int bam_smpl_get_sample_id(bam_smpl_t *bsmpl, int bam_id, bam1_t *bam_rec)
{
file_t *file = &bsmpl->files[bam_id];
if ( file->default_idx >= 0 ) return file->default_idx;
char *aux_rg = (char*) bam_aux_get(bam_rec, "RG");
aux_rg = aux_rg ? aux_rg+1 : "?";
int rg_id;
if ( khash_str2int_get(file->rg2idx, aux_rg, &rg_id)==0 ) return rg_id;
if ( khash_str2int_get(file->rg2idx, "?", &rg_id)==0 ) return rg_id;
return -1;
}
static int mplp_func(void *data, bam1_t *b)
{
extern int bam_realn(bam1_t *b, const char *ref);
extern int bam_prob_realn_core(bam1_t *b, const char *ref, int ref_len, int flag);
extern int bam_cap_mapQ(bam1_t *b, char *ref, int ref_len, int thres);
char *ref;
mplp_aux_t *ma = (mplp_aux_t*)data;
int ret, skip = 0, ref_len;
do {
int has_ref;
ret = ma->iter? sam_itr_next(ma->fp, ma->iter, b) : sam_read1(ma->fp, ma->h, b);
if (ret < 0) break;
// The 'B' cigar operation is not part of the specification, considering as obsolete.
// bam_remove_B(b);
if (b->core.tid < 0 || (b->core.flag&BAM_FUNMAP)) { // exclude unmapped reads
skip = 1;
continue;
}
if (ma->conf->rflag_require && !(ma->conf->rflag_require&b->core.flag)) { skip = 1; continue; }
if (ma->conf->rflag_filter && ma->conf->rflag_filter&b->core.flag) { skip = 1; continue; }
if (ma->conf->bed) { // test overlap
skip = !bed_overlap(ma->conf->bed, ma->h->target_name[b->core.tid], b->core.pos, bam_endpos(b));
if (skip) continue;
}
if (ma->conf->rghash) { // exclude read groups
uint8_t *rg = bam_aux_get(b, "RG");
skip = (rg && khash_str2int_get(ma->conf->rghash, (const char*)(rg+1), NULL)==0);
if (skip) continue;
}
if (ma->conf->flag & MPLP_ILLUMINA13) {
int i;
uint8_t *qual = bam_get_qual(b);
for (i = 0; i < b->core.l_qseq; ++i)
qual[i] = qual[i] > 31? qual[i] - 31 : 0;
}
if (ma->conf->fai && b->core.tid >= 0) {
has_ref = mplp_get_ref(ma, b->core.tid, &ref, &ref_len);
if (has_ref && ref_len <= b->core.pos) { // exclude reads outside of the reference sequence
fprintf(stderr,"[%s] Skipping because %d is outside of %d [ref:%d]\n",
__func__, b->core.pos, ref_len, b->core.tid);
skip = 1;
continue;
}
} else {
has_ref = 0;
}
skip = 0;
if (has_ref && (ma->conf->flag&MPLP_REALN)) bam_prob_realn_core(b, ref, ref_len, (ma->conf->flag & MPLP_REDO_BAQ)? 7 : 3);
if (has_ref && ma->conf->capQ_thres > 10) {
int q = bam_cap_mapQ(b, ref, ref_len, ma->conf->capQ_thres);
if (q < 0) skip = 1;
else if (b->core.qual > q) b->core.qual = q;
}
if (b->core.qual < ma->conf->min_mq) skip = 1;
else if ((ma->conf->flag&MPLP_NO_ORPHAN) && (b->core.flag&BAM_FPAIRED) && !(b->core.flag&BAM_FPROPER_PAIR)) skip = 1;
} while (skip);
return ret;
}
// Add a new region into a list sorted by start,end. On input the coordinates
// are 1-based, stored 0-based, inclusive.
