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
|
// Copyright 2004 The Trustees of Indiana University.
// Use, modification and distribution is subject to the Boost Software
// License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
// Authors: Douglas Gregor
// Andrew Lumsdaine
// This program performs betweenness centrality (BC) clustering on the
// actor collaboration graph available at
// http://www.nd.edu/~networks/database/index.html and outputs the
// result of clustering in Pajek format.
//
// This program mimics the BC clustering algorithm program implemented
// by Shashikant Penumarthy for JUNG, so that we may compare results
// and timings.
#include <boost/graph/bc_clustering.hpp>
#include <boost/graph/adjacency_list.hpp>
#include <boost/graph/graph_traits.hpp>
#include <fstream>
#include <iostream>
#include <string>
#include <boost/tokenizer.hpp>
#include <boost/lexical_cast.hpp>
#include <map>
using namespace boost;
struct Actor
{
Actor(int id = -1) : id(id) {}
int id;
};
typedef adjacency_list<vecS, vecS, undirectedS, Actor,
property<edge_centrality_t, double> > ActorGraph;
typedef graph_traits<ActorGraph>::vertex_descriptor Vertex;
typedef graph_traits<ActorGraph>::edge_descriptor Edge;
void load_actor_graph(std::istream& in, ActorGraph& g)
{
std::map<int, Vertex> actors;
std::string line;
while (getline(in, line)) {
std::vector<Vertex> actors_in_movie;
// Map from the actor numbers on this line to the actor vertices
typedef tokenizer<char_separator<char> > Tok;
Tok tok(line, char_separator<char>(" "));
for (Tok::iterator id = tok.begin(); id != tok.end(); ++id) {
int actor_id = lexical_cast<int>(*id);
std::map<int, Vertex>::iterator v = actors.find(actor_id);
if (v == actors.end()) {
Vertex new_vertex = add_vertex(Actor(actor_id), g);
actors[actor_id] = new_vertex;
actors_in_movie.push_back(new_vertex);
} else {
actors_in_movie.push_back(v->second);
}
}
for (std::vector<Vertex>::iterator i = actors_in_movie.begin();
i != actors_in_movie.end(); ++i) {
for (std::vector<Vertex>::iterator j = i + 1;
j != actors_in_movie.end(); ++j) {
if (!edge(*i, *j, g).second) add_edge(*i, *j, g);
}
}
}
}
template<typename Graph, typename VertexIndexMap, typename VertexNameMap>
std::ostream&
write_pajek_graph(std::ostream& out, const Graph& g,
VertexIndexMap vertex_index, VertexNameMap vertex_name)
{
out << "*Vertices " << num_vertices(g) << '\n';
typedef typename graph_traits<Graph>::vertex_iterator vertex_iterator;
for (vertex_iterator v = vertices(g).first; v != vertices(g).second; ++v) {
out << get(vertex_index, *v)+1 << " \"" << get(vertex_name, *v) << "\"\n";
}
out << "*Edges\n";
typedef typename graph_traits<Graph>::edge_iterator edge_iterator;
for (edge_iterator e = edges(g).first; e != edges(g).second; ++e) {
out << get(vertex_index, source(*e, g))+1 << ' '
<< get(vertex_index, target(*e, g))+1 << " 1.0\n"; // HACK!
}
return out;
}
class actor_clustering_threshold : public bc_clustering_threshold<double>
{
typedef bc_clustering_threshold<double> inherited;
public:
actor_clustering_threshold(double threshold, const ActorGraph& g,
bool normalize)
: inherited(threshold, g, normalize), iter(1) { }
bool operator()(double max_centrality, Edge e, const ActorGraph& g)
{
std::cout << "Iter: " << iter << " Max Centrality: "
<< (max_centrality / dividend) << std::endl;
++iter;
return inherited::operator()(max_centrality, e, g);
}
private:
unsigned int iter;
};
int main(int argc, char* argv[])
{
std::string in_file;
std::string out_file;
double threshold = -1.0;
bool normalize = false;
// Parse command-line options
{
int on_arg = 1;
while (on_arg < argc) {
std::string arg(argv[on_arg]);
if (arg == "-in") {
++on_arg; assert(on_arg < argc);
in_file = argv[on_arg];
} else if (arg == "-out") {
++on_arg; assert(on_arg < argc);
out_file = argv[on_arg];
} else if (arg == "-threshold") {
++on_arg; assert(on_arg < argc);
threshold = lexical_cast<double>(argv[on_arg]);
} else if (arg == "-normalize") {
normalize = true;
} else {
std::cerr << "Unrecognized parameter \"" << arg << "\".\n";
return -1;
}
++on_arg;
}
if (in_file.empty() || out_file.empty() || threshold < 0) {
std::cerr << "error: syntax is actor_clustering [options]\n\n"
<< "options are:\n"
<< "\t-in <infile>\tInput file\n"
<< "\t-out <outfile>\tOutput file\n"
<< "\t-threshold <value>\tA threshold value\n"
<< "\t-normalize\tNormalize edge centrality scores\n";
return -1;
}
}
ActorGraph g;
// Load the actor graph
{
std::cout << "Building graph." << std::endl;
std::ifstream in(in_file.c_str());
if (!in) {
std::cerr << "Unable to open file \"" << in_file << "\" for input.\n";
return -2;
}
load_actor_graph(in, g);
}
// Run the algorithm
std::cout << "Clusting..." << std::endl;
betweenness_centrality_clustering(g,
actor_clustering_threshold(threshold, g, normalize),
get(edge_centrality, g));
// Output the graph
{
std::cout << "Writing graph to file: " << out_file << std::endl;
std::ofstream out(out_file.c_str());
if (!out) {
std::cerr << "Unable to open file \"" << out_file << "\" for output.\n";
return -3;
}
write_pajek_graph(out, g, get(vertex_index, g), get(&Actor::id, g));
}
return 0;
}
|
