C++ ContigNode::contigIndex方法代码示例

/** Attempt to merge the paths specified in mergeQ with path.
 * @return the number of paths merged
 */
static unsigned mergePaths(const Lengths& lengths,
		ContigPath& path,
		deque<ContigNode>& mergeQ, set<ContigNode>& seen,
		const ContigPathMap& paths)
{
	unsigned merged = 0;
	deque<ContigNode> invalid;
	for (ContigNode pivot; !mergeQ.empty(); mergeQ.pop_front()) {
		pivot = mergeQ.front();
		ContigPathMap::const_iterator path2It
			= paths.find(pivot.contigIndex());
		if (path2It == paths.end())
			continue;

		ContigPath path2 = path2It->second;
		if (pivot.sense())
			reverseComplement(path2.begin(), path2.end());
		ContigPath consensus = align(lengths, path, path2, pivot);
		if (consensus.empty()) {
			invalid.push_back(pivot);
			continue;
		}

		appendToMergeQ(mergeQ, seen, path2);
		path.swap(consensus);
		if (gDebugPrint)
#pragma omp critical(cout)
			cout << get(g_contigNames, pivot)
				<< '\t' << path2 << '\n'
				<< '\t' << path << '\n';
		merged++;
	}
	mergeQ.swap(invalid);
	return merged;
}

/** Return the specified path. */
static ContigPath getPath(const ContigPathMap& paths, ContigNode u)
{
	ContigPathMap::const_iterator it = paths.find(u.contigIndex());
	assert(it != paths.end());
	ContigPath path = it->second;
	if (u.sense())
		reverseComplement(path.begin(), path.end());
	return path;
}

/** Find a path for the specified distance estimates.
 * @param out [out] the solution path
 */
static void handleEstimate(const Graph& g,
		const EstimateRecord& er, bool dirIdx,
		ContigPath& out)
{
	if (er.estimates[dirIdx].empty())
		return;

	ContigNode origin(er.refID, dirIdx);
	ostringstream vout_ss;
	ostream bitBucket(NULL);
	ostream& vout = opt::verbose > 0 ? vout_ss : bitBucket;
	vout << "\n* " << get(vertex_name, g, origin) << '\n';

	unsigned minNumPairs = UINT_MAX;
	// generate the reachable set
	Constraints constraints;
	for (Estimates::const_iterator iter
				= er.estimates[dirIdx].begin();
			iter != er.estimates[dirIdx].end(); ++iter) {
		ContigNode v = iter->first;
		const DistanceEst& ep = iter->second;
		minNumPairs = min(minNumPairs, ep.numPairs);
		constraints.push_back(Constraint(v,
					ep.distance + allowedError(ep.stdDev)));
	}

	vout << "Constraints:";
	printConstraints(vout, g, constraints) << '\n';

	ContigPaths solutions;
	unsigned numVisited = 0;
	constrainedSearch(g, origin, constraints, solutions, numVisited);
	bool tooComplex = numVisited >= opt::maxCost;
	bool tooManySolutions = solutions.size() > opt::maxPaths;

	set<ContigID> repeats = findRepeats(er.refID, solutions);
	if (!repeats.empty()) {
		vout << "Repeats:";
		for (set<ContigID>::const_iterator it = repeats.begin();
				it != repeats.end(); ++it)
			vout << ' ' << get(g_contigNames, *it);
		vout << '\n';
	}

	unsigned numPossiblePaths = solutions.size();
	if (numPossiblePaths > 0)
		vout << "Paths: " << numPossiblePaths << '\n';

	for (ContigPaths::iterator solIter = solutions.begin();
			solIter != solutions.end();) {
		vout << *solIter << '\n';

		// Calculate the path distance to each node and see if
		// it is within the estimated distance.
		map<ContigNode, int> distanceMap
			= makeDistanceMap(g, origin, *solIter);

