unsigned int LevenshteinDistance::static_distance(const Sequence &sequence1, const Sequence &sequence2, const unsigned int cost_sub, const unsigned int cost_indel) {
size_t n = sequence1.length();
size_t m = sequence2.length();
unsigned int distances[n+1][m+1];
// Distance from empty string to the position
for (size_t i = 0; i <= n; i++) {
distances[i][0] = i * cost_indel;
}
for (size_t j = 0; j <= m; j++) {
distances[0][j] = j * cost_indel;
}
for (size_t i = 1; i <= n; i++) {
for (size_t j = 1; j <= m; j++) {
unsigned int cost = 0;
if (sequence1.at(i-1) != sequence2.at(j-1)) {
cost = cost_sub;
}
distances[i][j] = std::min(std::min(distances[i-1][j-1] + cost, distances[i][j-1] + cost_indel), distances[i-1][j] + cost_indel);
}
}
unsigned int min_distance = distances[n][m];
return (min_distance);
}
double DeterminatorInterpolated::getPercentile(const Sequence &sequence, int percentile) const {
size_t size = sequence.size();
double coeff = 100.0 / size;
double calcIdx = percentile / coeff - .5;
if (calcIdx <= 0) return sequence.front();
if (calcIdx >= size - 1) return sequence.back();
size_t idx0 = (size_t) floor(calcIdx);
int v0 = sequence.at(idx0);
int v1 = sequence.at(idx0 + 1);
double p0 = 100.0 / size * (idx0);
double slope = size / 100.0 * (percentile - p0);
return v0 + (v1 - v0) * slope;
}
bool LevenshteinDistance::is_seq_insertable(const std::vector<Sequence> &seqs, const Sequence &sequence, const size_t n, const unsigned int min_dist) {
size_t n_elements = seqs.size();
size_t m = sequence.length();
unsigned int distances[n+1][m+1];
for (size_t i = 0; i <= n; i++) {
distances[i][0] = i * cost_indel_;
}
for (size_t j = 0; j <= m; j++) {
distances[0][j] = j * cost_indel_;
}
bool seq_is_insertable = true;
for (size_t seq1_index = 0; seq1_index < n_elements && seq_is_insertable; seq1_index++) {
Sequence sequence1 = seqs.at(seq1_index);
for (size_t i = 1; i <= n; i++) {
for (size_t j = 1; j <= m; j++) {
unsigned int cost = 0;
if (sequence1.at(i-1) != sequence.at(j-1)) {
cost = cost_sub_;
}
distances[i][j] = std::min(std::min(distances[i-1][j-1] + cost, distances[i][j-1] + cost_indel_), distances[i-1][j] + cost_indel_);
}
}
unsigned int min_distance = distances[n][m];
if (min_distance < min_dist)
seq_is_insertable = false;
}
return seq_is_insertable;
}
unsigned int LevenshteinDistance::static_min_seq_distance(const std::vector< Sequence > &seqs, const Sequence &sequence, const size_t n, const unsigned int cost_sub, const unsigned int cost_indel) {
size_t n_elements = seqs.size();
size_t m = sequence.length();
unsigned int global_min_dist = UINT_MAX;
unsigned int distances[n+1][m+1];
for (size_t i = 0; i <= n; i++) {
distances[i][0] = i * cost_indel;
}
for (size_t j = 0; j <= m; j++) {
distances[0][j] = j * cost_indel;
}
for (size_t seq1_index = 0; seq1_index < n_elements; seq1_index++) {
Sequence sequence1 = seqs.at(seq1_index);
for (size_t i = 1; i <= n; i++) {
for (size_t j = 1; j <= m; j++) {
unsigned int cost = cost_sub;
if (sequence1.at(i-1) != sequence.at(j-1)) {
cost = cost_sub;
}
distances[i][j] = std::min(std::min(distances[i-1][j-1] + cost, distances[i][j-1] + cost_indel), distances[i-1][j] + cost_indel);
}
}
unsigned int min_distance = distances[n][m];
if (min_distance < global_min_dist)
global_min_dist = min_distance;
}
return(global_min_dist);
}
