void HashGraph::RefreshVertices(unsigned minCount)
{
#pragma omp parallel for
for (int64 i = 0; i < (int64)table_size; ++i)
{
HashNode *node = table[i];
HashNode *prev = NULL;
while (node != NULL)
{
if (node->IsDead() || node->Count() < minCount)
{
#pragma omp atomic
--num_nodes;
if (prev == NULL)
{
table[i] = node->next;
FreeNode(node, omp_get_thread_num());
node = table[i];
}
else
{
prev->next = node->next;
FreeNode(node, omp_get_thread_num());
node = prev->next;
}
}
else
{
node->ClearStatus();
prev = node;
node = prev->next;
}
}
}
}
double HashGraph::AverageCoverage()
{
long long sum = 0;
uint64 valid = 0;
for (int64 i = 0; i < table_size; ++i)
{
for (HashNode *node = table[i]; node; node = node->next)
{
sum += node->Count();
++valid;
}
}
return 1.0 * sum / valid;
}
double HashGraph::MedianCoverage()
{
vector<int> v;
v.reserve(num_nodes);
for (int64 i = 0; i < table_size; ++i)
{
for (HashNode *node = table[i]; node; node = node->next)
{
v.push_back(node->Count());
}
}
nth_element(v.begin(), v.begin() + v.size()/2, v.end());
return *(v.begin() + v.size()/2);
}
int64 HashGraph::Assemble(vector<Contig> &contigs)
{
contigs.resize(0);
int tangle = 0;
omp_lock_t lockContigs;
omp_init_lock(&lockContigs);
#pragma omp parallel for
for (int64 i = 0; i < (int64)table_size; ++i)
{
for (HashNode *node = table[i]; node != NULL; node = node->next)
{
if (!node->Lock(omp_get_thread_num()))
continue;
Contig contig;
contig.SetContent(node->kmer, kmerLength);
contig.sum_coverage = node->Count();
contig.is_tangle = false;
if (!node->kmer.IsPalindrome())
{
for (int strand = 0; strand < 2; ++strand)
{
KmerNodeAdapter adp = GetNodeAdapter(contig.GetEndKmer());
KmerNodeAdapter next(NULL);
while (true)
{
if (!GetNextNodeAdapter(adp, next))
break;
if (next.GetLockID() == omp_get_thread_num() && IsLoop(contig, next))
break;
if (!next.LockPreempt(omp_get_thread_num()))
goto FAIL;
contig.AddNucleotide(BitOperation::bitToIndex[adp.OutEdges()]);
adp = next;
contig.sum_coverage += next.Count();
}
if (adp.OutDegree() == 0)
contig.is_tangle = true;
contig.ReverseComplement();
}
}
if (contig.is_tangle)
{
#pragma omp atomic
++tangle;
}
omp_set_lock(&lockContigs);
contigs.push_back(contig);
omp_unset_lock(&lockContigs);
FAIL:
;
}
}
omp_destroy_lock(&lockContigs);
ClearStatus();
LogMessage("tangle %d total %d\n", tangle, contigs.size());
return contigs.size();
}