Merge branch 'branch-24.04' into bug_edge_masking

.github/workflows/pr.yaml

@@ -162,7 +162,7 @@ jobs:
     with:
       build_type: pull-request
       script: ci/test_wheel_cugraph-pyg.sh
-      matrix_filter: map(select(.ARCH == "amd64" and .CUDA_VER == "11.8.0"))
+      matrix_filter: map(select(.ARCH == "amd64"))
   wheel-build-cugraph-equivariant:
     secrets: inherit
     uses: rapidsai/shared-workflows/.github/workflows/[email protected]

.github/workflows/test.yaml

@@ -76,7 +76,7 @@ jobs:
       date: ${{ inputs.date }}
       sha: ${{ inputs.sha }}
       script: ci/test_wheel_cugraph-pyg.sh
-      matrix_filter: map(select(.ARCH == "amd64" and .CUDA_VER == "11.8.0"))
+      matrix_filter: map(select(.ARCH == "amd64"))
   wheel-tests-cugraph-equivariant:
     secrets: inherit
     uses: rapidsai/shared-workflows/.github/workflows/[email protected]

ci/test_python.sh

@@ -166,7 +166,7 @@ if [[ "${RAPIDS_CUDA_VERSION}" == "11.8.0" ]]; then
       pylibcugraphops \
       cugraph \
       cugraph-dgl \
-      'dgl>=1.1.0.cu*' \
+      'dgl>=1.1.0.cu*,<=2.0.0.cu*' \
       'pytorch>=2.0' \
       'pytorch-cuda>=11.8'
 

ci/test_wheel_cugraph-dgl.sh

@@ -34,6 +34,6 @@ DGL_URL="https://data.dgl.ai/wheels/cu${PYTORCH_CUDA_VER}/repo.html"
 
 rapids-logger "Installing PyTorch and DGL"
 rapids-retry python -m pip install torch --index-url ${PYTORCH_URL}
-rapids-retry python -m pip install dgl --find-links ${DGL_URL}
+rapids-retry python -m pip install dgl==2.0.0 --find-links ${DGL_URL}
 
 python -m pytest python/cugraph-dgl/tests

cpp/CMakeLists.txt

@@ -423,6 +423,7 @@ add_library(cugraph_c
         src/c_api/core_number.cpp
         src/c_api/core_result.cpp
         src/c_api/extract_ego.cpp
+        src/c_api/ecg.cpp
         src/c_api/k_core.cpp
         src/c_api/hierarchical_clustering_result.cpp
         src/c_api/induced_subgraph.cpp

cpp/include/cugraph_c/community_algorithms.h

@@ -177,6 +177,41 @@ double cugraph_hierarchical_clustering_result_get_modularity(
  */
 void cugraph_hierarchical_clustering_result_free(cugraph_hierarchical_clustering_result_t* result);
 
+/**
+ * @brief     Compute ECG clustering
+ *
+ * @param [in]  handle        Handle for accessing resources
+ * @param [in/out] rng_state  State of the random number generator, updated with each call
+ * @param [in]  graph         Pointer to graph.  NOTE: Graph might be modified if the storage
+ *                            needs to be transposed
+ * @param [in]  min_weight    Minimum edge weight in final graph
+ * @param [in]  ensemble_size The number of Louvain iterations to run
+ * @param [in]  max_level     Maximum level in hierarchy for final Louvain
+ * @param [in]  threshold     Threshold parameter, defines convergence at each level of hierarchy
+ *                            for final Louvain
+ * @param [in]  resolution    Resolution parameter (gamma) in modularity formula.
+ *                            This changes the size of the communities.  Higher resolutions
+ *                            lead to more smaller communities, lower resolutions lead to
+ *                            fewer larger communities.
+ * @param [in]  do_expensive_check
+ *                            A flag to run expensive checks for input arguments (if set to true)
+ * @param [out] result        Output from the Louvain call
+ * @param [out] error         Pointer to an error object storing details of any error.  Will
+ *                            be populated if error code is not CUGRAPH_SUCCESS
+ * @return error code
+ */
+cugraph_error_code_t cugraph_ecg(const cugraph_resource_handle_t* handle,
+                                 cugraph_rng_state_t* rng_state,
+                                 cugraph_graph_t* graph,
+                                 double min_weight,
+                                 size_t ensemble_size,
+                                 size_t max_level,
+                                 double threshold,
+                                 double resolution,
+                                 bool_t do_expensive_check,
+                                 cugraph_hierarchical_clustering_result_t** result,
+                                 cugraph_error_t** error);
+
 /**
  * @brief     Compute ECG clustering of the given graph
  *
@@ -200,13 +235,13 @@ void cugraph_hierarchical_clustering_result_free(cugraph_hierarchical_clustering
  *                              be populated if error code is not CUGRAPH_SUCCESS
  * @return error code
  */
-cugraph_error_code_t cugraph_ecg(const cugraph_resource_handle_t* handle,
-                                 cugraph_graph_t* graph,
-                                 double min_weight,
-                                 size_t ensemble_size,
-                                 bool_t do_expensive_check,
-                                 cugraph_hierarchical_clustering_result_t** result,
-                                 cugraph_error_t** error);
+cugraph_error_code_t cugraph_legacy_ecg(const cugraph_resource_handle_t* handle,
+                                        cugraph_graph_t* graph,
+                                        double min_weight,
+                                        size_t ensemble_size,
+                                        bool_t do_expensive_check,
+                                        cugraph_hierarchical_clustering_result_t** result,
+                                        cugraph_error_t** error);
 
