1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
|
// Copyright (C) 2017-2025 Chris N. Richardson and Garth N. Wells
//
// This file is part of DOLFINx (https://www.fenicsproject.org)
//
// SPDX-License-Identifier: LGPL-3.0-or-later
#include "dolfinx_wrappers/caster_mpi.h"
#include <array>
#include <dolfinx/common/IndexMap.h>
#include <dolfinx/graph/AdjacencyList.h>
#include <dolfinx/graph/ordering.h>
#include <dolfinx/graph/partition.h>
#include <dolfinx/graph/partitioners.h>
#include <dolfinx/graph/utils.h>
#include <nanobind/nanobind.h>
#include <nanobind/ndarray.h>
#include <nanobind/operators.h>
#include <nanobind/stl/array.h>
#include <nanobind/stl/function.h>
#include <nanobind/stl/map.h>
#include <nanobind/stl/pair.h>
#include <nanobind/stl/string.h>
#include <nanobind/stl/tuple.h>
#include <nanobind/stl/vector.h>
#include <ranges>
#include <vector>
namespace nb = nanobind;
namespace
{
/// Wrap a C++ graph partitioning function as a Python-ready function
template <typename Functor>
auto create_partitioner_py(Functor&& p_cpp)
{
return [p_cpp](dolfinx_wrappers::MPICommWrapper comm, int nparts,
const dolfinx::graph::AdjacencyList<std::int64_t>& local_graph,
bool ghosting)
{ return p_cpp(comm.get(), nparts, local_graph, ghosting); };
}
template <typename T, typename U>
void declare_adjacency_list_init(nb::module_& m, std::string type)
{
std::string pyclass_name = std::string("AdjacencyList_") + type;
nb::class_<dolfinx::graph::AdjacencyList<T, U>>(m, pyclass_name.c_str(),
"Adjacency List")
.def(
"__init__",
[](dolfinx::graph::AdjacencyList<T, U>* a,
nb::ndarray<const T, nb::ndim<1>, nb::c_contig> adj)
{
std::vector<T> data(adj.data(), adj.data() + adj.size());
new (a) dolfinx::graph::AdjacencyList<T, U>(
dolfinx::graph::regular_adjacency_list<U>(std::move(data), 1));
},
nb::arg("adj").noconvert())
.def(
"__init__",
[](dolfinx::graph::AdjacencyList<T, U>* a,
nb::ndarray<const T, nb::ndim<2>, nb::c_contig> adj)
{
std::vector<T> data(adj.data(), adj.data() + adj.size());
new (a) dolfinx::graph::AdjacencyList<T, U>(
dolfinx::graph::regular_adjacency_list<U>(std::move(data),
adj.shape(1)));
},
nb::arg("adj").noconvert())
.def(
"__init__",
[](dolfinx::graph::AdjacencyList<T, U>* a,
nb::ndarray<const T, nb::ndim<1>, nb::c_contig> array,
nb::ndarray<const std::int32_t, nb::ndim<1>, nb::c_contig> displ)
{
std::vector<T> data(array.data(), array.data() + array.size());
std::vector<std::int32_t> offsets(displ.data(),
displ.data() + displ.size());
new (a) dolfinx::graph::AdjacencyList<T, U>(std::move(data),
std::move(offsets));
},
nb::arg("data").noconvert(), nb::arg("offsets"))
.def(
"links",
[](const dolfinx::graph::AdjacencyList<T, U>& self, int i)
{
std::span<const T> link = self.links(i);
return nb::ndarray<const T, nb::numpy>(link.data(), {link.size()});
},
nb::rv_policy::reference_internal, nb::arg("i"),
"Links (edges) of a node")
.def_prop_ro(
"array",
[](const dolfinx::graph::AdjacencyList<T, U>& self)
{
return nb::ndarray<const T, nb::numpy>(self.array().data(),
{self.array().size()});
},
nb::rv_policy::reference_internal)
.def_prop_ro(
"offsets",
[](const dolfinx::graph::AdjacencyList<T, U>& self)
{
return nb::ndarray<const std::int32_t, nb::numpy>(
self.offsets().data(), {self.offsets().size()});
},
nb::rv_policy::reference_internal)
.def_prop_ro("num_nodes", &dolfinx::graph::AdjacencyList<T, U>::num_nodes)
.def("__eq__", &dolfinx::graph::AdjacencyList<T, U>::operator==,
nb::is_operator())
.def("__repr__", &dolfinx::graph::AdjacencyList<T, U>::str)
.def("__len__", &dolfinx::graph::AdjacencyList<T, U>::num_nodes);
}
template <typename T, typename U>
void declare_adjacency_list(nb::module_& m, std::string type)
{
std::string pyclass_name = std::string("AdjacencyList_") + type;