		// Remove solutions whose distance estimates are not correct.
		unsigned validCount = 0, invalidCount = 0, ignoredCount = 0;
		for (Estimates::const_iterator iter
					= er.estimates[dirIdx].begin();
				iter != er.estimates[dirIdx].end(); ++iter) {
			ContigNode v = iter->first;
			const DistanceEst& ep = iter->second;
			vout << get(vertex_name, g, v) << ',' << ep << '\t';

			map<ContigNode, int>::iterator dmIter
				= distanceMap.find(v);
			if (dmIter == distanceMap.end()) {
				// This contig is a repeat.
				ignoredCount++;
				vout << "ignored\n";
				continue;
			}

			// translate distance by -overlap to match
			// coordinate space used by the estimate
			int actualDistance = dmIter->second;
			int diff = actualDistance - ep.distance;
			unsigned buffer = allowedError(ep.stdDev);
			bool invalid = (unsigned)abs(diff) > buffer;
			bool repeat = repeats.count(v.contigIndex()) > 0;
			bool ignored = invalid && repeat;
			if (ignored)
				ignoredCount++;
			else if (invalid)
				invalidCount++;
			else
				validCount++;
			vout << "dist: " << actualDistance
				<< " diff: " << diff
				<< " buffer: " << buffer
				<< " n: " << ep.numPairs
				<< (ignored ? " ignored" : invalid ? " invalid" : "")
				<< '\n';
		}

//.........这里部分代码省略.........

/** Identify paths subsumed by the specified path.
 * @param overlaps [out] paths that are found to overlap
 * @return the ID of the subsuming path
 */
static ContigID identifySubsumedPaths(const Lengths& lengths,
		ContigPathMap::const_iterator path1It,
		ContigPathMap& paths,
		set<ContigID>& out,
		set<ContigID>& overlaps)
{
	ostringstream vout;
	out.clear();
	ContigID id(path1It->first);
	const ContigPath& path = path1It->second;
	if (gDebugPrint)
		vout << get(g_contigNames, ContigNode(id, false))
			<< '\t' << path << '\n';

	for (ContigPath::const_iterator it = path.begin();
			it != path.end(); ++it) {
		ContigNode pivot = *it;
		if (pivot.ambiguous() || pivot.id() == id)
			continue;
		ContigPathMap::iterator path2It
			= paths.find(pivot.contigIndex());
		if (path2It == paths.end())
			continue;
		ContigPath path2 = path2It->second;
		if (pivot.sense())
			reverseComplement(path2.begin(), path2.end());
		ContigPath consensus = align(lengths, path, path2, pivot);
		if (consensus.empty())
			continue;
		if (equalIgnoreAmbiguos(consensus, path)) {
			if (gDebugPrint)
				vout << get(g_contigNames, pivot)
					<< '\t' << path2 << '\n';
			out.insert(path2It->first);
		} else if (equalIgnoreAmbiguos(consensus, path2)) {
			// This path is larger. Use it as the seed.
			return identifySubsumedPaths(lengths, path2It, paths, out,
					overlaps);
		} else if (isCycle(lengths, consensus)) {
			// The consensus path is a cycle.
			bool isCyclePath1 = isCycle(lengths, path);
			bool isCyclePath2 = isCycle(lengths, path2);
			if (!isCyclePath1 && !isCyclePath2) {
				// Neither path is a cycle.
				if (gDebugPrint)
					vout << get(g_contigNames, pivot)
						<< '\t' << path2 << '\n'
						<< "ignored\t" << consensus << '\n';
				overlaps.insert(id);
				overlaps.insert(path2It->first);
			} else {
				// At least one path is a cycle.
				if (gDebugPrint)
					vout << get(g_contigNames, pivot)
						<< '\t' << path2 << '\n'
						<< "cycle\t" << consensus << '\n';
				if (isCyclePath1 && isCyclePath2)
					out.insert(path2It->first);
				else if (!isCyclePath1)
					overlaps.insert(id);
				else if (!isCyclePath2)
					overlaps.insert(path2It->first);
			}
		} else {
			if (gDebugPrint)
				vout << get(g_contigNames, pivot)
					<< '\t' << path2 << '\n'
					<< "ignored\t" << consensus << '\n';
			overlaps.insert(id);
			overlaps.insert(path2It->first);
		}
	}
	cout << vout.str();
	return id;
}