 /**
  * @brief   Extract ego graphs

cpp/src/c_api/ecg.cpp

@@ -0,0 +1,154 @@
+/*
+ * Copyright (c) 2022-2024, NVIDIA CORPORATION.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <cugraph_c/community_algorithms.h>
+
+#include <cugraph/algorithms.hpp>
+#include <cugraph/detail/shuffle_wrappers.hpp>
+#include <cugraph/detail/utility_wrappers.hpp>
+#include <cugraph/graph_functions.hpp>
+
+#include <c_api/abstract_functor.hpp>
+#include <c_api/graph.hpp>
+#include <c_api/graph_helper.hpp>
+#include <c_api/hierarchical_clustering_result.hpp>
+#include <c_api/random.hpp>
+#include <c_api/resource_handle.hpp>
+#include <c_api/utils.hpp>
+
+#include <optional>
+
+namespace {
+
+struct ecg_functor : public cugraph::c_api::abstract_functor {
+  raft::handle_t const& handle_;
+  cugraph::c_api::cugraph_rng_state_t* rng_state_{nullptr};
+  cugraph::c_api::cugraph_graph_t* graph_{nullptr};
+  double min_weight_{0.1};
+  size_t ensemble_size_{10};
+  size_t max_level_{0};
+  double threshold_{0.001};
+  double resolution_{1};
+  bool do_expensive_check_{false};
+  cugraph::c_api::cugraph_hierarchical_clustering_result_t* result_{};
+
+  ecg_functor(::cugraph_resource_handle_t const* handle,
+              ::cugraph_rng_state_t* rng_state,
+              ::cugraph_graph_t* graph,
+              double min_weight,
+              size_t ensemble_size,
+              size_t max_level,
+              double threshold,
+              double resolution,
+              bool do_expensive_check)
+    : abstract_functor(),
+      handle_(*reinterpret_cast<cugraph::c_api::cugraph_resource_handle_t const*>(handle)->handle_),
+      rng_state_(reinterpret_cast<cugraph::c_api::cugraph_rng_state_t*>(rng_state)),
+      graph_(reinterpret_cast<cugraph::c_api::cugraph_graph_t*>(graph)),
+      max_level_(max_level),
+      threshold_(threshold),
+      resolution_(resolution),
+      do_expensive_check_(do_expensive_check)
+  {
+  }
+
+  template <typename vertex_t,
+            typename edge_t,
+            typename weight_t,
+            typename edge_type_type_t,
+            bool store_transposed,
+            bool multi_gpu>
+  void operator()()
+  {
+    if constexpr (!cugraph::is_candidate<vertex_t, edge_t, weight_t>::value) {
+      unsupported();
+    } else {
+      // ecg expects store_transposed == false
+      if constexpr (store_transposed) {
+        error_code_ = cugraph::c_api::
+          transpose_storage<vertex_t, edge_t, weight_t, store_transposed, multi_gpu>(
+            handle_, graph_, error_.get());
+        if (error_code_ != CUGRAPH_SUCCESS) return;
+      }
+
+      auto graph =
+        reinterpret_cast<cugraph::graph_t<vertex_t, edge_t, false, multi_gpu>*>(graph_->graph_);
+
+      auto graph_view = graph->view();
+
+      auto edge_weights = reinterpret_cast<
+        cugraph::edge_property_t<cugraph::graph_view_t<vertex_t, edge_t, false, multi_gpu>,
+                                 weight_t>*>(graph_->edge_weights_);
+
+      auto number_map = reinterpret_cast<rmm::device_uvector<vertex_t>*>(graph_->number_map_);
+
+      rmm::device_uvector<vertex_t> clusters(0, handle_.get_stream());
+
+      weight_t modularity;
+
+      std::tie(clusters, std::ignore, modularity) =
+        cugraph::ecg(handle_,
+                     rng_state_->rng_state_,
+                     graph_view,
+                     (edge_weights != nullptr)
+                       ? std::make_optional(edge_weights->view())
+                       : std::make_optional(cugraph::c_api::create_constant_edge_property(
+                                              handle_, graph_view, weight_t{1})
+                                              .view()),
+                     static_cast<weight_t>(min_weight_),
+                     ensemble_size_,
+                     max_level_,
+                     static_cast<weight_t>(threshold_),
+                     static_cast<weight_t>(resolution_));
+
+      rmm::device_uvector<vertex_t> vertices(graph_view.local_vertex_partition_range_size(),
+                                             handle_.get_stream());
+      raft::copy(vertices.data(), number_map->data(), vertices.size(), handle_.get_stream());
+
+      result_ = new cugraph::c_api::cugraph_hierarchical_clustering_result_t{
+        modularity,
+        new cugraph::c_api::cugraph_type_erased_device_array_t(vertices, graph_->vertex_type_),
+        new cugraph::c_api::cugraph_type_erased_device_array_t(clusters, graph_->vertex_type_)};
+    }
+  }
+};
+
+}  // namespace
+
+extern "C" cugraph_error_code_t cugraph_ecg(const cugraph_resource_handle_t* handle,
+                                            cugraph_rng_state_t* rng_state,
+                                            cugraph_graph_t* graph,
+                                            double min_weight,
+                                            size_t ensemble_size,
+                                            size_t max_level,
+                                            double threshold,
+                                            double resolution,
+                                            bool_t do_expensive_check,
+                                            cugraph_hierarchical_clustering_result_t** result,
+                                            cugraph_error_t** error)
+{
+  ecg_functor functor(handle,
+                      rng_state,
+                      graph,
+                      min_weight,
+                      ensemble_size,
+                      max_level,
+                      threshold,
+                      resolution,
+                      do_expensive_check);
+
+  return cugraph::c_api::run_algorithm(graph, functor, result, error);
+}

