void BuildGraph::print_setting (void) {
cout << "filename : " << filename_ << endl;
print.Vector("sentences", sentences_);
print.Dict("relscore", *relscore_);
cout << "winsize : " << winsize_ << endl;
print.Array("lambda", lamb_list_, lamb_size_);
print.Array("cutoff_path", cutoff_list_, cutoff_size_);
}
int main (void) {
string text_dirpath = string("/Users/KimKR/Desktop/NEXT_LAB/keyword/test");
string relscore_path = string("/Users/KimKR/Desktop/NEXT_LAB/keyword/rel_score/test/bm25_3_3");
vector<string> textfiles = vector<string>();
getdir(text_dirpath, textfiles);
int winsize = 2;
int cutoff_list[] = {3};
int cutoff_size = 1;
double lambda_list[1];
int lambda_size = 1;
Converter cv;
Print print;
// for(int i = 0; i < 21; ++i) {
// lambda_list[i] = i * 0.05;
// }
lambda_list[0] = 0.85;
clock_t all_begin = clock();
for(int fidx = textfiles.size() - 1; fidx >= 0; --fidx) {
string filename = textfiles[fidx];
vector<string> sentences = TextRead(text_dirpath + "/" + filename);
clock_t begin = clock();
for(int k = K_STR; k < K_STEP; ++k) {
double k1 = k * 0.1;
for(int b1 = 0; b1 < 1; ++b1) {
double b = 0.05 * b1;
string relfile_path = relscore_path + "_" + cv.to_string(k1, 2) + "_" + cv.to_string(b, 3) + "/" + filename;
map<string, double> relscore;
if(RelscoreRead(relfile_path, &relscore) == 0) continue;
BuildGraph bg(filename, sentences, &relscore, winsize, k1, b, lambda_list, lambda_size, cutoff_list, cutoff_size);
Graph G;
G.null_ = true;
bg.run(true, false, false, false);
print.Vector("top_score_candidates", bg.top_score_candidates_.at(3).at(0.85));
SaveGraph(bg.top_score_candidates_, bg.filename_, bg.k1_, bg.b_, bg.winsize_);
}
}
clock_t end = clock();
double elapsed_secs = double(end - begin) / CLOCKS_PER_SEC;
cout << filename << " : " << elapsed_secs << endl;
}
clock_t end = clock();
double elapsed_secs = double(end - all_begin) / CLOCKS_PER_SEC;
cout << "all finished" << " : " << elapsed_secs << endl;
getchar();
}
void BuildGraph::generate_candidates(void) {
for(int cidx = 0; cidx < cutoff_size_; ++cidx) {
int cutoff = cutoff_list_[cidx];
for(int lidx = 0; lidx < lamb_size_; ++lidx) {
double lamb = lamb_list_[lidx];
for(int sidx = 0; sidx < num_of_sentence_; ++sidx) {
vector<string> words = Split(sentences_[sidx], ';');
vector<string> candidate;
for(int widx = 0; widx < words.size(); ++widx) {
if (node_scores_[cutoff][lamb].count(words[widx])) {
candidate.push_back(words[widx]);
} else if (candidate.size() > 0) {
keywords_candidates_[cutoff][lamb].push_back(Join(candidate, ' '));
candidate.clear();
}
}
if (candidate.size() > 0) {
keywords_candidates_[cutoff][lamb].push_back(Join(candidate, ' '));
}
print.Vector("keywords_candidates_ : ", keywords_candidates_[cutoff][lamb]);
}
}
}
};
