Remove legacy renumbering

benchmarks/cugraph/pytest-based/bench_cugraph_uniform_neighbor_sample.py

@@ -65,7 +65,7 @@ def create_graph(graph_data):
         # FIXME: edgelist_df should have column names that match the defaults
         # for G.from_cudf_edgelist()
         G.from_cudf_edgelist(
-            edgelist_df, source="src", destination="dst", edge_attr="wgt", legacy_renum_only=True
+            edgelist_df, source="src", destination="dst", edge_attr="wgt"
         )
         num_verts = G.number_of_vertices()
 
@@ -94,7 +94,6 @@ def create_graph(graph_data):
             source="src",
             destination="dst",
             edge_attr="weight",
-            legacy_renum_only=True,
         )
 
     else:
@@ -135,7 +134,7 @@ def create_mg_graph(graph_data):
         # for G.from_cudf_edgelist()
         edgelist_df = dask_cudf.from_cudf(edgelist_df)
         G.from_dask_cudf_edgelist(
-            edgelist_df, source="src", destination="dst", edge_attr="wgt", legacy_renum_only=True
+            edgelist_df, source="src", destination="dst", edge_attr="wgt"
         )
         num_verts = G.number_of_vertices()
 
@@ -164,7 +163,6 @@ def create_mg_graph(graph_data):
             source="src",
             destination="dst",
             edge_attr="weight",
-            legacy_renum_only=True,
         )
 
     else:

benchmarks/cugraph/standalone/benchmark.py

@@ -1,4 +1,4 @@
-# Copyright (c) 2021-2022, NVIDIA CORPORATION.
+# Copyright (c) 2021-2023, 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
@@ -133,7 +133,6 @@ def run(self):
         # Algos with transposed=True : PageRank, Katz.
         # Algos with transposed=False: BFS, SSSP, Louvain, HITS,
         # Neighborhood_sampling.
-        # Algos supporting the legacy_renum_only: HITS, Neighborhood_sampling
         #
         for i in range(len(self.algos)):
             # set transpose=True when renumbering
@@ -151,7 +150,7 @@ def run(self):
                         self.algos[i][1]["alpha"] = katz_alpha
                     if hasattr(G, "compute_renumber_edge_list"):
                         G.compute_renumber_edge_list(
-                            transposed=True, legacy_renum_only=True)
+                            transposed=True)
                 else:
                     # FIXME: Pagerank still follows the old path. Update this once it
                     # follows the pylibcugraph/C path
@@ -166,7 +165,7 @@ def run(self):
                         G.compute_renumber_edge_list(transposed=False)
                     else:
                         G.compute_renumber_edge_list(
-                            transposed=False, legacy_renum_only=True)
+                            transposed=False)
                 self.__log("done.")
         # FIXME: need to handle individual algo args
         for ((algo, params), validator) in zip(self.algos, self.validators):

benchmarks/shared/build_cugraph_ucx/test_cugraph_sampling.py

@@ -64,7 +64,6 @@ def create_mg_graph(graph_data):
         source="src",
         destination="dst",
         edge_attr="weight",
-        legacy_renum_only=True,
     )
     return G
 

cpp/include/cugraph/detail/utility_wrappers.hpp

@@ -158,5 +158,21 @@ std::tuple<rmm::device_uvector<vertex_t>, rmm::device_uvector<edge_t>> filter_de
 template <typename data_t>
 bool is_sorted(raft::handle_t const& handle, raft::device_span<data_t> span);
 
+/**
+ * @brief Check if two device spans are equal.  Returns true if every element in the spans are
+ * equal.
+ *
+ * @tparam data_t type of data in span
+ * @param handle RAFT handle object to encapsulate resources (e.g. CUDA stream, communicator, and
+ * handles to various CUDA libraries) to run graph algorithms.
+ * @param span1 The span of data to compare
+ * @param span2 The span of data to compare
+ * @return true if equal, false if not equal
+ */
+template <typename data_t>
+bool is_equal(raft::handle_t const& handle,
+              raft::device_span<data_t> span1,
+              raft::device_span<data_t> span2);
+
 }  // namespace detail
 }  // namespace cugraph

cpp/src/c_api/legacy_spectral.cpp

@@ -307,9 +307,11 @@ struct analyze_clustering_ratio_cut_functor : public cugraph::c_api::abstract_fu
 
       weight_t score;
 