nb::class_<dolfinx::graph::AdjacencyList<T, U>>(m, pyclass_name.c_str(),
"Adjacency List")
.def_prop_ro(
"offsets",
[](const dolfinx::graph::AdjacencyList<T, U>& self)
{
return nb::ndarray<const std::int32_t, nb::numpy>(
self.offsets().data(), {self.offsets().size()});
},
nb::rv_policy::reference_internal)
.def_prop_ro("num_nodes", &dolfinx::graph::AdjacencyList<T, U>::num_nodes)
.def("__eq__", &dolfinx::graph::AdjacencyList<T, U>::operator==,
nb::is_operator())
.def("__len__", &dolfinx::graph::AdjacencyList<T, U>::num_nodes);
}
} // namespace
namespace dolfinx_wrappers
{
void graph(nb::module_& m)
{
declare_adjacency_list_init<std::int32_t, std::nullptr_t>(m, "int32");
declare_adjacency_list_init<std::int64_t, std::nullptr_t>(m, "int64");
declare_adjacency_list<std::tuple<int, std::size_t, std::int8_t>,
std::pair<std::int32_t, std::int32_t>>(
m, "int_sizet_int8__int32_int32");
using partition_fn
= std::function<dolfinx::graph::AdjacencyList<std::int32_t>(
MPICommWrapper, int,
const dolfinx::graph::AdjacencyList<std::int64_t>&, bool)>;
m.def(
"partitioner", []() -> partition_fn
{ return create_partitioner_py(dolfinx::graph::partition_graph); },
"Default graph partitioner");
#ifdef HAS_PTSCOTCH
m.def(
"partitioner_scotch",
[](double imbalance, int seed) -> partition_fn
{
return create_partitioner_py(dolfinx::graph::scotch::partitioner(
dolfinx::graph::scotch::strategy::none, imbalance, seed));
},
nb::arg("imbalance") = 0.025, nb::arg("seed") = 0,
"SCOTCH graph partitioner");
#endif
#ifdef HAS_PARMETIS
m.def(
"partitioner_parmetis",
[](double imbalance, std::array<int, 3> options) -> partition_fn
{
return create_partitioner_py(
dolfinx::graph::parmetis::partitioner(imbalance, options));
},
nb::arg("imbalance") = 1.02,
nb::arg("options") = std ::array<int, 3>({1, 0, 5}),
"ParMETIS graph partitioner");
#endif
#ifdef HAS_KAHIP
m.def(
"partitioner_kahip",
[](int mode = 1, int seed = 1, double imbalance = 0.03,
bool suppress_output = true) -> partition_fn
{
return create_partitioner_py(dolfinx::graph::kahip::partitioner(
mode, seed, imbalance, suppress_output));
},
nb::arg("mode") = 1, nb::arg("seed") = 1, nb::arg("imbalance") = 0.03,
nb::arg("suppress_output") = true, "KaHIP graph partitioner");
#endif
m.def("reorder_gps", &dolfinx::graph::reorder_gps, nb::arg("graph"));
m.def(
"comm_graph", [](const dolfinx::common::IndexMap& map, int root)
{ return dolfinx::graph::comm_graph(map, root); }, nb::arg("map"),
nb::arg("root") = 0,
"Build a graph representing parallel communication patterns.");
m.def(
"comm_graph_data",
[](dolfinx::graph::AdjacencyList<
std::tuple<int, std::size_t, std::int8_t>,
std::pair<std::int32_t, std::int32_t>>& g)
{
std::vector<std::tuple<int, int, std::map<std::string, std::size_t>>>
adj;
for (std::int32_t n = 0; n < g.num_nodes(); ++n)
{
for (auto [e, w, local] : g.links(n))
{
adj.emplace_back(n, e,
std::map<std::string, std::size_t>{
{"local", local}, {"weight", w}});
}
}
std::vector<
std::pair<std::int32_t, std::map<std::string, std::int32_t>>>
nodes;
std::ranges::transform(
g.node_data().value(), std::ranges::views::iota(0),
std::back_inserter(nodes),
[](auto data, auto n)
{
return std::pair(
n, std::map<std::string, std::int32_t>{
{"weight", data.first}, {"num_ghosts", data.second}});
});
return std::pair(std::move(adj), std::move(nodes));
},
"Build a graph edge and node data representing parallel communication "
"patterns. Can be used to creat NetworkX graphs.");
m.def(
"comm_to_json",
[](dolfinx::graph::AdjacencyList<
std::tuple<int, std::size_t, std::int8_t>,
std::pair<std::int32_t, std::int32_t>>& g)
{ return dolfinx::graph::comm_to_json(g); },
"Build a JSON string representation of a parallel communication "
"graph that can use used by build a NetworkX graph.");
}
} // namespace dolfinx_wrappers
|