static void _regions_add(bcf_sr_regions_t *reg, const char *chr, int start, int end)
{
if ( start==-1 && end==-1 )
{
start = 0; end = MAX_CSI_COOR-1;
}
else
{
start--; end--; // store 0-based coordinates
}
if ( !reg->seq_hash )
reg->seq_hash = khash_str2int_init();
int iseq;
if ( khash_str2int_get(reg->seq_hash, chr, &iseq)<0 )
{
// the chromosome block does not exist
iseq = reg->nseqs++;
reg->seq_names = (char**) realloc(reg->seq_names,sizeof(char*)*reg->nseqs);
reg->regs = (region_t*) realloc(reg->regs,sizeof(region_t)*reg->nseqs);
memset(®->regs[reg->nseqs-1],0,sizeof(region_t));
reg->seq_names[iseq] = strdup(chr);
reg->regs[iseq].creg = -1;
khash_str2int_set(reg->seq_hash,reg->seq_names[iseq],iseq);
}
region_t *creg = ®->regs[iseq];
// the regions may not be sorted on input: binary search
int i, min = 0, max = creg->nregs - 1;
while ( min<=max )
{
i = (max+min)/2;
if ( start < creg->regs[i].start ) max = i - 1;
else if ( start > creg->regs[i].start ) min = i + 1;
else break;
}
if ( min>max || creg->regs[i].start!=start || creg->regs[i].end!=end )
{
// no such region, insert a new one just after max
hts_expand(region1_t,creg->nregs+1,creg->mregs,creg->regs);
if ( ++max < creg->nregs )
memmove(&creg->regs[max+1],&creg->regs[max],(creg->nregs - max)*sizeof(region1_t));
creg->regs[max].start = start;
creg->regs[max].end = end;
creg->nregs++;
}
}
int bcf_sr_regions_seek(bcf_sr_regions_t *reg, const char *seq)
{
reg->iseq = reg->start = reg->end = -1;
if ( khash_str2int_get(reg->seq_hash, seq, ®->iseq) < 0 ) return -1; // sequence seq not in regions
// using in-memory regions
if ( reg->regs ) return 0;
// reading regions from tabix
if ( reg->itr ) tbx_itr_destroy(reg->itr);
reg->itr = tbx_itr_querys(reg->tbx, seq);
if ( reg->itr ) return 0;
return -1;
}
int ploidy_parse(const char *line, char **chr_beg, char **chr_end, reg_t *reg, void *payload, void *usr)
{
int i, ret;
ploidy_t *ploidy = (ploidy_t*) usr;
void *sex2id = ploidy->sex2id;
// Check for special case of default ploidy "* * * <sex> <ploidy>"
int default_ploidy_def = 0;
char *ss = (char*) line;
while ( *ss && isspace(*ss) ) ss++;
if ( ss[0]=='*' && (!ss[1] || isspace(ss[1])) )
default_ploidy_def = 1; // definition of default ploidy, chr="*"
else
{
// Fill CHR,FROM,TO
ret = regidx_parse_tab(line,chr_beg,chr_end,reg,NULL,NULL);
if ( ret!=0 ) return ret;
}
// Skip the fields already parsed by regidx_parse_tab
ss = (char*) line;
while ( *ss && isspace(*ss) ) ss++;
for (i=0; i<3; i++)
{
while ( *ss && !isspace(*ss) ) ss++;
if ( !*ss ) return -2; // wrong number of fields
while ( *ss && isspace(*ss) ) ss++;
}
if ( !*ss ) return -2;
// Parse the payload
char *se = ss;
while ( *se && !isspace(*se) ) se++;
if ( !*se || se==ss ) error("Could not parse: %s\n", line);
