Azure: fix conflict

.circleci/config.yml

@@ -76,6 +76,25 @@ jobs:
           ./configure CC=gcc-8 CXX=g++-8 --with-mpi=/usr/lib/openmpi
       - run: make -j12 distcheck
 
+  clang-format:
+    docker:
+      - image: *docker_image
+    steps:
+      - run: *install_common
+      - run: |
+          apt-get install --yes --force-yes sudo
+          sudo apt-get install --yes --force-yes curl
+          curl https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
+          sudo apt-add-repository "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-8 main"
+          sudo apt-get update
+          sudo apt-get install -y --no-install-recommends clang-format-8
+          sudo ln -s clang-format-8 /usr/bin/clang-format
+      - checkout
+      - run: |
+          ./autogen.sh
+          ./configure
+      - run: make clang-format
+
 workflows:
   version: 2
   abyss_builds:
@@ -85,3 +104,4 @@ workflows:
       - linux_gcc6
       - linux_gcc5
       - linux_clang6
+      - clang-format

.clang-format

@@ -0,0 +1,9 @@
+BasedOnStyle: Mozilla
+AlignAfterOpenBracket: AlwaysBreak
+AlignOperands: true
+ColumnLimit: 100
+ContinuationIndentWidth: 4
+IndentCaseLabels: false
+IndentWidth: 4
+TabWidth: 4
+UseTab: ForIndentation

Assembly/Options.cc

@@ -17,6 +17,9 @@ using namespace std;
 
 #define PROGRAM "ABYSS"
 
+#define STR_HELPER(x) #x
+#define STR(x) STR_HELPER(x)
+
 namespace opt {
 
 static const char VERSION_MESSAGE[] =
@@ -48,7 +51,7 @@ static const char USAGE_MESSAGE[] =
 "      --SS              assemble in strand-specific mode\n"
 "      --no-SS           do not assemble in strand-specific mode\n"
 "  -o, --out=FILE        write the contigs to FILE\n"
-"  -k, --kmer=N          the length of a k-mer (when -K is not set)\n"
+"  -k, --kmer=N          the length of a k-mer (when -K is not set) [<=" STR(MAX_KMER) "]\n"
 "                        or the span of a k-mer pair (when -K is set)\n"
 "  -K, --single-kmer=N   the length of a single k-mer in a k-mer pair\n"
 "  -t, --trim-length=N   maximum length of blunt contigs to trim [k]\n"

Bloom/Bloom.h

@@ -46,28 +46,28 @@ namespace Bloom {
 	/** Print a progress message after loading this many seqs */
 	static const unsigned LOAD_PROGRESS_STEP = 100000;
 	/** file format version number */
-	static const unsigned BLOOM_VERSION = 4;
+	static const unsigned BLOOM_VERSION = 5;
 
 	/** Return the hash value of this object. */
 	inline static size_t hash(const key_type& key)
 	{
 		if (key.isCanonical())
-			return hashmem(&key, sizeof key);
+			return hashmem(&key, key.bytes());
 
 		key_type copy(key);
 		copy.reverseComplement();
-		return hashmem(&copy, sizeof copy, 0);
+		return hashmem(&copy, copy.bytes(), 0);
 	}
 
 	/** Return the hash value of this object given seed. */
 	inline static size_t hash(const key_type& key, size_t seed)
 	{
 		if (key.isCanonical())
-			return hashmem(&key, sizeof key, seed);
+			return hashmem(&key, key.bytes(), seed);
 
 		key_type copy(key);
 		copy.reverseComplement();
-		return hashmem(&copy, sizeof copy, seed);
+		return hashmem(&copy, copy.bytes(), seed);
 	}
 
 	template <typename BF>
@@ -190,17 +190,6 @@ namespace Bloom {
 		return header;
 	}
 
-	//TODO: Bloom filter calculation methods
-	//static double calcApproxFPR(size_t numBucket, size_t numEntr,
-	//		unsigned numHash) const;
-	//static double calcRedunancyFPR(size_t numBucket, size_t numEntr,
-	//		unsigned numHash) const;
-	//static size_t calcOptimalSize(size_t numEle, double fpr) const;
-	//static size_t calcOptimalSize(size_t numEle, double fpr,
-	//		unsigned numHash) const;
-	//static unsigned calcOptimalNumHash(double fpr);
-	//static unsigned calcOptimalNumHash(size_t numBucket, size_t numEle);
-
 };
 
 #endif /* BLOOM_H_ */

BloomDBG/HashAgnosticCascadingBloom.h → Bloom/HashAgnosticCascadingBloom.h

@@ -5,7 +5,7 @@
 #ifndef HASH_AGNOSTIC_CASCADING_BLOOM_H
 #define HASH_AGNOSTIC_CASCADING_BLOOM_H 1
 
-#include "lib/bloomfilter/BloomFilter.hpp"
+#include "vendor/btl_bloomfilter/BloomFilter.hpp"
 #include <vector>
 
