Add graph data science algorithms for Typescript

cypher/DependsOn_Relationship_Weights/Add_fine_grained_weights_for_Typescript_internal_module_dependencies.cypher

@@ -0,0 +1,54 @@
+// Add fine grained weight properties for dependencies between internal Typescript modules
+
+// Get the top level dependency between a Typescript module and the external modules it uses
+ MATCH (source:TS:Module)-[moduleDependency:DEPENDS_ON]->(target:Module)
+// Exclude all targets where an ExternalModule was found that resolves to them
+// because those are covered in the fine grained weights for "ExternalModule"s.
+ WHERE NOT EXISTS { (target)<-[:RESOLVES_TO]-(resolvedTarget:ExternalModule) }
+  WITH source
+      ,target
+      ,moduleDependency
+OPTIONAL MATCH (source)-[elementDependency:DEPENDS_ON]->(elementType:TS)<-[:EXPORTS]-(target)
+  WITH source
+      ,target
+      ,moduleDependency
+      ,count(DISTINCT elementType.globalFqn)          AS elementTypeCount
+      ,sum(elementDependency.cardinality)             AS elementTypeCardinality
+      ,collect(DISTINCT elementType.globalFqn)[0..4]  AS elementTypeExamples
+OPTIONAL MATCH (source)-[abstractDependency:DEPENDS_ON]->(abstractType:TypeAlias|Interface)<-[:EXPORTS]-(target)
+  WITH source
+      ,target
+      ,moduleDependency
+      ,elementTypeCount        
+      ,elementTypeCardinality   
+      ,elementTypeExamples     
+      ,count(DISTINCT abstractType.globalFqn)         AS abstractTypeCount
+      ,sum(abstractDependency.cardinality)            AS abstractTypeCardinality
+      ,collect(DISTINCT abstractType.globalFqn)[0..4] AS abstractTypeExamples 
+// Set additional fine grained relationship properties (weights) to distinguish low and high coupling elements.
+// The "cardinality" property is similar to "weight" property for Java dependencies and comes from the jQAssistant Typescript Plugin.
+// - "abstractTypeCardinality" is the sum of all TypeAlias and Interface cardinality properties (if available)
+//   and corresponds to the "weightInterfaces" relationship property for Java.
+// - "declarationCardinality"  is the sum of all cardinality properties (including available Interface and TypeAlias)
+//   and is the same as the already existing "cardinality" of the moduleDependency. Thats why its left out.
+// - "lowCouplingElement25PercentWeight" subtracts 75% of the weights for abstract types like Interfaces and Type aliases
+//   to compensate for their low coupling influence. Not included "high coupling" elements like Functions and Classes 
+//   remain in the weight as they were. The same applies for "lowCouplingElement10PercentWeight" but with in a stronger manner.
+ SET moduleDependency.abstractTypeCount       = abstractTypeCount
+    ,moduleDependency.abstractTypeCardinality = abstractTypeCardinality
+    ,moduleDependency.lowCouplingElement25PercentWeight = 
+        toInteger(elementTypeCardinality - round(abstractTypeCardinality * 0.75))
+    ,moduleDependency.lowCouplingElement10PercentWeight = 
+        toInteger(elementTypeCardinality - round(abstractTypeCardinality * 0.90))
+RETURN source.globalFqn    AS sourceName
+      ,target.globalFqn    AS targetName
+      ,abstractTypeCount
+      ,elementTypeCount
+      ,moduleDependency.cardinality AS externalModuleCardinality
+      ,abstractTypeCardinality
+      ,elementTypeCardinality
+      ,moduleDependency.lowCouplingElement25PercentWeight
+      ,moduleDependency.lowCouplingElement10PercentWeight
+      ,abstractTypeExamples
+      ,elementTypeExamples
+ORDER BY sourceName ASC

cypher/Exploration/Explore_Typescript_elements_with_same_globalFqn.cypher

@@ -1,11 +1,12 @@
 // Explore Typescript elements with same globalFqn
 
  MATCH (ts1:TS)
- WHERE ts1.globalFqn CONTAINS '"'
+ WHERE ts1.globalFqn IS NOT NULL
  MATCH (ts2:TS)
- WHERE ts2.globalFqn CONTAINS '"'
-   AND (ts2.globalFqn = ts1.globalFqn
-   OR   toLower(ts2.globalFqn) = toLower(ts1.globalFqn))
+ WHERE ts2.globalFqn IS NOT NULL
+   AND ts2.globalFqn = ts1.globalFqn
    AND ts1 <> ts2
 RETURN labels(ts1), labels(ts2), count(*)
+      ,collect(ts1.globalFqn)[0..4] AS examples1
+      ,collect(ts2.globalFqn)[0..4] AS examples2
 LIMIT 30

