From 646df4803e76e692da86f0567c33c0b5a1ee6c79 Mon Sep 17 00:00:00 2001
From: Arun Sharma <arun@sharma-home.net>
Date: Fri, 1 May 2026 09:58:00 -0700
Subject: [PATCH] Avoid extra Leiden mapping copies

Use references to the fine-to-coarse mapping while the newly created
coarsened view owns it, then copy that mapping only once into
composedMapping for flattening. This removes avoidable O(n) vector
copies during each ParallelLeidenView coarsening handoff without
changing ownership or partition semantics.
---
 networkit/cpp/community/ParallelLeidenView.cpp | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)

diff --git a/networkit/cpp/community/ParallelLeidenView.cpp b/networkit/cpp/community/ParallelLeidenView.cpp
index 22a2f218c..890467545 100644
--- a/networkit/cpp/community/ParallelLeidenView.cpp
+++ b/networkit/cpp/community/ParallelLeidenView.cpp
@@ -261,7 +261,7 @@ void ParallelLeidenView::run() {
                 ParallelPartitionCoarseningView ppcView(*G, originalRefined);
                 ppcView.run();
                 auto newCoarsenedView = ppcView.getCoarsenedGraphView();
-                auto map = ppcView.getFineToCoarseNodeMapping();
+                const auto &map = ppcView.getFineToCoarseNodeMapping();
                 const auto &oldNodeMapping = currentCoarsenedView->getNodeMapping();
 
                 // Create new partition for the new coarsened view
@@ -269,8 +269,7 @@ void ParallelLeidenView::run() {
                 p.setUpperBound(result.upperBound());
 
                 // Map communities through the chain of coarsenings BEFORE moving map
-                // Capture map by value and other refs explicitly
-                G->parallelForNodes([map = map, &p, this, &oldNodeMapping](node originalNode) {
+                G->parallelForNodes([&map, &p, this, &oldNodeMapping](node originalNode) {
                     node newCoarseNode = map[originalNode];
                     // Find what community this original node belonged to
                     node oldCoarseNode = oldNodeMapping[originalNode];
@@ -279,7 +278,7 @@ void ParallelLeidenView::run() {
 
                 // Since ppcView is always built from *G, map already represents
                 // original -> current-coarse directly. No chain composition needed.
-                composedMapping = std::move(map);
+                composedMapping = map;
 
                 result = std::move(p);
                 currentCoarsenedView = newCoarsenedView;
@@ -289,16 +288,15 @@ void ParallelLeidenView::run() {
                 ParallelPartitionCoarseningView ppcView(*currentGraph, refined);
                 ppcView.run();
                 auto newCoarsenedView = ppcView.getCoarsenedGraphView();
-                auto map = ppcView.getFineToCoarseNodeMapping();
+                const auto &map = ppcView.getFineToCoarseNodeMapping();
 
                 // Maintain Partition, add every coarse Node to the community its fine Nodes were in
                 Partition p(newCoarsenedView->numberOfNodes());
                 p.setUpperBound(result.upperBound());
-                currentGraph->parallelForNodes(
-                    [map = map, &p, this](node u) { p[map[u]] = result[u]; });
+                currentGraph->parallelForNodes([&map, &p, this](node u) { p[map[u]] = result[u]; });
 
                 // ppcView is built from *G so map is already original -> coarse; just store it.
-                composedMapping = std::move(map);
+                composedMapping = map;
                 result = std::move(p);
 
                 currentCoarsenedView = newCoarsenedView;