cpp/src/c_api/legacy_ecg.cpp

@@ -31,19 +31,19 @@
 
 namespace {
 
-struct ecg_functor : public cugraph::c_api::abstract_functor {
+struct legacy_ecg_functor : public cugraph::c_api::abstract_functor {
   raft::handle_t const& handle_;
   cugraph::c_api::cugraph_graph_t* graph_;
   double min_weight_;
   size_t ensemble_size_;
   bool do_expensive_check_;
   cugraph::c_api::cugraph_hierarchical_clustering_result_t* result_{};
 
-  ecg_functor(::cugraph_resource_handle_t const* handle,
-              ::cugraph_graph_t* graph,
-              double min_weight,
-              size_t ensemble_size,
-              bool do_expensive_check)
+  legacy_ecg_functor(::cugraph_resource_handle_t const* handle,
+                     ::cugraph_graph_t* graph,
+                     double min_weight,
+                     size_t ensemble_size,
+                     bool do_expensive_check)
     : abstract_functor(),
       handle_(*reinterpret_cast<cugraph::c_api::cugraph_resource_handle_t const*>(handle)->handle_),
       graph_(reinterpret_cast<cugraph::c_api::cugraph_graph_t*>(graph)),
@@ -120,15 +120,16 @@ struct ecg_functor : public cugraph::c_api::abstract_functor {
 
 }  // namespace
 
-extern "C" cugraph_error_code_t cugraph_ecg(const cugraph_resource_handle_t* handle,
-                                            cugraph_graph_t* graph,
-                                            double min_weight,
-                                            size_t ensemble_size,
-                                            bool_t do_expensive_check,
-                                            cugraph_hierarchical_clustering_result_t** result,
-                                            cugraph_error_t** error)
+extern "C" cugraph_error_code_t cugraph_legacy_ecg(
+  const cugraph_resource_handle_t* handle,
+  cugraph_graph_t* graph,
+  double min_weight,
+  size_t ensemble_size,
+  bool_t do_expensive_check,
+  cugraph_hierarchical_clustering_result_t** result,
+  cugraph_error_t** error)
 {
-  ecg_functor functor(handle, graph, min_weight, ensemble_size, do_expensive_check);
+  legacy_ecg_functor functor(handle, graph, min_weight, ensemble_size, do_expensive_check);
 
   return cugraph::c_api::run_algorithm(graph, functor, result, error);
 }