cypher/Node_Embeddings/Node_Embeddings_0a_Query_Calculated.cypher

@@ -0,0 +1,18 @@
+// Query already calculated and written node embeddings on nodes with label in parameter $dependencies_projection_node including a communityId and centrality. Variables: dependencies_projection_node, dependencies_projection_write_property
+
+ MATCH (codeUnit)
+ WHERE $dependencies_projection_node IN LABELS(codeUnit)
+   AND codeUnit[$dependencies_projection_write_property] IS NOT NULL
+   // AND codeUnit.notExistingToForceRecalculation IS NOT NULL // uncomment this line to force recalculation
+OPTIONAL MATCH (artifact:Java:Artifact)-[:CONTAINS]->(codeUnit)
+   WITH *, replace(last(split(artifact.fileName, '/')), '.jar', '') AS artifactName
+OPTIONAL MATCH (projectRoot:Directory)<-[:HAS_ROOT]-(proj:TS:Project)-[:CONTAINS]->(codeUnit)
+   WITH *, last(split(projectRoot.absoluteFileName, '/')) AS projectName   
+ RETURN DISTINCT 
+        coalesce(codeUnit.fqn, codeUnit.globalFqn, codeUnit.fileName, codeUnit.signature, codeUnit.name) AS codeUnitName
+       ,coalesce(codeUnit.name, replace(last(split(codeUnit.fileName, '/')), '.jar', ''))                AS shortCodeUnitName
+       ,coalesce(artifactName, projectName)                                                              AS projectName
+       ,coalesce(codeUnit.communityLeidenId, 0)           AS communityId
+       ,coalesce(codeUnit.centralityPageRank, 0.01)       AS centrality
+       ,codeUnit[$dependencies_projection_write_property] AS embedding
+  ORDER BY communityId

cypher/Node_Embeddings/Node_Embeddings_1a_Fast_Random_Projection_Estimate.cypher

@@ -3,6 +3,7 @@
 CALL gds.fastRP.stream.estimate(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
   }
 )

cypher/Node_Embeddings/Node_Embeddings_1b_Fast_Random_Projection_Statistics.cypher

@@ -3,6 +3,7 @@
 CALL gds.fastRP.stats(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
   }
 )

cypher/Node_Embeddings/Node_Embeddings_1c_Fast_Random_Projection_Mutate.cypher

@@ -3,6 +3,7 @@
 CALL gds.fastRP.mutate(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,mutateProperty: $dependencies_projection_write_property
   }

cypher/Node_Embeddings/Node_Embeddings_1d_Fast_Random_Projection_Stream.cypher

@@ -3,15 +3,21 @@
 CALL gds.fastRP.stream(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
   }
 )
 YIELD nodeId, embedding
  WITH gds.util.asNode(nodeId) AS codeUnit
      ,embedding
-OPTIONAL MATCH (artifact:Artifact)-[:CONTAINS]->(codeUnit)
-RETURN DISTINCT coalesce(codeUnit.fqn, codeUnit.fileName, codeUnit.signature, codeUnit.name) AS codeUnitName
-     ,coalesce(codeUnit.communityLeidenId, 0) AS communityId
-     ,coalesce(codeUnit.centralityPageRank, 0.01) AS centrality
-     ,replace(last(split(artifact.fileName, '/')), '.jar', '') AS artifactName
-     ,embedding
+OPTIONAL MATCH (artifact:Java:Artifact)-[:CONTAINS]->(codeUnit)
+   WITH *, replace(last(split(artifact.fileName, '/')), '.jar', '') AS artifactName
+OPTIONAL MATCH (projectRoot:Directory)<-[:HAS_ROOT]-(proj:TS:Project)-[:CONTAINS]->(codeUnit)
+   WITH *, last(split(projectRoot.absoluteFileName, '/')) AS projectName   
+ RETURN DISTINCT 
+        coalesce(codeUnit.fqn, codeUnit.globalFqn, codeUnit.fileName, codeUnit.signature, codeUnit.name) AS codeUnitName
+       ,coalesce(codeUnit.name, replace(last(split(codeUnit.fileName, '/')), '.jar', '')) AS shortCodeUnitName
+       ,coalesce(artifactName, projectName) AS projectName
+       ,coalesce(codeUnit.communityLeidenId, 0) AS communityId
+       ,coalesce(codeUnit.centralityPageRank, 0.01) AS centrality
+       ,embedding

