void LodManager::checkAlignment(hal_size_t minQuery,
const string& path,
AlignmentConstPtr alignment)
{
if (alignment->getNumGenomes() == 0)
{
stringstream ss;
ss << "No genomes found in base alignment specified in " << path;
throw hal_exception(ss.str());
}
#ifndef NDEBUG
if (minQuery == 0)
{
vector<string> leafNames = alignment->getLeafNamesBelow(
alignment->getRootName());
string name = !leafNames.empty() ? leafNames[0] : alignment->getRootName();
const Genome* genome = alignment->openGenome(name);
bool seqFound = genome->containsDNAArray();
alignment->closeGenome(genome);
if (seqFound == false)
{
stringstream ss;
ss << "HAL file for highest level of detail (0) in " << path
<< "must contain DNA sequence information.";
throw hal_exception(ss.str());
}
}
#endif
}
static void printBranches(ostream& os, AlignmentConstPtr alignment)
{
const Genome* root = alignment->openGenome(alignment->getRootName());
set<const Genome*> genomes;
getGenomesInSubTree(root, genomes);
genomes.insert(root);
bool first = true;
for (set<const Genome*>::iterator i = genomes.begin(); i != genomes.end();
++i)
{
if ((*i)->getParent() != NULL)
{
if (!first)
{
os << " ";
}
else
{
first = false;
}
os << (*i)->getName();
}
}
os << endl;
}
void MappedSegmentColCompareTest::checkCallBack(AlignmentConstPtr alignment)
{
if (alignment->getNumGenomes() == 0)
{
return;
}
validateAlignment(alignment);
set<const Genome*> genomeSet;
hal::getGenomesInSubTree(alignment->openGenome(alignment->getRootName()),
genomeSet);
for (set<const Genome*>::iterator i = genomeSet.begin(); i != genomeSet.end();
++i)
{
const Genome* srcGenome = *i;
for (set<const Genome*>::iterator j = genomeSet.begin();
j != genomeSet.end(); ++j)
{
const Genome* tgtGenome = *j;
if (srcGenome->getSequenceLength() > 0 &&
tgtGenome->getSequenceLength() > 0)
{
_ref = srcGenome;
_tgt = tgtGenome;
createColArray();
createBlockArray();
compareArrays();
}
}
}
}
void printParent(ostream& os, AlignmentConstPtr alignment,
const string& genomeName)
{
if (genomeName != alignment->getRootName())
{
os << alignment->getParentName(genomeName) << endl;
}
}
void printBranchLength(ostream& os, AlignmentConstPtr alignment,
const string& genomeName)
{
if (genomeName != alignment->getRootName())
{
string parentName = alignment->getParentName(genomeName);
os << alignment->getBranchLength(parentName, genomeName) << endl;
}
}
void printGenomes(ostream& os, AlignmentConstPtr alignment)
{
const Genome* root = alignment->openGenome(alignment->getRootName());
set<const Genome*> genomes;
getGenomesInSubTree(root, genomes);
genomes.insert(root);
for (set<const Genome*>::iterator i = genomes.begin(); i != genomes.end();
++i)
{
if (i != genomes.begin())
{
os << ",";
}
os << (*i)->getName();
}
os << endl;
}
void hal::validateAlignment(AlignmentConstPtr alignment)
{
deque<string> bfQueue;
bfQueue.push_back(alignment->getRootName());
while (bfQueue.empty() == false)
{
string name = bfQueue.back();
bfQueue.pop_back();
if (name.empty() == false)
{
const Genome* genome = alignment->openGenome(name);
if (genome == NULL)
{
throw hal_exception("Failure to open genome " + name);
}
validateGenome(genome);
vector<string> childNames = alignment->getChildNames(name);
for (size_t i = 0; i < childNames.size(); ++i)
{
bfQueue.push_front(childNames[i]);
}
}
}
}
void printRootName(ostream& os, AlignmentConstPtr alignment)
{
os << alignment->getRootName() << endl;
}
void MafExport::convertEntireAlignment(ostream& mafStream,
AlignmentConstPtr alignment)
{
hal_size_t appendCount = 0;
size_t numBlocks = 0;
_mafStream = &mafStream;
_alignment = alignment;
writeHeader();
// Load in all leaves from alignment
vector<string> leafNames = alignment->getLeafNamesBelow(alignment->getRootName());
vector<const Genome *> leafGenomes;
for (hal_size_t i = 0; i < leafNames.size(); i++) {
const Genome *genome = alignment->openGenome(leafNames[i]);
assert(genome != NULL);
leafGenomes.push_back(genome);
}
ColumnIterator::VisitCache visitCache;
// Go through all the genomes one by one, and spit out any columns
// they participate in that we haven't seen.
for (hal_size_t i = 0; i < leafGenomes.size(); i++) {
const Genome *genome = leafGenomes[i];
ColumnIteratorConstPtr colIt = genome->getColumnIterator(NULL,
0,
0,
NULL_INDEX,
_noDupes,
_noAncestors);
colIt->setVisitCache(&visitCache);
for (;;) {
if (appendCount == 0) {
_mafBlock.initBlock(colIt, _ucscNames, _printTree);
assert(_mafBlock.canAppendColumn(colIt) == true);
}
if (_mafBlock.canAppendColumn(colIt) == false)
{
// erase empty entries from the column. helps when there are
// millions of sequences (ie from fastas with lots of scaffolds)
if (numBlocks++ % 1000 == 0)
{
colIt->defragment();
}
if (appendCount > 0)
{
mafStream << _mafBlock << '\n';
}
_mafBlock.initBlock(colIt, _ucscNames, _printTree);
assert(_mafBlock.canAppendColumn(colIt) == true);
}
_mafBlock.appendColumn(colIt);
appendCount++;
if (colIt->lastColumn()) {
// Have to break here because otherwise
// colIt->toRight() will crash.
break;
}
colIt->toRight();
}
// Copy over the updated visit cache information. This is a
// deep copy, so it's slow, but necessary to preserve the
// column iterator ownership of the visit cache
visitCache.clear();
ColumnIterator::VisitCache *newVisitCache = colIt->getVisitCache();
for(ColumnIterator::VisitCache::iterator it = newVisitCache->begin();
it != newVisitCache->end(); it++) {
visitCache[it->first] = new PositionCache(*it->second);
}
}
// if nothing was ever added (seems to happen in corner case where
// all columns violate unique), mafBlock ostream operator will crash
// so we do following check
if (appendCount > 0)
{
mafStream << _mafBlock << endl;
}
}