void graphTrimAndSmooth (StringGraph* pGraph, size_t trimLength, bool bIsGapPrecent)
{
pGraph->simplify();
SGTrimVisitor trimVisit("",trimLength);
SGSmoothingVisitor smoothingVisit(opt::maxIndelLength, opt::pBWT, bIsGapPrecent);
//SGSimpleBubbleVisitor sbVisit(opt::pBWT,opt::kmerLength,opt::maxBubbleGapDivergence, opt::maxBubbleDivergence, opt::maxIndelLength);
if (pGraph->visitP(trimVisit))
pGraph->simplify();
if (pGraph->visitP(smoothingVisit))
{
pGraph->simplify();
if (pGraph->visitP(trimVisit))
pGraph->simplify();
}
// SGDuplicateVisitor sgdup;
// if(pGraph->visit(sgdup))
// pGraph->simplify();
// if (pGraph->visit(sbVisit))
// {
// pGraph->simplify();
// if (pGraph->visitP(trimVisit))
// pGraph->simplify();
// }
}
void assemble()
{
Timer t("sga assemble");
StringGraph* pGraph = SGUtil::loadASQG(opt::asqgFile, opt::minOverlap, true);
if(opt::bExact)
pGraph->setExactMode(true);
pGraph->printMemSize();
// Visitor functors
SGTransitiveReductionVisitor trVisit;
SGGraphStatsVisitor statsVisit;
SGRemodelVisitor remodelVisit;
SGEdgeStatsVisitor edgeStatsVisit;
SGTrimVisitor trimVisit(opt::trimLengthThreshold);
SGBubbleVisitor bubbleVisit;
SGBubbleEdgeVisitor bubbleEdgeVisit;
SGContainRemoveVisitor containVisit;
SGValidateStructureVisitor validationVisit;
// Pre-assembly graph stats
std::cout << "[Stats] Input graph:\n";
pGraph->visit(statsVisit);
// Remove containments from the graph
std::cout << "Removing contained vertices from graph\n";
while(pGraph->hasContainment())
pGraph->visit(containVisit);
// Pre-assembly graph stats
std::cout << "[Stats] After removing contained vertices:\n";
pGraph->visit(statsVisit);
// Remove any extraneous transitive edges that may remain in the graph
if(opt::bPerformTR)
{
std::cout << "Removing transitive edges\n";
pGraph->visit(trVisit);
}
// Compact together unbranched chains of vertices
pGraph->simplify();
if(opt::bValidate)
{
std::cout << "Validating graph structure\n";
pGraph->visit(validationVisit);
}
// Remove dead-end branches from the graph
if(opt::numTrimRounds > 0)
{
std::cout << "Trimming bad vertices\n";
int numTrims = opt::numTrimRounds;
while(numTrims-- > 0)
pGraph->visit(trimVisit);
std::cout << "\n[Stats] Graph after trimming:\n";
pGraph->visit(statsVisit);
}
// Resolve small repeats
if(opt::resolveSmallRepeatLen > 0)
{
SGSmallRepeatResolveVisitor smallRepeatVisit(opt::resolveSmallRepeatLen);
std::cout << "Resolving small repeats\n";
int totalSmallRepeatRounds = 0;
while(pGraph->visit(smallRepeatVisit))
std::cout << "Finished small repeat resolve round " << totalSmallRepeatRounds++ << "\n";
std::cout << "\n[Stats] After small repeat resolution:\n";
pGraph->visit(statsVisit);
}
//
if(opt::coverageCutoff > 0)
{
std::cout << "Coverage visit\n";
SGCoverageVisitor coverageVisit(opt::coverageCutoff);
pGraph->visit(coverageVisit);
pGraph->visit(trimVisit);
pGraph->visit(trimVisit);
pGraph->visit(trimVisit);
}
// Peform another round of simplification
pGraph->simplify();
if(opt::numBubbleRounds > 0)
{
std::cout << "\nPerforming variation smoothing\n";
SGSmoothingVisitor smoothingVisit(opt::outVariantsFile, opt::maxBubbleGapDivergence, opt::maxBubbleDivergence, opt::maxIndelLength);
int numSmooth = opt::numBubbleRounds;
while(numSmooth-- > 0)
pGraph->visit(smoothingVisit);
pGraph->simplify();
}
pGraph->renameVertices("contig-");
std::cout << "\n[Stats] Final graph:\n";
pGraph->visit(statsVisit);
// Rename the vertices to have contig IDs instead of read IDs
//pGraph->renameVertices("contig-");
// Write the results
SGFastaVisitor av(opt::outContigsFile);
pGraph->visit(av);
pGraph->writeASQG(opt::outGraphFile);
delete pGraph;
}