ploidy->tmp_str.l = 0;
kputsn(ss,se-ss,&ploidy->tmp_str);
sex_ploidy_t *sp = (sex_ploidy_t*) payload;
if ( khash_str2int_get(sex2id, ploidy->tmp_str.s, &sp->sex) != 0 )
{
ploidy->nsex++;
hts_expand0(char*,ploidy->nsex,ploidy->msex,ploidy->id2sex);
ploidy->id2sex[ploidy->nsex-1] = strdup(ploidy->tmp_str.s);
sp->sex = khash_str2int_inc(ploidy->sex2id, ploidy->id2sex[ploidy->nsex-1]);
ploidy->sex2dflt = (int*) realloc(ploidy->sex2dflt,sizeof(int)*ploidy->nsex);
ploidy->sex2dflt[ploidy->nsex-1] = ploidy->dflt;
}
inline int regidx_push(regidx_t *idx, char *chr_beg, char *chr_end, uint32_t beg, uint32_t end, void *payload)
{
if ( beg > MAX_COOR_0 ) beg = MAX_COOR_0;
if ( end > MAX_COOR_0 ) end = MAX_COOR_0;
int rid;
idx->str.l = 0;
kputsn(chr_beg, chr_end-chr_beg+1, &idx->str);
if ( khash_str2int_get(idx->seq2regs, idx->str.s, &rid)!=0 )
{
// new chromosome
idx->nseq++;
int m_prev = idx->mseq;
hts_expand0(reglist_t,idx->nseq,idx->mseq,idx->seq);
hts_expand0(char*,idx->nseq,m_prev,idx->seq_names);
idx->seq_names[idx->nseq-1] = strdup(idx->str.s);
rid = khash_str2int_inc(idx->seq2regs, idx->seq_names[idx->nseq-1]);
}
int regidx_insert(regidx_t *idx, char *line)
{
if ( !line )
return _regidx_build_index(idx);
char *chr_from, *chr_to;
reg_t reg;
int ret = idx->parse(line,&chr_from,&chr_to,®,idx->payload,idx->usr);
if ( ret==-2 ) return -1; // error
if ( ret==-1 ) return 0; // skip the line
int rid;
idx->str.l = 0;
kputsn(chr_from, chr_to-chr_from+1, &idx->str);
if ( khash_str2int_get(idx->seq2regs, idx->str.s, &rid)!=0 )
{
idx->nseq++;
int m_prev = idx->mseq;
hts_expand0(reglist_t,idx->nseq,idx->mseq,idx->seq);
hts_expand0(char*,idx->nseq,m_prev,idx->seq_names);
idx->seq_names[idx->nseq-1] = strdup(idx->str.s);
rid = khash_str2int_inc(idx->seq2regs, idx->seq_names[idx->nseq-1]);
}
static void bsmpl_add_readgroup(bam_smpl_t *bsmpl, file_t *file, const char *rg_id, const char *smpl_name)
{
int ismpl = -1;
if ( smpl_name )
{
if ( khash_str2int_get(bsmpl->name2idx,smpl_name,&ismpl) < 0 )
{
// new sample
bsmpl->nsmpl++;
bsmpl->smpl = (char**) realloc(bsmpl->smpl,sizeof(char*)*bsmpl->nsmpl);
bsmpl->smpl[bsmpl->nsmpl-1] = strdup(smpl_name);
ismpl = khash_str2int_inc(bsmpl->name2idx,bsmpl->smpl[bsmpl->nsmpl-1]);
}
}
if ( !strcmp("*",rg_id) )
{
// all read groups in the bam treated as the same sample
file->default_idx = ismpl;
return;
}
if ( !file->rg2idx ) file->rg2idx = khash_str2int_init();
if ( khash_str2int_has_key(file->rg2idx,rg_id) ) return; // duplicate @RG:ID
khash_str2int_set(file->rg2idx, strdup(rg_id), ismpl);
}
int regidx_seq_nregs(regidx_t *idx, const char *seq)
{
int iseq;
if ( khash_str2int_get(idx->seq2regs, seq, &iseq)!=0 ) return 0; // no such sequence
