bool RemasterWithStructure(CCdCore* cd, string* msg)
{
static const string msgHeader = "Remastering CD to ";
CRef< CSeq_id > seqId;
cd->GetSeqIDForRow(0,0,seqId);
if (seqId->IsPdb())
return false;
AlignmentCollection ac(cd,CCdCore::USE_NORMAL_ALIGNMENT);
int nrows = ac.GetNumRows();
int i = 1;
for (; i < nrows; i++)
{
ac.GetSeqIDForRow(i,seqId);
if (seqId->IsPdb())
break;
}
if ( i < nrows)
{
ReMasterCdWithoutUnifiedBlocks(cd, i, true);
if (msg) {
*msg = msgHeader + seqId->AsFastaString();
}
return true;
}
else
return false;
}
int PurgeConsensusSequences(CCdCore* pCD, bool resetFields)
{
int nPurged = 0;
vector<int> consensusRows, consensusSeqListIds;
if (pCD) {
// First see if the master is a consensus; if so, remaster to 2nd row
if (pCD->UsesConsensusSequenceAsMaster()) {
ReMasterCdWithoutUnifiedBlocks(pCD, 1,true);
}
// Next, find all rows using 'consensus' as a seq_id and remove them.
nPurged = pCD->GetRowsWithConsensus(consensusRows);
if (nPurged) {
pCD->EraseTheseRows(consensusRows);
if (pCD->FindConsensusInSequenceList(&consensusSeqListIds)) {
for (int i = consensusSeqListIds.size() - 1; i >= 0 ; --i) {
pCD->EraseSequence(consensusSeqListIds[i]);
}
}
}
}
return nPurged;
}
void SeqSwapper::swapSequences()
{
int numNormal = m_cd->GetNumRows();
LOG_POST("Clustering");
vector< vector<int> * > clusters;
makeClusters(m_clusteringThreshold, clusters);
vector< pair<int, int> > replacementPairs;
set<int> structures;
LOG_POST("Clustering is done (made " << clusters.size() << " clusters)");
LOG_POST("Find replacements by BLAST in each cluster");
for (int i = 0; i < clusters.size(); i++)
{
vector<int>* cluster = clusters[i];
if (cluster) {
findReplacements(*cluster, replacementPairs, structures);
delete cluster;
}
}
LOG_POST("Done with BLAST in each cluster");
set<int> usedPendings;
vector<int> selectedNormalRows;
int newMaster = -1;
for (int p = 0; p < replacementPairs.size(); p++)
{
//debug
CRef< CSeq_id > seqId;
m_ac.GetSeqIDForRow(replacementPairs[p].first, seqId);
string nid = seqId->AsFastaString();
m_ac.GetSeqIDForRow(replacementPairs[p].second, seqId);
string pid = seqId->AsFastaString();
LOG_POST("replacing "<<nid<<" with "<<pid);
//take care of master replacement
if (replacementPairs[p].first == 0)
newMaster = replacementPairs[p].second - numNormal;
else
{
selectedNormalRows.push_back(replacementPairs[p].first);
usedPendings.insert(replacementPairs[p].second - numNormal);
}
}
m_cd->EraseTheseRows(selectedNormalRows);
if (structures.size() > 0)
{
LOG_POST("Adding "<<structures.size()<<" structures");
for (set<int>::iterator sit = structures.begin(); sit != structures.end(); sit++)
usedPendings.insert(*sit - numNormal);
}
if (newMaster >= 0)
promotePendingRows(usedPendings, &newMaster);
else
promotePendingRows(usedPendings);
//findStructuralPendings(structures);
if (newMaster > 0)
{
ReMasterCdWithoutUnifiedBlocks(m_cd, newMaster, true);
vector<int> rows;
rows.push_back(newMaster);
m_cd->EraseTheseRows(rows);
}
m_cd->ResetPending();
m_cd->EraseSequences();
}
