txgraph: Add CountDistinctClusters function (feature)

Author: sipa

src/test/fuzz/txgraph.cpp

@@ -526,6 +526,50 @@ FUZZ_TARGET(txgraph)
                 // these here without making more calls to real, which could affect its internal
                 // state. A full comparison is done at the end.
                 break;
+            } else if (!sel_sim.IsOversized() && command-- == 0) {
+                // CountDistinctClusters.
+                std::vector<TxGraph::Ref*> refs;
+                // Gather a list of up to 15 (or up to 255) Ref pointers.
+                auto count = provider.ConsumeIntegralInRange<size_t>(0, alt ? 255 : 15);
+                refs.resize(count);
+                for (size_t i = 0; i < count; ++i) {
+                    refs[i] = pick_fn();
+                }
+                // Their order should not matter, shuffle them.
+                std::shuffle(refs.begin(), refs.end(), rng);
+                // Invoke the real function.
+                auto result = real->CountDistinctClusters(refs, use_main);
+                // Build a vector with representatives of the clusters the Refs occur in in the
+                // simulated graph. For each, remember the lowest-index transaction SimPos in the
+                // cluster.
+                std::vector<DepGraphIndex> sim_reps;
+                for (auto ref : refs) {
+                    // Skip Refs that do not occur in the simulated graph.
+                    auto simpos = sel_sim.Find(ref);
+                    if (simpos == SimTxGraph::MISSING) continue;
+                    // Start with component equal to just the Ref's SimPos.
+                    auto component = SimTxGraph::SetType::Singleton(simpos);
+                    // Keep adding ancestors/descendants of all elements in component until it no
+                    // longer changes.
+                    while (true) {
+                        auto old_component = component;
+                        for (auto i : component) {
+                            component |= sel_sim.graph.Ancestors(i);
+                            component |= sel_sim.graph.Descendants(i);
+                        }
+                        if (component == old_component) break;
+                    }
+                    // Remember the lowest-index SimPos in component, as a representative for it.
+                    assert(component.Any());
+                    sim_reps.push_back(component.First());
+                }
+                // Remove duplicates from sim_reps.
+                std::sort(sim_reps.begin(), sim_reps.end());
+                sim_reps.erase(std::unique(sim_reps.begin(), sim_reps.end()), sim_reps.end());
+                // Compare the number of deduplicated representatives with the value returned by
+                // the real function.
+                assert(result == sim_reps.size());
+                break;
             } else if (command-- == 0) {
                 // DoWork.
                 real->DoWork();

src/txgraph.cpp

@@ -454,6 +454,7 @@ class TxGraphImpl final : public TxGraph
     GraphIndex GetTransactionCount(bool main_only = false) noexcept final;
     bool IsOversized(bool main_only = false) noexcept final;
     std::strong_ordering CompareMainOrder(const Ref& a, const Ref& b) noexcept final;
+    GraphIndex CountDistinctClusters(std::span<const Ref* const> refs, bool main_only = false) noexcept final;
 
     void SanityCheck() const final;
 };
@@ -1781,6 +1782,33 @@ std::strong_ordering TxGraphImpl::CompareMainOrder(const Ref& a, const Ref& b) n
     return entry_a.m_main_lin_index <=> entry_b.m_main_lin_index;
 }
 
+TxGraph::GraphIndex TxGraphImpl::CountDistinctClusters(std::span<const Ref* const> refs, bool main_only) noexcept
+{
+    size_t level = GetSpecifiedLevel(main_only);
+    ApplyDependencies(level);
+    auto& clusterset = GetClusterSet(level);
+    Assume(clusterset.m_deps_to_add.empty());
+    // Build a vector of Clusters that the specified Refs occur in.
+    std::vector<Cluster*> clusters;
+    clusters.reserve(refs.size());
+    for (const Ref* ref : refs) {
+        if (ref == nullptr) continue;
+        if (GetRefGraph(*ref) == nullptr) continue;
+        Assume(GetRefGraph(*ref) == this);
+        auto cluster = FindCluster(GetRefIndex(*ref), level);
+        if (cluster != nullptr) clusters.push_back(cluster);
+    }
+    // Count the number of distinct elements in clusters.
+    std::sort(clusters.begin(), clusters.end());
+    Cluster* last{nullptr};
+    GraphIndex ret{0};
+    for (Cluster* cluster : clusters) {
+        ret += (cluster != last);
+        last = cluster;
+    }
+    return ret;
+}
+
 void Cluster::SanityCheck(const TxGraphImpl& graph, int level) const
 {
     // There must be an m_mapping for each m_depgraph position (including holes).

src/txgraph.h

@@ -149,6 +149,11 @@ class TxGraph
     /** Compare two transactions according to their order in the main graph. Both transactions must
      *  be in the main graph. The main graph must not be oversized. */
     virtual std::strong_ordering CompareMainOrder(const Ref& a, const Ref& b) noexcept = 0;
+    /** Count the number of distinct clusters that the specified transactions belong to. If
+     *  main_only is false and a staging graph exists, staging clusters are counted. Otherwise,
+     *  main clusters are counted. Refs that do not exist in the queried graph are ignored. The
+     *  queried graph must not be oversized. */
+    virtual GraphIndex CountDistinctClusters(std::span<const Ref* const>, bool main_only = false) noexcept = 0;
 
     /** Perform an internal consistency check on this object. */
     virtual void SanityCheck() const = 0;