 /**
@@ -26,9 +26,11 @@
 class HashAgnosticCascadingBloom
 {
   public:
-
 	/** Default constructor */
-	HashAgnosticCascadingBloom() : m_k(0), m_hashes(0) {}
+	HashAgnosticCascadingBloom()
+	  : m_k(0)
+	  , m_hashes(0)
+	{}
 
 	/**
 	 * Constructor.
@@ -37,8 +39,9 @@ class HashAgnosticCascadingBloom
 	 * @param levels number of levels in Cascading Bloom filter
 	 * @param k k-mer size
 	 */
-	HashAgnosticCascadingBloom(size_t size, unsigned hashes,
-		size_t levels, unsigned k) : m_k(k), m_hashes(hashes)
+	HashAgnosticCascadingBloom(size_t size, unsigned hashes, size_t levels, unsigned k)
+	  : m_k(k)
+	  , m_hashes(hashes)
 	{
 		m_data.reserve(levels);
 		for (unsigned i = 0; i < levels; i++)
@@ -50,16 +53,10 @@ class HashAgnosticCascadingBloom
 	 * files.  This is used to make BloomFilter support the
 	 * same interface as HashAgnosticCascadingBloom.
 	 */
-	HashAgnosticCascadingBloom(const string& bloomPath)
-	{
-		loadFilter(bloomPath);
-	}
+	HashAgnosticCascadingBloom(const string& bloomPath) { loadFilter(bloomPath); }
 
 	/** Destructor */
-	~HashAgnosticCascadingBloom()
-	{
-		clear();
-	}
+	~HashAgnosticCascadingBloom() { clear(); }
 
 	/** Return k-mer size used by Bloom filter. */
 	unsigned getKmerSize() const { return m_k; }
@@ -74,6 +71,9 @@ class HashAgnosticCascadingBloom
 		return m_data.back()->getFilterSize();
 	}
 
+	/** Return the size of the bit array. */
+	size_t getFilterSize() const { return size(); }
+
 	/** Return the number of elements with count >= levels. */
 	size_t popcount() const
 	{
@@ -82,16 +82,10 @@ class HashAgnosticCascadingBloom
 	}
 
 	/** Return number of levels in cascading Bloom filter */
-	unsigned levels() const
-	{
-		return m_data.size();
-	}
+	unsigned levels() const { return m_data.size(); }
 
 	/** Return the estimated false positive rate */
-	double FPR() const
-	{
-		return pow((double)popcount()/size(), m_hashes);
-	}
+	double FPR() const { return pow((double)popcount() / size(), m_hashes); }
 
 	/**
 	 * Return true if the element with the given hash values
@@ -145,8 +139,7 @@ class HashAgnosticCascadingBloom
 	}
 
 	/** Operator for writing the Bloom filter to a stream */
-	friend std::ostream& operator<<(std::ostream& out,
-		const HashAgnosticCascadingBloom& o)
+	friend std::ostream& operator<<(std::ostream& out, const HashAgnosticCascadingBloom& o)
 	{
 		assert(o.m_data.size() > 0);
 		assert(o.m_data.back() != NULL);
@@ -166,7 +159,6 @@ class HashAgnosticCascadingBloom
 	}
 
   private:
-
 	/** Free all allocated memory and reset parameters to defaults */
 	void clear()
 	{

Bloom/Makefile.am

@@ -2,7 +2,8 @@ bin_PROGRAMS = abyss-bloom
 
 abyss_bloom_CPPFLAGS = -I$(top_srcdir) \
 	-I$(top_srcdir)/Common \
-	-I$(top_srcdir)/DataLayer
+	-I$(top_srcdir)/DataLayer \
+	-I$(top_srcdir)/vendor
 
 abyss_bloom_CXXFLAGS = $(AM_CXXFLAGS) $(OPENMP_CXXFLAGS)
 
@@ -18,4 +19,5 @@ abyss_bloom_SOURCES = bloom.cc \
 	ConcurrentBloomFilter.h \
 	CascadingBloomFilter.h \
 	CascadingBloomFilterWindow.h \
-	RollingBloomDBGVisitor.h
+	RollingBloomDBGVisitor.h \
+	HashAgnosticCascadingBloom.h

Bloom/RollingBloomDBGVisitor.h

@@ -2,7 +2,9 @@
 #define _ROLLING_BLOOM_DBG_VISITOR_H_ 1
 
 #include "Common/UnorderedMap.h"
+#include "Common/UnorderedSet.h"
 #include "Graph/BreadthFirstSearch.h"
+#include "vendor/btl_bloomfilter/BloomFilter.hpp"
 
