//
// selectCandidates: build a list of candidates
//
// Notes:
// For primary detection.
void
SignalClassifier::classifySignals(Activity *act)
{
SystemType origin = act->getMode();
Debug(DEBUG_NEVER, 1, "classify signals");
params = act->getActivityParams();
int32_t maxCandidates = params.maxNumberOfCandidates;
recentRfi = act->getRecentRfiMask();
testSignal = act->getTestSignalMask();
operations = params.operations;
act->setCandidatesOverMax(candidatesOverMax = 0);
SuperClusterer *superClusterer = act->getSuperClusterer();
Assert(superClusterer);
// scan the list of signals
for (int32_t i = 0; i < superClusterer->getCount(); i++) {
// get the next signal
ClusterTag tag = superClusterer->getNthMainSignal(i);
CwClusterer *cwClusterer = tag.holder->getCw();
PulseClusterer *pulseClusterer = tag.holder->getPulse();
if (cwClusterer) {
Debug(DEBUG_NEVER, i, "cw clusterer");
CwPowerSignal& signal = cwClusterer->getNth(tag.index);
classifySignal(act, signal.sig, superClusterer, i, maxCandidates);
signal.sig.containsBadBands = applyBadBands(act, signal.sig);
if (signal.sig.sigClass == CLASS_CAND)
act->addCwCandidate(&signal, origin);
act->addCwSignal(&signal);
}
else if (pulseClusterer) {
Debug(DEBUG_NEVER, i, "pulse cluster");
PulseSignalHeader *signal = &pulseClusterer->getNth(tag.index);
// cout << "SignalClassifier pulse " << *signal;
// Pulse *p = (Pulse *) (signal + 1);
// for (int j = 0; j < signal->train.numberOfPulses; ++j)
// cout << "pulse " << j << p[j];
classifySignal(act, signal->sig, superClusterer, i, maxCandidates);
signal->sig.containsBadBands = applyBadBands(act, signal->sig);
if (signal->sig.sigClass == CLASS_CAND)
act->addPulseCandidate(signal, origin);
act->addPulseSignal(signal);
// displayPulseSignal(*signal);
}
}
act->setCandidatesOverMax(candidatesOverMax);
}
void PhylogenyViewer::gatherKmerObservations(){
/* set to true to use only assembled kmers */
bool useOnlyAssembledKmer=false;
GridTableIterator iterator;
iterator.constructor(m_subgraph,m_parameters->getWordSize(),m_parameters);
//* only fetch half of the iterated things because we just need one k-mer
// for any pair of reverse-complement k-mers
int parity=0;
map<CoverageDepth,LargeCount> frequencies;
while(iterator.hasNext()){
#ifdef ASSERT
assert(parity==0 || parity==1);
#endif
Vertex*node=iterator.next();
Kmer key=*(iterator.getKey());
if(parity==0){
parity=1;
}else if(parity==1){
parity=0;
continue; // we only need data with parity=0
}
// check for assembly paths
/* here, we just want to find a path with
* a good progression */
Direction*a=node->m_directions;
bool nicelyAssembled=false;
while(a!=NULL){
int progression=a->getProgression();
if(progression>= CONFIG_NICELY_ASSEMBLED_KMER_POSITION){
nicelyAssembled=true;
}
a=a->getNext();
}
if(useOnlyAssembledKmer && !nicelyAssembled){
continue; // the k-mer is not nicely assembled...
}
#ifdef ASSERT
assert(nicelyAssembled || !useOnlyAssembledKmer);
#endif
// at this point, we have a nicely assembled k-mer
int kmerCoverage=node->getCoverage(&key);
VirtualKmerColorHandle color=node->getVirtualColor();
set<PhysicalKmerColor>*physicalColors=m_colorSet->getPhysicalColors(color);
vector<TaxonIdentifier> taxons;
// get a list of taxons associated with this kmer
for(set<PhysicalKmerColor>::iterator j=physicalColors->begin();
j!=physicalColors->end();j++){
PhysicalKmerColor physicalColor=*j;
PhysicalKmerColor nameSpace=physicalColor/COLOR_NAMESPACE_MULTIPLIER;
if(nameSpace==COLOR_NAMESPACE_PHYLOGENY){
PhysicalKmerColor colorForPhylogeny=physicalColor % COLOR_NAMESPACE_MULTIPLIER;
#ifdef ASSERT
if(m_colorsForPhylogeny.count(colorForPhylogeny)==0){
//cout<<"Error: color "<<colorForPhylogeny<<" should be in m_colorsForPhylogeny which contains "<<m_colorsForPhylogeny.size()<<endl;
}
#endif
//assert(m_colorsForPhylogeny.count(colorForPhylogeny)>0);
// this means that this genome is not in the taxonomy tree
if(m_genomeToTaxon.count(colorForPhylogeny)==0){
if(m_warnings.count(colorForPhylogeny)==0){
cout<<"Warning, color "<<colorForPhylogeny<<" is not stored, "<<m_genomeToTaxon.size()<<" available. This means that you provided a genome sequence that is not classified in the taxonomy."<<endl;
#ifdef VERBOSE
for(map<GenomeIdentifier,TaxonIdentifier>::iterator i=m_genomeToTaxon.begin();i!=m_genomeToTaxon.end();i++){
cout<<" "<<i->first<<"->"<<i->second;
}
cout<<endl;
#endif
}
m_warnings.insert(colorForPhylogeny);
continue;
}
#ifdef ASSERT
assert(m_genomeToTaxon.count(colorForPhylogeny)>0);
#endif
TaxonIdentifier taxon=m_genomeToTaxon[colorForPhylogeny];
taxons.push_back(taxon);
}
}
classifySignal(&taxons,kmerCoverage,node,&key);
int count=taxons.size();
frequencies[count]++;
}
/*
*
* TODO: move this in colored operation files
*
cout<<endl;
cout<<"Taxon frequencies (only one DNA strand selected)"<<endl;
cout<<"Count Frequency"<<endl;
for(map<CoverageDepth,LargeCount>::iterator i=frequencies.begin();i!=frequencies.end();i++){
cout<<""<<i->first<<" "<<i->second<<endl;
}
cout<<"Taxon observations"<<endl;
*/
//showObservations(&cout);
m_gatheredObservations=true;
m_syncedTree=false;
m_unknownSent=false;
m_messageReceived=true;
m_messageSent=false;
m_countIterator=m_taxonObservations.begin();
}
