Add community metrics with Modularity&Conductance

Author: JohT

cypher/Community_Detection/Community_Detection_9_Community_Metrics.cypher

@@ -0,0 +1,44 @@
+// Community Metrics
+
+  CALL gds.conductance.stream(
+ $dependencies_projection + '-cleaned', {
+     relationshipWeightProperty: $dependencies_projection_weight_property
+    ,communityProperty: $dependencies_projection_write_property
+})
+ YIELD community AS communityId, conductance
+  WITH collect({communityId: communityId, conductance: conductance}) AS conductances
+  CALL gds.modularity.stream(
+ $dependencies_projection + '-cleaned', {
+     relationshipWeightProperty: $dependencies_projection_weight_property
+    ,communityProperty: $dependencies_projection_write_property
+})
+ YIELD communityId, modularity
+  WITH conductances
+      ,collect({communityId: communityId, modularity: modularity}) AS modularities
+ MATCH (member)
+ WHERE member[$dependencies_projection_write_property] IS NOT NULL
+   AND $dependencies_projection_node IN LABELS(member)
+  WITH conductances
+      ,modularities
+      ,member[$dependencies_projection_write_property]                               AS communityId
+      ,coalesce(member.fqn, member.fileName, member.name)                            AS memberName
+      ,coalesce(member.name, replace(last(split(member.fileName, '/')), '.jar', '')) AS shortMemberName
+  WITH conductances
+      ,modularities
+      ,communityId
+      ,count(DISTINCT memberName)        AS memberCount
+      ,collect(DISTINCT shortMemberName) AS shortMemberNames
+      ,collect(DISTINCT memberName)      AS memberNames
+      ,reduce(memberConductance = 0, conductance IN conductances | 
+        CASE conductance.communityId WHEN communityId THEN conductance.conductance 
+        ELSE memberConductance END)      AS conductance
+      ,reduce(memberModularity = 0, modularity IN modularities | 
+        CASE modularity.communityId WHEN communityId THEN modularity.modularity 
+        ELSE memberModularity END)       AS modularity
+ RETURN communityId
+       ,conductance
+       ,modularity
+       ,memberCount
+       ,shortMemberNames[0..9] AS someMemberNamesShort
+       ,memberNames[0..9]      AS someMemberNames
+ORDER BY communityId ASCENDING

scripts/reports/CommunityCsv.sh

@@ -114,7 +114,7 @@ detectCommunitiesWithLouvain() {
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_10_Delete_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_11_Add_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
 
-    queryCommunitiyModularities "${@}" "${writePropertyName}"
+    calculateCommunityMetrics "${@}" "${writePropertyName}"
 }
 
 # Community Detection using the Leiden Algorithm
@@ -160,7 +160,7 @@ detectCommunitiesWithLeiden() {
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_10_Delete_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_11_Add_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
 
-    queryCommunitiyModularities "${@}" "${writePropertyName}"
+    calculateCommunityMetrics "${@}" "${writePropertyName}"
 }
 
 # Community Detection using the Weakly Connected Components Algorithm
@@ -196,6 +196,8 @@ detectCommunitiesWithWeaklyConnectedComponents() {
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_9_Write_Mutated.cypher" "${@}" "${writePropertyName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_10_Delete_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_11_Add_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
+
+    calculateCommunityMetrics "${@}" "${writePropertyName}"
 }
 
 # Community Detection using the Label Propagation Algorithm
@@ -231,6 +233,8 @@ detectCommunitiesWithLabelPropagation() {
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_9_Write_Mutated.cypher" "${@}" "${writePropertyName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_10_Delete_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_11_Add_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
+
+    calculateCommunityMetrics "${@}" "${writePropertyName}"
 }
 
 # Community Detection using the K-Core Decomposition Algorithm
@@ -265,6 +269,8 @@ detectCommunitiesWithKCoreDecomposition() {
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_9_Write_Mutated.cypher" "${@}" "${writePropertyName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_10_Delete_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_11_Add_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
+
+    calculateCommunityMetrics "${@}" "${writePropertyName}"
 }
 
 # Community Detection using the Approximate Maximum k-cut Algorithm
@@ -301,9 +307,11 @@ detectCommunitiesWithApproximateMaximumKCut() {
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_9_Write_Mutated.cypher" "${@}" "${writePropertyName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_10_Delete_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
     execute_cypher "${PROJECTION_CYPHER_DIR}/Dependencies_11_Add_Label.cypher" "${@}" "${writePropertyName}" "${writeLabelName}"
+
+    calculateCommunityMetrics "${@}" "${writePropertyName}"
 }
 
-# Query modularity for Leiden communities
+# Calculates community metrics including "Modularity" and "Conductance"
 # 
 # Required Parameters:
 # - dependencies_projection=...
@@ -312,7 +320,7 @@ detectCommunitiesWithApproximateMaximumKCut() {
 #   Name of the property that contains the communitiy id
 # - dependencies_projection_weight_property=...
 #   Name of the node property that contains the dependency weight. Example: "weight"
-queryCommunitiyModularities() {
+calculateCommunityMetrics() {
     local COMMUNITY_DETECTION_CYPHER_DIR="${CYPHER_DIR}/Community_Detection"
 
     local nodeLabel
@@ -322,7 +330,7 @@ queryCommunitiyModularities() {
     propertyName=$( extractQueryParameter "dependencies_projection_write_property" "${@}")
 
     # Stream to CSV
-    execute_cypher "${COMMUNITY_DETECTION_CYPHER_DIR}/Community_Detection_7d_Modularity_Members.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}_${propertyName}_Communities_Modularities.csv"
+    execute_cypher "${COMMUNITY_DETECTION_CYPHER_DIR}/Community_Detection_9_Community_Metrics.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}_${propertyName}_Community_Metrics.csv"
 }
 
 # Write modularity for Leiden communities