cypher/Node_Embeddings/Node_Embeddings_1e_Fast_Random_Projection_Write.cypher

@@ -3,6 +3,7 @@
 CALL gds.fastRP.write(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,writeProperty: $dependencies_projection_write_property
   }

cypher/Node_Embeddings/Node_Embeddings_2a_Hash_GNN_Estimate.cypher

@@ -6,9 +6,11 @@ CALL gds.beta.hashgnn.stream.estimate(
      ,iterations: 3
      ,generateFeatures: {
          dimension: toInteger($dependencies_projection_embedding_dimension) * 4
-        ,densityLevel: 1
+        ,densityLevel: 3
      }
      ,outputDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,neighborInfluence: 0.9
+     ,randomSeed: 30
   }
 )
  YIELD requiredMemory, nodeCount, relationshipCount, bytesMin, bytesMax, heapPercentageMin, heapPercentageMax, treeView

cypher/Node_Embeddings/Node_Embeddings_2c_Hash_GNN_Mutate.cypher

@@ -6,9 +6,11 @@ CALL gds.beta.hashgnn.mutate(
      ,iterations: 3
      ,generateFeatures: {
          dimension: toInteger($dependencies_projection_embedding_dimension) * 4
-        ,densityLevel: 1
+        ,densityLevel: 3
      }
      ,outputDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,neighborInfluence: 0.9
+     ,randomSeed: 30
      ,mutateProperty: $dependencies_projection_write_property
   }
 )

cypher/Node_Embeddings/Node_Embeddings_2d_Hash_GNN_Stream.cypher

@@ -6,17 +6,24 @@ CALL gds.beta.hashgnn.stream(
      ,iterations: 3
      ,generateFeatures: {
          dimension: toInteger($dependencies_projection_embedding_dimension) * 4
-        ,densityLevel: 1
+        ,densityLevel: 3
      }
      ,outputDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,neighborInfluence: 0.9
+     ,randomSeed: 30
   }
 )
 YIELD nodeId, embedding
  WITH gds.util.asNode(nodeId) AS codeUnit
      ,embedding
-OPTIONAL MATCH (artifact:Artifact)-[:CONTAINS]->(codeUnit)
-RETURN DISTINCT coalesce(codeUnit.fqn, codeUnit.fileName, codeUnit.signature, codeUnit.name) AS codeUnitName
+OPTIONAL MATCH (artifact:Java:Artifact)-[:CONTAINS]->(codeUnit)
+   WITH *, replace(last(split(artifact.fileName, '/')), '.jar', '') AS artifactName
+OPTIONAL MATCH (projectRoot:Directory)<-[:HAS_ROOT]-(proj:TS:Project)-[:CONTAINS]->(codeUnit)
+   WITH *, last(split(projectRoot.absoluteFileName, '/')) AS projectName   
+ RETURN DISTINCT 
+        coalesce(codeUnit.fqn, codeUnit.globalFqn, codeUnit.fileName, codeUnit.signature, codeUnit.name) AS codeUnitName
+       ,coalesce(codeUnit.name, replace(last(split(codeUnit.fileName, '/')), '.jar', '')) AS shortCodeUnitName
+       ,coalesce(artifactName, projectName) AS projectName
      ,coalesce(codeUnit.communityLeidenId, 0) AS communityId
      ,coalesce(codeUnit.centralityPageRank, 0.01) AS centrality
-     ,replace(last(split(artifact.fileName, '/')), '.jar', '') AS artifactName
      ,embedding

cypher/Node_Embeddings/Node_Embeddings_3a_Node2Vec_Estimate.cypher

@@ -4,6 +4,7 @@ CALL gds.node2vec.write.estimate(
  $dependencies_projection + '-cleaned', {
      ,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
      ,iterations: 3
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,writeProperty: $dependencies_projection_write_property
   }

cypher/Node_Embeddings/Node_Embeddings_3c_Node2Vec_Mutate.cypher

@@ -4,6 +4,7 @@ CALL gds.node2vec.mutate(
  $dependencies_projection + '-cleaned', {
      ,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
      ,iterations: 3
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,mutateProperty: $dependencies_projection_write_property
   }

