pio_status_t pio_open_ex(struct pio_file_t *plink_file, const char *fam_path, const char *bim_path, const char *bed_path)
{
int error = 0;
int num_samples = 0;
int num_loci = 0;
if( fam_open( &plink_file->fam_file, fam_path ) == PIO_OK )
{
num_samples = fam_num_samples( &plink_file->fam_file );
}
else
{
error = 1;
}
if( bim_open( &plink_file->bim_file, bim_path ) == PIO_OK )
{
num_loci = bim_num_loci( &plink_file->bim_file );
}
else
{
error = 1;
}
if( bed_open( &plink_file->bed_file, bed_path, num_loci, num_samples ) != PIO_OK )
{
error = 1;
}
if( error == 0 )
{
return PIO_OK;
}
else
{
fam_close( &plink_file->fam_file );
bim_close( &plink_file->bim_file );
bed_close( &plink_file->bed_file );
return PIO_ERROR;
}
}
region_table_t *parse_regions_from_bed_file(char *filename, const char *url, const char *species, const char *version) {
bed_file_t *file = bed_open(filename);
if (file == NULL) {
return NULL;
}
region_table_t *regions_table = new_region_table_from_ws(url, species, version);
int ret_code = 0;
size_t max_batches = 20, batch_size = 2000;
list_t *read_list = (list_t*) malloc (sizeof(list_t));
list_init("batches", 1, max_batches, read_list);
#pragma omp parallel sections
{
// The producer reads the bed file
#pragma omp section
{
LOG_DEBUG_F("Thread %d reads the BED file\n", omp_get_thread_num());
ret_code = bed_read_batches(read_list, batch_size, file);
list_decr_writers(read_list);
if (ret_code) {
LOG_FATAL_F("Error while reading BED file %s (%d)\n", filename, ret_code);
}
}
// The consumer inserts regions in the structure
#pragma omp section
{
list_item_t *item = NULL;
bed_batch_t *batch;
bed_record_t *record;
region_t *regions_batch[REGIONS_CHUNKSIZE];
int avail_regions = 0;
while ( item = list_remove_item(read_list) ) {
batch = item->data_p;
// For each record in the batch, generate a new region
for (int i = 0; i < batch->records->size; i++) {
record = batch->records->items[i];
region_t *region = region_new(strndup(record->sequence, record->sequence_len),
record->start, record->end, strndup(&record->strand, 1),
NULL);
LOG_DEBUG_F("region '%s:%u-%u'\n", region->chromosome, region->start_position, region->end_position);
regions_batch[avail_regions++] = region;
// Save when the recommended size is reached
if (avail_regions == REGIONS_CHUNKSIZE) {
insert_regions(regions_batch, avail_regions, regions_table);
for (int i = 0; i < avail_regions; i++) {
free(regions_batch[i]);
}
avail_regions = 0;
}
}
bed_batch_free(batch);
list_item_free(item);
}
// Save the remaining regions that did not fill a batch
if (avail_regions > 0) {
insert_regions(regions_batch, avail_regions, regions_table);
for (int i = 0; i < avail_regions; i++) {
free(regions_batch[i]);
}
avail_regions = 0;
}
}
}
list_free_deep(read_list, NULL);
bed_close(file, 1);
return regions_table;
}