return idx->seq[iseq].nreg;
}
static void bcf_sr_sort_set(bcf_srs_t *readers, sr_sort_t *srt, const char *chr, int min_pos)
{
if ( !srt->grp_str2int )
{
// first time here, initialize
if ( !srt->pair )
{
if ( readers->collapse==COLLAPSE_NONE ) readers->collapse = BCF_SR_PAIR_EXACT;
bcf_sr_set_opt(readers, BCF_SR_PAIR_LOGIC, readers->collapse);
}
bcf_sr_init_scores(srt);
srt->grp_str2int = khash_str2int_init();
srt->var_str2int = khash_str2int_init();
}
int k;
khash_t(str2int) *hash;
hash = srt->grp_str2int;
for (k=0; k < kh_end(hash); k++)
if ( kh_exist(hash,k) ) free((char*)kh_key(hash,k));
hash = srt->var_str2int;
for (k=0; k < kh_end(hash); k++)
if ( kh_exist(hash,k) ) free((char*)kh_key(hash,k));
kh_clear(str2int, srt->grp_str2int);
kh_clear(str2int, srt->var_str2int);
srt->ngrp = srt->nvar = srt->nvset = 0;
grp_t grp;
memset(&grp,0,sizeof(grp_t));
// group VCFs into groups, each with a unique combination of variants in the duplicate lines
int ireader,ivar,irec,igrp,ivset,iact;
for (ireader=0; ireader<readers->nreaders; ireader++) srt->vcf_buf[ireader].nrec = 0;
for (iact=0; iact<srt->nactive; iact++)
{
ireader = srt->active[iact];
bcf_sr_t *reader = &readers->readers[ireader];
int rid = bcf_hdr_name2id(reader->header, chr);
grp.nvar = 0;
hts_expand(int,reader->nbuffer,srt->moff,srt->off);
srt->noff = 0;
srt->str.l = 0;
for (irec=1; irec<=reader->nbuffer; irec++)
{
bcf1_t *line = reader->buffer[irec];
if ( line->rid!=rid || line->pos!=min_pos ) break;
if ( srt->str.l ) kputc(';',&srt->str);
srt->off[srt->noff++] = srt->str.l;
size_t beg = srt->str.l;
for (ivar=1; ivar<line->n_allele; ivar++)
{
if ( ivar>1 ) kputc(',',&srt->str);
kputs(line->d.allele[0],&srt->str);
kputc('>',&srt->str);
kputs(line->d.allele[ivar],&srt->str);
}
if ( line->n_allele==1 )
{
kputs(line->d.allele[0],&srt->str);
kputsn(">.",2,&srt->str);
}
// Create new variant or attach to existing one. But careful, there can be duplicate
// records with the same POS,REF,ALT (e.g. in dbSNP-b142)
char *var_str = beg + srt->str.s;
int ret, var_idx = 0, var_end = srt->str.l;
while ( 1 )
{
ret = khash_str2int_get(srt->var_str2int, var_str, &ivar);
if ( ret==-1 ) break;
var_t *var = &srt->var[ivar];
if ( var->vcf[var->nvcf-1] != ireader ) break;
srt->str.l = var_end;
kputw(var_idx, &srt->str);
var_str = beg + srt->str.s;
var_idx++;
}
if ( ret==-1 )
{
ivar = srt->nvar++;
hts_expand0(var_t,srt->nvar,srt->mvar,srt->var);
srt->var[ivar].nvcf = 0;
khash_str2int_set(srt->var_str2int, strdup(var_str), ivar);
free(srt->var[ivar].str); // possible left-over from the previous position
}
var_t *var = &srt->var[ivar];
var->nalt = line->n_allele - 1;
var->type = bcf_get_variant_types(line);
srt->str.s[var_end] = 0;
if ( ret==-1 )
var->str = strdup(var_str);
