graph_clustering.hh 4.86 KB
Newer Older
1
2
// graph-tool -- a general graph modification and manipulation thingy
//
Tiago Peixoto's avatar
Tiago Peixoto committed
3
// Copyright (C) 2006-2015 Tiago de Paula Peixoto <tiago@skewed.de>
4
5
6
7
8
9
10
11
12
13
14
//
// This program is free software; you can redistribute it and/or
// modify it under the terms of the GNU General Public License
// as published by the Free Software Foundation; either version 3
// of the License, or (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
// GNU General Public License for more details.
//
15
// you should have received a copy of the GNU General Public License
16
17
18
19
20
// along with this program. If not, see <http://www.gnu.org/licenses/>.

#ifndef GRAPH_CLUSTERING_HH
#define GRAPH_CLUSTERING_HH

21
22
#include "config.h"

23
#include "hash_map_wrap.hh"
24
25
#include <boost/mpl/if.hpp>

Tiago Peixoto's avatar
Tiago Peixoto committed
26
#ifndef __clang__
27
28
#include <ext/numeric>
using __gnu_cxx::power;
Tiago Peixoto's avatar
Tiago Peixoto committed
29
30
31
32
33
34
35
#else
template <class Value>
Value power(Value value, int n)
{
    return pow(value, n);
}
#endif
36

37
38
39
40
41
42
namespace graph_tool
{
using namespace boost;

// calculates the number of triangles to which v belongs
template <class Graph>
43
pair<int,int>
44
45
get_triangles(typename graph_traits<Graph>::vertex_descriptor v, const Graph &g)
{
46
47
    typedef typename graph_traits<Graph>::vertex_descriptor vertex_t;

48
    typedef gt_hash_set<vertex_t> set_t;
49
50
51

    set_t neighbour_set;

52
    neighbour_set.resize(out_degree(v, g));
53

54
    size_t triangles = 0;
55

56
57
    typename graph_traits<Graph>::adjacency_iterator n, n_end;
    for (tie(n, n_end) = adjacent_vertices(v, g); n != n_end; ++n)
58
    {
59
        if (*n == v) // no self-loops
60
            continue;
61
62
        neighbour_set.insert(*n);
    }
63

64
65
66
67
    for (tie(n, n_end) = adjacent_vertices(v, g); n != n_end; ++n)
    {
        typename graph_traits<Graph>::adjacency_iterator n2, n2_end;
        for (tie(n2, n2_end) = adjacent_vertices(*n, g); n2 != n2_end; ++n2)
68
        {
69
            if (*n2 == *n) // no self-loops
70
                continue;
71
72
            if (neighbour_set.find(*n2) != neighbour_set.end())
                ++triangles;
73
74
        }
    }
75
76

    size_t k = out_degree(v, g);
77
78
79
80
81
82
83
84
    return make_pair(triangles/2,(k*(k-1))/2);
}


// retrieves the global clustering coefficient
struct get_global_clustering
{
    template <class Graph>
85
    void operator()(const Graph& g, double& c, double& c_err) const
86
87
88
89
90
91
92
    {
        size_t triangles = 0, n = 0;
        pair<size_t, size_t> temp;

        int i, N = num_vertices(g);

        #pragma omp parallel for default(shared) private(i,temp) \
93
            schedule(runtime) if (N > 100) reduction(+:triangles, n)
94
95
96
97
98
99
100
        for (i = 0; i < N; ++i)
        {
            typename graph_traits<Graph>::vertex_descriptor v = vertex(i, g);
            if (v == graph_traits<Graph>::null_vertex())
                continue;

            temp = get_triangles(v, g);
101
            triangles += temp.first;
102
103
            n += temp.second;
        }
104
        c = double(triangles) / n;
105
106
107

        // "jackknife" variance
        c_err = 0.0;
108
        double cerr = 0.0;
109

110
        #pragma omp parallel for default(shared) private(i,temp) \
111
            schedule(runtime) if (N > 100) reduction(+:cerr)
112
113
114
115
116
117
118
        for (i = 0; i < N; ++i)
        {
            typename graph_traits<Graph>::vertex_descriptor v = vertex(i, g);
            if (v == graph_traits<Graph>::null_vertex())
                continue;

            temp = get_triangles(v, g);
119
            double cl = double(triangles - temp.first) / (n - temp.second);
120

121
            cerr += power(c - cl, 2);
122
        }
123
        c_err = sqrt(cerr);
124
125
126
127
128
129
130
    }
};

// sets the local clustering coefficient to a property
struct set_clustering_to_property
{
    template <class Graph, class ClustMap>
131
    void operator()(const Graph& g, ClustMap clust_map) const
132
    {
133
        typedef typename property_traits<ClustMap>::value_type c_type;
134
135
        typename get_undirected_graph<Graph>::type ug(g);
        int i, N = num_vertices(g);
136

137
        #pragma omp parallel for default(shared) private(i) schedule(runtime) if (N > 100)
138
139
140
141
142
        for (i = 0; i < N; ++i)
        {
            typename graph_traits<Graph>::vertex_descriptor v = vertex(i, g);
            if (v == graph_traits<Graph>::null_vertex())
                continue;
143

144
145
146
147
            pair<size_t,size_t> triangles = get_triangles(v,ug); // get from ug
            double clustering = (triangles.second > 0) ?
                double(triangles.first)/triangles.second :
                0.0;
148

149
            clust_map[v] = c_type(clustering);
150
151
        }
    }
152

153
154
155
156
    template <class Graph>
    struct get_undirected_graph
    {
        typedef typename mpl::if_
Tiago Peixoto's avatar
Tiago Peixoto committed
157
158
159
160
           <std::is_convertible<typename graph_traits<Graph>::directed_category,
                                directed_tag>,
            const UndirectedAdaptor<Graph>,
            const Graph& >::type type;
161
162
163
    };
};

164
} //graph-tool namespace
165
166

#endif // GRAPH_CLUSTERING_HH