diff --git a/dockerfiles/scripts/compress_graph.sh b/dockerfiles/scripts/compress_graph.sh --- a/dockerfiles/scripts/compress_graph.sh +++ b/dockerfiles/scripts/compress_graph.sh @@ -60,11 +60,13 @@ { # Build a function (MPH) that maps node names to node numbers in # lexicographic order (output: .mph) + echo -e "# MPH step\n" && java_cmd it.unimi.dsi.sux4j.mph.GOVMinimalPerfectHashFunction \ --zipped $compr_graph_path.mph --temp-dir $tmp_dir \ $graph_path.nodes.csv.gz && # Build the graph in BVGraph format (output: .{graph,offsets,properties}) + echo -e "\n# BV compress step\n" && java_cmd it.unimi.dsi.big.webgraph.ScatteredArcsASCIIGraph \ --function $compr_graph_path.mph --temp-dir $tmp_dir \ --zipped $compr_graph_path-bv < $graph_path.edges.csv.gz && @@ -73,10 +75,12 @@ --list $compr_graph_path-bv && # Find a better permutation using a BFS traversal order (output: .order) + echo -e "\n# BFS step\n" && java_cmd it.unimi.dsi.law.big.graph.BFS \ $compr_graph_path-bv $compr_graph_path.order && # Permute the graph accordingly + echo -e "\n# Permute step\n" && java_cmd it.unimi.dsi.big.webgraph.Transform mapOffline \ $compr_graph_path-bv $compr_graph_path \ $compr_graph_path.order $batch_size $tmp_dir && @@ -84,9 +88,11 @@ --list $compr_graph_path && # Compute graph statistics (output: .{indegree,outdegree,stats}) + echo -e "\n# Stats step\n" && java_cmd it.unimi.dsi.big.webgraph.Stats $compr_graph_path && # Create transposed graph (to allow backward traversal) + echo -e "\n# Transpose step\n" && java_cmd it.unimi.dsi.big.webgraph.Transform transposeOffline \ $compr_graph_path $compr_graph_path-transposed \ $batch_size $tmp_dir &&