int mvcf = var->mvcf;
var->nvcf++;
hts_expand0(int*, var->nvcf, var->mvcf, var->vcf);
if ( mvcf != var->mvcf ) var->rec = (bcf1_t **) realloc(var->rec,sizeof(bcf1_t*)*var->mvcf);
var->vcf[var->nvcf-1] = ireader;
var->rec[var->nvcf-1] = line;
grp.nvar++;
hts_expand(var_t,grp.nvar,grp.mvar,grp.var);
grp.var[grp.nvar-1] = ivar;
}
char *grp_key = grp_create_key(srt);
int ret = khash_str2int_get(srt->grp_str2int, grp_key, &igrp);
if ( ret==-1 )
{
igrp = srt->ngrp++;
hts_expand0(grp_t, srt->ngrp, srt->mgrp, srt->grp);
free(srt->grp[igrp].var);
srt->grp[igrp] = grp;
srt->grp[igrp].key = grp_key;
khash_str2int_set(srt->grp_str2int, grp_key, igrp);
memset(&grp,0,sizeof(grp_t));
}
else
free(grp_key);
srt->grp[igrp].nvcf++;
}
free(grp.var);
// initialize bitmask - which groups is the variant present in
for (ivar=0; ivar<srt->nvar; ivar++)
{
srt->var[ivar].mask = kbs_resize(srt->var[ivar].mask, srt->ngrp);
kbs_clear(srt->var[ivar].mask);
}
for (igrp=0; igrp<srt->ngrp; igrp++)
{
for (ivar=0; ivar<srt->grp[igrp].nvar; ivar++)
{
int i = srt->grp[igrp].var[ivar];
kbs_insert(srt->var[i].mask, igrp);
}
}
// create the initial list of variant sets
for (ivar=0; ivar<srt->nvar; ivar++)
{
ivset = srt->nvset++;
hts_expand0(varset_t, srt->nvset, srt->mvset, srt->vset);
varset_t *vset = &srt->vset[ivset];
vset->nvar = 1;
hts_expand0(var_t, vset->nvar, vset->mvar, vset->var);
vset->var[vset->nvar-1] = ivar;
var_t *var = &srt->var[ivar];
vset->cnt = var->nvcf;
vset->mask = kbs_resize(vset->mask, srt->ngrp);
kbs_clear(vset->mask);
kbs_bitwise_or(vset->mask, var->mask);
int type = 0;
if ( var->type==VCF_REF ) type |= SR_REF;
else
{
if ( var->type & VCF_SNP ) type |= SR_SNP;
if ( var->type & VCF_MNP ) type |= SR_SNP;
if ( var->type & VCF_INDEL ) type |= SR_INDEL;
if ( var->type & VCF_OTHER ) type |= SR_OTHER;
}
var->type = type;
}
#if DEBUG_VSETS
debug_vsets(srt);
#endif
// initialize the pairing matrix
hts_expand(int, srt->ngrp*srt->nvset, srt->mpmat, srt->pmat);
hts_expand(int, srt->nvset, srt->mcnt, srt->cnt);
memset(srt->pmat, 0, sizeof(*srt->pmat)*srt->ngrp*srt->nvset);
for (ivset=0; ivset<srt->nvset; ivset++)
{
varset_t *vset = &srt->vset[ivset];
for (igrp=0; igrp<srt->ngrp; igrp++) srt->pmat[ivset*srt->ngrp+igrp] = 0;
srt->cnt[ivset] = vset->cnt;
}
// pair the lines
while ( srt->nvset )
{
#if DEBUG_VSETS
fprintf(stderr,"\n");
debug_vsets(srt);
#endif
int imax = 0;
for (ivset=1; ivset<srt->nvset; ivset++)
if ( srt->cnt[imax] < srt->cnt[ivset] ) imax = ivset;
int ipair = -1;
uint32_t max_score = 0;
for (ivset=0; ivset<srt->nvset; ivset++)
{
if ( kbs_logical_and(srt->vset[imax].mask,srt->vset[ivset].mask) ) continue; // cannot be merged
uint32_t score = pairing_score(srt, imax, ivset);
