int half_sw_banded(const Ts1 &s1, const Ts2 &s2, int match, int mismatch, int indel, const Tf &lizard_tail, int max_indels = 6) { using std::min; using std::max; using seqan::length; using std::vector; const int INF = 1005000; if (max_indels == 0) { return half_hamming(s1, s2, match, mismatch, lizard_tail); } int len1 = length(s1), len2 = length(s2); // Cache lengths because of length(s) computation maybe not O(1) e.g. for char*-strings vector<int> base(2*max_indels + 1, -INF); for (int i2 = max(0, len1 - max_indels); i2 <= min(len1 + max_indels, len2); ++i2) { int ind = i2 + max_indels - len1; base[ind] = lizard_tail(len2 - i2); } vector<int> new_base(2*max_indels + 1, -INF); for (int i1 = len1 - 1; i1 >= 0; --i1) { for (int i2 = i1 + max_indels; i2 >= i1 - max_indels; --i2) { int inx = max_indels + i2 - i1; if ((i2 < 0) || (i2 > len2)) { new_base[inx] = -INF; } else if (i2 == len2) { new_base[inx] = lizard_tail(len1 - i1); } else { int m = (s1[i1] == s2[i2]) ? match : mismatch; if (inx == 0) new_base[inx] = max( { m + base[inx], indel + new_base[inx + 1]} ); else if (inx == 2*max_indels) new_base[inx] = max( { m + base[inx], indel + base[inx - 1] } ); else new_base[inx] = max( { m + base[inx], indel + base[inx - 1], indel + new_base[inx + 1]} ); } } std::swap(base, new_base); } return max(-INF, base[max_indels]); // Score is always finite due to possibility to align using mismatches }
int half_hamming(const Ts1 &s1, const Ts2 &s2, int match, int mismatch, const Tf &lizard_tail) { int len1 = length(s1), len2 = length(s2); // Cache lengths because of length(s) computation maybe not O(1) e.g. for char*-strings size_t len = std::min<size_t>(len1, len2); size_t res = 0; for (size_t i = 0; i < len; ++i) { res += (s1[i] == s2[i]) ? match : mismatch; } res += lizard_tail(std::abs(len1 - len2)); return res; }
void add(const T &s, size_t id, const Tf &toIndex) { if (!root) { root.reset(new TrieNode); } typename TrieNode::pointer_type p = this->root.get(); for (size_t i = 0; i < length(s); ++i) { size_t el = toIndex(s[i]); assert((0 <= el) && (el < p->children.size())); if (!p->children[el]) { p->children[el] = new TrieNode(); } p = p->children[el]; } if (!p->ids) { p->ids = new std::vector<size_t>; } p->ids->push_back(id); }