BOOST_FIXTURE_TEST_CASE(test_serialisation2,Fixture) {
dijkstra(g,s,p,d);
string str = my_serialize2(d);
vector<int> vect = deserialize2(str);
BOOST_CHECK_EQUAL_COLLECTIONS(d.begin(),d.end(),vect.begin(),vect.end());
}
int main(int argc, char *argv[]) {
// Process the command-line arguments.
const char *vocab_path = "vocab.out";
const char *map_path = "map_id_label.txt";
const char *cats_path = "cats.out";
bool ova = true;
{
int i;
for (i = 1; i < argc; ++i) {
if (!strcmp(argv[i], "-h")) {
goto usage;
}
else if (!strcmp(argv[i], "-v")) {
vocab_path = argv[++i];
}
else if (!strcmp(argv[i], "-m")) {
map_path = argv[++i];
}
else if (!strcmp(argv[i], "-c")) {
++i;
if (!strcmp(argv[i], "ova")) {
ova = true;
}
else if (!strcmp(argv[i], "ovo")) {
ova = false;
}
else {
std::cerr << argv[0] << ": Unsupported classifier: `" << argv[i]
<< "'\n";
goto err;
}
}
else {
break;
}
}
if (i < argc)
cats_path = argv[i++];
if (i != argc)
goto usage;
if (!vocab_path || !map_path || !cats_path)
goto usage;
}
{
// Load the vocabulary.
std::cout << "Loading vocabulary...\n";
vocab_type vocab;
{
std::ifstream fs(vocab_path, std::ios::binary);
deserialize2(vocab, fs);
}
// Load the category map.
std::cout << "Loading category map...\n";
std::map< int, std::string > cat_map;
{
std::ifstream fs(map_path);
for (std::string line; std::getline(fs, line);) {
std::istringstream ss(line);
int i;
std::string label;
ss >> i;
ss.get(); // ','
std::getline(ss, label);
cat_map[i] = label;
}
}
// Load the category classifier.
std::cout << "Loading classifier...\n";
df_type df;
{
std::ifstream fs(cats_path, std::ios::binary);
if (ova)
deserialize2(df.get< ova_df_type >(), fs);
else
deserialize2(df.get< ovo_df_type >(), fs);
}
// Extract features for all input files.
std::vector< std::string > paths;
std::string path;
while (std::getline(std::cin, path))
paths.push_back(path);
#pragma omp parallel for schedule(dynamic)
for (typename std::vector< std::string >::size_type i = 0;
i < paths.size(); ++i) {
const std::string &path = paths[i];
// Extract the features.
image_type image;
load_svg(path.c_str(), image);
image = 1. - image;
std::vector< feature_desc_type > descs;
extract_descriptors(image, descs);
feature_hist_type hist;
feature_hist(descs, vocab, hist);
const int cat = ova ? df.get< ova_df_type >()(hist) :
df.get< ovo_df_type >()(hist);
assert(cat);
#pragma omp critical
{
std::cout << path << ' ' << cat_map[cat] << '\n';
}
}
}
return 0;
usage:
std::cerr << "Usage: " << argv[0]
<< " [-v vocab-file] [-m map-file] [-c classifier] [cats-file]\n";
err:
return 1;
}
