,15 @@ object TraversalExtraction {
               if (defParent.get.id != typeSymbol.owner.id)
                 Warning.logParentNotOwner(global)(defParent.get, typeSymbol.owner)
 
-            parentTypeSymbols.foreach(s =>
-              extractedModel.add(typeSymbol.id, "extends", s.id))
+            parentTypeSymbols.foreach(s => {
+              extractedModel.add(typeSymbol.id, "extends", s.id)
+              println
+              println(s"Symbol ${typeSymbol.ownerChain.reverse}")
+              println(s"Performs Type extension of ${s.ownerChain.reverse}. TypeParams: ${s.typeParams}")
+              println("signatureString: " + s.signatureString)
+              println("typeSignature: " + s.typeSignature)
+              println
+            })
 
             val traverser = new ExtractionTraversal(Some(tree.tpe.typeSymbol))
             body foreach { tree => traverser.traverse(tree) }

compiler-plugin/src/main/scala/org/canve/compilerPlugin/normalization/CrossProjectNormalize.scala

@@ -12,10 +12,10 @@ import org.canve.simpleGraph._
 trait DataReader {
 
   type LoadedGraph = SimpleGraph[SymbolCompilerId, ExtractedSymbol, ExtractedSymbolRelation]
-  type ReIndexedGraph = SimpleGraph[QualifiedID, ExtractedSymbol, ExtractedSymbolRelation]
+  type ReIndexedGraph = SimpleGraph[FurtherQualifiedID, ExtractedSymbol, ExtractedSymbolRelation]
 
   /*
-   * builds canve nodes and edges from a canve data directory
+   * builds graph from a canve data directory
    */
   def readCanveDirData(dir: File): ManagedExtractedGraph = {
 
@@ -62,9 +62,9 @@ object CrossProjectNormalizer extends DataReader {
 
     projectGraphs.foreach { graph =>
       graph.vertexIterator.foreach { v => val symbol = v.data
-        aggregateGraph.vertex(symbol.qualifiedId) match {
+        aggregateGraph.vertex(symbol.qualifiedIdAndSignature) match {
 
-          case None => aggregateGraph ++ aggregateGraph.Vertex(key = symbol.qualifiedId, data = symbol)
+          case None => aggregateGraph ++ aggregateGraph.Vertex(key = symbol.qualifiedIdAndSignature, data = symbol)
 
           case Some(v) => maybeMerge(aggregateGraph, aggregateGraphSymbol = v.data, sameKeyedSymbol = symbol)
 
@@ -88,15 +88,15 @@ object CrossProjectNormalizer extends DataReader {
       case (ExternallyDefined, ExternallyDefined) => // do nothing
 
       case (ExternallyDefined, ProjectDefined) => 
-        aggregateGraph -- aggregateGraphSymbol.qualifiedId 
-        aggregateGraph ++ aggregateGraph.Vertex(key = sameKeyedSymbol.qualifiedId, data = sameKeyedSymbol)
+        aggregateGraph -- aggregateGraphSymbol.qualifiedIdAndSignature 
+        aggregateGraph ++ aggregateGraph.Vertex(key = sameKeyedSymbol.qualifiedIdAndSignature, data = sameKeyedSymbol)
         println("deduplicated one node")
 
-      case (ProjectDefined, ProjectDefined) => 
+      case (ProjectDefined, ProjectDefined) =>  
         // TODO: This exception message doesn't help zoom in on the project names without logs excavation.
         //       Might be solved by adding the project name to each symbol being compared, somewhere before. 
         println(DataNormalizationException(
-          "Two subprojects seem to define the same qualified name:" + 
+          "Two symbols seem to define the same qualified name:" + 
           s"\n$aggregateGraphSymbol and" +
           s"\n$sameKeyedSymbol"))
     }
@@ -110,9 +110,9 @@ object CrossProjectNormalizer extends DataReader {
     return
 
     println(DataNormalizationException(
-      s"""Two symbols having the same Qualified ID are unexpectedly different:
-      |$s1" +  
-      |$s2" +  
+      s"""Two symbols expected to be the same are unexpectedly different:
+      |$s1 
+      |$s2  
       |This may indicate the normalization model is relying on a false assumption.""".stripMargin)) 
   }
 }