void operator() (Sequence& seq){
int used_windows_size=windows_size;
if (windows_size==0 || windows_size>seq.getDataSize()){ // if windows size == 0 then we use the full read length as windows
used_windows_size=seq.getDataSize();
}
if(not keep_low_complexity and not is_high_complexity(seq,kmer_size)){return;}
bool exists;
unsigned char count;
itKmer->setData (seq.getData());
vector<int> values; // For each position: number of occurrences of the kmer starting at this position.
// vector<int> covered_positions; // DEPRECATED [OLD WAY FOR COMPUTING SHARE KMER POSITIONS, FASTER BUT NON IMPLEMENTED WITH WINDOWS SIZE METHODS (max_populated_window)]
vector<bool> position_shared = vector<bool>(seq.getDataSize()); // boolean vector. A position is true if it's covered by at least a shared kmer.
for (int pos=0;pos<seq.getDataSize();pos++) position_shared[pos]=false;
int position=0;
for (itKmer->first(); !itKmer->isDone(); itKmer->next()){
quasiDico->get_value((*itKmer)->value().getVal(),exists,count);
if(!exists) {
count=0;
}
values.push_back(count);
// if (count>0) covered_positions.push_back(position); // DEPRECATED [OLD WAY FOR COMPUTING SHARE KMER POSITIONS, FASTER BUT NON IMPLEMENTED WITH WINDOWS SIZE METHODS (max_populated_window)]
if (count>0) { // TODO: OPTIMIZABLE.
for (int pos=position;pos<position+kmer_size && pos<=seq.getDataSize();pos++) position_shared[pos]=true;
}
position++;
}
// float percentage_shared_positions = 100*number_positions_covered_shared_kmer(covered_positions, seq.getDataSize())/float(seq.getDataSize()); // DEPRECATED [OLD WAY FOR COMPUTING SHARE KMER POSITIONS, FASTER BUT NON IMPLEMENTED WITH WINDOWS SIZE METHODS (max_populated_window)]
const int mpw = max_populated_window(position_shared,used_windows_size);
const float percentage_span_kmer = 100*mpw/float(used_windows_size);
// if (percentage_shared_positions !=percentage_span_kmer){cout<<percentage_shared_positions<< " == " <<percentage_span_kmer<<" ?"<<endl; exit(1);} // TO REMOVE
float mean;
int median, min, max;
if(mean_median_min_max(values, mean, median, min, max)){
string toPrint (to_string(seq.getIndex())+" "+to_string(mean)+" "+to_string(median)+" "+to_string(min)+" "+to_string(max)+" "+to_string(percentage_span_kmer));
// toPrint.append(" ");
// for(int i=0;i<seq.getDataSize() ;i++){
//
// if (position_shared[i]) {
// toPrint.append("1");
// }
// else toPrint.append("0");
// }
toPrint.append("\n");
if (percentage_span_kmer>=threshold) {
synchro->lock();
// bv->set(seq.getIndex());
fwrite(toPrint.c_str(), sizeof(char), toPrint.size(), outFile);
synchro->unlock ();
}
synchro->unlock ();
}
else{
string toPrint (to_string(seq.getIndex())+" none\n");
synchro->lock();
fwrite(toPrint.c_str(), sizeof(char), toPrint.size(), outFile);
synchro->unlock ();
}
}
bool operator () (Sequence& seq) const { return indexes.find (seq.getIndex()) != indexes.end(); }
