tg_lcc.gsql

CREATE DISTRIBUTED QUERY tg_lcc (STRING v_type, STRING e_type,INT top_k=100,BOOL print_results = True, STRING result_attribute = "",
  STRING file_path = "", BOOL display_edges = FALSE)  SYNTAX V1 { 

    /*
    First Author: <First Author Name>
    First Commit Date:  <First Commit Date>

    Recent Author: <Recent Commit Author Name>
    Recent Commit Date: <Recent Commit Date>


    Repository:
        https://github.com/tigergraph/gsql-graph-algorithms/tree/master/algorithms/Community

    Maturity:
        Production

    Description: 
        The Local Clustering Coefficient algorithm computes the local clustering coefficient 
        for each node in the graph. 
        lcc = Number_trangles/((n-1)n/2)
        Here n is the outdegreeof vertex.

    Publications:
        NA

    TigerGraph Documentation:
        https://docs.tigergraph.com/graph-ml/current/community-algorithms/local-clustering-coefficient

    Parameters:
        v_type:
            vertex types to traverse
        print_results:
            If True, print JSON output
        e_type:
            edge types to traverse
        result_attribute:
            INT attribute to store results to
        top_k:
            report only this many top scores
        file_path:
            file to write CSV output to
        display_edges:
            If True, output edges for visualization
    */
   
  TYPEDEF TUPLE<VERTEX Vertex_ID, FLOAT score> Vertex_Score;
  HeapAccum<Vertex_Score>(top_k, score DESC) @@top_scores_heap;
  SumAccum<FLOAT> @sum_tri; #number of trangles
  SumAccum<FLOAT> @sum_lcc; #lcc value
  SetAccum<int> @neighbors_set; #neighbors set
  OrAccum @or_self_con; #check if the vertex is self-connect
  SetAccum<EDGE> @@edge_set;
  FILE f (file_path);
  # Here we compute the intersection for 2 points on the triangle.
  
  Start = {v_type};
  Start = SELECT s 
          FROM Start:s-(e_type)-v_type:t
          ACCUM 
              IF getvid(s) != getvid(t) THEN 
                  t.@neighbors_set += getvid(s)
              ELSE   
                  t.@or_self_con+=TRUE 
              END;# check id the vertex is self-connect
                           
  Start = SELECT s 
          FROM Start:s-(e_type)-v_type:t
          WHERE s.outdegree(e_type)>1
          ACCUM 
              s.@sum_tri+=COUNT((s.@neighbors_set INTERSECT t.@neighbors_set))
          POST-ACCUM 
              IF s.@or_self_con  AND s.outdegree(e_type)<3 THEN 
                      s.@sum_lcc+=0
              ELSE IF s.@or_self_con AND s.outdegree(e_type)>2 THEN 
                      s.@sum_lcc+= (((s.@sum_tri+1-s.outdegree(e_type)))/((s.outdegree(e_type)-2)*(s.outdegree(e_type)-1)))
              ELSE 
                      s.@sum_lcc+= ((s.@sum_tri)/((s.outdegree(e_type)-0)*(s.outdegree(e_type)-1)))
              END;
        
  #output
  Start = SELECT s 
          FROM Start:s
          # Calculate Closeness Centrality for each vertex
	  POST-ACCUM 
	      IF result_attribute != "" THEN 
	          s.setAttr(result_attribute, s.@sum_lcc) 
              END,
	      IF print_results THEN 
	          @@top_scores_heap += Vertex_Score(s, s.@sum_lcc) 
              END,
	      IF file_path != "" THEN 
	          f.println(s, s.@sum_lcc) 
	      END;
	      
  IF file_path != "" THEN
      f.println("Vertex_ID", "lcc");
  END;

  IF print_results THEN
      PRINT @@top_scores_heap AS top_scores;
      IF display_edges THEN
          PRINT Start[Start.@sum_lcc];
	  Start = SELECT s
		  FROM Start:s -(e_type:e)-:t
		  ACCUM @@edge_set += e;
		  PRINT @@edge_set;
      END;
  END;
}