gapped_search_result
search(const gapped_pattern& pat) const
{
gapped_search_result res;
const vector<string_type>& s = pat.subpatterns;
size_type min_gap;
size_type max_gap;
std::cerr << "REGEX ::: " << pat.raw_regexp << std::endl;
min_gap = s[0].size() + pat.gaps[0].first;
max_gap = s[0].size() + pat.gaps[0].second;
auto last_size = s[s.size() - 1].size();
// get ranges
vector<sdsl::int_vector_const_iterator<sdsl::int_vector<0>>::const_iterator> its;
vector<sdsl::int_vector_const_iterator<sdsl::int_vector<0>>::const_iterator> its2;
for (auto sx : s) {
size_type sp, ep;
forward_search(m_text.begin(), m_text.end(), m_sa, 0, m_sa.size()-1, sx.begin(), sx.end(), sp, ep);
its.push_back(m_sa.begin() + sp);
its2.push_back(m_sa.begin() + ep + 1);
}
vector<pii> spans1, spans2;
for (auto it = its[0]; it != its2[0]; ++it)
spans1.emplace_back(*it, *it);
auto pspan1 = &spans1;
auto pspan2 = &spans2;
// incremental search
for (size_t i = 1; i < its.size(); ++i) {
std::sort(pspan1->begin(), pspan1->end());
pspan2->clear();
for (auto it = its[i]; it != its2[i]; ++it) {
auto pos = *it;
for (auto pot_match = lower_bound(pspan1->begin(), pspan1->end(), make_pair(pos - max_gap, pos));
pot_match != pspan1->end() && pot_match->first <= pos - min_gap;
++pot_match)
pspan2->emplace_back(pos, pot_match->second);
}
std::swap(pspan1, pspan2);
}
std::sort(pspan1->begin(), pspan1->end(), [](const pii &left, const pii &right) { return left.second < right.second; });
for (auto it = pspan1->begin(); it != pspan1->end(); ++it) {
res.positions.push_back(it->second);
auto end = it->first + last_size;
while (it != pspan1->end() && it->second < end)
++it;
}
return res;
}
index_bstree(collection& col)
{
sdsl::load_from_file(m_text, col.file_map[consts::KEY_TEXT]);
sdsl::csa_wt<sdsl::wt_huff<>, 1> csa;
sdsl::construct(csa, col.file_map[consts::KEY_TEXT], 0);
m_sa = sdsl::int_vector<>(csa.size(), 0, sdsl::bits::hi(csa.size() - 1) + 1);
std::copy(csa.begin(), csa.end(), m_sa.begin());
}
rank_tfidf(cache_config& cconfig) {
load_from_cache(num_terms, surf::KEY_COLLEN, cconfig);
if (!cache_file_exists(surf::KEY_DOC_LENGTHS, cconfig)){
surf::construct_doc_lengths<sdsl::int_alphabet_tag::WIDTH>(cconfig);
}
load_from_cache(doc_lengths, surf::KEY_DOC_LENGTHS, cconfig);
num_docs = doc_lengths.size();
std::cerr<<"num_docs = "<<num_docs<<std::endl;
auto min_itr = std::min_element(doc_lengths.begin(),doc_lengths.end());
min_doc_len = *min_itr;
std::cerr<<"min_doc_len = "<<min_doc_len<<std::endl;
}