cypher/Node_Embeddings/Node_Embeddings_3d_Node2Vec_Stream.cypher

@@ -4,15 +4,21 @@ CALL gds.node2vec.stream(
  $dependencies_projection + '-cleaned', {
      ,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
      ,iterations: 3
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
   }
 )
 YIELD nodeId, embedding
  WITH gds.util.asNode(nodeId) AS codeUnit
      ,embedding
-OPTIONAL MATCH (artifact:Artifact)-[:CONTAINS]->(codeUnit)
-RETURN DISTINCT coalesce(codeUnit.fqn, codeUnit.fileName, codeUnit.signature, codeUnit.name) AS codeUnitName
+OPTIONAL MATCH (artifact:Java:Artifact)-[:CONTAINS]->(codeUnit)
+   WITH *, replace(last(split(artifact.fileName, '/')), '.jar', '') AS artifactName
+OPTIONAL MATCH (projectRoot:Directory)<-[:HAS_ROOT]-(proj:TS:Project)-[:CONTAINS]->(codeUnit)
+   WITH *, last(split(projectRoot.absoluteFileName, '/')) AS projectName   
+ RETURN DISTINCT 
+        coalesce(codeUnit.fqn, codeUnit.globalFqn, codeUnit.fileName, codeUnit.signature, codeUnit.name) AS codeUnitName
+       ,coalesce(codeUnit.name, replace(last(split(codeUnit.fileName, '/')), '.jar', '')) AS shortCodeUnitName
+       ,coalesce(artifactName, projectName) AS projectName
      ,coalesce(codeUnit.communityLeidenId, 0) AS communityId
      ,coalesce(codeUnit.centralityPageRank, 0.01) AS centrality
-     ,replace(last(split(artifact.fileName, '/')), '.jar', '') AS artifactName
      ,embedding

cypher/Node_Embeddings/Node_Embeddings_3e_Node2Vec_Write.cypher

@@ -4,6 +4,7 @@ CALL gds.node2vec.write(
  $dependencies_projection + '-cleaned', {
      ,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
      ,iterations: 3
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,writeProperty: $dependencies_projection_write_property
   }

cypher/Typescript_Enrichment/Add_RESOLVES_TO_relationship_for_matching_declarations.cypher

@@ -1,13 +1,13 @@
 // Adds a relation "RESOLVES_TO" from a Typescript element to an external declaration if their global fully qualified names match.
 // Inspired by https://github.com/jQAssistant/jqa-java-plugin/blob/f092122b62bb13d597840b64b73b2010bd074d1f/src/main/resources/META-INF/jqassistant-rules/java-classpath.xml#L5
+// Related to https://github.com/jqassistant-plugin/jqassistant-typescript-plugin/issues/35
 
 MATCH (element:TS&!ExternalDeclaration)
 WHERE  element.globalFqn IS NOT NULL
 MATCH (external:TS&ExternalDeclaration)
 WHERE  element.globalFqn IS NOT NULL 
-  AND (element.globalFqn = external.globalFqn
-   OR  toLower(element.globalFqn) = toLower(external.globalFqn))
-  AND element <> external
+  AND  element.globalFqn = external.globalFqn
+  AND  element <> external
  CALL {  WITH element, external
         MERGE (external)-[:RESOLVES_TO]->(element)
       } IN TRANSACTIONS

cypher/Typescript_Enrichment/Add_RESOLVES_TO_relationship_for_matching_modules.cypher

@@ -1,13 +1,14 @@
 // Adds a relation "RESOLVES_TO" from an external module to a module if their global fully qualified names match.
 // Inspired by https://github.com/jQAssistant/jqa-java-plugin/blob/f092122b62bb13d597840b64b73b2010bd074d1f/src/main/resources/META-INF/jqassistant-rules/java-classpath.xml#L5
+// Related to https://github.com/jqassistant-plugin/jqassistant-typescript-plugin/issues/35
 
 MATCH (module:TS:Module)
 WHERE  module.globalFqn IS NOT NULL
 MATCH (externalModule:TS:ExternalModule)
 WHERE module.globalFqn IS NOT NULL
-  AND (toLower(module.globalFqn) = toLower(externalModule.globalFqn)
-   OR  toLower(module.globalFqn) = split(toLower(externalModule.globalFqn), '/index.')[0]
-   OR  toLower(externalModule.globalFqn) = split(toLower(module.globalFqn), '/index.')[0]
+  AND (module.globalFqn = externalModule.globalFqn
+   OR  module.globalFqn = split(externalModule.globalFqn, '/index.')[0]
+   OR  externalModule.globalFqn = split(module.globalFqn, '/index.')[0]
       )
   AND module <> externalModule
  CALL {  WITH module, externalModule