// fprintf(stderr,"score: %d %d, logic=%d \t..\t %u\n", imax,ivset,srt->pair,score);
if ( max_score < score ) { max_score = score; ipair = ivset; }
}
// merge rows creating a new variant set this way
if ( ipair!=-1 && ipair!=imax )
{
imax = merge_vsets(srt, imax, ipair);
continue;
}
push_vset(srt, imax);
}
srt->chr = chr;
srt->pos = min_pos;
}
int bcf_sr_regions_next(bcf_sr_regions_t *reg)
{
if ( reg->iseq<0 ) return -1;
reg->start = reg->end = -1;
reg->nals = 0;
// using in-memory regions
if ( reg->regs )
{
while ( reg->iseq < reg->nseqs )
{
reg->regs[reg->iseq].creg++;
if ( reg->regs[reg->iseq].creg < reg->regs[reg->iseq].nregs ) break;
reg->iseq++;
}
if ( reg->iseq >= reg->nseqs ) { reg->iseq = -1; return -1; } // no more regions left
region1_t *creg = ®->regs[reg->iseq].regs[reg->regs[reg->iseq].creg];
reg->start = creg->start;
reg->end = creg->end;
return 0;
}
// reading from tabix
char *chr, *chr_end;
int ichr = 0, ifrom = 1, ito = 2, is_bed = 0, from, to;
if ( reg->tbx )
{
ichr = reg->tbx->conf.sc-1;
ifrom = reg->tbx->conf.bc-1;
ito = reg->tbx->conf.ec-1;
if ( ito<0 ) ito = ifrom;
is_bed = reg->tbx->conf.preset==TBX_UCSC ? 1 : 0;
}
int ret = 0;
while ( !ret )
{
if ( reg->itr )
{
// tabix index present, reading a chromosome block
ret = tbx_itr_next(reg->file, reg->tbx, reg->itr, ®->line);
if ( ret<0 ) { reg->iseq = -1; return -1; }
}
else
{
if ( reg->is_bin )
{
// Waited for seek which never came. Reopen in text mode and stream
// through the regions, otherwise hts_getline would fail
hts_close(reg->file);
reg->file = hts_open(reg->fname, "r");
if ( !reg->file )
{
fprintf(stderr,"[%s:%d %s] Could not open file: %s\n", __FILE__,__LINE__,__FUNCTION__,reg->fname);
reg->file = NULL;
bcf_sr_regions_destroy(reg);
return -1;
}
reg->is_bin = 0;
}
// tabix index absent, reading the whole file
ret = hts_getline(reg->file, KS_SEP_LINE, ®->line);
if ( ret<0 ) { reg->iseq = -1; return -1; }
}
ret = _regions_parse_line(reg->line.s, ichr,ifrom,ito, &chr,&chr_end,&from,&to);
if ( ret<0 )
{
fprintf(stderr,"[%s:%d] Could not parse the file %s, using the columns %d,%d,%d\n", __FILE__,__LINE__,reg->fname,ichr+1,ifrom+1,ito+1);
return -1;
}
}
if ( is_bed ) from++;
*chr_end = 0;
if ( khash_str2int_get(reg->seq_hash, chr, ®->iseq)<0 )
{
fprintf(stderr,"Broken tabix index? The sequence \"%s\" not in dictionary [%s]\n", chr,reg->line.s);
exit(1);
}
*chr_end = '\t';
reg->start = from - 1;
reg->end = to - 1;
return 0;
}
/*
The logic of this function is a bit complicated because we want to work
also with broken bams containing read groups that are not listed in the
header. The desired behavior is as follows:
- when -G is given, read groups which are not listed in the header must
be given explicitly using the "?" symbol in -G.
