RDKit::INT_VECT HierarchicalClusterPicker::pick(const double *distMat,
unsigned int poolSize,
unsigned int pickSize) const {
PRECONDITION(distMat,"bad distance matrix");
RDKit::VECT_INT_VECT clusters = this->cluster(distMat, poolSize, pickSize);
CHECK_INVARIANT(clusters.size() == pickSize, "");
// the last step: find a representative element from each of the
// remaining clusters
RDKit::INT_VECT picks;
for (unsigned int i = 0; i < pickSize; i++) {
int pick;
double minSumD2 = RDKit::MAX_DOUBLE;
for (RDKit::INT_VECT_CI cxi1 = clusters[i].begin();
cxi1 != clusters[i].end(); ++cxi1 ) {
int curPick = (*cxi1);
double d2sum = 0.0;
for (RDKit::INT_VECT_CI cxi2 = clusters[i].begin();
cxi2 != clusters[i].end(); ++cxi2) {
if (cxi1 == cxi2) {
continue;
}
double d = getDistFromLTM(distMat, curPick, (*cxi2));
d2sum += (d*d);
}
if (d2sum < minSumD2) {
pick = curPick;
minSumD2 = d2sum;
}
}
picks.push_back(pick);
}
return picks;
}