 #include <iostream>
 
@@ -11,32 +13,44 @@
  * a bounded breadth first traversal. An instance of this class may be passed
  * as an argument to the `breadthFirstSearch` function.
  */
-template <typename GraphT>
+template<typename GraphT>
 class RollingBloomDBGVisitor : public DefaultBFSVisitor<GraphT>
 {
-public:
-
+  public:
 	typedef typename boost::graph_traits<GraphT>::vertex_descriptor VertexT;
 	typedef typename boost::graph_traits<GraphT>::edge_descriptor EdgeT;
 
 	typedef unordered_map<VertexT, unsigned> DepthMap;
 	typedef typename DepthMap::const_iterator DepthMapConstIt;
 	typedef typename DepthMap::iterator DepthMapIt;
 
-	typedef std::vector<std::pair<std::string, unordered_set<VertexT> > >
-		KmerProperties;
+	typedef std::vector<std::pair<std::string, unordered_set<VertexT>>> KmerProperties;
 	typedef typename KmerProperties::const_iterator KmerPropertiesIt;
 
+	typedef std::vector<std::pair<std::string, BloomFilter*>> BloomProperties;
+	typedef typename BloomProperties::const_iterator BloomPropertiesIt;
+
 	/** Constructor */
-	RollingBloomDBGVisitor(const VertexT& root, unsigned maxDepth,
-		const KmerProperties& kmerProperties, std::ostream& out)
-		: m_root(root), m_maxDepth(maxDepth),
-		m_kmerProperties(kmerProperties), m_out(out)
+	template<typename VertexSetT>
+	RollingBloomDBGVisitor(
+	    const VertexSetT& roots,
+	    unsigned maxDepth,
+	    const KmerProperties& kmerProperties,
+	    const BloomProperties& bloomProperties,
+	    std::ostream& out)
+	  : m_maxDepth(maxDepth)
+	  , m_kmerProperties(kmerProperties)
+	  , m_bloomProperties(bloomProperties)
+	  , m_out(out)
 	{
-		m_depthMap[root] = 0;
+		typedef typename VertexSetT::const_iterator RootIt;
+
+		for (RootIt it = roots.begin(); it != roots.end(); ++it)
+			m_depthMap[*it] = 0;
 
 		/* start directed graph (GraphViz) */
-		m_out << "digraph " << root.kmer().c_str() << " {\n";
+		m_out << "digraph "
+		      << " {\n";
 	}
 
 	/** Called when graph search completes */
@@ -64,8 +78,7 @@ class RollingBloomDBGVisitor : public DefaultBFSVisitor<GraphT>
 
 		DepthMapIt childIt;
 		bool inserted;
-		boost::tie(childIt, inserted) = m_depthMap.insert(
-			std::make_pair(child, parentDepth + 1));
+		boost::tie(childIt, inserted) = m_depthMap.insert(std::make_pair(child, parentDepth + 1));
 		assert(inserted);
 
 		/*
@@ -90,13 +103,22 @@ class RollingBloomDBGVisitor : public DefaultBFSVisitor<GraphT>
 		assert(depthIt != m_depthMap.end());
 		unsigned depth = depthIt->second;
 
-	    m_out << "depth=" << depth;
+		m_out << "depth=" << depth;
 
-		for (KmerPropertiesIt it = m_kmerProperties.begin();
-			it != m_kmerProperties.end(); ++it) {
+		for (KmerPropertiesIt it = m_kmerProperties.begin(); it != m_kmerProperties.end(); ++it) {
 			if (it->second.find(v) != it->second.end())
 				m_out << "," << it->first;
 		}
+
+		size_t hashes[MAX_HASHES];
+		v.rollingHash().getHashes(hashes);
+
+		for (BloomPropertiesIt it = m_bloomProperties.begin(); it != m_bloomProperties.end();
+		     ++it) {
+			if (it->second->contains(hashes))
+				m_out << "," << it->first;
+		}
+
 		m_out << "];\n";
 
 		return BFS_SUCCESS;
@@ -126,8 +148,7 @@ class RollingBloomDBGVisitor : public DefaultBFSVisitor<GraphT>
 	/** Return if the current edge is a forward edge */
 	bool isForwardEdge(const EdgeT& e, const GraphT& g)
 	{
-		assert(source(e, g) == m_currentVertex
-			|| target(e, g) == m_currentVertex);
+		assert(source(e, g) == m_currentVertex || target(e, g) == m_currentVertex);
 		return source(e, g) == m_currentVertex;
 	}
 
@@ -138,20 +159,17 @@ class RollingBloomDBGVisitor : public DefaultBFSVisitor<GraphT>
 		const VertexT& v = target(e, g);
 
 		/* declare edge (GraphViz) */
-		m_out << '\t' << u.kmer().c_str() << " -> "
-			<< v.kmer().c_str() << ";\n";
+		m_out << '\t' << u.kmer().c_str() << " -> " << v.kmer().c_str() << ";\n";
 	}
 
-protected:
-
+  protected:
 	VertexT m_currentVertex;
 	DepthMap m_depthMap;
 
-	const VertexT& m_root;
 	const unsigned m_maxDepth;
 	const KmerProperties& m_kmerProperties;
+	const BloomProperties& m_bloomProperties;
 	std::ostream& m_out;
 };
 
-
 #endif