用途
高速な文字列検索を行う. イメージはgrep.
計算量
構築:
クエリ:
使い方
構築
SuffixArray sufa(s);検索
auto res = sufa.lower_upper_bound(t);一致部分として[res.first,res.second)が得られる.
実装(WIP)
struct SuffixArray { vector<int> SA; const string s; SuffixArray(const string &str) : s(str) { SA.resize(s.size()); iota(begin(SA), end(SA), 0); sort(begin(SA), end(SA), [&](int a, int b) { return s[a] == s[b] ? a > b : s[a] < s[b]; }); vector<int> classes(s.size()), c(s.begin(), s.end()), cnt(s.size()); for (int len = 1; len < s.size(); len <<= 1) { for (int i = 0; i < s.size(); i++) { if (i > 0 && c[SA[i - 1]] == c[SA[i]] && SA[i - 1] + len < s.size() && c[SA[i - 1] + len / 2] == c[SA[i] + len / 2]) { classes[SA[i]] = classes[SA[i - 1]]; } else { classes[SA[i]] = i; } } iota(begin(cnt), end(cnt), 0); copy(begin(SA), end(SA), begin(c)); for (int i = 0; i < s.size(); i++) { int s1 = c[i] - len; if (s1 >= 0) SA[cnt[classes[s1]]++] = s1; } classes.swap(c); } } int operator[](int k) const { return SA[k]; } size_t size() const { return s.size(); } bool lt_substr(const string &t, int si = 0, int ti = 0) { int sn = (int)s.size(), tn = (int)t.size(); while (si < sn && ti < tn) { if (s[si] < t[ti]) return true; if (s[si] > t[ti]) return false; ++si, ++ti; } return si >= sn && ti < tn; } int lower_bound(const string &t) { int low = -1, high = (int)SA.size(); while (high - low > 1) { int mid = (low + high) / 2; if (lt_substr(t, SA[mid])) low = mid; else high = mid; } return high; } pair<int, int> lower_upper_bound(string &t) { int idx = lower_bound(t); int low = idx - 1, high = (int)SA.size(); t.back()++; while (high - low > 1) { int mid = (low + high) / 2; if (lt_substr(t, SA[mid])) low = mid; else high = mid; } t.back()--; return {idx, high}; } void output() { for (int i = 0; i < size(); i++) { cout << i << ": " << s.substr(SA[i]) << endl; } }};Verify
//TODO