From c0210c8d552e07b63c28358a5856a77b26af8daa Mon Sep 17 00:00:00 2001
From: Marin Tomic <marin.tomic@memgraph.io>
Date: Wed, 28 Mar 2018 13:53:00 +0200
Subject: [PATCH] Real-world like graph generator

---
 .../macro_benchmark/clients/graph_500_bfs.cpp | 62 +++++++++++++
 tests/manual/graph_500_generate_snapshot.cpp  | 91 +++++++++++++++++++
 2 files changed, 153 insertions(+)
 create mode 100644 tests/macro_benchmark/clients/graph_500_bfs.cpp
 create mode 100644 tests/manual/graph_500_generate_snapshot.cpp

diff --git a/tests/macro_benchmark/clients/graph_500_bfs.cpp b/tests/macro_benchmark/clients/graph_500_bfs.cpp
new file mode 100644
index 000000000..68053a373
--- /dev/null
+++ b/tests/macro_benchmark/clients/graph_500_bfs.cpp
@@ -0,0 +1,62 @@
+#include <memory>
+#include <random>
+#include <shared_mutex>
+#include <vector>
+
+#include "gflags/gflags.h"
+
+#include "long_running_common.hpp"
+#include "stats/stats.hpp"
+#include "stats/stats_rpc_messages.hpp"
+#include "threading/sync/rwlock.hpp"
+
+class Graph500BfsClient : public TestClient {
+ public:
+  Graph500BfsClient(int id) : TestClient(), rg_(id) {
+    auto result = Execute("MATCH (n:Node) RETURN count(1)", {}, "NumNodes");
+    CHECK(result) << "Read-only query should not fail";
+    num_nodes_ = result->records[0][0].ValueInt();
+  }
+
+ private:
+  std::mt19937 rg_;
+  int num_nodes_;
+
+  void Step() override {
+    std::uniform_int_distribution<int64_t> dist(0, num_nodes_ - 1);
+
+    int start = -1;
+    do {
+      start = dist(rg_);
+      auto result = Execute(
+          "MATCH (n:Node {id: $id})-->(m) WHERE m != n "
+          "RETURN count(m) AS degree",
+          {{"id", start}}, "GetDegree");
+      CHECK(result) << "Read-only query should not fail";
+      if (result->records[0][0].ValueInt() > 0) {
+        break;
+      }
+    } while (true);
+
+    auto result =
+        Execute("MATCH path = (n:Node {id: $id})-[*bfs]->() RETURN count(1)",
+                {{"id", start}}, "Bfs");
+    CHECK(result) << "Read-only query should not fail!";
+  }
+};
+
+int main(int argc, char **argv) {
+  gflags::ParseCommandLineFlags(&argc, &argv, true);
+  google::InitGoogleLogging(argv[0]);
+
+  std::vector<std::unique_ptr<TestClient>> clients;
+  for (int i = 0; i < FLAGS_num_workers; ++i) {
+    clients.emplace_back(std::make_unique<Graph500BfsClient>(i));
+  }
+
+  RunMultithreadedTest(clients);
+
+  stats::StopStatsLogging();
+
+  return 0;
+}
diff --git a/tests/manual/graph_500_generate_snapshot.cpp b/tests/manual/graph_500_generate_snapshot.cpp
new file mode 100644
index 000000000..326028b9f
--- /dev/null
+++ b/tests/manual/graph_500_generate_snapshot.cpp
@@ -0,0 +1,91 @@
+#include <random>
+
+#include "gflags/gflags.h"
+#include "glog/logging.h"
+#include "json/json.hpp"
+
+#include "snapshot_generation/snapshot_writer.hpp"
+
+DEFINE_int32(num_workers, 1,
+             "Number of distributed workers (including master)");
+DEFINE_string(dir, "tmp",
+              "Directory for storing workers durability directories.");
+DEFINE_string(config, "", "Path to config JSON file");
+
+/**
+ * Config file should be defined as follows:
+ * {
+ *   "scale": 10,
+ *   "edge_factor": 16,
+ *   "probabilities": [0.57, 0.19, 0.19]
+ * }
+ */
+
+using namespace snapshot_generation;
+
+int main(int argc, char *argv[]) {
+  google::ParseCommandLineFlags(&argc, &argv, true);
+  google::InitGoogleLogging(argv[0]);
+
+  nlohmann::json config;
+  {
+    std::ifstream config_file(FLAGS_config);
+    config_file >> config;
+  }
+
+  int64_t N = 1LL << config["scale"].get<int>();
+  int64_t M = config["edge_factor"].get<double>() * N;
+
+  const double A = config["probabilities"][0];
+  const double B = config["probabilities"][1];
+  const double C = config["probabilities"][2];
+  const double D = 1 - (A + B + C);
+
+  std::vector<std::pair<int64_t, int64_t>> edges(M);
+
+  std::mt19937_64 gen(std::random_device{}());
+  std::uniform_real_distribution<double> dist(0, 1);
+
+  for (int i = 0; i < M; ++i) {
+    for (int j = 0; j < config["scale"]; ++j) {
+      if (dist(gen) > A + B) {
+        edges[i].first |= 1 << j;
+        if (dist(gen) > C / (C + D)) {
+          edges[i].second |= 1 << j;
+        }
+      } else {
+        if (dist(gen) > A / (A + B)) {
+          edges[i].second |= 1 << j;
+        }
+      }
+    }
+  }
+
+  std::vector<int64_t> vertex_labels(N);
+  std::iota(vertex_labels.begin(), vertex_labels.end(), 0);
+  std::random_shuffle(vertex_labels.begin(), vertex_labels.end());
+
+  GraphState state(FLAGS_num_workers);
+
+  state.CreateIndex("Node", "id");
+
+  std::vector<gid::Gid> vertices;
+  vertices.reserve(N);
+
+  for (int i = 0; i < N; ++i) {
+    vertices.emplace_back(state.CreateNode(i % FLAGS_num_workers, {"Node"},
+                                           {{"id", vertex_labels[i]}}));
+  }
+
+  std::random_shuffle(edges.begin(), edges.end());
+  for (int i = 0; i < M; ++i) {
+    auto e = edges[i];
+    VLOG(1) << vertex_labels[e.first] << " " << vertex_labels[e.second];
+    state.CreateEdge(vertices[e.first], vertices[e.second], "Edge", {});
+  }
+
+  LOG(INFO) << fmt::format("nodes = {}, edges = {}", N, M);
+  WriteToSnapshot(state, FLAGS_dir);
+
+  return 0;
+}