/**
* Gets records for the most recent position and fills up the buffer from file i.
* returns true if buffer is filled or it is not necessary to fill buffer.
* returns false if no more records are found to fill buffer
*/
void BCFSyncedStreamReader::fill_buffer(int32_t i)
{
//not necessary to fill buffer
if (buffer[i].size()>=2)
return;
int32_t pos1 = buffer[i].size()==0 ? 0 : bcf_get_pos1(buffer[i].front());
if (ftypes[i]==FT_BCF_GZ)
{
bcf1_t *v = get_bcf1_from_pool();
bool populated = false;
while (itrs[i] && bcf_itr_next(vcfs[i], itrs[i], v) >= 0)
{
populated = true;
bcf_unpack(v, BCF_UN_STR);
buffer[i].push_back(v);
insert_into_pq(i, v);
if (pos1==0)
{
pos1 = bcf_get_pos1(v);
}
if (bcf_get_pos1(v)!=pos1)
{
break;
}
v = get_bcf1_from_pool();
populated = false;
}
if (!populated)
store_bcf1_into_pool(v);
}
else if (ftypes[i]==FT_VCF_GZ)
{
while (itrs[i] && tbx_itr_next(vcfs[i], tbxs[i], itrs[i], &s) >= 0)
{
bcf1_t *v = get_bcf1_from_pool();
vcf_parse(&s, hdrs[i], v);
bcf_unpack(v, BCF_UN_STR);
buffer[i].push_back(v);
insert_into_pq(i, v);
if (pos1==0)
{
pos1 = bcf_get_pos1(v);
}
if (bcf_get_pos1(v)!=pos1)
{
break;
}
}
}
}
/**
* Ensures that buffer for each file contains at least records of 2 different positions
* Updates the latest position. [store latest and second latest]
* returns false when all files are read through.
* Note that these bcf1_t memory allocation are handled by BCFSyncedStreamReader.
*/
bool BCFSyncedStreamReader::read_next_position(std::vector<bcfptr*>& current_recs)
{
//put records in pool
for (uint32_t i=0; i<current_recs.size(); ++i)
{
store_bcf1_into_pool(current_recs[i]->v);
delete(current_recs[i]);
}
current_recs.clear();
//process records in priority queue or initialize next interval if pq is empty
//initialize_next_interval tops up the pq
if (pq.size()!=0 || initialize_next_interval())
{
//dequeue pqueue most recent position and return it
bcfptr* variant = pq.top();
bcfptr* cvariant = variant;
while (bcfptr_scmp(cvariant, variant)==0)
{
bcfptr *b = pq.top();
current_recs.push_back(b);
buffer[b->file_index].remove(b->v);
fill_buffer(b->file_index);
pq.pop();
if (pq.size()==0)
{
break;
}
else
{
cvariant = pq.top();
}
}
//current_pos1 = current_recs.front().pos1;
return true;
}
else //end of contig or eof for all files
{
return false;
}
}
/**
* Gets records for the most recent position and fills up the buffer from file i.
* returns true if buffer is filled or it is not necessary to fill buffer.
* returns false if no more records are found to fill buffer
*/
void BCFSyncedReader::fill_buffer(int32_t i)
{
if (buffer[i].size()>=2)
return;
if (random_access)
{
int32_t pos1 = buffer[i].size()==0 ? 0 : bcf_get_pos1(buffer[i].front());
if (ftypes[i].format==bcf)
{
bcf1_t *v = get_bcf1_from_pool();
bool populated = false;
while (itrs[i] && bcf_itr_next(files[i], itrs[i], v)>=0)
{
populated = true;
bcf_unpack(v, BCF_UN_STR);
//check to ensure order
if (!buffer[i].empty())
{
if (!bcf_is_in_order(buffer[i].back(), v))
{
fprintf(stderr, "[E:%s:%d %s] VCF file not in order: %s\n", __FILE__, __LINE__, __FUNCTION__, file_names[i].c_str());
exit(1);
}
}
buffer[i].push_back(v);
insert_into_pq(i, v);
if (pos1==0)
{
pos1 = bcf_get_pos1(v);
}
if (bcf_get_pos1(v)!=pos1)
{
break;
}
v = get_bcf1_from_pool();
populated = false;
}
if (!populated)
store_bcf1_into_pool(v);
}
else if (ftypes[i].format==vcf)
{
while (itrs[i] && tbx_itr_next(files[i], tbxs[i], itrs[i], &s)>=0)
{
bcf1_t *v = get_bcf1_from_pool();
vcf_parse(&s, hdrs[i], v);
bcf_unpack(v, BCF_UN_STR);
//check to ensure order
if (!buffer[i].empty())
{
if (!bcf_is_in_order(buffer[i].back(), v))
{
fprintf(stderr, "[E:%s:%d %s] VCF file not in order: %s\n", __FILE__, __LINE__, __FUNCTION__, file_names[i].c_str());
exit(1);
}
}
buffer[i].push_back(v);
insert_into_pq(i, v);
if (pos1==0)
{
pos1 = bcf_get_pos1(v);
}
if (bcf_get_pos1(v)!=pos1)
{
break;
}
}
}
}
else
{
int32_t rid = buffer[i].size()==0 ? -1 : bcf_get_rid(buffer[i].front());
int32_t pos1 = buffer[i].size()==0 ? 0 : bcf_get_pos1(buffer[i].front());
bcf1_t *v = get_bcf1_from_pool();
bool populated = false;
while (bcf_read(files[i], hdrs[i], v)>=0)
{
populated = true;
bcf_unpack(v, BCF_UN_STR);
//check to ensure order
if (!buffer[i].empty())
{
if (!bcf_is_in_order(buffer[i].back(), v))
{
fprintf(stderr, "[E:%s:%d %s] VCF file not in order: %s\n", __FILE__, __LINE__, __FUNCTION__, file_names[i].c_str());
exit(1);
}
}
buffer[i].push_back(v);
insert_into_pq(i, v);
if (rid==-1)
{
rid = bcf_get_rid(v);
pos1 = bcf_get_pos1(v);
}
if (bcf_get_rid(v)!=rid || bcf_get_pos1(v)!=pos1)
{
break;
}
v = get_bcf1_from_pool();
populated = false;
}
if (!populated)
store_bcf1_into_pool(v);
}
}