Otherwise:
- if the bam has no header, all reads in the file are assigned to a
single sample named after the file
- if there is at least one sample defined in the header, reads with no
read group id or with a read group id not listed in the header are
assigned to the first sample encountered in the header
*/
int bam_smpl_add_bam(bam_smpl_t *bsmpl, char *bam_hdr, const char *fname)
{
bsmpl->nfiles++;
bsmpl->files = (file_t*) realloc(bsmpl->files,bsmpl->nfiles*sizeof(file_t));
file_t *file = &bsmpl->files[bsmpl->nfiles-1];
memset(file,0,sizeof(file_t));
file->fname = strdup(fname);
file->default_idx = -1;
if ( bsmpl->ignore_rg || !bam_hdr )
{
// The option --ignore-RG is set or there is no BAM header: use the file name as the sample name
bsmpl_add_readgroup(bsmpl,file,"*",file->fname);
return bsmpl->nfiles-1;
}
void *bam_smpls = khash_str2int_init();
int first_smpl = -1, nskipped = 0;
const char *p = bam_hdr, *q, *r;
while (p != NULL && (q = strstr(p, "@RG")) != 0)
{
char *eol = strchr(q + 3, '\n');
if (q > bam_hdr && *(q - 1) != '\n') { // @RG must be at start of line
p = eol;
continue;
}
p = q + 3;
if ((q = strstr(p, "\tID:")) != 0) q += 4;
if ((r = strstr(p, "\tSM:")) != 0) r += 4;
if (r && q)
{
char *u, *v;
int ioq, ior;
for (u = (char*)q; *u && *u != '\t' && *u != '\n'; ++u);
for (v = (char*)r; *v && *v != '\t' && *v != '\n'; ++v);
ioq = *u; ior = *v; *u = *v = '\0';
// q now points to a null terminated read group id
// r points to a null terminated sample name
if ( !strcmp("*",q) || !strcmp("?",q) )
error("Error: the read group IDs \"*\" and \"?\" have a special meaning in the mpileup code. Please fix the code or the bam: %s\n", fname);
int accept_rg = 1;
if ( bsmpl->sample_list )
{
// restrict samples based on the -s/-S options
char *name = khash_str2str_get(bsmpl->sample_list,r);
if ( bsmpl->sample_logic==0 )
accept_rg = name ? 0 : 1;
else if ( !name )
accept_rg = 0;
else
r = name;
}
if ( accept_rg && bsmpl->rg_list )
{
// restrict readgroups based on the -G option, possibly renaming the sample
accept_rg = bsmpl_keep_readgroup(bsmpl,file,q,&r);
}
if ( accept_rg )
bsmpl_add_readgroup(bsmpl,file,q,r);
else
{
bsmpl_add_readgroup(bsmpl,file,q,NULL); // ignore this RG but note that it was seen in the header
nskipped++;
}
if ( first_smpl<0 )
khash_str2int_get(bsmpl->name2idx,r,&first_smpl);
if ( !khash_str2int_has_key(bam_smpls,r) )
khash_str2int_inc(bam_smpls,strdup(r));
*u = ioq; *v = ior;
}
else
break;
p = eol;
}
int nsmpls = khash_str2int_size(bam_smpls);
khash_str2int_destroy_free(bam_smpls);
const char *smpl_name = NULL;
int accept_null_rg = 1;
if ( bsmpl->rg_list && !bsmpl_keep_readgroup(bsmpl,file,"?",&smpl_name) ) accept_null_rg = 0;
if ( bsmpl->sample_list && first_smpl==-1 ) accept_null_rg = 0;
if ( !accept_null_rg && first_smpl==-1 )
{
// no suitable read group is available in this bam: ignore the whole file.
free(file->fname);
if ( file->rg2idx ) khash_str2int_destroy_free(file->rg2idx);
bsmpl->nfiles--;
return -1;
}
if ( !accept_null_rg ) return bsmpl->nfiles-1;
if ( nsmpls==1 && !nskipped )
{
file->default_idx = first_smpl;
return bsmpl->nfiles-1;
}
if ( !smpl_name ) smpl_name = first_smpl==-1 ? file->fname : bsmpl->smpl[first_smpl];
bsmpl_add_readgroup(bsmpl,file,"?",smpl_name);
return bsmpl->nfiles-1;
}