int bns_intv2rid(const bntseq_t *bns, int64_t rb, int64_t re)
{
int is_rev, rid_b, rid_e;
if (rb < bns->l_pac && re > bns->l_pac) return -2;
rid_b = bns_pos2rid(bns, bns_depos(bns, rb, &is_rev));
rid_e = bns_pos2rid(bns, bns_depos(bns, re, &is_rev) - 1);
return rid_b == rid_e? rid_b : -1;
}
int bns_intv2rid(const bntseq_t *bns, int64_t rb, int64_t re)
{
int is_rev, rid_b, rid_e;
if (rb < bns->l_pac && re > bns->l_pac) return -2;
assert(rb <= re);
rid_b = bns_pos2rid(bns, bns_depos(bns, rb, &is_rev));
rid_e = rb < re? bns_pos2rid(bns, bns_depos(bns, re - 1, &is_rev)) : rid_b;
return rid_b == rid_e? rid_b : -1;
}
uint8_t *bns_fetch_seq(const bntseq_t *bns, const uint8_t *pac, int64_t *beg, int64_t mid, int64_t *end, int *rid)
{
int64_t far_beg, far_end, len;
int is_rev;
uint8_t *seq;
if (*end < *beg) *end ^= *beg, *beg ^= *end, *end ^= *beg; // if end is smaller, swap
assert(*beg <= mid && mid < *end);
*rid = bns_pos2rid(bns, bns_depos(bns, mid, &is_rev));
far_beg = bns->anns[*rid].offset;
far_end = far_beg + bns->anns[*rid].len;
if (is_rev) { // flip to the reverse strand
int64_t tmp = far_beg;
far_beg = (bns->l_pac<<1) - far_end;
far_end = (bns->l_pac<<1) - tmp;
}
*beg = *beg > far_beg? *beg : far_beg;
*end = *end < far_end? *end : far_end;
seq = bns_get_seq(bns->l_pac, pac, *beg, *end, &len);
if (seq == 0 || *end - *beg != len) {
fprintf(stderr, "[E::%s] begin=%ld, mid=%ld, end=%ld, len=%ld, seq=%p, rid=%d, far_beg=%ld, far_end=%ld\n",
__func__, (long)*beg, (long)mid, (long)*end, (long)len, seq, *rid, (long)far_beg, (long)far_end);
}
assert(seq && *end - *beg == len); // assertion failure should never happen
return seq;
}
int bns_cnt_ambi(const bntseq_t *bns, int64_t pos_f, int len, int *ref_id)
{
int left, mid, right, nn;
if (ref_id) *ref_id = bns_pos2rid(bns, pos_f);
left = 0; right = bns->n_holes; nn = 0;
while (left < right) {
mid = (left + right) >> 1;
if (pos_f >= bns->ambs[mid].offset + bns->ambs[mid].len) left = mid + 1;
else if (pos_f + len <= bns->ambs[mid].offset) right = mid;
else { // overlap
if (pos_f >= bns->ambs[mid].offset) {
nn += bns->ambs[mid].offset + bns->ambs[mid].len < pos_f + len?
bns->ambs[mid].offset + bns->ambs[mid].len - pos_f : len;
} else {
nn += bns->ambs[mid].offset + bns->ambs[mid].len < pos_f + len?
bns->ambs[mid].len : len - (bns->ambs[mid].offset - pos_f);
}
break;
}
}
return nn;
}
void mem_reg2ovlp(const mem_opt_t *opt, const bntseq_t *bns, const uint8_t *pac, bseq1_t *s, mem_alnreg_v *a)
{
int i;
kstring_t str = {0,0,0};
for (i = 0; i < a->n; ++i) {
const mem_alnreg_t *p = &a->a[i];
int is_rev, rid, qb = p->qb, qe = p->qe;
int64_t pos, rb = p->rb, re = p->re;
pos = bns_depos(bns, rb < bns->l_pac? rb : re - 1, &is_rev);
rid = bns_pos2rid(bns, pos);
assert(rid == p->rid);
pos -= bns->anns[rid].offset;
kputs(s->name, &str); kputc('\t', &str);
kputw(s->l_seq, &str); kputc('\t', &str);
if (is_rev) qb ^= qe, qe ^= qb, qb ^= qe; // swap
kputw(qb, &str); kputc('\t', &str); kputw(qe, &str); kputc('\t', &str);
kputs(bns->anns[rid].name, &str); kputc('\t', &str);
kputw(bns->anns[rid].len, &str); kputc('\t', &str);
kputw(pos, &str); kputc('\t', &str); kputw(pos + (re - rb), &str); kputc('\t', &str);
ksprintf(&str, "%.3f", (double)p->truesc / opt->a / (qe - qb > re - rb? qe - qb : re - rb));
kputc('\n', &str);
}
s->sam = str.s;
}