-      if (cugraph::detail::is_sorted(handle_,
-                                     raft::device_span<vertex_t const>{
-                                       vertices_->as_type<vertex_t const>(), vertices_->size_})) {
+      if (cugraph::detail::is_equal(
+            handle_,
+            raft::device_span<vertex_t const>{vertices_->as_type<vertex_t const>(),
+                                              vertices_->size_},
+            raft::device_span<vertex_t const>{number_map->data(), number_map->size()})) {
         cugraph::ext_raft::analyzeClustering_ratio_cut(
           legacy_graph_view, n_clusters_, clusters_->as_type<vertex_t>(), &score);
       } else {
@@ -321,6 +323,15 @@ struct analyze_clustering_ratio_cut_functor : public cugraph::c_api::abstract_fu
         raft::copy(
           tmp_c.data(), clusters_->as_type<vertex_t>(), clusters_->size_, handle_.get_stream());
 
+        cugraph::renumber_ext_vertices<vertex_t, false>(
+          handle_,
+          tmp_v.data(),
+          tmp_v.size(),
+          number_map->data(),
+          graph_view.local_vertex_partition_range_first(),
+          graph_view.local_vertex_partition_range_last(),
+          false);
+
         cugraph::c_api::detail::sort_by_key(
           handle_,
           raft::device_span<vertex_t>{tmp_v.data(), tmp_v.size()},
@@ -403,9 +414,11 @@ struct analyze_clustering_edge_cut_functor : public cugraph::c_api::abstract_fun
 
       weight_t score;
 
-      if (cugraph::detail::is_sorted(handle_,
-                                     raft::device_span<vertex_t const>{
-                                       vertices_->as_type<vertex_t const>(), vertices_->size_})) {
+      if (cugraph::detail::is_equal(
+            handle_,
+            raft::device_span<vertex_t const>{vertices_->as_type<vertex_t const>(),
+                                              vertices_->size_},
+            raft::device_span<vertex_t const>{number_map->data(), number_map->size()})) {
         cugraph::ext_raft::analyzeClustering_edge_cut(
           legacy_graph_view, n_clusters_, clusters_->as_type<vertex_t>(), &score);
       } else {
@@ -417,6 +430,15 @@ struct analyze_clustering_edge_cut_functor : public cugraph::c_api::abstract_fun
         raft::copy(
           tmp_c.data(), clusters_->as_type<vertex_t>(), clusters_->size_, handle_.get_stream());
 
+        cugraph::renumber_ext_vertices<vertex_t, false>(
+          handle_,
+          tmp_v.data(),
+          tmp_v.size(),
+          number_map->data(),
+          graph_view.local_vertex_partition_range_first(),
+          graph_view.local_vertex_partition_range_last(),
+          false);
+
         cugraph::c_api::detail::sort_by_key(
           handle_,
           raft::device_span<vertex_t>{tmp_v.data(), tmp_v.size()},
@@ -499,9 +521,11 @@ struct analyze_clustering_modularity_functor : public cugraph::c_api::abstract_f
 
       weight_t score;
 
-      if (cugraph::detail::is_sorted(handle_,
-                                     raft::device_span<vertex_t const>{
-                                       vertices_->as_type<vertex_t const>(), vertices_->size_})) {
+      if (cugraph::detail::is_equal(
+            handle_,
+            raft::device_span<vertex_t const>{vertices_->as_type<vertex_t const>(),
+                                              vertices_->size_},
+            raft::device_span<vertex_t const>{number_map->data(), number_map->size()})) {
         cugraph::ext_raft::analyzeClustering_modularity(
           legacy_graph_view, n_clusters_, clusters_->as_type<vertex_t>(), &score);
       } else {
@@ -513,6 +537,15 @@ struct analyze_clustering_modularity_functor : public cugraph::c_api::abstract_f
         raft::copy(
           tmp_c.data(), clusters_->as_type<vertex_t>(), clusters_->size_, handle_.get_stream());
 
+        cugraph::renumber_ext_vertices<vertex_t, false>(
+          handle_,
+          tmp_v.data(),
+          tmp_v.size(),
+          number_map->data(),
+          graph_view.local_vertex_partition_range_first(),
+          graph_view.local_vertex_partition_range_last(),
+          false);
+
         cugraph::c_api::detail::sort_by_key(
           handle_,
           raft::device_span<vertex_t>{tmp_v.data(), tmp_v.size()},

cpp/src/detail/utility_wrappers.cu

@@ -206,5 +206,26 @@ template bool is_sorted(raft::handle_t const& handle, raft::device_span<int32_t
 template bool is_sorted(raft::handle_t const& handle, raft::device_span<int64_t> span);
 template bool is_sorted(raft::handle_t const& handle, raft::device_span<int64_t const> span);
 
+template <typename data_t>
+bool is_equal(raft::handle_t const& handle,
+              raft::device_span<data_t> span1,
+              raft::device_span<data_t> span2)
+{
+  return thrust::equal(handle.get_thrust_policy(), span1.begin(), span1.end(), span2.begin());
+}
+
+template bool is_equal(raft::handle_t const& handle,
+                       raft::device_span<int32_t> span1,
+                       raft::device_span<int32_t> span2);
+template bool is_equal(raft::handle_t const& handle,
+                       raft::device_span<int32_t const> span1,
+                       raft::device_span<int32_t const> span2);
+template bool is_equal(raft::handle_t const& handle,
+                       raft::device_span<int64_t> span1,
+                       raft::device_span<int64_t> span2);
+template bool is_equal(raft::handle_t const& handle,
+                       raft::device_span<int64_t const> span1,
+                       raft::device_span<int64_t const> span2);
+
 }  // namespace detail
 }  // namespace cugraph

docs/cugraph/source/api_docs/structure.rst

@@ -102,4 +102,3 @@ Other
    :toctree: api/
 
    cugraph.hypergraph
-   cugraph.structure.shuffle

python/cugraph-dgl/cugraph_dgl/dataloading/utils/extract_graph_helpers.py

@@ -44,7 +44,6 @@ def create_cugraph_graph_from_edges_dict(
             source="_SRC_",
             destination="_DST_",
             edge_attr=["wgt", "_EDGE_ID_", "etp"],
-            legacy_renum_only=True,
             renumber=True,
         )
     else:
@@ -53,7 +52,6 @@ def create_cugraph_graph_from_edges_dict(
             source="_SRC_",
             destination="_DST_",
             edge_attr=["wgt", "_EDGE_ID_", "etp"],
-            legacy_renum_only=True,
             renumber=True,
         )
     return G

python/cugraph-pyg/cugraph_pyg/data/cugraph_store.py

@@ -409,15 +409,13 @@ def __construct_graph(
                 source="src",
                 destination="dst",
                 edge_attr=["w", "eid", "etp"],
-                legacy_renum_only=True,
             )
         else:
             graph.from_cudf_edgelist(
                 df,
                 source="src",
                 destination="dst",
                 edge_attr=["w", "eid", "etp"],
-                legacy_renum_only=True,
             )
 
         return graph
@@ -500,9 +498,8 @@ def _get_edge_index(self, attr: CuGraphEdgeAttr) -> Tuple[TensorType, TensorType
         if attr.layout != EdgeLayout.COO:
             raise TypeError("Only COO direct access is supported!")
 
-        # Currently, graph creation enforces that legacy_renum_only=True
-        # is always called, and the input vertex ids are always of integer
-        # type.  Therefore, it is currently safe to assume that for MG
+        # Currently, graph creation enforces that input vertex ids are always of
+        # integer type.  Therefore, it is currently safe to assume that for MG
         # graphs, the src/dst col names are renumbered_src/dst
         # and for SG graphs, the src/dst col names are src/dst.
         # This may change in the future if/when renumbering or the graph

python/cugraph-service/server/cugraph_service_server/testing/benchmark_server_extension.py

@@ -95,15 +95,13 @@ def create_graph_from_rmat_generator(
             source="src",
             destination="dst",
             edge_attr="weight",
-            legacy_renum_only=True,
         )
     else:
         G.from_cudf_edgelist(
             edgelist_df,
             source="src",
             destination="dst",
             edge_attr="weight",
-            legacy_renum_only=True,
         )
 
     return G

python/cugraph/cugraph/__init__.py

@@ -13,6 +13,7 @@
 
 from cugraph.community import (
     ecg,
+    induced_subgraph,
     ktruss_subgraph,
     k_truss,
     louvain,

python/cugraph/cugraph/community/CMakeLists.txt

@@ -1,5 +1,5 @@
 # =============================================================================
-# Copyright (c) 2022, NVIDIA CORPORATION.
+# Copyright (c) 2022-2023, 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
@@ -13,11 +13,8 @@
 # =============================================================================
 
 set(cython_sources
-    ecg_wrapper.pyx
     ktruss_subgraph_wrapper.pyx
     leiden_wrapper.pyx
-    spectral_clustering_wrapper.pyx
-    subgraph_extraction_wrapper.pyx
 )
 
 set(linked_libraries cugraph::cugraph)

python/cugraph/cugraph/community/__init__.py

@@ -1,4 +1,4 @@
-# Copyright (c) 2019-2022, NVIDIA CORPORATION.
+# Copyright (c) 2019-2023, 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
@@ -22,6 +22,7 @@
     analyzeClustering_ratio_cut,
 )
 from cugraph.community.subgraph_extraction import subgraph
+from cugraph.community.induced_subgraph import induced_subgraph
 from cugraph.community.triangle_count import triangle_count
 from cugraph.community.ktruss_subgraph import ktruss_subgraph
 from cugraph.community.ktruss_subgraph import k_truss